Decennial Aarhus conferences - Department of Computer Science
Transcription
Decennial Aarhus conferences - Department of Computer Science
CRITICAL2015 ALTERNATIVES 5THDECENNIAL AARHUS CONFERENCE CONFERENCE PROCEEDINGS Conference Papers Critical Alternatives Peer-reviewed papers Proceedings of The Fifth Decennial Aarhus Conference 17- 21 August 2015, Aarhus, Denmark Edited by Shaowen Bardzell, Susanne Bødker, Ole Sejer Iversen, Clemens N. Klokmose and Henrik Korsgaard In cooperation with ACM SIGCHI Title: Critical Alternatives Proceedings of The Fifth Decennial Aarhus Conference Editors: Shaowen Bardzell, Susanne Bødker, Ole Sejer Iversen, Clemens N. Klokmose and Henrik Korsgaard Publisher: Computer Science, Aarhus University, Denmark Printed by SUN-TRYK, Aarhus, Denmark Copyright © remains with the contributing authors. Publication rights licensed to Aarhus University and ACM. Following the conference, the proceedings will be made available through the Aarhus Series on Human Centered Computing and ACM Digital Library. i Foreword 1975-1985-1995-2005 — the decennial Aarhus conferences have traditionally been instrumental for setting new agendas for critically engaged thinking about information technology. The conference series is fundamentally interdisciplinary and emphasizes thinking that is firmly anchored in action, intervention, and scholarly critical practice. With the title Critical Computing – between sense and sensibility, the 2005 edition of the conference marked that computing was rapidly seeping into everyday life. In 2015, we see critical alternatives in alignment with utopian principles—that is, the aspiration that life might not only be different but also radically better. At the same time, radically better alternatives do not emerge out of nowhere: they emerged from contested analyses of the mundane present and demand both commitment and labor to work towards them. Critical alternatives matter and make people reflect. The fifth decennial Aarhus conference, Critical Alternatives, in 2015 aims to set new agendas for theory and practice in computing for quality of human life. While the early Aarhus conferences, from 1975 and onwards, focused on computing in working life, computing today is influencing most parts of human life (civic life, the welfare state, health, learning, leisure, culture, intimacy, ...), thereby calling for critical alternatives from a general quality of life perspective. The papers selected for the conference have undergone a meticulous reviewing process looking at methodical soundness as well as potentials for the creating alternatives and provoking debate. Among 71 full and short paper submissions 21 were accepted. The accepted papers span a broad range of positions and concerns ranging from play to politics. We would like to express great thanks for help and support to the numerous people who have contributed to making the conference possible. In particular we want to thank Marianne Dammand and Ann Mølhave for secretarial help, including registration and hotel arrangements. We want to thank the center for Participatory IT (PIT) and the Department of Computer Science, University of Aarhus for providing resources for the planning and operation of the conference. We hope that the conference will inspire critical and alternative thinking and action for the next decennium. Welcome Olav W. Bertelsen and Kim Halskov Conference Chairs Shaowen Bardzell, Susanne Bødker, Ole Sejer Iversen and Clemens N. Klokmose Chairs of the Program Henrik Korsgaard Proceedings Chair ii Conference committee Conference co-chairs: Olav W. Bertelsen & Kim Halskov Communications co-chairs: Joanna Saad-Sulonen & Eva Eriksson Program co-chairs: Shaowen Bardzell & Ole Iversen Student volunteer co-chairs: Tobias Sonne & Nervo Verdezoto Paper co-chairs: Critical practices track co-chairs: Ole Sejer Iversen, Susanne Bødker & Shaowen Lone Koefoed Hansen & Jeff Bardzell Bardzell Demonstrations co-chairs: Short paper chair: Niels Olof Bouvin & Erik Grönvall Clemens Nylandsted Klokmose Social chair: Workshop co-chairs: Mads Møller Jensen Peter Dalsgaard & Anna Vallgårda AV chairs: Proceedings chair: Philip Tchernavskij & Nick Nielsen Henrik Korsgaard Web chair: Nicolai Brodersen Hansen Program committee Alan Blackwell, Cambridge University, UK Jonas Fritsch, Aarhus University, Denmark Anna Vallgårda, ITU, Copenhagen, Denmark Kia Höök, Mobile Life, Sweden Anne Marie Kanstrup, University of Aalborg, Denmark Liam Bannon, University of Limerick, Ireland Carla Simone, Milano University, Italy Liesbeth Huybrechts, University of Hasselt, Belgium Carman Neustaedter, Simon Fraser University, Canada Marcus Foth, Queensland University of Technology, Australia Christopher le Dantec, Georgia Tech, USA Mathias Korn, IUPUI, USA Christopher Frauenberger, Vienna University of Technology, Rachel Smith, Aarhus University, Denmark Austria Ron Wakkary, Simon Fraser University, Canada Claus Bossen, Aarhus University, Denmark Sampsa Hyysalo, Aalto University, Finland Dag Svanæs, NTNU, Norway Shad Gross, Indiana University, USA Deborah Tatar, Virginia Tech, USA Silvia Lindtner, University of Michigan, USA Eli Blevis, Indiana University, USA Steve Harrison, Virginia Tech, USA Elisa Giaccardi, TU Delft, The Netherlands Stuart Reeves, University of Nottingham, UK Erik Grönvall, ITU Copenhagen, Denmark Tuck Leong, University of Sydney, Australia Gilbert Cockton, Northumbria University, UK Volkmar Pipek, University of Siegen, Germany Gopinaath Kannabiran, Indiana University, USA Wendy Ju, Stanford, USA Helena Karasti, University of Oulu, Finland Yanki Lee, HKDI DESIS Lab for Social Design Research, Hong Irina Shklovski, ITU Copenhagen, Denmark Kong John Vines, Newcastle University, UK iii Participatory Trajectories 1 Revisiting the UsersAward Programme from a Value Sensitive Design 5 On Creating and Sustaining Alternatives: The case of Danish Telehealth 17 Computing and the Common. Hints of a new utopia in Participatory Design 21 Concordance: A Critical Participatory Alternative in Healthcare IT Perspective Åke Walldius, Jan Gulliksen & Yngve Sundblad Morten Kyng Maurizio Teli Erik Gronvall, Nervo Verdezoto, Naveen Bagalkot & Tomas Sokoler Keeping Secrets 25 Networked Privacy Beyond the Individual: Four Perspectives to ‘Sharing’ 29 Personal Data: Thinking Inside the Box Airi Lampinen Amir Chaudhry, Jon Crowcroft, Hamed Haddadi, Heidi Howard, Anil Madhavapeddy & Richard Mortier Deconstructing Norms 33 Deconstructivist Interaction Design: Interrogating Expression and Form 37 In Search of Fairness: Critical Design Alternatives for Sustainability 41 Why Play? Examining the Roles of Play in ICTD 53 Double Binds and Double Blinds: Evaluation Tactics in Critically Martin Murer, Verena Fuchsberger & Manfred Tscheligi Somya Joshi & Teresa Cerratto Pargman Pedro Ferreira Oriented HCI Vera Khovanskaya, Eric P. S. Baumer & Phoebe Sengers The Alternative Rhetorics of HCI 65 Note to Self: Stop Calling Interfaces “Natural” 69 Gaza Everywhere: exploring the applicability of a rhetorical lens in HCI 73 Human-computer interaction as science Lone Koefoed Hansen & Peter Dalsgaard Omar Sosa-Tzec, Erik Stolterman & Martin A. Siegel Stuart Reeves iv Charismatic Materiality 85 Designed in Shenzhen: Shanzhai Manufacturing and Maker Entrepreneurs 97 Material Speculation: Actual Artifacts for Critical Inquiry 109 Charismatic Technology Silvia Lindtner, Anna Greenspan & David Li Ron Wakkary, William Odom, Sabrina Hauser, Garnet Hertz & Henry Lin Morgan G. Ames Critical subjects and subjectivities 121 An Anxious Alliance 133 The User Reconfigured: On Subjectivities of Information 145 Creating Friction: Infrastructuring Civic Engagement in Everyday Life Kaiton Williams Jeffrey Bardzell & Shaowen Bardzell Matthias Korn & Amy Voida Past and Future 157 Not The Internet, but This Internet: How Othernets Illuminate Our Feudal 169 Interacting with an Inferred World: The Challenge of Machine Learning for Internet Paul Dourish Humane Computer Interaction Alan F. Blackwell v Revisiting the UsersAward Programme from a Value Sensitive Design Perspective Åke Walldius KTH Royal Inst. of Technology SE-100 44 Stockholm, Sweden aakew@kth.se Jan Gulliksen KTH Royal Inst. of Technology SE-100 44 Stockholm, Sweden gulliksen@kth.se Yngve Sundblad KTH Royal Inst. of Technology SE-100 44 Stockholm, Sweden y@kth.se The UsersAward (UA) programme was launched in 1998, initiated by the LO (Swedish Trade Union Confederation) in cooperation with the TCO (Swedish Confederation for Professional Employees) and a group of researchers from KTH (as research coordinator), Uppsala University, Gävle University, and Luleå Technical University. ABSTRACT The goal of the UsersAward (UA) programme is to develop and maintain a strategy for enhancing the quality of workplace software through on-going user-driven quality assessment. Key activities are development of sets of quality criteria, as the USER CERTIFIED 2002 and 2006 instruments, and performing large domain specific user satisfaction surveys building on these quality criteria. In 2005 we performed a first analysis of the values that inform the criteria and procedure making up the 2002 instrument, using the Value Sensitive Design methodology. This paper is a follow-up of that study. We report on new types of stakeholders having engaged with the UA programme and reflect on how the conceptual considerations and explicit values of the programme have shifted as a consequence. The UsersAward programme follows the “Scandinavian tradition” of involving users in IT development for use at workplaces. In the seminal Utopia project in the 1980s the focus was on user involvement in the design and development of the IT support [7]. The investigations and opinion making activities the UA programme has performed since its inception (domain specific user surveys, software certifications, prize competitions) indicate that the users also have to participate in the procurement, deployment, periodic screenings and further development of the software [9]. The motivation for the first analysis of the UA programme from the Value Sensitive Design (VSD) perspective was ”to understand how the principled and systematic approach to designing for human values that Value Sensitive Design offers can further the understanding of when and how different stakeholders can contribute to IT design, development, and deployment in a sustainable and mutually beneficial way” [10]. The aim of this follow-up paper is to report on new participants in, and activities and results of, the UA programme since 2005 and to reflect on how some of the conceptual considerations and explicit values have shifted as a consequence of the programme’s development. Keywords UsersAward programme, user satisfaction surveys, userdriven certification of software, workplace IT, Participatory Design, Value Sensitive Design ACM Classification Keywords H.5.3 Group and Organization Interfaces INTRODUCTION The Human-Computer Interaction (HCI) and Information Systems research communities are moving from studying design and deployment of IT into the workplace to the digitalization of society as a whole. IT is no longer considered a separate component in working life, but one of the driving forces for social and technical innovation. The digitalization process means that a number of work tasks are disappearing [3]. Often it is the physically heavy routine tasks and jobs requiring no or little education that are affected. In the future the demands for highly educated workers will increase and consequently new requirements are put on the working staff, with increased pressure and stress as a consequence. This makes it important to consider a wider spectrum of values affected by the introduction of IT in work and societal contexts. We briefly summarize the UA programme, VSD as a critical screening perspective and the main findings regarding how a better understanding of direct and indirect stakeholder values involved may benefit the future development of the UA programme. After that summary we will report on how we have tried to apply the findings from the initial paper. Since 2005, there has been a growing public awareness of how low quality IT systems negatively affect the work environment in several sectors of society. We argue that this increased awareness provides future challenges and possibilities for HCI research in general and for the UA programme in particular. We conclude the paper by briefly describing how the participation of new stakeholders has resulted in new methods for investigating how software can Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21317 1 benefit both user self-direction and the economic health of organisations using the software. investigated workplaces meet the required level of agreement with the 29 statements. If successful a detailed protocol is published, with quotes from the users’ free-form comments on all pertinent issues. The 2006 revision of the instrument was prompted by an interest from German unions and researchers, who, after having replicated surveys based on the UC 2002 instrument [6], argued for an adaptation of some of the quality criteria and more realistic levels of acceptance in respect to the difference between German and Scandinavian work organisation and team autonomy. In 2005, three software packages had received the USER CERTIFIED 2002 label. Between 2005 and 2009 three more packages passed the user satisfaction screenings. Furthermore, two of the packages applied for, and passed, new USER CERTIFIED screenings in 2007 and 2009 respectively. THE USERSAWARD PROGRAMME Between 1998 and 2011 the UA programme developed and performed a unique combination of industry wide user surveys, user conferences, targeted design projects, yearly IT Prize contests, and, as the main research challenge, a Users’ certification instrument and procedure for workplace software packages [5]. The 2005 VSD screening of the programme focused on the certification instrument [10]. We briefly outline the rationale and procedure for the USER CERTIFIED 2002 and 2006 instruments. The rationale behind the User Certified instrument The inspiration for the UA programme was the successful certification programme for display units, TCO’92, launched by TCO in 1992 through a broad cooperation between TCO, researchers and environmental organisations. This certification programme had been regularly upgraded (TCO’95, TCO’99, TCO CERTIFIED) and had by 2002 put its label on more than 200 million display units worldwide, [2]. VALUE SENSITIVE DESIGN Value Sensitive Design (VSD) is a theoretically grounded approach to design of technology that seeks to account for human values in a principled and comprehensive manner throughout the design process [4]. The approach is today an acknowledged approach in HCI. Here is a brief summary. LO wanted the UA initiative to develop a similar model for the workplace software market. The goal was to develop a method that was user-driven, both in the sense that it was initiated by, and developed in cooperation with, Sweden's two largest employee organizations (2+1.3 million members), and in the sense that the certificate each software package received was based on end-users from at least two different workplaces who, after having operated the software for more than nine months, had given it their seal of approval. In VSD values are viewed neither as inscribed into technology nor as simply transmitted by social forces. People and social systems affect technological development, and technologies shape individual behaviour and social systems. It is a tripartite methodology, consisting of conceptual, empirical, and technical investigations, which are applied iteratively and integratively. Conceptual investigations comprise philosophically informed analyses of the central constructs and issues under investigation. Empirical investigations focus on the human response to the technical artefact, and on the larger social context in which the technology is situated, using methods from the social sciences. Technical investigations focus on the design and performance of the technology itself. Applying the USER CERTIFIED 2002 and 2006 instruments We briefly summarize the USER CERTIFIED instrument, its design and our experience from applying it in eight certification processes. For more comprehensive descriptions see [5, 9]. A key aspect of Value Sensitive Design is its focus on both direct and indirect stakeholders. The direct stakeholders include the users of the system in question, the system developers, and the managers of both users and developers. The indirect stakeholders are not direct users of the system, but nevertheless affected by the system, either positively or negatively. For example, the direct stakeholders for a hospital scheduling system might be doctors, nurses, and other hospital personnel. Two important classes of indirect stakeholders would be the patients and their families – even though they don’t use the system directly, they are strongly affected by it. The assessment procedure starts by asking a software provider, who applies for certification, to fill out a selfdeclaration regarding the software and its intended use and to suggest three workplaces where the user satisfaction of the package can be assessed. The main activity is the set of interviews and questionnaire surveys that the evaluation team carries out at the three workplaces. 29 quality criteria in the certification questionnaire are presented as statements to be confirmed on a value scale from 1 (total dismissal) to 6 (total agreement). At each of the three workplaces, three end-users and three representatives from management are interviewed separately, based on the questionnaire. Also 10% (or at least 10) of the software’s end-users answer a user version of the questionnaire. RESULTS FROM THE 2005 VSD SCREENING Furthermore, VSD makes a distinction between explicitly supported values (i.e., ones that we explicitly adopt and support throughout the design of the system), and stakeholder values (i.e., ones that are important to some but not necessarily all of the diverse stakeholders for that system). In the 2005 screening we found that “the principal explicitly supported values of the UsersAward programme The questionnaire covers six areas (totally 29 statements): Overall benefits, Deployment method, Technical features, Worktask support, Communicative support, Local assessment. The users are considered satisfied as a whole, and a certificate is issued, when at least two of the 2 itself are transparency and fairness: transparency, because we want the process by which software packages are certified to be open and understandable; and fairness, because we want the certification assessment to be made in an unbiased manner.” [10] Unionen, has made surveys in 2008, 2010, 2012, 2013 and 2015 [8]. And in 2013 the union for civil and municipal servants, Vision, made their first survey. These surveys are based on UA surveys of industry, health care and banks performed in 2002, 2004, and 2005 respectively, surveys which in turn were based on the UA certification values and criteria [5]. For the evaluated systems, the screening found that “the values the programme attempts to foster are all related to human welfare and human flourishing. They include: competency development for the individual, the team, and the organization as a whole (in particular opportunities for exploration and learning); enhanced degree of self-direction for individual workers and teams; supporting flexible, selfdirected communication within and between work teams; and economic health of the organization using the system.” VSD calls on the investigators to consider indirect as well as direct stakeholders, and harms in addition to benefits. The 2005 screening of the UA programme resulted in a recommendation “that the certification instrument articulates in a more systematic way questions about who are the indirect as well as the direct stakeholders, and about the harms as well as benefits of the system for the different stakeholder groups.” [10] The empirical investigations proposed in the 2005 screening aim at clarifying “how well the UsersAward programme supports the values of transparency and fairness [and] how well it fosters the values listed above in the systems being evaluated – and whether there are other values that should be added to the list, or values that should be clarified or subsumed.” [10] Related to increased activism of individual unions was a shift of ownership of the UA programme. As of this writing, TCO Development has just acquired the programme, methods and databases. With its 20 years experience of environmental labelling of IT hardware (http://tcodevelopment.com) a new challenge for the UA research and development now is to complement the TCO Certified label designed to support the values of hardware ergonomics and the social responsibility of IT manufacturers with the UsersAward label for IT workplace software usability. This, also, calls for a reorientation of the UA research. Are UA’s declared explicit values of 2005 still relevant? For the UA programme as a whole, we interpret an increased union activism and a more resourceful owner as benefitting an increased awareness of concrete usability problems at Swedish workplaces. Other UA initiatives we regard as supporting the transparency and fairness of the programme are the pilot study mentioned above [1], and two international research workshops arranged around the concept of User-driven IT Design and Quality Assurance (UITQ), which led to, among other things, two replications of UA certification studies by a German research group [5]. CONTEXTUAL CHANGES AND RESEARCH SINCE 2005 The 2005 VSD screening exercise was very important for our own understanding of the UA activities. We were fortunate to have Alan Borning, one of the most knowledgeable proponents of the approach, as co-author. However, the closest we have been to a renewed meta level screening is a small (109 respondents) survey study carried out in 2009 in order to gauge the needs of different stakeholders interested in the programme and responding to the survey – buying companies, user organisations, public agencies, municipalities, IT providers, IT consultants and universities [1]. Although the study was not done along the VSD lines of thought it resulted in some reconsiderations of the UA values which we will briefly mention below. In the absence of a systematic VSD analysis the following points of self-assessment can be seen as preparation for such a study when we (or, preferably, external reviewers) get the means of conducting one. First, we will account for two major changes in the environment of the UA programme since 2005. Then we will try to pinpoint what this has meant for the adherence to UA’s explicit values in the UA practice and to our reconsiderations of these values. The growing series of union reports, which typically cover a randomised sample of about 2 000 users responding to about 30 statements, clearly show that users experience no or only a slow increase in the explicit values in systems that UA promote – competency development, self-direction, and economic health of the organisation [8]. The most alarming results concern the lack of self-direction – users are not allowed to participate in the deployment of new systems in the four areas problematized in the surveys: how work tasks should be transformed, how functionality should be prioritised, and how education and training should be delivered in order to have the systems support the new ways of working [9]. Under the heading “Four out of ten think that the IT-systems are difficult to use” the Union 2015 report concludes: “Here is a potential for cost reductions for the employers and reduced stress and frustration for the users.” [8] A further indication of the relevance of the values declared in 2005 is the fact that the Swedish health promoting agency Prevent, jointly owned and managed by the employer and employee organisations, has initiated a programme against what they term “TechnoPanic” (http://www.prevent.se/TeknikPanik). This we interpret as a sign that the direct interest of organised users with respect to self-direction and competence is shared by the employer organisations with their more direct interest in the Changes in the UA programme environment The first major change regarding the context for the UA programme is that the individual trade unions have started to perform IT user satisfaction surveys of their own memberships. The biggest union for white collar workers, 3 economic health of the member companies they serve through the Prevent health programme. Development’s focus on challenging hardware and software providers to live up to high quality standards has underlined the importance of investigating how investments in usability may result in better work environment as well as a sustained or increased level of economic health for the organisations that use good quality software. RECONSIDERING EXPLICIT VALUES The survey study in 2009 shed some light on one of the questions raised in 2005 – which other stakeholders should be involved in user-driven IT quality assurance? The results indicated that both user organisations and universities ranked higher as trustworthy institutions than public agencies and private companies [1]. The non-profit company TCO Development (TCO-D, new owner of UA) represents a special case, as it is in turn owned by a trade union confederation. But more than that, the clients of TCO-D are not only the federated unions and their membership. TCO-D also addresses management representatives of hardware and software providers directly with certification services. This in turn moves the economic health of companies, specifically through more selfdirection for the users, to the forefront as an explicit value for a renewed UA programme. ACKNOWLEDGMENTS This research was supported in part by Swedish Agency for Innovation Systems (VINNOVA). REFERENCES 1. Bengtsson, L., Ivarsen, O., Lind, T., Olve, N., Sandblad, B., Sundblad, Y., Walldius, Å. (2009). Nya initiativ för användbarhetsarbetet en pejling av behovet hos UsersAwards olika interessentgrupper. TR TRITACSC-MDI 2009 2. Boivie, P.E. (2007). Global standard – how computer displays worldwide got the TCO logo, Premiss (2007). 3. Frey, C. B. & M. A. Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” Oxford University manuscript, 2013. Anticipating the TCO-D future involvement in UA, the UA research 2009-2011 focused on understanding economic impacts of managing usability. A new format for analysing and presenting results from certification was developed [5]. The instrument applies strategy maps, a tool developed by Kaplan & Norton for facilitating iterative, collaborative deliberation among stakeholders on strategic issues. As of this writing, we have only used usability strategy maps in certification protocols and in minor field studies. The most visible foregrounding of usability efforts contributing to economic health is the new survey question “How much time would you save per day if the IT-systems worked the way you wanted?” This question has yielded stable member estimations of around 25 minutes per day, resulting in assessments of overall potential savings of more than 10000 MSEK a year for the Unionen membership alone [8]. Through collaboration within the Nordic eHealth Network (NeRN), mandated by the Nordic Council of Ministers, this question has in turn been taken up in similar user satisfaction surveys in some of the other Nordic countries [11]. 4. Friedman, B. (ed.) Human Values and the Design of Computer Technology, Cambridge University Press and CSLI, New York, NY and Stanford, CA, 1995. 5. Ivarsen O., Lind T., Olve N.-G., Sandblad B., Sundblad Y. & Walldius Å. (2011). Slutrapport UsersAward2 – utvecklad kvalitetssäkring av IT-användning, MID, CSC, KTH. 6. Prümper, J., Vöhringer-Kuhn, T. & Hurtienne, J.: UsersAward – First results of a Pilot Study in Germany, In Proceedings of the UITQ 2005 workshop. 7. Sundblad, Y.: UTOPIA - Participatory Design from Scandinavia to the World, in Proceedings History of Nordic Computing 3, pp. 176-186. Springer Verlag, Heidelberg (2010). 8. Unionen (2015). Tjänstemännens IT-miljö 2014 – Lyssna på oss som ska använda det”, Unionen. 9. Walldius, Å., Sundblad, Y., Sandblad, B., Bengtsson, L., Gulliksen, J. (2009). User certification of Workplace Software – Assessing both Artefact and Usage, BIT (Behaviour & Information Technology), vol.28, no.2, pp.101-120. (2009) CONCLUSION VSD is usually employed in guiding the design of individual information systems. In the preliminary screening of the UA programme in 2005, we used VSD to inform the design of a programme intended to impact the design of computer systems – in other words, we were working one level removed from the design of the individual IT system. In this revisit of the UA programme, we conclude that the programme’s explicit values are the same as 2005, although there has been a shift towards understanding the mutual interplay between user selfdirection and economic values of IT usage. TCO 10. Walldius, Å., Sundblad, Y., Borning, A. (2005). A First Analysis of the UsersAward Programme from a Value Sensitive Design Perspective, in Bertelsen, O., N., O., Bouvin, P., G., Krogh and M., Kyng (eds.), Proceedings of the Fourth Decennial Aarhus Conference August 2024, 2005, pp. 199-202 11.www.norden.org/en/publications/publikationer/2013-522 4 On Creating and Sustaining Alternatives: The case of Danish Telehealth Morten Kyng Aarhus University & The Alexandra Institute Aabogade 34; 8200 Aarhus N; Denmark mkyng@cs.au.dk platform, called OpenTele. OpenTele is already the most used telehealth platform in Denmark and in January 2015 the five Danish regions, who own the public hospitals, decided to increase its use over the next few years. ABSTRACT This paper presents and discusses an initiative aimed at creating direct and long lasting influence on the use and development of telemedicine and telehealth by healthcare professionals, patients and citizens. The initiative draws on ideas, insights, and lessons learned from Participatory Design (PD) as well as from innovation theory and software ecosystems. Last, but not least, the ongoing debate on public finances/economy versus tax evasion by major private companies has been an important element in shaping the vision and creating support for the initiative. This vision is about democratic control, about structures for sustaining such control beyond initial design and implementation and about continued development through Participatory Design projects. We see the “middle element”, the structures for sustaining democratic control beyond initial design and implementation as the most important and novel contribution of the paper. 4S began as an effort, initiated by researchers, to get the most out of a set of IT infrastructure tools, primarily in terms of benefits for public health service providers and private vendors. 4S then developed into what it is today: an organization facilitating direct and long lasting influence by healthcare professionals, patients and citizens. The key people involved in the development of 4S are researchers in Participatory Design (PD), IT managers from the public hospital sector, and – for the last year – doctors and other healthcare personnel working at public hospitals. In recent months people from patients organizations have joined the initiative, and they expect to play a key role in the future. For the PD researchers 4S was originally an initiative quite separate from their work in PD. That work was about democracy, mainly in terms of doing design in ways that allowed participating users to contribute in ways that made the outcome more attuned to their interest. 4S was about national infrastructure, open source software tools and software ecosystems. However, as it turned out, 4S could be developing into a very successful mechanism for sustaining democratic user influence. Author Keywords Contexts for design; sustaining results; control; Participatory Design; software eco-systems; innovation; Open Source; telehealth, telemedicine, healthcare technology. ACM Classification Keywords D.2.10 Design, Methodologies, D.2.m Miscellaneous H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. One might say that Scandinavian Participatory Design (PD) began as an effort towards extending concerns for workplace democracy to cover IT [2], and gradually, over more than four decades, focused more and more on techniques for PD, see e.g. [3-5]. In contrast, 4S began as an effort to get the most out of a set of IT infrastructure tools, primarily in terms of benefits for private vendors and public health service providers [6]. From there, 4S developed into an organization facilitating and sustaining direct influence on the use and development of telehealth by healthcare professionals, patients and citizens. In doing so 4S has added a new type of context for PD in Denmark. INTRODUCTION This paper is about how users of technology may achieve direct and long lasting influence on the IT-systems they use. The case presented and discussed is telemedicine and telehealth in Denmark. The main mechanism for achieving this influence is a foundation controlled by healthcare professionals, patients, citizens, public healthcare organizations and researchers. The foundation, called 4S [1], controls two sets of open source software: an ITinfrastructure toolbox and a telemedicine/telehealth1 In our view, the case of 4S, especially how 4S changed over time, presents important lessons for those who want to develop and sustain critical, democratic alternatives, and for those who want to contribute to the continued developments of Participatory Design (PD) – especially those who are interested in the relations between PD and its larger setting Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark 1 From now on we just use the term telehealth, and take it to include telemedicine. DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21297 5 in society. Therefore we describe the development of 4S in quite some detail. This description involves quite a few elements: organizations, projects and software. To assist the reader these are listed below: providers, e.g. hospitals and general practitioners, and across systems from different vendors. And through such sharing make it simpler for the health service providers to create better services. The aim of 4S was to support expanding use of the toolbox after the end of the Net4Care project, i.e. to sustain the results of the project. Looking at other similar projects, both research projects and public, open source projects, we believed this to be a major challenge: The results of research projects, like Palcom [7], are not being used nearly as much as intended and thus not developed after the end of the project. Or public open source create animosity due to lack of good documentation, lack of tutorials, and/or a flawed governing structure [6]. Regions Denmark has five regions whose main task is to administer the public hospitals Municipalities There are 98 municipalities. Their main tasks include administration of prevention, elder care, e.g. nursing homes, and public schools. OpenTele An open source telemedicine/-health platform developed by three of the five Danish regions and used in two large national projects. The Second Idea: Open Source Telemedicine Platform The Net4Care infrastructure toolbox, and the architecture on which it was based, had many merits [6], but it was difficult to create sufficient interest among health service providers and vendors. To them Net4Care was just another research project without backing from strong public or private players. Thus we were heading down the road towards “not being used nearly as much as intended”. Two things changed this. First we suggested that 4S should also have governance of an open source telehealth platform, called OpenTele. OpenTele was being developed by three of the five Danish regions as part of two large telehealth projects. And the regions saw 4S as a way to increase the likelihood of sustaining OpenTele after the end of these two projects. The inclusion of OpenTele, as was our intension, dramatically increased the interest in 4S among both these three regions and the healthcare professionals using the OpenTele system. Secondly, the public authority responsible for health IT, The National eHealth Authority, published a so-called reference architecture for telehealth [8]. The Net4Care infrastructure toolbox was fully aligned with this reference architecture. Obviously this dramatically increased the interest in 4S among health IT companies that wanted to understand and use the reference architecture. Net4Care A research project that developed a proposal for a national infrastructure for telehealth. Net4Care is also the name of an infrastructure toolbox developed by the project. 4S A foundation with members from the three regions that developed OpenTele, a municipality, the state level and from OpenTele users and research. The researchers also participated in the Net4Care project. 4S governs two sets of open source software: the Net4Care infrastructure toolbox and the OpenTele telehealth platform 4S has a board, a coordinator, a software group and healthcare forums. The rest of the paper is organized as follows: First we present 4S: how it began, the steps toward what it is today, and what 4S is currently working with. Secondly, we look at why 4S might work, i.e. what are the characteristics of telehealth and the societal context that make it possible for an organization like 4S to succeed. Then, in the Discussion section, we look at 4S in relation to ongoing debates in PD on democracy and sustaining results – and we briefly reflect on why it took so long to develop the 4S initiative. We conclude the paper by a short discussion of what we have learned from 4S. Then, in June 2013, the 4S foundation outlined above was formally established, and became copyright holder of the two sets of open source software: the OpenTele platform for telehealth and the Net4Care infrastructure toolbox. Implementing 4S In the following months we focused on four areas: interactions with stakeholders, developing the role of the board, involvement of healthcare professionals and patients, and handling governance of the open source software. THE 4S FOUNDATION – AND HOW IT WAS DEVELOPED The First Idea: Shared, Public Infrastructure Developing the blueprint for the 4S organization began in 2011. The idea was to create an organization to promote and further develop an open source IT-infrastructure toolbox. The toolbox was developed as part of a research project on national infrastructures for telehealth and named Net4Care after the project [6]. The aim of the toolbox was to make it easier for an IT-company to effectively and safely support sharing of health data across health service Interactions with main stakeholders: Private IT-companies and users of OpenTele The first task that 4S focused on was creating links with private IT-companies that were potential users of the infrastructure toolbox. To this end we set up an experiment to investigate and illustrate the effectiveness of the toolbox. This resulted in a public demonstration of sharing data across systems of different types, e.g. homecare telehealth 6 and electronic patient records, and from different vendors at a conference in 2013 [9]. One of the results was quite strong support from The Confederation of Danish Industry. Our second task was creating links with the users of OpenTele. This proved to be more difficult than expected, primarily because the two projects using OpenTele were behind schedule, and for that reason were closely followed by their funding agencies. The top-level project managers did not see dialogue with 4S as justifiable amidst difficulties in catching up, and thus did not want to facilitate cooperation between project participants and 4S. However, one of the sub-projects, the one on complicated pregnancies, decided that cooperation was worthwhile and over the next two years this cooperation became the catalyst for the current high involvement of healthcare professionals and patients in 4S. Figure 1 and 2 illustrates being monitored at the hospital and at home for a woman with a complicated pregnancy. Being monitored at the hospital typically means hospitalization for eight to ten weeks. This is one of the reasons why many prefer tele-monitoring. Figure 2. Monitoring at home. A major task that the board wanted to work with was the creation of synergy in the continued development of OpenTele. This development was distributed across the two large projects and one of these projects consisted of five almost independent sub-projects. Thus on the one hand coordination and creating synergy might be difficult. On the other hand, the composition of the board made the members believe that they could orchestrate this. As it turned out, the “project logic” overruled most of the efforts aimed at creating synergy. This project logic emerged form time schedules so tight that coordination often was impossible and that analysis of alternatives never happened. On the positive side, the problems in creating synergy have increased the willingness on behalf of the regions to give more power to 4S in the development of plans for the future of the OpenTele platform. Developing the role of the board: From project focus to national goals The board of 4S consisted of IT-managers from the three regions developing OpenTele, a representative from the organizations that developed the Net4Care infrastructure toolbox, a medical representative from the organizations that used OpenTele, and two representatives from the national level. Shortly after the formalization of 4S an ITmanager from a large municipality joined the board. A PD researcher from the organizations that developed the infrastructure toolbox had the role of CEO on a part time basis. Involvement of healthcare professionals and patients Democratic control by healthcare professionals and patients/citizens was not on the original agenda of 4S, but it was the intention to use 4S to supplement the current influence by managers and IT developers with input from users. 4S chose to make this visible through the organizational structure, which consists of the abovementioned board, a software group and healthcare forums. Especially the forum on complicated pregnancies was an active participant in fieldwork-based evaluations and in organizing workshops with pregnant women and healthcare professionals to generate suggestions for improvements of OpenTele, cf. Figure 3 below. This forum was also responsible for organizing the first national workshop for healthcare professionals on the use of telehealth for complicated pregnancies. Figure 1. Monitoring at the hospital. 7 revised modules for the OpenTele platform, e.g. a module supporting the use of smart watches or a better visualization of different health data. However, nothing happened, and in the spring of 2014 only the company that developed the original OpenTele solution was developing on OpenTele. Therefore 4S decided that the research group behind the infrastructure toolbox should step in and develop a small number of new modules as well as some modifications to existing modules in close cooperation with healthcare professionals and patients, cf. Figure 4 below. Figure 3. Workshop with pregnant women, and women who had recently given birth, to generate suggestions for improvements of OpenTele. Through this we learned a lot about what was needed to facilitate a multi-vendor strategy and – even more important – we realized that 4S had the potential to become an organizational frame for democratic influence in the area of telehealth through the control of the open source platform OpenTele and the possibilities for continued development of the platform based on Participatory Design (PD). Briefly stated we advocate PD as a way to develop great contributions with respect to both IT and organization. We advocate 4S as a way of keeping control over the software and its development. And we advocate open source as a way of creating competition among different solutions and providers – ranging from research groups to private companies – in a way that will enhance quality and keep the costs down. Governance and handling of the open source software The last area that we worked with concerned all the traditional stuff related to governance of open source software, such as setting up a code repository and developing processes for quality control when modifying or adding software, including bug-tracking. It also included documentation, tutorials and discussion forums. In addition we wanted to develop standard contracts to be used by healthcare providers when contracting work based on the 4S software. When we look at the ability to realize these goals there is a marked difference between the infrastructure toolbox and the OpenTele platform. With respect to the toolbox all the relevant goals have been met. This is primarily due to the fact that the group behind the formation of 4S also developed the toolbox and that this group actively promotes 4S and involves others in the continued development of the toolbox. This includes involving both private companies and national organizations working with telehealth. In addition no tight timetables are currently involved. Supporting a Societal Trend: The Next Tasks, Opportunities and Challenges In January 2015 the five Danish regions decided to implement infrastructure for telehealth within two years, and to do so based on the national reference architecture, thus increasing the importance of the 4S infrastructure toolbox. Secondly, they decided that the main elements of the OpenTele platform should form the basis for telehealth solutions. These decisions open new possibilities of influence for 4S. At the same time the decisions imply that several different actors will want to exercise influence on 4S. Thus those interested in the democratic aspects of 4S should step up their efforts in order not to be marginalized. When IT comes to OpenTele, the situation is – for several reasons – very different. First of all the OpenTele software uploaded to the 4S code repository has so far been several steps behind the current versions used by the two telehealth projects. The three regions paying for OpenTele want 4S to have the most recent code. However, transfer of real governance – as opposed to formal governance, described in documents – and transfer of code from the company originally developing OpenTele to 4S was never part of the “project logic” mentioned above. Thus the regions have so far failed to allocated the necessary resources and “decision power” to speed up this transfer. And the private company responsible for the versions used by the telehealth projects has currently no real incentives to change this. The situation is, however, not satisfactory, and the 4S board have recently decided to set up a timetable for finishing the transfer. Figure 4. Workshop with health personnel and a pregnant women to generate suggestions for improvements in the use of OpenTele at the hospital and improvements to OpenTele itself. Realizing the Democratic Potential It was the intension of all 4S board members that a number of different companies should begin to develop new or 8 First of all, as part of this increased focus on promoting “democracy and PD”, the key people behind the democratic aspects of 4S, i.e. PD researchers and healthcare professionals, realized that “creating great solutions via PD” and “keeping control over software via 4S” wasn’t enough to create the interest and engagement that was needed. Based on our discussion with different healthcare professionals working at public hospitals that experienced continual budget cuts, we realized that many saw 4S as a way of “fighting back”. If not literally, then in terms of providing solutions and opportunities for development that wasn’t depending on private companies and their profits. So this became the new story that made 4S attractive to many healthcare workers. And – as discussed in the next section – a story that was aligned with a growing societal trend of “fighting back” when tax cuts and public budget cuts were presented by governments as the only way forward. WHY 4S MIGHT SUCCEED In this section we first look at telehealth using different disciplinary perspectives that we find to be important in understanding the possibilities open to 4S. These perspectives include innovation, software ecosystems, open source, public finances, and Participatory Design. Based on this understanding we then look at the different stakeholders in telehealth in Denmark and in 4S. We discuss their interests and how different aspect of 4S may make them positive, neutral or negative. Innovation Telehealth, including telemedicine, has been around for many years. However, there are no widely used set-ups, no widely recognized good business cases and no big players with a strong, successful product, cf. e.g. [10-12]. Thus we may characterize telehealth as an emerging area in the sense that there are no so called dominant designs [13]. We may also say that telehealth is in the fluid stage. In this stage different actors experiment with new functionalities, user involvement is common and solutions are often flexible but ineffective [14]. To make an analogy from the area of products, we may look at the tablet computer. In 1989 GRiD Systems released the fist commercially available tablet-type portable computer [15]. But the fluid stage continued for more than two decades and only ended when Apple introduced the iPad, which quickly established itself as the leader among a few dominant designs. During the fluid stage the basic concepts or designs change quite often, new players have a reasonable chance of success and cooperation with users increases the chance of success [13]. Drawing on C. Christens notion of disruptive technology [16] we also observe that new players have especially good chances of success in the cases where they develop new types of solutions to a new market and these solutions only later become interesting in one or more big, established markets. Thus in the area of telehealth, a small new organization like 4S may have a much better chance than a similar effort directed towards say Electronic Patient Records, an established area dominated by multi billion dollar companies like Epic Systems Corporation [17]. And if solutions in the telehealth area develop in ways that over time make them attractive also in relation to other kinds of hospital-based healthcare services, then theories like those of C. Christensen tell us that small players like 4S have a reasonable chance of entering the bigger market of say hospital IT systems. As part of developing this aspect of 4S we have begun to orchestrate critical debate, sharing of experiences and ideas among healthcare professionals, patients, citizens, public healthcare organizations and researchers. In addition 4S facilitate cooperation among them – also when they do not agree. As it turned out, the decision by the five regions to implement infrastructure for telehealth caused major disputes to develop with the municipalities that are responsible for public home care. In this situation, 4S succeeded in organizing a series of workshops and meetings to develop a compromise. However, looking at the current options for the above-mentioned groups, we find that especially patients and citizens as well as healthcare professionals need new venues and opportunities. 4S has begun doing this through our health forums, cf. [1], and to our knowledge 4S is the only Danish organization facilitating this kind of activity. In addition to the activities organized by the health forums we have just begun the process of having a representative from a patient organization join the board. Secondly, 4S is developing its role as an active partner in creation of PD projects within telehealth aimed at adding to and improving the 4S software. However, the explicit focus on PD is new and we only have a few results so far. Thirdly, 4S continues to improve its open source code repositories and processes for accessing and delivering code. These processes facilitate the transfer of results of successful, relevant PD projects to the open source software controlled by 4S. This includes documentation, quality control and CE Marking according to e.g. the Medical Device Directive of the European Union. Software ecosystems Finally, 4S continue to provide and improve resources that facilitate use of the 4S software. The plan for the next 18 months includes texts on telehealth for patients and healthcare professionals, material on open source business models for customers and suppliers, as well as blue prints for contracts covering development, maintenance and operation. Our first ideas on creating 4S to promote and further develop an open source IT-infrastructure toolbox was subsequently rephrased, analyzed, and further developed using concepts from the area of software ecosystems. The research field of software ecosystems has emerged as the study of complex interactions between software frameworks and software architectures on one hand, and 9 organizations, users, customers, developers, and businesses on the other. Based on a recent literature review [18] we define a ‘software ecosystem’ as: “the interaction of a set of actors on top of a common technological platform that results in a number of software solutions or services”. Further, ‘‘Each actor is motivated by a set of interests or business models and connected to the rest of the actors and the ecosystem as a whole with symbiotic relationships, while, the technological platform is structured in a way that allows the involvement and contribution of the different actors.’’ Well-known examples of software ecosystems are the Apple iOS and Google Android ecosystems respectively. The telehealth ecosystem in Denmark is a very different story, but all the same we were able to use the software ecosystem concepts to analyze the status quo, identify the lack of connectivity as a key issue, and to argue for the possibility to improve the ecosystem through the creation of 4S and an active guardianship of the Net4Care infrastructure toolbox. We also learned that 4S should pay specific attention to the “interests or business models” of all major actors. Regarding the infrastructure toolbox itself, this worked quite well. The toolbox was seen as a useful, and it did not compete with any existing products. However, our inclusion of the telehealth platform OpenTele, challenged our relationship with many private vendors, which viewed OpenTele as a direct competitor to their closed source solutions. Subsequently we have worked directly on developing business models that illustrate how different kinds of vendors may benefit from both the infrastructure toolbox and the telehealth platform. As of mid 2015 several of the main, private actors of different kinds are positive or neutral towards 4S including the OpenTele platform as illustrated by a recent press release from the Danish Regions co-signed by the Confederation of Danish Industry [19]. For more on 4S in a software ecosystem perspective se [6]. important because this made it possible for the company to develop new versions for new markets without having to negotiate with any owners of the software. Looking back, the decision to use open source has proven to be very important in the development of support for the 4S vision of democratic control and PD. This is discussed below. Public finances and tax evasion The healthcare systems around the world are under increased pressure: populations are growing, the percentage of elderly and old people is increasing, and the number of possible treatments is growing, as are the costs of medicine. In addition the expected level of public service is growing in countries like Denmark. As a result current models for financing public health are in general considered not to be sustainable, see e.g. [20]. In this situation many governments look to telehealth as a way to keep expenditure under control. However, so far such expectations have not been validated, se e.g. [10, 11]. The most common conclusion seems to be that telehealth is not one specific thing, and that the way different elements are put together and integrated in context is what results in bad, neutral or god outcomes for the different stakeholders [21]. Following this line of thought we find that 4S offers two important possibilities: First of all the infrastructure toolbox and the OpenTele platform make it possible to tailor many different kinds of telehealth solutions, and to integrate these with other types of health IT systems in many different ways. Secondly, since both the toolbox and OpenTele are open source, it is possible to use and modify them in experiments, and then – based on the 4S processes for quality control – transfer results into the 4S code base and use them in large-scale operation. In addition, the debate on financing the modern well-fare state has changed in recent years. For many people it is no longer only a question of how to control public spending, but also a question of how to counteract unfair rules and regulations as well as systematic tax evasion by big international corporations – tax evasion on a scale that threatens national coherence through extreme inequality, cf. [22] and, for a discussion of these issues related to innovation, [23]. As it turned out, this changed public climate makes it possible to tell the story about 4S as a way of providing a democratic alternative to big corporations with flawed tax morale. In many cases, where we have discussed the possibility to use OpenTele with hospital personnel, this aspect of 4S has played an important role. Especially in situations where hospital budgets are being reduced, the idea of having a non-profit, public organization like 4S governing open source software is very attractive to many hospital employees. Open Source Open source was chosen for both the infrastructure toolbox and the OpenTele platform early on, i.e. prior to the formation of 4S. With respect to the toolbox is was a natural choice for modules that we intended to be used in as many telehealth – and other health-IT – systems as possible. It was also an approach we had used earlier in EU research projects developing infrastructure tools cf. [7]. Concerning the OpenTele platform, the decision to use open source was made by the three regions financing the development. One reason for this decision was their experiences with licensebased payment models for some commercial telemedicine systems. While these models were acceptable in small to medium sized pilots, they presented – in the view of the regions – an unacceptable price tag when scaling up. Open source was seen as a way of getting rid of the license fees. In addition, open source was seen as a way to realize a multi vendor strategy that would support both better quality and competitive prices. For the company developing the first versions of OpenTele, open source was probably Participatory Design Participatory Design has, since is emergence in the early 1970s, been seen as a vehicle for democratization of tech- 10 nology, although views has varied quite a bit on how successful PD has been and are in this respect, see. e.g. [24-26] teaching material for the unions was an important element. In addition it turned out that existing system development methods were inadequate for this endeavor, and thus – by necessity – the researchers in the Utopia project began to develop tools and techniques for PD [31]. In relation to our discussion of 4S and why it might succeed we’ll first look at the issue of visions and then consider partners and results. Looking at this list of results, it is not surprising that the trade unions was a very strong partner in the first two generations of project. They co-created the venues for debate and dissemination, and they were partners in a comprehensive educational effort based on the trade union educational programs for shop stewards. An example on this is the final report from the Utopia projects. It was written for graphical workers, printed in 70.000 copies and distributed by the Nordic unions to all their members. Visions Participatory Design, especially in the Scandinavian varieties, had from its formation in the early 1970s and the following two decades a strong and appealing vision based on work place democracy and more broadly democratization of technology. If we look at the activities and publications from this period it is obvious that this vision played a very important role when people chose to work with PD, see e.g. the volumes of the Scandinavian Journal of Information Systems cited above [24-26]. Since the late 1990s focus in many, probably in most, PD projects has been on how best to involve people (users/potential users) in the development of ICT in ways that allow them to promote their own interests in relation to the ICT being developed. Focus is often on the project level, on PD tools and processes, and it seems that focus on structures outside the individual projects has decreased, see e.g. [5, 24-26, 32]. From the 1990s and on “work place democracy” faded and only the more abstract “democratizat¬ion of technology” remained. This change in vision meant that PD changed from being part of something that was important for trade unions and their members to something important for those who were involved in a PD project, and to PD researchers. One might characterize this as a change from political to academic. This development is also reflected in [27], where Shapiro argues that PD theoretically has the potential to produce better systems. This is a very important point for PD researches, but we have found it difficult to use this kind of argument to create engagement among patients and healthcare personnel. For example we found it difficult to convince users that PD is much better than agile development methods like say SCRUM [28]. At least two kinds of arguments have been used as explanations: the changing role of trade unions and “normalization” of use of technology [4]. The first argument assumes that the weakening role of trade unions over the last decades makes it difficult for PD researchers and/or trade unions to continue the kind of cooperation that the first and second generation of projects represented. In our view this is not substantiated: trade unions still have comprehensive course programs, they engage in struggles for organizing labor of different nationalities, including tracing opaque ownership relations, debates on the impact of globalization etc. We find it more likely that “normalization” plays an important role. Normalization refers to the changes in the use of technology at the workplace that have occurred since the early 1970s. Back then IT was new to the shop floor and the first generation of projects played an important role in helping the unions to understand potential implications of IT and how to deal with them. The vision of 4S – democratic control, structures for sustaining such control, and continued development through Participatory Design projects – has created strong engagement from healthcare professionals and patients, much stronger than the abstract “democratization of technology.” Partners and results In Scandinavia the first generation of projects (e.g. The Norwegian Iron and Metal workers project [29] focused on how local trade unions might influence managementcontrolled development and deployment of IT systems at the individual workplace in order to promote worker interests [2, 4]. An important part of this work consisted of developing courses for local union people, including the teaching material used. Another element was national dataagreements that regulated the way management introduced ICT systems [3, 4]. When we consider the Utopia project, it took place at a time when information technology was about to dramatically change the newspaper industry – and once more this was a project that played an important role in helping the trade unions understand the potential implications and to develop a design of a system that represented an alternative to the major commercial products being developed. During the 1990s and in the following decade the use of IT has become business as usual in many industries, and in our In the second generation of projects (e.g. The UTOPIA project [30]) focus was on worker-controlled design of IT systems from a trade union perspective, both to generate demands when negotiating with management and to form the basis for development of specific systems by commercial companies. And once more, developing 11 view, this is a primary reason why the projects of the first generations are difficult to “re-invent”2. line of reasoning Concrete Consensus, and describe it in the following way: When we look at 4S in this light we see that a strong partner like the unions is missing. However, 4S is addressing an area where the use of technology is emerging and about to dramatically change treatment, rehabilitation and prevention. Thus 4S is producing a number of results, which makes 4S very relevant to large numbers of healthcare professionals, citizens and citizen organizations as well as to patients and patient organizations. Thus we find that 4S has the potential to expand its role as a part of the discourse on telehealth among these groups and to facilitate development of e.g. courses on telehealth for home care nurses and different patient groups. In addition 4S is currently facilitating a small, but growing number of PD projects aimed at improving telehealth, for patients/citizens and healthcare professionals, as well as for the public healthcare providers. The results of these projects are sustained by transferring the software to the 4S OpenTele code base and by making the relevant documentation, organizational models, course material etc. available through 4S. • The system and organization of work/activities will have positive effects for both users/employees and for buyers/employers (representing the owners of the organization using the system). • It is possible to find or establish one or more companies that will implement and market the system. • A market exists for the system, i.e. there are companies/individuals who will buy it. The Utopia project designed a system that probably fulfilled the first bullet, failed with the second, and thus did not provide real information on the third [30]. In our analysis of the possibilities of 4S to succeed in providing substantial improvements for patients/citizens and healthcare personnel, we use the above notion of concrete consensus as a way of understanding how the interests of patients/citizens and healthcare personnel may co-exist with the interests of other stakeholders. The users The primary users are patients/citizens and healthcare personnel. Both these groups are experiencing major changes affecting their daily life or work when telehealth is introduced. However, the consequences vary depending on the solution in question. Currently no obviously good solutions covering both IT and organization exist, but both groups, and especially the patients, give the current OpenTele solutions high rankings. The groups may influence the development of OpenTele through 4S, both via the board and via PD projects. Concerning the 4S PD projects we see the same possibilities and pitfalls as in other PD projects in the healthcare area. These include asymmetric power relations and often the lack of a strong network and other resources on the patient side, se e.g. [3336]. When it comes to the board we expect patient/citizen organizations to play a major role, since the individual patient or citizen do not have the resources, and since these organizations already are actively involved in promoting patient/citizen interests in relation to health and welfare technology, including telehealth. Thus it will be the responsibility of these organizations to safeguard against patients/citizens becoming hostages for other interests. With respect to healthcare personnel we are not sure how they will utilize the board. Currently the healthcare personnel participate based on their involvement in telehealth projects and as telehealth experts recognized by their peers. At the moment the professional organizations, e.g. the Danish Medical Association, are not actively involved in the area of telehealth. Thus the healthcare personnel currently involved in 4S also see 4S as a way of developing policies on behalf of their respective professional organizations. The Stakeholders in Telemedicine As the next part of analyzing why 4S might be a success we consider the different stakeholders and their interests in 4S. In doing so we also summarize the interests of each group across the disciplinary discussion in the previous subsections. However, to set the scene we first look at some of the key ideas and assumptions from the first generations of Scandinavian PD projects. First of all we note that the first Scandinavian PD projects, like the Iron and Metal Workers project [29] and the Utopia project [30] aimed at improving work by changing the way computers were used at the workplace. To do so they focused on cooperation between researchers and trade unions. However, to achieve the improvements aimed at, the idea was that the trade unions would subsequently use the knowledge and other results from the projects to negotiate such improvements with management and owners, both locally and centrally. In [5], Kyng labels this 2 Early examples on suggestions for where to look for new possibilities for PD include working in developing countries and with non-profit organizations. Both these lines of work have resulted in new projects, but not on a scale like the first generations of Scandinavian projects. Recently several new forms of extremely individualized labor are emerging in e.g. programming, translation and transportation. It has, in discussions among PD researchers, been suggested that PD researchers should look into making a new kind of PD-project to help understand what is going on and how to counteract negative consequences. To our knowledge such a project hasn’t been initiated yet. 12 OpenTele platform, this is being accepted as a fact of life, i.e. as a decision made by the regions, not by 4S. In this situation many private companies see 4S as making positive contributions, e.g. in terms of suggesting business models for different types of companies and developing processes for accessing and delivering code that make use of the OpenTele platform attractive for a number of different types of companies. The providers of health services The hospitals and other providers of health services in Denmark are expected to introduce telehealth in the next few years and to reduce costs and increase quality by doing so. However, as previously mentioned, there are no hard data supporting that they can achieve this. Thus many of the providers of health services see a need for building more knowledge and developing and modifying telehealth solutions based on experiences. And for these reasons they engage in 4S. At the same time they acknowledge that the users of telehealth should play a key role in this, and therefore welcome the emphasis that 4S place on the influence of there groups. In summary, 4S promotes OpenTele solutions, but aim for co-existence with other solutions, not for confrontation. Actually 4S actively facilitate co-existence based on datasharing and integration through the infrastructure toolbox. And in doing so it seems that 4S is creating support from most of the stakeholders and that only a rather small minority is now negative towards 4S. In the future we will see differences between the different user groups and between these groups and the managers representing the providers of health services when it comes to specific solutions. So far we believe that the possibilities offered to the users by 4S, will make it easier for them to promote their own interests – also, and maybe especially, in the face of disagreements with management. This being said, quite a few people, both in the public and the private sector, are skeptical towards open source, and we expect this skepticism to continue to influence also how they view 4S. Finally, it should be noted that the vision of 4S means different things to different people. To many, and probably to the most enthusiastic supporters, it is important that 4S provides a democratic alternative to big corporations with flawed morale concerning tax. To others, 4S is important as a way of sustaining results of PD projects. And to managers at hospitals and regional IT departments, 4S is a way of combining development of cost-effective solutions with user involvement and acceptance. The public authorities financing health services In Denmark, the Ministry of Health and the municipalities finance health services. As mentioned they expect telehealth to deliver substantial savings, but they too know that there are no hard data supporting this. At the same time, they consider the economic models for the primary, commercial telehealth solutions to be too expensive when scaling up to the national level. Thus, for the same reasons as the providers of health services, they are positive towards 4S. They also see 4S as way to develop business models and cost structures that are beneficial for the different public stakeholders. At the same time they expect the 4S infrastructure toolbox to speed up data sharing and the integration of telehealth solutions with other types of health IT and thus pave the way for better economy in telehealth. DISCUSSION 4S today is emerging as an organization that sustains democratic control by healthcare professionals and patients/citizens. However, that 4S ended up in this way is at least as surprising to the researchers who took the initiative to create 4S as it was to the researchers creating the Utopia project, that the main heritage of that project probably is contributions to PD tools and techniques. In the following we discuss these two issues: Fist we look at sustaining democratic control and related work in PD. Then we briefly discuss the issue of surprise, and we do so under the heading “Why this took a long time”. Suppliers of health IT The last group of stakeholders we consider are companies providing health IT. Several of these companies have made substantial investments in the area of telehealth and many of them find that the public customers should just buy their solutions and get on with the job of providing telehealth services3. However, this is not really happening. The markets for telehealth are small and fragmented, and most of the customers postpone purchases. In this situation the 4S infrastructure toolbox is seen as a way to integrate proprietary solutions from different vendors and thus support growing market integration. In addition the contributions from 4S to building knowledge among public customers is also seen as a way to speed up their decisions on buying telehealth solutions. When it comes to the On Participatory Design and sustaining democracy Following the first generations of Scandinavian PD projects, PD research has focused primarily on PD tools and techniques, see e.g. [3, 5, 32]. The PDC 2012 proceedings, [32] does contain a session on values and politics, ibid pp. 29-40. However, of the three papers in the session two are about methods and techniques and the third is about how children with profound disabilities can participate in formative design processes. All three papers represent interesting and valid contributions, but they are not related to politics in the sense of the first Scandinavian PD projects. From time to time this development results in discussions of how to reinforce or even reinstate ideals on 3 Personal communication with managers from several big, international IT companies from 2008 to the present. 13 democracy and political action. This was discussed e.g. by Kyng in [4], where he pointed to “normalization” of the use of IT at workplaces as one of the reasons of declining cooperation with Trade Unions in PD. In [3] Bødker et al list a number of changes in technology, management, unions and research which they all see as contributing to the shift from political to pragmatic. They also explain the huge success of PD as largely depending on participation being a basic epistemological (knowledge theoretical) principle. At the same time, they are not willing to let go of the political aspects of the past. They discuss the “consumer movement” as one promising possibility, and mention the computer equipment certification by TCO, the Swedish Confederation of professional employees, see also [37]. The most vivid debate of the development of PD research is probably the 2002 and 2003 issues of the Scandinavian Journal of Information Systems, where the titles of the papers include “P for Political”, “D for Democracy”, and “A for Alternatives”. In “P for Political” [38], it is argued that “PD must encompass work motivated in political conscience, …not only participatory design.” An almost alternative view is presented by Kyng in [5], who argues that the politics in PD relies on external factors, e.g. Trade Unions. When we then look at where 4S may be positioned is this debate, we find that the 4S initiative has developed into being political, in the sense that it facilitates democratic control over IT solutions by users and their organizations. However, the mechanism for this control, the 4S organization, is not PD or a result of PD, but an “external factor” that uses PD to develop solutions in the interests of users. And 4S in turn, is itself dependent on the growing societal trend of “fighting back” when tax cuts and public budget cuts are presented by governments as the only way forward. that developed a system for hospital operation room logistics. The system received very positive evaluations by health care personnel [40]. Among other things they pointed to improved work environment. Later a privately owned company was formed to turn the system into a product, and that company was recently sold for almost 20 million dollars. The current web-site for the company emphasizes improved productivity, and do not mention work environment [41]. Why this took a long time Looking back at the history of 4S, we note that it took us a long time to move from the initial ideas on 4S to the current vision and organization. And when we began to think about this, we found it especially surprising that it took us so long to make the connection between 4S and PD. However, today we believe that this delay is understandable. First of all we note that it is common knowledge that people tend to stick proven ways of doing things for (too) long, se also [13]. One might say that the success and tradition of PD research made us blind to other ways of doing than continuing current PD research with its focus on PD methods, techniques and tools – and on the concept of a project as the frame for doing things. Secondly, we note that we were equally slow in understanding and developing the democratic potential of 4S. In retrospect one might say that we were not looking for democracy, and thus we did not see it for a very long time. In addition, the researchers involved in the creation of 4S, adhere to the notion that politics in PD relies mainly on external factors. Thus we did not want to push for the “fighting back” part of the 4S vision, but to develop it together with key players among the users. What about sustaining results? If this understanding will help others to move faster and more effectively when creating alternatives and mechanisms to sustain them we are not quite sure. However, a simple lesson for PD researchers could be that there is more to democracy and politics than PD and projects, a subject that we expand a bit upon in the section below. A different line of reflection in PD is represented by the investigations into sustaining results of PD projects. The paper by Iversen and Dindler [39] reviews a number of papers on sustaining PD initiatives and identifies four different forms of sustaining: maintaining, scaling, replicating, and evolving. They go on to discuss how to increase sustainability based on the more nuanced understanding that the different forms represent. The paper concludes, “in general, the sustainability perspective does not fundamentally change the PD toolbox, but may influence the choice of method, how they are modified and when they are used.” p. 166. We agree with the main points presented in the paper. However, we note that the paper is about sustaining of the results of PD projects. And we find that the options discussed for improving sustainability of results are framed by the notion of a project. In contrast to this approach, the 4S initiative sets up a new, potentially permanent, organization to sustain results – and to frame PD projects. The 4S approach also solves another issue related to sustaining results: that of keeping control of results over time. The issue is illustrated by a PD project CONCLUSIONS The story of 4S is quite specific: IT is about telehealth in Denmark and how an organization set up to govern open source software developed into a frame for sustaining democratic control and PD projects. On the other hand we find that the story illustrates some key points in the discussions on PD and democracy. First of all 4S illustrates how PD may become part of a societal trend, and that the specific aspects of that trend are important for the outcome. In our case fighting back when tax cuts and public budget cuts are presented by governments as the only way forward. 14 Secondly, 4S illustrates how proper support structures for the participants/users and their “group interests” are important and at the same time that developing these structures are not, or at least need not be, part of the results of PD, at lest not explicit, consciously executed PD. REFERENCES 1. 2. 3. Thirdly, we find that 4S illustrates that “Democracy” or “Democratic control” often are too general categories to create enthusiasm and engagement, there has to be a manifestation of a more concrete vision, ‘problem’ and ‘solution’. 4. When we turn to the possibilities for doing something analogue in other countries we are mildly positive: in countries with sufficiently similar political systems and health systems, it seems worthwhile to investigate if and how cooperation between 4S and matching organizations might be feasible. On a very practical level, we are talking with the Greek organization “SciFY – Open Source software/hardware for social benefit” and the Faculty of medicine, University of Thessaly, about using OpenTele to deliver telemedicine in small villages in rural areas where access to medical assistance is difficult. 5. 6. 7. When it comes to possibilities for doing something analogue in another specific area, a first step could be to carry out an analysis of that area, similar to the one presented in the section Why 4S might succeed. And then, based on this initial understanding, develop a more targeted analysis, tailored to the specifics of the area in question. 8. FUTURE WORK IN THE PD RESEARCH ARENA 10. 9. Up until now the 4S initiative have been about creating and developing the 4S organization and using “state-of-the art” PD within this framework. However, we expect that the future activities of 4S will provide numerous opportunities for PD research. Issues we would like to address include PD processes for going from successful research prototype via clinical trials to tested and certified software. And from small-scale pilot to large-scale daily operation. Furthermore we see a strong need for PD to cooperate with other disciplines to encompass work on business models, e.g. for starts-ups working with the 4S software, and on the socalled business cases used by e.g. public providers of health services to evaluate new service proposals. Up until now these business cases have mainly been using quite standard economic concepts and models. Hopefully PD can contribute to the development of supplementary concepts and models. 11. 12. 13. 14. 15. 16. ACKNOWLEDGMENTS We thank all the people involved in the 4S initiative, and especially the healthcare personnel and the pregnant women in the sub-project on complicated pregnancies at Aarhus University Hospital. In addition we thank the reviewers for their comments. 17. 18. 19. 15 4S webpage. Available from: http://4S-online.dk. Ehn, P. and M. Kyng, The Collective Resource Approach to Systems Design, in Computers and Democracy. 1987, Avebury: Aldershot, UK. p. 17-57. Bødker, S., et al., Co-operative Design—perspectives on 20 years with ‘the Scandinavian IT Design Model’, in Proceedings of NordiCHI 2000. 2000: Stockholm, Sweden. p. 22-24. Kyng, M., Users and computers: A contextual approach to design of computer artifacts. Scandinavian Journal of Information Systems, 1998. 10(1&2): p. 7-44. Kyng, M., Bridging the Gap Between Politics and Techniques: On the next practices of participatory design. Scandinavian Journal of Information Systems, 2010. 22(1): p. 49-68. Christensen, H.B., et al., Analysis and design of software ecosystem architectures – Towards the 4S telemedicine ecosystem. Special issue on Software Ecosystems, 2014. 56(11): p. 1476-1492. PalCom project webpage. Available from: http://www.ist-palcom.org/. Statens Serum, I. and I.T. National Sundheds, Referencearkitektur for opsamling af helbredsdata hos borgeren. 2013, National Sundheds-IT: København. The UNIK partnership conference. 2013; Available from: http://www.partnerskabetunik.dk/media/38318/opdate ret invitation.pdf. Steventon, A., et al., Effect of telehealth on use of secondary care and mortality: findings from the Whole System Demonstrator cluster randomised trial. BMJ (Clinical research ed.), 2012. 344: p. e3874. EU project Renewing Health. Available from: http://www.renewinghealth.eu/en/. the Telecare Nord project on COPD. Available from: http://www.rn.dk/Sundhed/Til-sundhedsfaglige-ogsamarbejdspartnere/TeleCare-Nord. Tidd, J. and J. Bessant, Managing innovation : integrating technological, market and organizational change. Vol. 4th. 2009, New York: Wiley. 638. Abernathy, W. and J. Utterback, Patterns of industrial innovation. MIT Technology Review, 1978(6): p. 4047. The BYTE awards : GRID System's GRiDPad. BYTE Magazine, 1990. 15(1): p. 285-286. Christensen, C.M., The Innovator's Dilemma : When New Technologies Cause Great Firms to Fail. 1997, Harward: Harvars Business Press. 225. Epic Systems Corporation. Available from: http://en.wikipedia.org/wiki/Epic_Systems. Manikas, K. and K.M. Hansen, Software ecosystems – A systematic literature review. Journal of Systems and Software, 2013. 86(5): p. 1294-1306. Fælles vej frem for telemedicin til glæde for både patienter og erhvervsliv. 02.02.2015; Available from: 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. http://www.regioner.dk/aktuelt/nyheder/2015/februar/f %C3%A6lles+vej+frem+for+telemedicin+til+gl%C3 %A6de+for+b%C3%A5de+patienter+og+erhvervsliv. Porter, M. June 6, 2006, Mandag Morgen: Copenhagen, Denmark. Goodwin, N., Jumping the gun in the telehealth steeplechase. 2012. Piketty, T., Capital in the Twenty-First Century. 2014: Harvard University Press. Mazzucato, M., The entrepreneurial state : debunking public vs. private sector myths. Anthem frontiers of global political economy. 2013, London: Anthem Press. 237. Scandinavian J of Information Systems, 1998. 10. Scandinavian J of Information Systems, 2003. 15. Scandinavian J of Information Systems, 2010. 22. Shapiro, D., Participatory design: the will to succeed, in Proceedings of the 4th decennial conference on Critical computing: between sense and sensibility. 2005, ACM: Aarhus, Denmark. p. 29-38. Schwaber, K. Controlled Chaos : Living on the Edge 1996. Nygaard, K., The ‘Iron and metal project’: trade union participation, in Computers Dividing Man and Work – Recent Scandinavian Research on Planning and Computers from a Trade Union Perspective., Å. Sandberg, Editor. 1979, Swedish Center for Working Life, Demos Project report no. 13, Utbildningsproduktion: Malmö, Sweden. p. 94-107. Bødker, S., et al., A Utopian Experience. Proceedings of the 1986 Conference on Computers and Democracy, 1987: p. 251-278. Ehn, P. and M. Kyng, Cardboard Computers: Mocking-it-up or Hands-on the Future, in Design at work: Cooperative design of computer systems, J. Greenbaum and M. Kyng, Editors. 1991, Lawrence Erlbaum Associates, Inc.: Hillsdale, New Jersey, USA. p. 169-195. 32. Proceedings of PDC 2012, The Participatory Design Conference. 2012: Roskilde, Denmark. 33. Grönvall, E. and M. Kyng, Beyond Utopia: reflections on participatory design in home-based healthcare with weak users, in Proceedings of the 29th Annual European Conference on Cognitive Ergonomics. 2011, ACM: Rostock, Germany. p. 189-196. 34. Christensen, L.R. and E. Grönvall, Challenges and Opportunities for Collaborative Technologies for Home Care Work, in ECSCW 2011: Proceedings of the 12th European Conference on Computer Supported Cooperative Work, 24-28 September 2011, Aarhus Denmark, S. Bødker, et al., Editors. 2011, Springer London. p. 61-80. 35. Fitzpatrick, G. and G. Ellingsen, A Review of 25 Years of CSCW Research in Healthcare: Contributions, Challenges and Future Agendas. Computer Supported Cooperative Work (CSCW), 2012: p. 1-57. 36. Bertelsen, W.O., et al., Therapeutic Strategies - a Challenge for User Involvement in Design, in Workshop at NordiChi 2010, The 6th Nordic Conference on Human-Computer Interaction. 2010: Reykjavik, Iceland. 37. Gulliksen, J., et al., Key principles for user-centred systems design. Behaviour & Information Technology, 2003. 22(6): p. 397-409. 38. Beck, E.E., P for Political: Participation is Not Enough. Scandinavian Journal of Information Systems, 2002. 14(1): p. 77-92. 39. Iversen, O.S. and C. Dindler, Sustaining participatory design initiatives. CoDesign: International Journal of CoCreation in Design and the Arts, 2014. 10(3-4): p. 153-170. 40. iHospital. Det Interaktive Hospital. 2005; Available from: http://www.ihospital.dk/. 41. Cetrea. Clinical Logostics. 2015; Available from: http://www.cetrea.com/index.php/en/. 16 Computing and the Common. Hints of a new utopia in Participatory Design Maurizio Teli Department of Information Engineering and Computer Science, University of Trento Trento, Italy maurizio.teli@unitn.it K.4.0 Computers and Society General. In order to explain how a new utopia for PD can take place, this statement is organized as follows. Firstly, I will begin with a preliminary discussion of critical theories of contemporary societies, including an understanding of the role of digital technologies and the common. Secondly, I will present the recent design contributions on computing and the commons. Both will pave the way for a theoretical clarification of the difference between the commons and the common. Thirdly, I will discuss the social and institutional conditions enabling a common-oriented utopia in PD. In conclusion, I will connect critical theory, design contributions, and the social and institutional conditions, to sketch out future directions based on my proposal. INTRODUCTION A SHORT SUMMARY OF CRITICAL THEORY TODAY ABSTRACT In this statement, I draw upon the need of Participatory Design to engage with new utopias. I point to contemporary critical theories and to concurrent social conditions that make possible to identify the construction of the common as a possible utopia. In conclusion, I suggest that forms of community-based participatory design could be actual practices supporting such utopia. Author Keywords Participatory design; commons; common; utopia. ACM Classification Keywords During the Aarhus conference of 2005, Dan Shapiro [21] elaborated on Participatory Design [PD] as a political movement in a phase of economic stability and growth. It is common knowledge that since then things have dramatically changed, due to the effect of “The Great Recession”, the economic crisis that has been haunting the Western countries during the past eight years. According to Ehn [8], PD originated in a specific social context characterized by strong trade unions as the main social ally of PD. In the light of the current societal situation, I propose an update of the politics of PD along similar lines. Such update is needed because the economic crisis has pointed out that “business as usual” is not a viable practice from the point of view of the aspirations toward a just and sustainable society. In fact, the crisis has shown the societal limitations of the steady accumulation of capital and it has made evident the need for forms of renewal of the bases on which current societies are tied together [12]. In this changing context, PDers could contribute to the shaping of contemporary societies by collectively engaging in the definition of new utopias that could inform their actions. In this respect, both keynote speakers at the XIII Participatory Design Conference, Pelle Ehn and Shaowen Bardzell, articulated reflections on the roles of past and future utopias for research and practice in PD. Their reflections opened up a space for new directions in utopian thinking in computing. While the pioneers of PD where siding with the workers in the factory system, nowadays PD practitioners are called to an updated understanding of the context and the identification of new social allies (something Dearden et al. tried to do in Aarhus 2005 focusing on agencies promoting emancipation [6]). Critical analyses of the current situation provide a useful lens to reposition the politics of PD and here I present three influential analyses that are looking at the computing-society relationship. This statement takes on the challenging task of engaging with the construction of a possible utopia oriented to the common. The construction of such utopia is possible nowadays by leveraging on two concurrent social conditions: the existence of a group of highly-skilled precarious workers; and the existence of institutional opportunities for scholars in the design discipline to connect that group with the common through research funding. Sociologist Christian Fuchs focuses on the notion of “mode of production” as developed by Karl Marx. Specifically, he summarizes the complex of working activities bringing to the production of digital technologies through the concept of digital labor [9]. The interesting part of Fuchs's update of Marxist theory is the stress on how, in the different places of digital production (from the African mines to Facebook), relations of production involve ownership, coercion, allocation/distribution, and the division of labor. Moreover, he distinguishes between work and labor, the former being the activity of transformation of the world, while the latter Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21318 17 being the commodification of work through the job relation. This distinction is something particularly significant in the age of social media and other digital technologies, in which work is algorithmically commodified. My proposal for a renewed utopia for PD is to orient PD practices toward the common, being aware of the current mode of production and of the role of netarchical capitalism. That implies the quest for new social allies. Before identifying them, I discuss how PD is already engaging with the commons. Michel Bauwens, a scholar and an activist, helps clarify the way through which contemporary capitalism is articulated. Bauwens contribution is particularly interesting in its definition of contemporary capitalism as “netarchical” [2], defined as the “brand of capital that embraces the peer to peer revolution […] It is the force behind the immanence of peer to peer.” [p. 7]. As the factory was the locus of construction of workers solidarity in industrial capitalism, the netarchically enabled peer-to-peer collaboration in commons-based project can be the place where the prevalence of the commons on the market can be elaborated and politically constructed. For both Bauwens and Fuchs, the possibility of collaborative work is strengthened by contemporary digital technologies but it is also in the digital domain that such collaborative work can become a form of labor and contribute to the accumulation of capital. PARTICIPATORY DESIGN, THE COMMONS, AND THE COMMON Although less oriented to the labor-capital conflict than before [10], in PD there has been a growing attention to the commons, a peculiar institutional arrangement discussed extensively by Elinor Ostrom [17]. The commons are institutional arrangements for managing shared resources that are not based on private nor state property. Specifically, in the digital domain, a commons can be defined through a legal protection that favor the availability to third parties and a form of collective ownership and that entails distributed governance in the management of the interactive artifact. Digital commons (like Wikipedia or Free Software) are characterized by specific organizational traits, such as voluntary participation and contribution. In those institutional settings, many participants contribute without a necessity of doing that (as in a job relation) but out of their commitment to the project [3]. A similar concern is shared by the philosophers Michael Hardt and Toni Negri [11], two of the key thinkers of the stream of Marxism known as Autonomous Marxism (AM). AM relies upon Marx’s “Fragment on the machines”, that stresses how productive forces evolve through the expansion of the “general intellect”, a form of collective and distributed knowledge. As knowledge is a growing part of the production process, the life of the knowledge producers itself is turned into a source of value, independently from the classical labor relation, in what AM defines as the “life theory of value” [16]. In such a perspective, AM shares the preoccupation of Fuchs for infinite exploitation of the social media users, for example, and the opportunities that Bauwens sees in the forms of peer-to-peer collaboration. Another reason that makes AM interesting is the inclusion of authors like Spinoza, Foucault, and Deleuze as inspirational sources. Drawing upon them, Hardt and Negri develop a peculiar anthropology, based on three fundamental elements: the anthropological priority of freedom over power, the latter seen as a containing force; the social priority of the multitude of the poors, whose actions institutionalized powers react to; and the centrality of affect in the development of social life, with the possibility of forms of love and hate to take the stage in historical development. The domain on which freedom, the multitude, and affect operate is that of the common. The common, without an “s”, is intended as the ensemble of the material and symbolic elements that tie together human beings. The perspective of Negri and Hardt has the privilege to enrich the analysis of the political economy, as the one discussed by Fuchs and Bauwens, with a perspective on the subjects of social life, characterized by affects, a desire for freedom, and the capability to act collectively. From a design perspective, pointing to the commons implies not only to understand how Intellectual Property Rights affect design [14], but also to acknowledge that designing a commons entails the social processes of maintenance and governance of a commons, that is a process of commoning [13]. In the PD tradition, the attention to the commons as a specific form of production has been gaining momentum, framing the roles of the users and designers [15], understanding collaborative production in specific places [19], or trying to include in the design process the political implications of Free Software [5]. All this work has proven effective in discussing some of the implications for the design practices of a commons perspective. However, these are still lacking a perspective able to politically scale to the societal dimensions described by Fuchs, Bauwens or Autonomous Marxists. My suggestion is to supplement the framework proposed by Elinor Ostrom, who focused on the actual institutional arrangements related to the management of a specific resource, to locate the specific commons in the wider perspective of Hardt and Negri's accent on the common, as the ensemble of the material and symbolic elements that tie together human beings. In Hardt and Negri's reading, the common can actually be nourished or dispossesed, and the actual forms of capitalism are drawing precisely on forms of accumulation by dispossession, in which value is extracted out of the collaborative capabilities of people. In this context, PD can locate itself as a progressive force by strengthening social practices and social groups that nourish the common, and by identifying both relevant social allies and practical means. Due to the centrality of 18 knowledge in contemporary society, I argue that a relevant ally can be identified in the “The Fifth Estate”, a group composed by highly-skilled precarious workers. Moreover, design projects with this group are possible thanks to specific narratives used by the funding agencies, in particular by the European Union. HORIZON 2020 AND DIGITAL SOCIAL INNOVATION In the early times of PD, the presence of strong trade unions constituted an opportunity for designers to rethink their practices in terms of political positioning. Today the alliance with the Fifth Estate could be similarly challenging. Moreover, it can be fruitfully achieved by leveraging current narratives in the European Union funding strategies. THE FIFTH ESTATE IN SOUTHERN EUROPE Recently, Allegri and Ciccarelli [1] have proposed an analytical category, named the “Fifth Estate”, to include highly-skilled precarious and freelance workers who are fully or partially excluded from accessing welfare security and benefits in Southern European welfare states. In fact, the main research policy instrument in the EU is Horizon 2020 which basic narrative regards the innovation addressing social and environmental issues as the key component for the European development in the globalized world. In this framework, a couple of specific lines of funding are relevant to understand how the research policies of the European Union frame the theme of interest here, that is the commons in the digital world, the “Onlife Initiative” and CAPS (Collective Awareness Platforms for Sustainability and Social Innovation2). In the last decade in Southern Europe that social group has grown and has included an increasing number of university graduates, as a result of the worsening labour market conditions [Eurostat data until 20121]. Looking at the age group 30-34, we see that at the European level there are about 35% of people who completed tertiary education, with a growth of approximately 15% of graduates in the past 15 years. Enrollment in tertiary education is also growing, with an overall difference of about 2.5 million students between 2003 and 2012. Contemporaneously, in Southern European countries the level of employment three years after graduation has dramatically decreased (-14% in Italy and -23% in Spain). In Greece, the country that is worse off, in 2012 only 40% of the university graduates had a job three years after graduation. Moreover, the growth of part-time jobs during the crisis together with a decrease in full-time jobs, suggests that the quality of available jobs is decreasing too. The more relevant is what the EU refers to as CAPS, a very small line of funding, counting approximately 35 million Euros per year. Nevertheless, this line of funding is particularly interesting as it revolves around the collective distribution of social power through the deployment of technologies. On the IEEE Technology and Society Magazine, Fabrizio Sestini [20], the reference person in the European Commission for this line of funding, states that “The ultimate goal is to foster a more sustainable future based on a low-carbon, beyond GDP economy, and a resilient, cooperative democratic community.”. Examples of project already funded through this line include, for example, D-Cent and P2PValue. D-Cent3 is a project working with social movements (like the Spanish Indignados) to build technologies for direct democracy and economic empowerment, as digital social currencies. P2PValue4 focuses on value in peer production, in connection with new forms of cooperative organizations and significant forms of activism. Allegri and Ciccarelli [1] argue that the Fifth Estate could be, in contemporary capitalism, what the Fourth Estate was in industrial capitalism: one of the leading forces capable of articulating new perspectives of wealth distribution. The emergence of the Fifth Estate is characterized by new forms of commons-based practices like bottom-up cooperation, solidarity, and civic cooperation. In fact, the growth of phenomena like co-working spaces (more than 100 only in Italy [18]) and of new funding strategies like crowdfunding (in May 2014 there were 41 active platforms only in Italy! [4]), suggests that the Fifth Estate is actually experimenting with new forms of collaboration. Moreover, this EU narrative frames innovation as “digital social innovation”, a collaborative form of innovation based on the co-creation of knowledge, technologies, and services. The stress on democracy by Sestini and the focus on collaboration in relation to digital social innovation, clearly refer to organizational forms that differ from the traditional bureaucratic organizations or from the networked enterprises typical of the period between the end of the 1990s and the beginning of the 2000s. To summarize, the described EU narrative can actually constitute an opportunity window for common-oriented PD project to be funded and conducted. To put it briefly, the Fifth Estate is engaging in commonsbased forms of association that can nourish the common. Therefore, we can argue that one of the potential allies for contemporary PD is the Fifth Estate, as one of the social groups able to characterize future progressive change (it is not by chance that some commentators are pointing to that social group as the backbone of Syriza in Greece and Podemos in Spain). 2 http://ec.europa.eu/research/participants/portal/desktop/en/opportunities/h2020 /topics/9082-ict-10-2015.html#tab1 3 http://dcentproject.eu/ 1 http://ec.europa.eu/eurostat/web/main/home 4 http://www.p2pvalue.eu/ 19 Participatory Design. Routledge, Oxford, 2012, 182– 209. CONCLUSION: PRACTICING UTOPIA Through an understanding of contemporary capitalism as something providing tensions between forms of social collaboration and of accumulation by dispossession, I articulate a potential update of PD positioning in relation to the social allies PD could talk to and to the practical means to conduct projects. I identify the significant social subject in the “Fifth Estate”, a social group of highly-skilled precarious workers, and the practical means to fund and conduct projects in the European Union drive toward digital social innovation. The ensemble of the discussed theories, design perspectives, and social conditions, is what constitutes a design space oriented to the common. Ehn, P. Scandinavian design: On participation and skill. In P.S. Adler and T.A. Winograd, eds., Usability. Oxford University Press, Inc., 1992, 96–132. 9. Fuchs, C. Digital Labour and Karl Marx. Routledge, 2014. 10. Halskov, K. and Brodersen Hansen, N.. The diversity of participatory design research practice at PDC 2002– 2012. International Journal of Human-Computer Studies 74, 2015 81–92. http://doi.org/10.1016/j.ijhcs.2014.09.003 11. Hardt, M. and Negri, A. Commonwealth. Belknap Press, Cambridge, Mass, 2009. From the point of view of PD practices, community-based PD [7] looks like the most interesting methodological starting point toward common-oriented PD, as it entangles the capability to intercept the diverse and distributed character of the Fifth Estate and it can be easily aligned with the definition of the funding agencies. 12. Harvey, D. Seventeen Contradictions and the End of Capitalism. Oxford University Press, Oxford; New York, 2014. 13. Marttila, S., Botero, A., and Saad-Sulonen, J. Towards commons design in participatory design. Proceedings of the 13th Participatory Design Conference, ACM (2014), 9–12. Summarizing, I provided hints of a new utopia in PD based on the idea of nourishing the common, the ensemble of the material and symbolic elements that tie together human beings. Such utopia can potentially make the PD community a strong actor in the construction of a more just and sustainable society. 14. Marttila, S. and Hyyppä, K. Rights to remember?: how copyrights complicate media design. Proceedings of the 8th Nordic Conference on Human-Computer Interaction: Fun, Fast, Foundational, ACM (2014), 481–490. REFERENCES 1. 8. Allegri, G. and Ciccarelli, R. Il Quinto Stato. Perché il lavoro indipendente è il nostro futuro. Precari, autonomi, free lance per una nuova società. Ponte alle Grazie, Milan, 2013. 2. Bauwens, M. The political economy of peer production. post-autistic economics review, 37 (2006), 33–44. 15. Marttila, S., Nilsson, E.M., and Seravalli, A. Opening Production: Design and Commons. In P. Ehn, E.M. Nilsson and R. Topgaard, eds., Making Futures: Marginal Notes on Innovation, Design, and Democracy. The MIT Press, Cambridge, MA, 2014, 87–98. 3. Benkler, Y. Practical Anarchism Peer Mutualism, Market Power, and the Fallible State. Politics & Society 41, 2 (2013), 213–251. 16. Morini, C. and Fumagalli, A. Life put to work: Towards a life theory of value. Ephemera: theory & politics in organization 10, 3/4 (2010), 234–252. 4. Castrataro, D. and Pais, I. Analisi delle Piattaforme Italiane di Crowdfunding. Italian Crowdfunding Network, 2014. 17. Ostrom, E. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press, 1990. 5. Clement, A., McPhail, B., Smith, K.L., and Ferenbok, J. Probing, Mocking and Prototyping: Participatory Approaches to Identity Infrastructuring. Proceedings of the 12th Participatory Design Conference: Research Papers - Volume 1, ACM (2012), 21–30. 18. Rete Cowo. Tutti gli spazi Cowo, città per città. 2015. http://www.coworkingproject.com/coworkingnetwork/. Dearden, A., Walker, S., and Watts, L. Choosing Friends Carefully: Allies for Critical Computing. Proceedings of the 4th Decennial Conference on Critical Computing: Between Sense and Sensibility, ACM (2005), 133–136. 20. Sestini, F. Collective awareness platforms: Engines for sustainability and ethics. Technology and Society Magazine, IEEE 31, 4 (2012), 54–62. 6. 7. 19. Seravalli, A. Making Commons: attempts at composing prospects in the opening of production. 2014. 21. Shapiro, D. Participatory Design: The Will to Succeed. Proceedings of the 4th Decennial Conference on Critical Computing: Between Sense and Sensibility, ACM (2005), 29–38. DiSalvo, C., Clement, A., and Pipek, V. Participatory design for, with, and by communities. In J. Simonsen and T. Robertson, eds., International Handbook of 20 Concordance: A Critical Participatory Alternative in Healthcare IT Erik Grönvall IT Uni. of Copenhagen Rued Langgaards Vej 7 Copenhagen, Denmark erig@itu.dk Nervo Verdezoto Aarhus University Aabogade 34 D, 8200 Aarhus, Denmark nervo@cs.au.dk Naveen Bagalkot Srishti School of Art, Design and Tech., Bangalore, India naveen@srishti.ac.in Tomas Sokoler IT Uni. of Copenhagen Rued Langgaards Vej 7 Copenhagen, Denmark sokoler@itu.dk collaboration between patients and healthcare professionals in the negotiation of care activities. The HCI field has also gone through a transformation where the perception and role of the ‘user’ and other stakeholders in design has changed. People have gone from being perceived as human factors, to users, to design collaborators, and lately as active participants in design processes that design-for-future-use (rather than design-for-use) [3-5]. Ongoing design processes include a shift away from design-for-use towards designing for design after design. Here, infrastructuring is used within the design literature (see e.g. Ehn [5] and Seravelli [16]), to explore how to design meta-designs, or designing for design after design, and how diverse stakeholders through constantly ongoing design processes design-in-use rather than designing for use before use. ABSTRACT The healthcare sector is undergoing large changes in which technology is given a more active role in both in-clinic and out-of-clinic care. Authoritative healthcare models such as compliance and adherence which relies on asymmetric patient-doctor relationships are being challenged as society, patient roles and care contexts transforms, for example when care activities move into non-clinical contexts. Concordance is an alternative model proposed by the medical field that favours an equal and collaborative patient-doctor relationship in the negotiation of care. Similarly, HCI researchers have applied diverse models of engagement in IT design ranging from authoritative models (e.g. perceiving people as human factors to design for) to more democratic design processes (e.g. Participatory Design). IT design has also been crafted as on-going processes that are integrated parts of everyday use. Based on the best practice of participation from the medical and the HCI fields, we identify critical alternatives for healthcare design. These alternatives highlight opportunities with ongoing design processes in which the design of care regimens and care IT are perceived as one process. Reflecting on the above transformations, and our own experiences working with healthcare IT and design, we describe perspectives on participation originated from both the medical and the HCI field. From the medical field, we take the transformation from compliance and adherence to concordance to enlighten the perspectives on participation. In contrast to compliance and adherence that represent authoritative models of care where the patient is perceived as a passive receiver of care instructions [13], concordancebased care emphasizes an active patient, co-responsible in defining and ensuring their own care plans. From the HCI field, we account for 'ongoing design processes' and 'participatory design' as design ideals. Based on these ideals from the medical (i.e. concordance) and design (i.e. PD and infrastructuring) fields, and in particular their roles in contemporary healthcare IT projects, we show how bridging disciplinary boundaries could offer critical alternatives on how to design for future care and care IT. Author Keywords Concordance; Critical Alternative; Participatory Design; Healthcare; Ongoing design; Infrastructuring ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION During the last decade the doctor-patient relationship has been redefined. It has undergone a transformation from being defined by authoritative care models towards models with a more balanced relationship between care providers and receivers. Information Technology (IT) systems for healthcare and their design processes may either reinforce the authoritative care models or facilitate a more equal PARTICIPATION FROM A HEALTHCARE AND DESIGN PERSPECTIVE Considering the shifting notions of participation within HCI and healthcare, it seems important to examine the roles that various actors may assume in healthcare IT design and use and how participation is shaping current and future care practices. We will now describe participation and how it has been perceived and changed in healthcare as well as in healthcare IT design processes. Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21315 21 Participation in Concordance Healthcare: From Compliance is moved out of the clinical consultations and into the people’s everyday settings and routines, the discussion about patients’ and doctors’ roles and responsibilities become even more critical. to Compliance has traditionally been the prevailing relationship between a doctor and her patients. Compliance measures to what degree a patient follows a prescribed treatment, without considering the patient as a partner in the delivery of care [13]. However, as reported in the medical literature (e.g. [13]), a prescription that a patient disagrees with already at the medical consultation will have less or no value. Therefore a shift from compliance to adherence was initiated within the medical practice. Adherence extends compliance by also considering to what degree a patient agrees with the proposed treatment plan. However, both compliance and adherence have been criticized within the medical field as being too authoritative [9]. The medical field has proposed concordance as a more democratic model to shape the doctor-patient relationship. Concordance emphasizes equal participation and active doctor-patient collaboration at the clinic in both the definition and fulfilment of a care regimen [9]. In doing so, concordance promotes an ongoing negotiation between the patient and healthcare professional through a democratic and open-ended patient-professional collaboration process [9]. As such, concordance involves a political stance where the patient and healthcare professional are considered equal peers in the treatment. Participation Healthcare in Design: Participatory Design in Within HCI, there has been a shift during the last 30 years from perceiving people as human factors to consider them as actors within the design process [3, 7, 10]. Today many HCI practitioners acknowledge the benefits of involving different stakeholders in design processes. Participatory Design (PD) is one design approach based on an active and equal involvement of different stakeholders throughout the design process. Initially, much of the PD work in healthcare focused on initiatives coming from professional settings to support the development process of technology focusing on both patient’s and health professional’s needs to improve clinical practice. For example, developing Electronic Patient Record systems to support clinical consultations and medical practices [18]. In recent years, the increasing move of care from the hospital to the home has required the application of diverse PD methods in non-professional settings to actively involve patients in the design process, uncovering their specific needs in relation to their treatments and their everyday life [7, 12, 24]. In nonprofessional settings, uncovered challenges for patients to perform care activities includes the lack of patient’s motivation to perform the prescribed treatment, the individual mental and physical capabilities, the individual’s perception of technology, the dynamic use context, the aesthetics of the home and of care technology, and the distribution of control over the setting [7]. From a medical perspective, concordance is an activity situated at the clinic, as part of the medical consultation. However, in an earlier work by the authors, it is argued that the implementation of care activities in everyday life situations extends beyond the clinic. The negotiation of care is situated in everyday life contexts and should be considered as an ongoing negotiation [2]. Our previous work showed how care receivers negotiate the integration of prescribed assistive tools into their home environment and the impact of the care regimen on them and their families’ everyday lives. For example, a woman rejected a prescribed stepping board, as she preferred to use her own stairs for her physiotherapy. Another example was a spouse that embedded objects from the home environment into the patient’s rehabilitation activities to support the prescribed exercises. These examples illustrate the patients’ ongoing negotiations with others, objects and the environments to manage their out-of-clinic care. The examples also extend the role of concordance to include more than the in-clinic consultation activities, embracing also the ongoing healthcare practices that take place in everyday life. An ever-growing number of healthcare IT design and research projects have applied participatory methods to engage patients and doctors in designing healthcare IT systems for later deployment and use. Contemporary PD research do not only consider how to design ready-to-use systems but also investigates meta-designs and designing for design after design (i.e. infrastructuring) to allow people to participate in ongoing design processes (design-in-use) [5]. As such, infrastructuring calls for attention to the active role of people as designers, through use, to develop support for practices that could not be envisioned before use [16]. While participatory healthcare IT researchers has eagerly adopted the ideals of active participation of patients and doctors in the design process (e.g., [7, 22]), they have in general not considered the role of patients’ active participation in shaping their own care regimen after the initial design process has ended, nor designed for such participation. In other words, many healthcare IT designs have adopted the medical ideals of compliance and adherence, designing healthcare IT solutions that ensure or support a patient to follow a prescribed regimen (such as tools for remote-monitoring [8, 12, 17] and medical reminders [14, 23]). Herein lays a paradox. It seems that Meanwhile, concordance is by no means a settled debate in the medical discourses. Its focus on changing the responsibilities of the doctors and patients is critical, and it is argued that the consequences of increased patient responsibility for their own care needs further exploration from a societal, moral and ethical perspective. In particular, Segal states that healthcare professionals should intervene if patients with life-threatening or very contagious diseases refuse antibiotics or other treatments [15]. As concordance 22 healthcare IT projects, while driven by methodological approaches that in their very essence build on models that value an active participation of different stakeholders (i.e., patients), are stuck with authoritative de-contextualized models concerning the relationship between doctor, patient, and treatment. Participatory design driven healthcare IT projects do exist that have investigated alternative strategies for care, including a renegotiated patient-physician relation and collaboration [1]. Nevertheless, with only few recent exceptions (e.g. [2]) concordance has not been a design goal in healthcare HCI projects [6]. Indeed, the development in the medical field towards concordance as an alternative to compliance and adherence has largely been overlooked by the healthcare IT design community. Instead, the HCI community has contributed for example with ideas of gamification and persuasive design to support the existing compliance and adherence models [14]. one. These two development-strands represent critical alternatives to the dominant PD practice in healthcare where different stakeholders’ co-design technology for later use under the dominance of compliance and adherence. Critical Alternatives: Concordance ALTERNATIVES Infrastructuring and Both PD and concordance involve a political stance where diverse stakeholders collaborate on democratic and equal terms. In particular, both approaches empower the citizen, being a (future) user of technology or receiver of care, to make an impact on his or her situation through active participation. Indeed, the philosophy behind PD and concordance bear much resemblance and, if thought of, there may not be such a large conceptual difference between co-designing one's own care (i.e. concordance) and co-designing the technology that will support the very same care. Or stated differently, should the design of healthcare IT and the design of one’s care regimen be kept as two separate activities or should they be perceived as one? As the introduction of new technology into a given practice will also change the practice [11], an ongoing design of both IT and care seems favorable. Considering the two strands presented above, strand A may be more straightforward to implement and control (as compared to strand B) as there are two separate phases. Phase one focusing on healthcare technology design, and then in phase two individualized care activities are designed based on the technology from phase one. These two phases are tightly connected but clearly separated. In strand B, both the technology and care design are perceived as one ongoing design process based on the idea of infrastructuring. In strand B new possibilities exists (as compared to strand A) to challenge and inform both the PD and the medical fields. Applying an ongoing design approach (design-in-use) may be a way to discuss participation, facilitating care practices where both care and the care IT are co-designed over time. In sum, concordance and PD in concert can provide an alternative approach for designing healthcare IT, informing and strengthening each other. To critically rethink citizen participation in care and healthcare IT design, and the relationship between designing individualized care regimens and designing healthcare IT to support these regimens, may open up for new strategies to improve the quality of care. Also it may inform both the healthcare and design research fields. CRITICAL PARTICIPATORY HEALTHCARE IT DESIGN PD, FOR As presented above, many PD healthcare IT projects have been informed by a medical perspective but have not considered the unexpected ways of how people appropriate technology in relation to everyday practices [2]. However, there are critical alternatives to consider when designing future healthcare IT solutions within both contemporary Participatory Design and medical research fields. We are referring in particular to, firstly how PD research has investigated ongoing design processes, supported by infrastructuring [5], and secondly how the medical science has turned towards concordance as an alternative to patient compliance and adherence [9]. People negotiate and appropriate care activities and healthcare IT to make them part of their everyday lives [2, 20]. Rather than designing for a specific use, it may therefore be advantageous to support practices of appropriation and negotiation. CONCLUSION In this paper, we have presented critical alternatives for future healthcare designs. The proposed alternatives move away from the authoritarian ideals of compliance and adherence towards a more participatory and collaborative doctor-patient relationship. This move implies that the separation of the design and the use of healthcare IT will dissolve. Through ongoing design processes, supported by concordance and infrastructuring, the act of designing individualized care regimens and designing healthcare IT to support these regimens may become one. The HCI research community should better understand the effects on the involved stakeholders (e.g. patients), not only working with concordance as a design ideal, but also working with ongoing design processes where both care regimens and healthcare technology are designed in concert. However, There are at least two possible developments in which PD research and the notion of concordance may join forces to provide a critical alternative for future healthcare designs. The first strand (A) is based upon two sequential phases. In phase one PD is used to co-design healthcare IT that aim for concordance rather than compliance and adherence (as we have previously argued in [2]). In the second phase (and as an extension to [2]) the design outcome of phase one is used by patients to negotiate and integrate concordance-based care practices into their everyday lives. The second strand (B) is an infrastructuring approach where the design of healthcare IT and the design of a patient’s care regimen are conceptually considered equal and where both the design of IT and care is ongoing and perceived as 23 participation as discussed in this paper will require rather resourceful participants, something that should be taken into consideration in future work. These critical alternatives do not only have the potential to point to new directions for healthcare IT design targeting an individual patient. For example in Europe there is an increased interest on service co-creation within healthcare and patient involvement in designing healthcare services, as seen in national health policies and NGO initiatives [19, 21]. It could prove useful to investigate ongoing design processes as described in this paper as a strategy for future healthcare services. The proposed critical alternatives can inspire designers and researchers to reconsider their design processes when designing out-of-clinic healthcare and technology to better support concordance-based care practices in people’s everyday life. 10. Iversen, O. S., Kanstrup, A. M. and Petersen, M. G., A visit to the 'new Utopia': revitalizing democracy, emancipation and quality in co-operative design, In Proc. NordiCHI 2004, ACM Press(2004), 171-179. 11. Kaptelinin, V. and Bannon, L. J. Interaction Design Beyond the Product: Creating Technology-Enhanced Activity Spaces. Human–Computer Interaction, 2011, 27(3):277-309. 12. Kusk, K., Nielsen, D. B., Thylstrup, T., et al., Feasibility of using a lightweight context-aware system for facilitating reliable home blood pressure selfmeasurements, In Proc. PervasiveHealth 2013, IEEE (2013), 236-239. 13. Moore, K. N. Compliance or Collaboration? the Meaning for the Patient. Nursing Ethics, 1995, 2(1):7177. 14. Oliveira, R. d., Cherubini, M. and Oliver, N., MoviPill: improving medication compliance for elders using a mobile persuasive social game, In Proc. UBICOMP 2010, ACM Press(2010), 251-260. 15. Segal, J. “Compliance” to “Concordance”: A Critical View. Journal of Medical Humanities, 2007, 28(2):8196. 16. Seravalli, A., Infrastructuring for opening production, from participatory design to participatory making?, In Proc. PDC 2012, ACM Press(2012), 53-56. 17. Siek, K., Khan, D., Ross, S., et al. Designing a Personal Health Application for Older Adults to Manage Medications: A Comprehensive Case Study. Journal of Medical Systems, 2011, 35(5):1099-1121. 18. Simonsen, J. and Hertzum, M., A Regional PD Strategy for EPR Systems: Evidence-Based IT Development, In Proc. PDC 2006-Vol II, 2006, 125-228. 19. Spencer, M., Dineen, R. and Phillips, A., Co-producing services – Co-creating health, NHS WALES, 2013. 20. Storni, C. Multiple Forms of Appropriation in SelfMonitoring Technology: Reflections on the Role of Evaluation in Future Self-Care. International Journal of Human-Computer Interaction, 2010, 26(5):537-561. 21. The health fundation, Co-creating Health, from: http://www.health.org.uk/programmes/co-creatinghealth, 2008. 22. Uzor, S., Baillie, L. and Skelton, D., Senior designers: empowering seniors to design enjoyable falls rehabilitation tools, In Proc. CHI 2012, ACM Press(2012), 1179-1188. 23. Vitality - GlowCaps, from: www.vitality.net, 2012. 24. Aarhus, R., Ballegaard, S. and Hansen, T., The eDiary: Bridging home and hospital through healthcare technology, In Proc ECSCW 2009, Springer(2009), 6383. ACKNOWLEDGMENTS We like to thank colleagues from both the medical and HCI fields in formulating the thoughts expressed in this paper. REFERENCES 1. Andersen, T., Bjørn, P., Kensing, F., et al. Designing for collaborative interpretation in telemonitoring: Reintroducing patients as diagnostic agents. International Journal of Medical Informatics, 2011, 80(8):e112-e126. 2. Bagalkot, N. L., Grönvall, E. and Sokoler, T., Concordance: design ideal for facilitating situated negotiations in out-of-clinic healthcare, In Proc. EA CHI '14, ACM Press(2014), 855-864. 3. Bannon, L., From human factors to human actors: the role of psychology and human-computer interaction studies in system design, In Design at work: cooperative design of computer systems, Lawrence Erlbaum Associates, Inc., Hillsdale, New Jersey, USA, 1991. 4. Dantec, C. L. and DiSalvo, C. Infrastructuring and the Formation of Publics in Participatory Design. Social Studies of Science, 2013, 43(2): 241-264. 5. Ehn, P., Participation in design things, In Proc. PDC 2008, ACM Press(2008), 92-101. 6. Fitzpatrick, G. and Ellingsen, G. A Review of 25 Years of CSCW Research in Healthcare: Contributions, Challenges and Future Agendas. Computer Supported Cooperative Work (CSCW), 2012, 22(4-6):609-665. 7. Grönvall, E. and Kyng, M. On participatory design of home-based healthcare. Cognition, Technology & Work, 2013, 15(4):389-401. 8. Grönvall, E. and Verdezoto, N., Beyond SelfMonitoring: Understanding Non-functional Aspects of Home-based Healthcare Technology, In Proc. UbiComp 2013, ACM Press(2013), 587-596. 9. Horne, R., Weinman, J., Barber, N., et al. Concordance, adherence and compliance in medicine taking. London: NCCSDO, 2005. The columns on the last page should be of approximately equal length. Remove these two lines from your final version. 24 Networked Privacy Beyond the Individual: Four Perspectives to ‘Sharing’ Airi Lampinen Mobile Life Centre, Stockholm University Stockholm, Sweden airi@mobilelifecentre.org rise of mainstream social media (for a review, see [8]). Rather than attempting a general discussion of privacy, this paper narrows in on a more limited set of issues by addressing interpersonal aspects of networked privacy. ABSTRACT Synthesizing prior work, this paper provides conceptual grounding for understanding the dialectic of challenges and opportunities that social network sites present to social life. With the help of the framework of interpersonal boundary regulation, this paper casts privacy as something people do, together, instead of depicting it as a characteristic or a possession. I illustrate interpersonal aspects of networked privacy by outlining four perspectives to ‘sharing’. These perspectives call for a rethink of networked privacy beyond an individual’s online endeavors. In the following, I deploy the framework of interpersonal boundary regulation [1] in order to approach privacy as something people do together, i.e. a collaborative activity, instead of depicting it as a characteristic of a piece of content or a possession of an individual. I provide examples of how the negotiation of accessibility and inaccessibility that characterizes social relationships plays out in the context of social network sites by outlining four perspectives to ‘sharing’ that have emerged from prior empirical studies. These perspectives call for a rethink of privacy in our networked age beyond the individual and his/her online endeavors. Author Keywords Privacy; interpersonal boundary regulation; social network site; Facebook; Last.fm; Couchsurfing ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. SOCIAL NETWORK SITES Ellison and boyd [4, p. 158] characterize a social network site as ‘a networked communication platform in which participants (1) have uniquely identifiable profiles that consist of user-supplied content, content provided by other users, and/or system-provided data; (2) can publicly articulate connections that can be viewed and traversed by others; and (3) can consume, produce, and/or interact with streams of user-generated content provided by their connections on the site.’ Nowadays, social network sites are commonly accessed through a variety of devices and platforms, such as applications on mobile devices. People interact with social network sites not only when they access them directly; rather, they may come across integrated features almost anywhere. For instance, Facebook’s ‘Like’ feature, which allows people to indicate with a click their approval and enjoyment of pieces of content, is now nearly ubiquitous on the Web. INTRODUCTION Social network sites (SNSs), such as Facebook, MySpace, LinkedIn, and Instagram, present people with novel opportunities to maintain social ties; craft an online presence; and, as a result, gain access to social validation and meaningful feedback [14]. As these platforms have become part of the everyday life for millions of people, severe concerns have been raised regarding their implications for social interaction and societies at large. This paper provides grounding for understanding the dialectic of challenges and opportunities that social network sites present to daily social life by building on an understanding of networked privacy that has been developed in and around Human–Computer Interaction over the past decade (for early work, see e.g. [10]). It is worth noting that the nature of privacy is a complex and long-standing question that has attracted the attention of scholars in several disciplines already well before the The widespread adoption of social network sites into everyday interaction affects sociality more broadly than just in terms of the activities that take place on these platforms. Ubiquitous access to social network sites and participation in social interaction via them is of course not universal. However, social network sites have become a pervasive part of social life for many and, at least in large parts of the Western world, even those who do not use them in their day-to-day life are embedded in social settings that are shaped by their existence. Typically even those who opt out Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21300 25 of using SNSs have, and indeed must have, an opinion on them. Finally, refusing to join a social network site does not mean that one would not be referred to or featured in the service (for instance, in photos or textual anecdotes that others share). The widespread adoption of SNSs challenges customary mechanisms for regulating interpersonal boundaries. For instance, when it comes to social interaction via these services, it is not reasonable to rely solely on the supportive structures that time and space can provide. In 2003, Palen and Dourish [10] paved the way for ongoing research into interpersonal networked privacy by raising the issue of boundary regulation in an increasingly networked world. While the boundary regulation framework originates in Altman’s considerations of social life in physical settings [1], offline activities have often been bypassed when examining boundary regulation in the context of SNSs. Yet, the social interaction that takes place in SNSs is embedded in the wider fabric of social relationships that surround individuals. Boundary regulation efforts must therefore be understood holistically, since the broad adoption of these platforms weaves them tightly into everyday life, with consequences also to those who can or will not use them. INTERPERSONAL BOUNDARY REGULATION The framework of boundary regulation settings [1] was first introduced by social psychologist Irwin Altman in the 1970s. Over the past decades, it has been developed further, most prominently as Communication Privacy Management theory [11,12]. Interpersonal boundary regulation refers to the negotiation of accessibility and inaccessibility that characterizes social relationships [2]. Interpersonal boundary regulation is a core process of social life: Interpersonal boundaries are constantly regulated through negotiations that draw lines of division between self and others, and ‘us’ versus ‘them’. They are used to structure how and with whom people interact. When successful, interpersonal boundary regulation allows people to come to terms with who they are and how they relate with one another. In contrast, less successful or failing efforts to regulate boundaries are experienced as conflict, confusion, and clashes in expectations, that is, as boundary turbulence [11] that necessitates re-negotiating rules of regulation. Finally, the idea that not just individuals, but groups, too, need to regulate their boundaries was a part of Altman’s [1] original characterization of privacy as boundary regulation. Yet, this notion of group privacy has received scarce scholarly attention within HCI, even though Petronio and colleagues [11,12] have elaborated explicitly on how individuals and groups make decisions over revealing or concealing private information, proposing that individuals develop and use rules collaboratively to control information flows. In contrast to Altman whose work focuses mainly on personal boundaries, Petronio pays particular attention to how people manage collectively co-owned information and collective boundaries, such as those within families and groups of friends. Both Altman and Petronio suggest that individuals need to regulate interpersonal boundaries proactively in order to achieve a desired degree of privacy. As individuals, groups, and societies, we regulate access to social interaction both by how physical spaces are built and through the behaviors that take place in them. Boundary regulation practices abound from making or avoiding eye contact to closing or opening a door. Boundaries of professional and leisure life are negotiated in intricate ways that range from what people wear and eat to with whom, about what, and in what kind of language they converse [9]. These practices are applied to achieve contextually desirable degrees of social interaction as well as to build and sustain relations with others and with the self. FOUR PERSPECTIVES TO ‘SHARING’ This section illustrates different interpersonal aspects of networked privacy by outlining four perspectives to ‘sharing’ that takes place in the context of SNSs. The notion of ‘sharing’ is grounded in prior empirical work on networked privacy [5,6,7,13] where it has emerged as a central way in which people talk about their engagement with SNSs and their efforts to manage what gets published, to whom, and with what kinds of implications. I draw on prior qualitative, interpretative analyses of ongoing sensemaking surrounding social life in the context of SNSs and of the emerging practices for regulating interpersonal boundaries in settings wherein (1) people may share content with multiple groups at once, (2) others may share content on their behalf, (3) sharing can be done via automated mechanisms, and (4) people may share as a group. Instead of treating ‘sharing’ as a monolithic, well-defined activity, I consider and contrast the sharing of manually selected digital content; the streaming of automatically tracked behavioral information; and, acts of sharing that directly challenge the hypothetical online–offline divide. While interpersonal boundaries are not just an analogue of physical demarcations, they are, in part, determined by physical structures. For instance, walls, doors, windows, and furniture shape access to various spaces and to the social interaction that can take place in them. Access to visual and auditory information is limited by what is physically possible [3]: Only people within a restricted geographical radius at any given time can see and hear what is being said and done by co-present others. Those witnessing an incident in an unmediated situation can tell others about it, carrying it further, but there are only so many people who can observe an event first-hand. Sharing one’s experiences and observations with people who are not present requires effort. Moreover, in face-to-face situations, those doing and saying things typically have a fairly good sense of who can see and hear their actions, as they can observe their audience in a shared temporal and spatial setting. Someone could, of course, be eavesdropping out of sight, but having such an unintended audience would be exceptional and violate commonly held rules of decorum. 26 where the main content that is being shared is information about music listening, is an instance of a class of services that allow users to stream behavioral data via automated sharing mechanisms. Automated sharing mechanisms are promoted as a means to make sharing increasingly effortless and authentic in its presumed completeness, but research indicates that much work can go into regulating what is being shared via them [13]. Interpersonal boundary regulation efforts in the presence of automated sharing mechanisms entail decisions that extend beyond choices over sharing per se and affect, instead, how participants behave in the first place: Uski & Lampinen [15] note that while norms concerning manual sharing on Facebook focus mainly on what and how one should share, the norms sanctioning automated sharing on Last.fm target primarily music listening, that is, the ‘actual’ behavior. In other words, when it comes to interpersonal boundary regulation in the presence of automated sharing, there are, first, interpersonal boundary regulation practices that regulate what is publicized, and second, efforts to regulate what to do and, consequently, what kind of data will be available for sharing later on. These efforts show how users’ agency to regulate interpersonal boundaries is restricted neither to adjusting the privacy settings provided within an SNS nor to selecting from among the choices services propose to their users in the user interface. Finally, it is worth noting that users engage in efforts to regulate sharing not only when the sharing mechanism risks publicity of content that they do not want to have in their profiles but also when the sharing mechanism fails to publicize content that it should have and that users want to appear in their profiles [13, 15]. This illustrates well how boundary regulation is not solely a matter of restricting access but also one of providing it. Sharing with Multiple Groups Social network sites may bring about group co-presence, a situation in which many groups important to an individual are simultaneously present in one context and their presence is salient for the individual [6]. Group co-presence makes it difficult to keep traditionally separate social spheres apart from one another. Maintaining a broad social network in an SNS may lead to a sense of having to present oneself and one’s social connections consistently with everyone. In an early study [6], participants reported few social tensions related to group co-presence on Facebook per se, but the authors argue that group co-presence was unproblematic only insofar as it was made unproblematic with the aid of behavioral and mental practices. The challenges related to this tendency of SNSs to flatten diverse audiences into a single group have since been well documented, often under the rubric of context collapse (for an overview, see [16]). A central challenge in sharing with multiple groups at once lies in regulating interpersonal boundaries in a situation where it is hard to keep track of to whom access is provided, or to figure out whether or when these audiences interact with content that is made available to them. Sharing on Behalf of Others While individuals can regulate interpersonal boundaries on their own, ultimately their success always relies on others’ support of their efforts to draw boundaries in a certain way. Interpersonal boundary regulation is best understood as cooperative on two levels: First, cooperation is manifested in continuous, subtle acts of contesting or supporting others’ efforts: maintaining a boundary or making a successful claim to identity requires that others affirm one’s actions and support the definition of oneself that is put forward. Lampinen et al. [7] report on Facebook users reckoning on how and by whom the content that they shared would be viewed and interpreted. Some pondered whether the content they were sharing might inadvertently create challenges for a friend. Participants’ efforts to cooperate included considerate acts of sharing, discretion in self-censorship, and benevolent interpretation. Second, in response to the need to deal with content disclosed by others and with one’s power to share on behalf of others, individuals cooperated occasionally more explicitly, too. This involved coming into agreement on shared codes of conduct through explicit negotiation of what to share, with whom, and under what conditions [7]. While interpersonal boundary regulation was not readily discussed as explicit cooperation, participants placed a strong emphasis on being trustworthy and considerate [7]. Indicating a willingness to take others into consideration and an awareness of reliance on others’ consideration, these participants saw this was not only as something desired of others but also as a standard they strove to live up to. Sharing as a Group Shifting the focus further beyond mainstream social network sites, consider Couchsurfing, a service where members can engage in hospitality exchange by hosting visitors or by staying with others as guests. In a recent study Lampinen [5] examined how multi-person households share accounts and regulate access to their domestic sphere as they welcome ‘couchsurfers’. Profiles on Couchsurfing serve to present two interwoven but distinguishable aspects of households: the domestic space and the people who live in it. The study touches upon group privacy by illustrating how, beyond sharing credentials, account sharing involves complex negotiations over how to present multiple people in a single profile, how to coordinate the households’ outfacing communication and decisions over whether to grant potential visitors access to the home, as well as how to share the benefits of a good reputation in a fair way. Here, sharing as a group entails multi-faceted boundary regulation that takes place not only in online interaction but also in the course of interacting with guests face-to-face. Thus, the study challenges the hypothetical online—offline divide with its focus on the practice of hosting where ’sharing’ online and offline are tightly interwoven. Sharing via Automated Mechanisms Providing a contrast to the aforementioned examples of manual sharing on Facebook, the music-centered Last.fm, 27 4. Ellison, N. B., & boyd, D. M. (2013). Sociality through social network sites. In W. H. Dutton (ed.), The Oxford handbook of Internet studies (pp. 151– 172). Oxford University Press. CONCLUSION Although social network sites have characteristics that disrupt conventional premises of interpersonal boundary regulation, boundaries are regulated through co-operative processes also in their presence. In fact, the use of social network sites may even make the performative nature of social life more visible than is usual or desirable. Increased awareness of the work that goes into achieving appropriate social interaction and sustaining meaningful relationships may feel uncomfortable as it challenges the illusion of the effortless authenticity of everyday socializing. 5. Lampinen, A. (2014). Account Sharing in the Context of Networked Hospitality Exchange. In Proceedings of the 2014 Conference on Computer-Supported Cooperative Work (pp. 499–501). ACM. 6. Lampinen, A., Tamminen, S., & Oulasvirta, A. (2009). All my people right here, right now: Management of group co-presence on a social networking site. In Proceedings of the 2009 International Conference on Supporting Group Work (pp. 281–290). ACM. This paper invites us to consider the limitations inherent in conceiving of privacy management as an individual-level endeavor to do with control over information by synthesizing research that illustrates how people regulate both individual and collective boundaries in cooperation with one other, with consideration to others, and in line with the affordances of different technologies and relevant social norms. When the tools and features of a service create constraints in day-to-day life, people may opt to lessen their engagement with the service [6,7] or to change the way they behave in order to manage what is being shared [7,13]. 7. Lampinen, A., Lehtinen, V., Lehmuskallio A., & Tamminen, S. (2011). We're in it together: Interpersonal management of disclosure in social network services. In Proceedings of the 2011 Annual Conference on Human Factors in Computing Systems (pp. 3217–3226). ACM. 8. Newell, P. B. (1995). Perspectives on privacy. Journal of Environmental Psychology, 15(2), 87–104. 9. Nippert-Eng, C. E. (1996). Home and work: Negotiating boundaries through everyday life. Chicago: University of Chicago Press. To design technologies and policies that are supportive of people’s boundary regulation aims and practices in an increasingly networked world, we first need to understand those pursuits and the reasoning behind them. By highlighting the cooperative nature of these everyday efforts to ‘make the world work’, this paper calls for a rethink of networked privacy beyond the individual and his/her online endeavors. 10. Palen, L., & Dourish, P. (2003). Unpacking privacy for a networked world. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 129-136). ACM. 11. Petronio, S. S. (2002). Boundaries of privacy: Dialectics of disclosure. Albany, NY: State University of New York Press. 12. Petronio, S. (2013). Brief status report on communication privacy management theory. Journal of Family Communication, 13(1), 6-14. ACKNOWLEDGMENTS This work has been supported by the Mobile Life VINN Excellence Centre. I thank my colleagues at Mobile Life, University of Helsinki, Helsinki Institute for Information Technology HIIT, University of California, Berkeley, and Microsoft Research New England. 13. Silfverberg, S., Liikkanen, L.A., & Lampinen, A. (2011). ‘I'll press play, but I won’t listen’: Profile work in a music-focused social network service. In Proceedings of the 2011 Conference on ComputerSupported Cooperative Work (pp. 207–216). ACM. REFERENCES 1. Altman, I. (1975). The environment and social behavior: Privacy, personal space, territory, crowding. Monterey, CA: Brooks/Cole Pub. Co. 14. Stern, S. (2008). Producing sites, exploring identities: Youth online authorship. In D. Buckingham (ed.), Youth, identity, and digital media (pp. 95–117). Cambridge, MA: MIT Press. 2. Altman, I., & Gauvain, M. (1981). A cross-cultural and dialectic analysis of homes. In Spatial representation behavior across the life span: Theory and application. New York: Academic Press. 15. Uski, S., & Lampinen, A. (2014). Social norms and selfpresentation on social network sites: Profile work in action. New Media & Society. 3. boyd, d. (2008). Why youth ♥ social network sites: The role of networked publics in teenage social life. In D. Buckingham (ed.), Youth, identity, and digital media (pp. 119–142). Cambridge, MA: MIT Press. 16. Vitak, J. (2012). The impact of context collapse and privacy on social network site disclosures. Journal of Broadcasting & Electronic Media, 56(4), 451-470. 28 Personal Data: Thinking Inside the Box Amir Chaudhry, Jon Crowcroft, Heidi Howard, Anil Madhavapeddy, Richard Mortier University of Cambridge first.last@cl.cam.ac.uk Hamed Haddadi Queen Mary University of London hamed@eecs.qmul.ac.uk they said they desired (We Want Privacy, but Can’t Stop Sharing, Kate Murphy, New York Times, 2014-10-05). This paradox implies dissatisfaction about what participants received in return for exposing so much about themselves online and yet, “they continued to participate because they were afraid of being left out or judged by others as unplugged and unengaged losers”. This example also indicates the inherently social nature of much “personal” data: it is impractical to withdraw from all online activity just to protect one’s privacy [3]. ABSTRACT We are in a ‘personal data gold rush’ driven by advertising being the primary revenue source for most online companies. These companies accumulate extensive personal data about individuals with minimal concern for us, the subjects of this process. This can cause many harms: privacy infringement, personal and professional embarrassment, restricted access to labour markets, restricted access to best value pricing, and many others. There is a critical need to provide technologies that enable alternative practices, so that individuals can participate in the collection, management and consumption of their personal data. In this paper we discuss the Databox, a personal networked device (and associated services) that collates and mediates access to personal data, allowing us to recover control of our online lives. We hope the Databox is a first step to re-balancing power between us, the data subjects, and the corporations that collect and use our data. Context sensitivity, opacity of data collection and drawn inferences, trade of personal data between third parties and data aggregators, and recent data leaks and privacy infringements all motivate means to engage with and control our personal data portfolios. However, technical constraints that ignore the interests of advertisers and analytics providers and so remove or diminish revenues supporting our “free” services and applications, will fail [12, 25]. A host of other motivations and uses for such a Databox have been presented elsewhere [20, 17, 9]. These include privacyconscious advertising, market research, personal archives, and analytical approaches to mental and physical health by combining data from different sources such as wearable devices, Electronic Health Records, and Online Social Networks. All these examples point to a need for individuals to have tools that allow them to take more explicit control over the collection and usage of their data and the information inferred from their online activities. Author Keywords Personal Data; Computer Systems; Privacy ACM Classification Keywords Information systems applications; Computing platforms THE PERSONAL DATA ECOSYSTEM Many Internet businesses rely on extensive, rich data collected about their users, whether to target advertising effectively or as a product for sale to other parties. The powerful network externalities that exist in a rich dataset collected about a large set of users make it difficult for truly competitive markets to form. A concrete example can be seen in the increasing range and reach of the information collected about us by third-party websites, a space dominated by just two or three players [7]. This dominance has a detrimental effect on the wider ecosystem: online service vendors find themselves at the whim of large platform and API providers, hampering innovation and distorting markets. DATABOX: A USER-CENTRIC ALTERNATIVE To address this need we propose the Databox, enabling individuals to coordinate the collection of their personal data, and to selectively and transiently make those data available for specific purposes. Following the Human-Data Interaction model [18], a Databox assists in provision of: • Legibility: means to inspect and reflect on “our” data, to understand what is being collected and how it is processed. Personal data management is considered an intensely personal matter however. Dourish argues that individual attitudes towards personal data and privacy are very complex and context dependent [5]. A recent three-year study showed that the more people disclosed on social media, the more privacy Copyright 2015 is held by the author(s). University and ACM. Derek McAuley University of Nottingham first.last@nottingham.ac.uk • Agency: means to manage “our” data and access to it, enabling us to act effectively in these systems as we see fit. • Negotiability: means to navigate data’s social aspects, by interacting with other data subjects and their policies. We do not envisage Databoxes entirely replacing dedicated, application-specific services such as Facebook and Gmail. Such sites that provide value will continue receiving personal data to process, in exchange for the services they offer. Databox simply provides its user with means to understand, Publication rights licensed to Aarhus 5th Decennial Aarhus Conference on Critical Alternatives August 17 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21312 29 control and negotiate access by others to their data across a range of sites, including such currently dominant players. As a physical object it offers a range of affordances that purely virtual approaches cannot, such as located, physical interactions based on its position and the user’s proximity. forms of cooperation, by third-party services and data aggregators. The Databox can be used as a central point for negotiating such data access and release rights. Controlled Access. Users must have fine-grained control over the data made available to third parties. At the very least, the Databox must be selectively queryable, though more complex possibilities include supporting privacy-preserving data analytics techniques, such as differential privacy [6] and homomorphic encryption [21]. A key feature of the Databox is its support for revocation of previously granted access. In systems where grant of access means that data can be copied elsewhere, it is effectively impossible to revoke access to the data accessed. In contrast, a Databox can grant access to process data locally without allowing copies to be taken of raw data unless that is explicitly part of the request. Subsequent access can thus easily be revoked [16]. A challenge is then to enable users to make informed decisions concerning the impact of releasing a given datum as this requires an understanding of the possible future information-states of all third parties that might access the newly released datum. One way to simplify this is to release data only after careful and irreversible aggregation of results to a degree that de-anonymisation becomes impossible. More complex decisions will require an on-going dialogue between the user and their Databox, to assist in understanding the impact of their decisions. Nor is the Databox oriented solely to privacy and prevention of activities involving personal data. It enables new applications that combine data from many silos to draw inferences presently unavailable. By redressing the extreme asymmetries in power relationships in the current personal data ecosystem, the Databox opens up a range of market and social approaches to how we conceive of, manage, cross-correlate and exploit “our” data to improve “our” lives. FEATURES OF A DATABOX What features must a Databox provide to achieve these aims? We answer in four parts: it must be a trusted platform providing facilities for data management of data at rest for the data subjects as well as controlled access by other parties wishing to use their data, and supporting incentives for all parties. Trusted Platform. Your Databox coordinates, indexes, secures and manages data about you and generated by you. These data can remain in many locations, but it is the Databox that holds the index and delegates the means to access that data. It must thus be highly trusted: the range of data at its disposal is potentially far more intrusive – as well as more useful – when compared to data available to traditional data silos. Thus, although privacy is not the primary goal of the Databox, there are clear requirements on the implementation of the Databox to protect privacy [11]. Trust in the platform requires strong security, reliable behaviour and consistent availability. All of the Databox’s actions and behaviours must be supported by pervasive logging, with associated tools, so that users and (potentially) third-party auditors can build trust that the system is operating as expected and, should something unforeseen happen, the results can at least be tracked. We envisage such a platform as having a physical component, perhaps in the form-factor of an augmented home broadband router, under the direct physical control of the individual. Thus, while making use of and collating data from remote cloud services, it would also manage data that the individual would not consider releasing to any remote cloud platform. Supporting Incentives. A consequence of the controlled access envisioned above is that users may deny third-party services access to data. The Databox thus must enable services alternate means to charge the user: those who wish to pay through access to their data may do so, while those who do not may pay through more traditional financial means. One possible expression of this would be to enable the Databox to make payments, tracing them alongside data flows to and from different third-party services made available via some form of app store. Commercial incentives include having the Databox act as a gateway to personal data currently in other silos, and as an exposure reduction mechanism for commercial organisation. This removes their need to be directly responsible for personal data, with all the legal costs and constraints that entail, instead giving control over to the data subject. This is particularly relevant for international organisations that must be aware of many legal frameworks. A simple analogy is online stores’ use of payment services (e.g., PayPal, Google Wallet) to avoid the overhead of Payment Card Infrastructure compliance. Data Management. A Databox must provide means for users to reflect upon the data it contains, enabling informed decision-making about their behaviours, and particularly whether to delegate access to others. As part of these interactions, and to support trust in the platform, users must be able to edit and delete data via their Databox as a way to handle the inevitable cases where bad data is discovered to have been inferred and distributed. Similarly, it may be appropriate for some data to not exhibit the usual digital tendency of perfect record. Means to enable the Databox automatically to forget data that are no longer relevant or have become untrue may increase trust in the platform by users [15]. Even if data has previously been used, it may still need to be “put beyond use” [2]. Concepts such as the European Union’s Right to be Forgotten require adherence to agreed protocols and other BARRIERS TO ADOPTION Many past systems provide some or all of these features, but none have really been successful due, we claim, to fundamental barriers that have yet to be coherently addressed. Trust. The growth of third party cloud services means users must trust, not only the service they are directly using, but also any infrastructure providers involved, as well as other parties such as local law enforcement. If the Databox is to take such a central place in our online lives, it must be trusted. Two key aspects stand out here for the Databox: (i) the need to trust that it will protect the user against breach of data due to attacks such as inference across different datasets; and 30 (ii) the need to trust that the software running on it is trustworthy and not acting maliciously. Two features of the design support this. First, all keys remain with the user themselves such that not even a Databox provider can gain access without permission. Second, by making a physical artefact a key component in a user’s Databox, e.g., a low-energy computing device hosted in their home, physical access controls – including turning it off or completely disconnecting it from all networks – can be applied to ensure data cannot leak. While this minimises the need to trust third parties, it increases trust placed in the software: we mitigate this by using open-source software built using modern languages (OCaml) on platforms (Xen) that limit the Trusted Computing Base, mitigating several classes of attack to which existing software is vulnerable. creasingly complex ecosystem [7]. Considering the churn experienced in the personal data startup space, with so many new but typically short-lived entrants, it seems that few truly viable models have yet been discovered. Our belief is that the power of personal data can only be realised when proper consideration is given to its social character, and it can be legibly combined with data from external sources. In this case, we might anticipate many potential business models [22]. Unfortunately, these approaches typically entail lodging all personal data into cloud-hosted services giving rise to concerns about privacy and future exploitation or mistaken release of data. In contrast, your Databox retains data – particularly control over that data – locally under your sole control. From a technology point of view, the general approach of Databox is that of “privacy by design”, though it remains to be seen if it can be successful in a space such as this, where policy and technology need to co-evolve. In order to sell personal data, there needs to be a method for determining the marginal rate of substitution (the rate at which the consumer is willing to substitute one good for another) for personal data. The sale of personal data and insights derived from it is the key utility in this ecosystem, and individuals’ preferences are the fundamental descriptors and success indicators. Usability. Personal data is so complex and rich that treating it homogeneously is almost always a mistake and, as noted above, user preferences in this space are complex: socially derived and context dependent. A very broad range of intents and requirements must be captured and expressed in machineactionable form. We will build on techniques developed in the Homework platform [19] which prototyped and deployed a range of novel task-specific interfaces that assisted users in the complex business of managing their home networks. Deciding which devices should be able to share in and access the digital footprint, even before considering sharing with other people, makes it even harder. Issues such as mixed, sometimes proprietary data formats, high variability in datum sizes, the multiplicity of standards for authentication to different systems to access data, lack of standard data processing pipelines and tools, and myriad other reasons make this job complex and time consuming. In addition, most data is inherently shared in that it implicates more than one individual and thus ownership and the concomitant right to grant access is not always clear. E.g., Use of cloud email services like Gmail: even if a user opts out by not using Gmail, there is a high chancethat a recipient of their email is using Gmail. Governments and regulatory bodies have attempted to impose regulatory frameworks that force the market to recognise certain individual rights. Unfortunately, legal systems are not sufficiently agile to keep up with the rapid pace of change in this area. Attempts at self-regulation such as the Do Not Track headers1 are ineffective, with only an insignificant fraction of services in compliance [7]. It is even possible that there may be a shift towards consumer protection legislation, as opposed to current prevalence of informed consent [13]. SUMMARY We are in an era where aggregation and analysis of personal data fuels large, centralised, online services. The many capabilities offered by these services have come at significant cost: lost of privacy and numerous other detrimental effects on the well-being of individuals and society. Many have commented that people simple do not see the need for technologies like this until they suffer some kind of harm from the exploitation of their data. On the other hand, it has been argued that privacy is negotiated through collective dynamics, and hence society reacts to the systems that are developed and released [10]. We speculate that data management may become a mundane activity in which we all engage, directly or through some representative, to a greater or lesser extent. Cost. There are a range of incentives that must align for the success of such a platform. The day-to-day costs of running a Databox have to be acceptable to users. Similarly, costs that third-parties incur when accessing the system will have to be recouped, including perhaps recompensing for access to data that previously they would have simply gathered themselves. It remains to be seen how this can be done in practice: Are users willing and able to pay in practice? What will be the response of users when offered pay-for versions of previously free-to-use services? There is some evidence that some users will be willing to make this trade-off [1], but studies also show that the situation is complex [24]. We have proposed the Databox as a technical platform that would provide means to redress this imbalance, placing the user in a position where they can understand, act and negotiate in this socio-technical system. By acting as a coordination point for your data, your Databox will provide means for you to reflect on your online presence, restore to you agency over your data, and enable a process of negotiation with other parties concerning your data. Even if the Databox as currently conceived is not a perfect solution, only THE WIDER ENVIRONMENT In response to growing awareness about how our data is processed, many startups have formed in recent years aiming to put users explicitly in control of their personal data or metadata. They typically provide platforms through which users can permit advertisers and content providers to enjoy metered access to valuable personal data, e.g., OpenPDS [4]. In exchange, users may benefit by receiving a portion of the monetary value generated from their data as it is traded in an in- 1 31 http://donottrack.us/ by taking initial, practical steps can we elicit the necessary knowledge to improve the state-of-the-art. We do not believe further progress can be made without focused effort on the practical development and deployment of the technologies involved. E.g., Before addressing the complex problems of comanaging data [3], a Databox that enables personal data to be collated and reflected upon will allow individuals to explore workflows managing both their own data and, through ad hoc social interaction, data involving other stakeholders. 10. Gürses, S. Can you engineer privacy? Commun. ACM 57, 8 (Aug. 2014), 20–23. 11. Haddadi, H., Hui, P., and Brown, I. Mobiad: private and scalable mobile advertising. Proc. ACM MobiArch (2010). 12. Leontiadis, I., Efstratiou, C., Picone, M., and Mascolo, C. Don’t kill my ads! balancing privacy in an ad-supported mobile application market. Proc. ACM HotMobile (2012). 13. Luger, E., and Rodden, T. An informed view on consent for UbiComp. In Proc. ACM UBICOMP (2013), 529–538. 14. Madhavapeddy, A., Mortier, R., Rotsos, C., Scott, D., Singh, B., Gazagnaire, T., Smith, S., Hand, S., and Crowcroft, J. Unikernels: Library operating systems for the cloud. In Proc. ACM ASPLOS (Mar. 16–20 2013). 15. Mayer-Schonberger, V. Delete: The Virtue of Forgetting in the Digital Age. Princeton University Press, 2009. 16. McAuley, D., Mortier, R., and Goulding, J. The Dataware Manifesto. In Proc. IEEE International Conf. on Communication Systems and Networks (COMSNETS) (January 2011). Invited paper. 17. Mortier, R., Greenhalgh, C., McAuley, D., Spence, A., Madhavapeddy, A., Crowcroft, J., and Hand, S. The personal container, or your life in bits. Proc. Digital Futures (2010). 18. Mortier, R., Haddadi, H., Henderson, T., McAuley, D., and Crowcroft, J. Human-data interaction: The human face of the data-driven society. SSRN (Oct. 1 2014). http://dx.doi.org/10.2139/ssrn.2508051. 19. Mortier, R., Rodden, T., Tolmie, P., Lodge, T., Spencer, R., Crabtree, A., Sventek, J., and Koliousis, A. Homework: Putting interaction into the infrastructure. In Proc. ACM UIST (2012), 197–206. 20. Mun, M., Hao, S., Mishra, N., Shilton, K., Burke, J., Estrin, D., Hansen, M., and Govindan, R. Personal data vaults: A locus of control for personal data streams. In Proc. ACM CoNEXT (2010), 1–12. 21. Naehrig, M., Lauter, K., and Vaikuntanathan, V. Can homomorphic encryption be practical? In Proc. ACM Cloud Computing Security Workshop (2011), 113–124. 22. Ng, I. C. Engineering a Market for Personal Data: The Hub-of-all-Things (HAT), A Briefing Paper. WMG Service Systems Working Paper Series (2014). 23. Rotsos, C., Howard, H., Sheets, D., Mortier, R., Madhavapeddy, A., Chaudhry, A., and Crowcroft, J. Lost in the edge: Finding your way with Signposts. In Proc. USENIX FOCI (Aug. 13 2013). 24. Skatova, A., Johal, J., Houghton, R., Mortier, R., Bhandari, N., Lodge, T., Wagner, C., Goulding, J., Crowcroft, J., and Madhavapeddy, A. Perceived risks of personal data sharing. In Proc. Digital Economy: Open Digital (Nov. 2013). 25. Vallina-Rodriguez, N., Shah, J., Finamore, A., Grunenberger, Y., Papagiannaki, K., Haddadi, H., and Crowcroft, J. Commercial break: Characterizing mobile advertising. In Proc. ACM IMC (2012). Thus we have begun development of underlying technologies for Databox: Nymote (http://nymote.org) and its constituent components of MirageOS [14], Irmin [8] and Signpost [23]. In addition, the community is developing methodologies for indexing and tracking the personal data held about us by third parties. However, the successful widespread deployment of such a platform will require that we tackle many significant issues of trust, usability, complexity and cost in ways that are transparent and scalable. Resolving questions such as those above requires that we develop and study Databoxes in-thewild, in partnership with individuals, consumer rights groups, privacy advocates, the advertising industry, and regulators. Acknowledgements. Work supported in part by the EU FP7 UCN project, grant agreement no 611001. Dr Haddadi visited Qatar Computing Research Institute during this work. REFERENCES 1. Acquisti, A., John, L. K., and Loewenstein, G. What is privacy worth? Journal of Legal Studies 42, 2 (2013), 249–274. 2. Brown, I., and Laurie, B. Security against compelled disclosure. In Proc. IEEE ACSAC (Dec 2000), 2–10. 3. Crabtree, A., and Mortier, R. Human data interaction: Historical lessons from social studies and CSCW. In Proc. ECSCW (Oslo, Norway, Sept. 19–23 2015). 4. de Montjoye, Y.-A., Shmueli, E., Wang, S. S., and Pentland, A. S. openpds: Protecting the privacy of metadata through safeanswers. PLoS ONE 9, 7 (07 2014), e98790. 5. Dourish, P. What we talk about when we talk about context. PUC 8, 1 (Feb. 2004), 19–30. 6. Dwork, C. Differential privacy. In Automata, Languages and Programming, M. Bugliesi, B. Preneel, V. Sassone, and I. Wegener, Eds., vol. 4052 of LNCS. Springer, 2006, 1–12. 7. Falahrastegar, M., Haddadi, H., Uhlig, S., and Mortier, R. Anatomy of the third-party web tracking ecosystem. CoRR abs/1409.1066 (2014). 8. Gazagnaire, T., Chaudhry, A., Crowcroft, J., Madhavapeddy, A., Mortier, R., Scott, D., Sheets, D., and Tsipenyuk, G. Irmin: a branch-consistent distributed library database. In Proc. ICFP OCaml User and Developer Workshop (Sept. 2014). 9. Guha, S., Reznichenko, A., Tang, K., Haddadi, H., and Francis, P. Serving ads from localhost for performance, privacy, and profit. In ACM Workshop on Hot Topics in Networks (2009). 32 Deconstructivist Interaction Design: Interrogating Expression and Form Martin Murer Center for HCI University of Salzburg martin.murer@sbg.ac.at Verena Fuchsberger Center for HCI University of Salzburg verena.fuchsberger@sbg.ac.at According to Vallgårda [23], giving form to computational materials in interaction design is a practice that comprises three elements: the physical form, the temporal form, and the interaction gestalt. As the computer is no longer at the center of attention, and as interaction design is no longer just a matter of interface design, Vallgårda proposed to establish a form-giving practice of interaction design. In order to develop such a practice, we need to explore and detail the according expressional vocabulary. Interaction design though does not start to develop its form-giving practice and expressional repertoire from scratch. As Hallnäs and Redström [10] argue, human-computer interaction has to some extent adopted rudimentary aesthetics from other areas, such as graphic design. Those design disciplines are highly relevant in HCI, for instance, in screen-dominated interaction design. This is considered insufficient, as the aesthetics ABSTRACT In this paper, we propose deconstructivist interaction design in order to facilitate the differentiation of an expressional vocabulary in interaction design. Based on examples that illustrate how interaction design critically explores (i.e., deconstructs) its own expressional repertoire, we argue that there are commonalities with deconstructivist phases in related design disciplines to learn from. Therefore, we draw on the role and characteristics of deconstructivism in the history of architecture, graphic design, and fashion. Afterwards, we reflect on how interaction design is already a means of deconstruction (e.g., in critical design). Finally, we discuss the potential of deconstructivism for form-giving practices, resulting in a proposal to extend interaction design’s expressional vocabulary of giving form to computational material by substantiating a deconstructivist perspective. [. . . ] of disciplines dominated by design-by-drawing tells us very little about the computational aspects of this new material we are working with. [10, p. 106] Author Keywords Deconstructivism; Interaction Design; Styles; Aesthetics ACM Classification Keywords We propose to look at historically evolved practices that enriched the aesthetic vocabulary of related design disciplines, such as architecture, graphic design, and fashion. In particular, we draw attention to deconstructive phases, as those share characteristics with the contemporary shift in interaction design from a pure focus on function to an increased attention to expressional possibilities (e.g., in form-driven ixd research [12]). We will depict the relation between deconstructivism and established design disciplines and discuss how interaction design is already in the service of deconstruction (e.g., in critical design). Finally, we will conclude with a proposal for substantiating a deconstructive perspective in interaction design that might enrich its expressional repertoire by interrogating the form of computational material. H.5.m. Information Interfaces and Presentation (e.g. HCI): Miscellaneous INTRODUCTION A motion controlled fur monkey makes a spectator realize, that she has been acting like a chimpanzee for the last couple of minutes [20]. Strangers sitting on a public bench unconsciously slip closer as the bench slowly changes its shape [13]. A whole room responds to the movements of a person by moving, collapsing, or expanding [22]. A lightbulb throws a shadow on a wall and all of a sudden the shadow grows wings and starts to fly [15]. These four encounters with interactive technology, all presented at recent premier HCI venues, share certain characteristics that are denotive to a current strand in interaction design, which shifts attention from function to expression. At first glance, the salient communality is that the four artefacts bear a surprising momentum, they are exceptional, rather a piece than a system. Less salient is that they all critically explore form-giving in interaction design by focusing their exploration on computation as a design material, they deconstruct our assumptions about how to interact with a particular element in our reality. Copyright 2015 is held by the author(s). University and ACM. Manfred Tscheligi Center for HCI University of Salzburg manfred.tscheligi@sbg.ac.at DECONSTRUCTIVISM AND DESIGN In a 2013 event the New York Museum of Modern Arts (MoMA) celebrated the 25th anniversary of a milestone exhibition in architecture: Philip Johnson and Mark Wigley had curated “Deconstructivist Architecture”, showcasing the work of seven at the time of the exhibition contemporary architects (see e.g., [11]). The curators did not attempt to define a style; they rather aimed to present then contemporal architecture of similar approaches. Still, it was that very exhibition that has coined the term deconstructivism as a framing for a highly influential development in architecture and many other disciplines in succession. The visual appearance of deconstructivist styles can be characterized by controlled Publication rights licensed to Aarhus 5th Decennial Aarhus Conference on Critical Alternatives August 17 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21313 33 chaos, unpredictability and distortion. Underneath its skin, deconstruction is not about a style or a movement. Rather, its proponents understand their work as an opposition to the ordered rationality of postmodernism. developing future interactive systems are reasonable and desirable outcomes of such approaches. Interaction design in the service of deconstruction The background of our proposal is, on one hand, constituted by the origin of deconstructivist movements, such as philosophy, architecture and graphic design. On the other hand, there is also related work in interaction design and HCI, which to varying extents includes deconstructivist elements as part of a method or thinking. With HCI becoming more implicated in culture [4], recent practices and concepts can be identified that incorporate deconstruction as a means to raise awareness for societal changes and issues, such as critical theory, critical design, or feminist design. Origins of Deconstructivist Movements Being based on Jacques Derrida’s philosophic concept of deconstruction, deconstructivism in architecture asks questions about modernism by “reexamining its own language, materials, and processes,” [16] as summarised by Wigley in his 1988 catalogue essay: A deconstructive architect is [. . . ] not one who dismantles buildings, but one who locates the inherent dilemmas within buildings. The deconstructive architect puts the pure forms of the architectural tradition on the couch and identifies the symptoms of a repressed impurity. The impurity is drawn to the surface by a combination of gentle coaxing and violent torture: the form is interrogated. [16] The study object of critical theory is, broadly speaking, any aspect of culture, exploring the constructedness of knowledge [1]. For HCI, Bardzell [1] suggests ways for criticism to contribute to its practice. Critical design [8] is an approach to provocation, which offers sources and strategies to inspire designs, such as techniques to radically rethink fundamental concepts. Those strategies aim at reconfiguring sociocultural norms in more aesthetic or social ways and may stimulate demand for such designs [3]. From this point of view, the deconstructive moment is de- and reconstructing knowledge. It was that MoMA exhibition in 1988 that got deconstructivism “catapulted into design press” [16], with a lasting impact not only on architecture. Other disciplines such as graphic design (e.g., [18]) or fashion (e.g., [14]) soon augmented existing streams with those formal concepts and mechanisms. Deconstructivism became a kind of cliché of a certain formal style, characterized by aggressive arrangements, sharp edges, fragmentation, etc. Deconstruction got phrased more generally as means “[. . . ] to deform a rationally structured space so that the elements within that space are forced into new relationships” [18, p. 122]. In graphic design and fashion the notion of deconstruction since changed from a style to an activity of critical form-giving, becoming a central element in any design practice [16, 14]. It is no longer en vogue but rather considered a zeitgeist (e.g., [6]). The contemporary relevance of deconstructing is though still present. As Derrida phrased it in a 1994 interview in The New York Times Magazine, there is some element in deconstruction “that belongs to the structure of history or events. It started before the academic phenomenon of deconstruction, and it will continue with other names.” [21] Feminism can also be considered from a deconstructivist point of view. Following the political agenda of critiquing power (e.g., [5]), its aim is to deconstruct social structures. In HCI, for instance, feminist thinking supports the awareness and accountability of its social and cultural consequences [2]. Buchmüller argues that one common goal of feminism is to “[C]hange the situation/position of the researched by offering them critical ways of thinking, new ways of expression as well as new opportunities of action.” [5, p. 175]. This means that feminism seeks to initiate social change by changing social structures. Deconstruction in a feminist tradition is, thus, applying deconstructive procedures to dominant gender structures [25]. The deconstructive element is related to the artifact [1] which initiates questioning a social setting. Besides critical design and feminism, a further notion of deconstructing the social by means of technology can be found in critical making. It emphasizes critique and expression rather than technical sophistication and function [17]. Making is understood as an activity, which is a vehicle for critically engaging with the world [9]. The constructive processes are considered as the site for analysis, the shared acts of making are more important than the evokative artifacts. Thus, the prototypes are a means to an end, which achieve value by creating, discussing and reflecting on them [17]. All these deconstructive movements have passed their peak of attention, but they left their respective disciplines with an enriched set of formal vocabulary. Taking this deconstructivist perspective, it may be worthwhile to consider how it relates to the goal of establishing a richer expressional vocabulary for form-giving practices in interaction design. A deconstructivist lens can help us understand what kind of deconstruction we already perform in interaction design, how we invert existing approaches and thinking, and how this framing might support both the practice and theory in order to understand and link their approaches to experiential knowledge creation. Similar to architecture, graphic design, and fashion, interaction design is in the first place concerned with construction. However, notions of deconstructivism, such as provocation, critical theory and design are pervading interaction design, being a mode of questioning the current state of the art. Both creating an enhanced understanding as well as imagining and DECONSTRUCTIVIST INTERACTION DESIGN With deconstructivist interaction design we seek to frame critical practices that emphasize the expressional possibilities interaction design provide. This is not mutually exclusive from other critical practices in human-computer interaction and design, i.e., a particular design or artefact may well satisfy multiple attributions. While, e.g., critical design aims to “introduce both designers and users to new ways of looking at the world and the role that designed objects can play for them 34 in it” [19, p. 51], we want to draw attention to the form-giving and material aspects. As Hallnäs and Redström [10] argue, in HCI-related research and design practice aesthetical decisions tend to be hidden by concerns for practical functionality, usability, user requirements etc. In line with Dunne [7], they argue, that those decisions are hidden, “not because they are not made, but because the expressions of the things we design become mere consequences of other concerns.” [10, p. 105] They propose a program for experimental design of computational things that turns the classical leitmotif “form follows function” upside down. While not explicated by Hallnäs and Redström, we argue that their program bears a deconstructive momentum: As a critical activity the classical design agenda is questioned by proposing an alternative leitmotif, i.e., function resides in the expression of things. tegration of physical and temporal form into a coherent entity while breaking with the assumptions of how reality behaves. Following Vallgårda’s trinity of forms (physical form, temporal form, interaction gestalt) in interaction design practice [23], the core, the interplay between the three, is what Leigh et al. [15] challenged in their design. Deconstructivist interaction design, however, may also deconstruct single elements as well. In order to question the physical form, the related design disciplines provide a whole spectrum of practices and vocabulary to draw on, as they already intensively explored the expressional repertoire of physical forms, for instance, in architecture, pottery, or statuary. Challenging, articulating, and expressing the temporal form or the interaction gestalt, however, is an exercise interaction design is necessarily engaged with in interrogating the interplay of the three formgiving elements. Looking through a deconstructivist lens at interaction design practices and the related artefacts, whether they are meant to be deconstructive or not (or, for instance, following a critical, feminist or any other approach), may support this exercise and contribute to the articulation of formal expressional vocabulary. As “form follows function” used to be the prevailing (structuralist) leitmotif of interface design, “function resides in the expression of things” may become a (deconstructive) leitmotif of interaction design. It would guide explorations that question the expressional qualities of materials and artefacts. It implies engagement with the expressional repertoire, its articulation and vocabulary. Furthermore, deconstructive design can help to extend the existing repertoire by critically challenging and examining what is already known. DISCUSSION AND CONCLUSION Starting from architectural deconstructivism, we argue that framing interaction design examples and strands from a deconstructivist perspective supports reflection on the act of taking apart form, function, meaning, concepts, or artifacts to enhance our knowledge and expressional vocabulary for a form-giving practice in interaction design. Exemplary Deconstructivist Interaction Design In the following we describe one of the artefacts highlighted in the very beginning of this paper in order to elaborate on how it bears a deconstructive stance. Remnance of Form by Leigh and colleagues [15] is an interactive installation that aims to explore the “dynamic tension between an object and its shadow.” [15, p. 411] Composed of a light, projection and tracking technology, a shadow can detach itself from its former role. The relationship between a light bulb, an object and its shadow is constantly modulated and distorted. A person interacting with Remnance of Form can move the light bulb and the object around in physical (and temporal) space. Depending on one’s movements in respect to the installation, the shadow, as a digital alteration of reality, is programmatically shaped into different forms and behaviours (i.e., changes shape, grows wings and flies away, shows fear). The reflection we presented is meant to be a starting point for a discussion on deconstructive interaction design by trying to consider current design practices through a deconstructivist lens. In order to strengthen this framing, further material and design studies and examinations from this perspective will be needed that support the sophistication of (deconstructive) form-giving practices. While there are obvious parallels to deconstructive phases in other disciplines, deconstructive interaction design underlies multiple deviations that are constituted in the constant advancement of computational technology. This prompts interaction design to continuously challenge and deconstruct established computational forms (e.g., as the predominate aesthetics of GUIs brought about a deconstruction of form, questioning whether “a machine needed to be visible at all.” [24, p. 67]) Remnance of Form deconstructs our perception and assumptions about interacting with reality. It is critical in the sense that it deforms the elements of what used to be a rationally structured space (i.e., our previous assumptions about the interaction with a light bulb, an object and its shadow) and forces them into new relationships. Arguably, its critical momentum does though not seek to “disrupt or transgress social and cultural norms” [3] as framed by critical design. Rather it is the interdependence of the form elements of interaction design [23] that are interrogated and explored at once. This adds a particular expression to our interaction design vocabulary, one that can easily be interpreted as being magical through dissolving the formerly stable associations between lights, objects and their shadows. The interaction gestalt of Remnance of Form as it unfolds through interaction, its disruptive and disturbing nature, was brought on by a balanced in- In that realm we do not consider deconstructive interaction design being tied to a political agenda, as in critical or feminist theory; rather, we aim to emphasize the act of questioning (material) form and functions. We do not oppose deconstruction to construction; instead, we see deconstruction as a means to gain an additional kind of knowledge, one, that allows to derive multiple functions out of form, both conceptually and practically. The framing of deconstructivist interaction design is no opposition to the epistemological value of prevailing critical practices in human-computer interaction, nor to constructive design research. Rather, we aim to draw attention to the potential of critical form-giving as means to enrich and further detail the expressional vocabulary of interaction design practice. Drawing on a notion Johnson used to 35 describe the then contemporary architecture, deconstructivist interaction design can be described as a “contemporary artistic phenomenon that derives its forms from constructivism and yet deviates from it.” [11] 12. Heekyoung Jung and Erik Stolterman. 2012. Digital Form and Materiality: Propositions for a New Approach to Interaction Design Research. In Proc. NordiCHI’12. ACM, 645–654. DOI: http://dx.doi.org/10.1145/2399016.2399115 ACKNOWLEDGMENTS 13. Sofie Kinch, Erik Grönvall, Marianne Graves Petersen, and Majken Kirkegaard Rasmussen. 2014. Encounters on a Shape-changing Bench: Exploring Atmospheres and Social Behaviour in Situ. In Proc. TEI’14. ACM, New York, NY, USA, 233–240. DOI: The financial support by the Austrian Federal Ministry of Science, Research and Economy and the National Foundation for Research, Technology and Development is gratefully acknowledged (Christian Doppler Laboratory for “Contextual Interfaces”). http://dx.doi.org/10.1145/2540930.2540947 14. Gizem Kiziltunali. 2012. A deconstructive system: fashion. Konferenssiesitelmä. 4th Global Conference, Inter-Disciplinary. Net. REFERENCES 1. Jeffrey Bardzell. 2009. Interaction criticism and aesthetics. In Proc. CHI 2009 (CHI ’09). ACM, New York, NY, USA, 10. DOI: 15. Sang-won Leigh, Asta Roseway, and Ann Paradiso. 2015. Remnance of Form: Altered Reflection of Physical Reality. In Proc. TEI’15. ACM, New York, NY, USA, 2. DOI: http://dx.doi.org/10.1145/1518701.1519063 2. Shaowen Bardzell and Jeffrey Bardzell. 2011. Towards a feminist HCI methodology: social science, feminism, and HCI. In Proc. CHI 2011. ACM, 10. DOI: http://dx.doi.org/10.1145/2677199.2690874 http://dx.doi.org/10.1145/1978942.1979041 16. Ellen Lupton and J Abbott Miller. 1994. Deconstruction and graphic design: history meets theory. Visible language 28 (1994), 346–346. 3. Shaowen Bardzell, Jeffrey Bardzell, Jodi Forlizzi, John Zimmerman, and John Antanitis. 2012. Critical design and critical theory: the challenge of designing for provocation. In Proc. DIS ’12. ACM, 10. DOI: 17. Matt Ratto. 2011. Critical Making: Conceptual and Material Studies in Technology and Social Life. The Information Society 27, 4 (2011), 252–260. DOI: http://dx.doi.org/10.1145/2317956.2318001 4. Mark Blythe, Jeffrey Bardzell, Shaowen Bardzell, and Alan Blackwell. 2008. Critical issues in interaction design. In Proc. BCS-HCI ’08. BCS, UK, 2. http: http://dx.doi.org/10.1080/01972243.2011.583819 18. Timothy Samara. 2005. Making and breaking the grid: a graphic design layout workshop. Rockport Publishers. //dl.acm.org/citation.cfm?id=1531826.1531880 19. Phoebe Sengers, Kirsten Boehner, Shay David, and Joseph ’Jofish’ Kaye. 2005. Reflective Design. In Proc. CC ’05. ACM, New York, NY, USA, 10. DOI: 5. S. Buchmueller. 2012. How can Feminism Contribute to Design? Reflexions about a feminist framework for design research and practice. In Design Research Society (DRS) International Conference. http://dx.doi.org/10.1145/1094562.1094569 20. Jan M. Sieber and Ralph Kistler. 2013. Monkey Business. In Proc. TEI’14. ACM, New York, NY, USA, 2. DOI: 6. Chuck Byrne and Martha Witte. 2001. A brave new world: understanding deconstruction. Graphic Design History (2001), 245. http://dx.doi.org/10.1145/2540930.2555203 7. Anthony Dunne. 1999. Hertzian tales. MIT press London. 21. Mitchell Stephens. 1994. Jacques Derrida and Deconstruction. The New York Times Magazine, January 23. (1994). 8. Anthony Dunne and Fiona Raby. 2001. Design noir: The secret life of electronic objects. Springer. 22. Anna Vallgårda. 2014a. The Dress Room: Responsive Spaces and Embodied Interaction. In Proc. NordiCHI’14. ACM, New York, NY, USA, 10. DOI: 9. Shannon Grimme, Jeffrey Bardzell, and Shaowen Bardzell. 2014. ”We’Ve Conquered Dark”: Shedding Light on Empowerment in Critical Making. In Proc. NordiCHI’14. ACM, New York, NY, USA, 10. DOI: http://dx.doi.org/10.1145/2639189.2639254 23. Anna Vallgårda. 2014b. Giving Form to Computational Things: Developing a Practice of Interaction Design. Personal Ubiquitous Comput. 18, 3 (March 2014), 577–592. DOI: http://dx.doi.org/10.1145/2639189.2641204 10. Lars Hallnäs and Johan Redström. 2002. Abstract information appliances: methodological exercises in conceptual design of computational things. In Proc. DIS’02. ACM, New York, NY, USA, 12. DOI: http://dx.doi.org/10.1007/s00779-013-0685-8 24. Mikael Wiberg and Erica Robles. 2010. Computational compositions: Aesthetics, materials, and interaction design. International Journal of Design 4, 2 (2010), 65–76. http://dx.doi.org/10.1145/778712.778730 11. Philip Johnson and Mark Wigley. 1988. Deconstructivist Architecture: The Museum of Modern Art, New York. Little, Brown. 25. Peter V Zima. 2002. Deconstruction and critical theory. Continuum International Publishing Group. 36 In Search of Fairness: Critical Design Alternatives for Sustainability Somya Joshi Stockholm University Borgarfjordsgatan 12 somya@dsv.su.se Teresa Cerratto Pargman Stockholm University Borgarfjordsgatan 12 tessy@dsv.su.se levels of human experience [5], but are also deeply concerned by environmental [1], socio-political [13], social sustainability [4] and ecological concerns [20]. Latest developments in Sustainable HCI (SHCI) clearly indicate an interest into the development and design (or undesign) of technologies relying less on instrumental purposes of efficiency connected with corporate profit [17] and more prone on volitional and value-laden aspects underlying people’s use of technologies [6,2]. Our paper follows from the above, focusing primarily on the social value laden enterprise and its processes of activism and co-creation that inform critical alternatives within the SHCI discourse. We do so via the illustrative lens of a case study (Fairphone [10]) that enables us to conceptualise technological design and consumption from a holistic perspective of social ecology. The paper contributes with four critical design alternatives towards sustainability. ABSTRACT Does fairness as an ideal fit within the broader quest for sustainability? In this paper we consider alternative ways of framing the wicked problem of sustainability. One that moves away from the established preference within HCI, towards technological quick-fixes. We adopt a critical lens to challenge the belief that by merely changing practices at an individual level one can do away with unsustainability. This thinking, we argue, is flawed for many reasons, but mostly because of the wickedness of the sustainability problem. By analyzing the case of Fairphone, we illustrate how it is possible to imagine and design change at a broader level of community engagement, when it comes to concerns of fairness and sustainability. We contribute to a deeper understanding of how social value laden enterprises along with open technological design can shape sustainable relationships between our environment and us. Author Keywords THEORETICAL UNDERPINNINGS Sustainability; Open Technologies; Critical Alternatives; Transactions; Social-Ecology. Our interest in the social ecology perspective [16,21] is due to its holistic understanding of the interplay between natural-ecological and socio-semiotic dimensions. In particular, we turn our attention to how this interplay of dimensions can be reflected in the design process. This relationship, we believe, is key to further elaborate on issues pertaining to the design and building of computer systems [7]. In particular, this specific relationship between material (natural) and human (constructed) facets of human-environment systems is referred to as transactions [21]. By transactional relationships, Stokols et al. [21] refer to continuous, bidirectional and mutually influencing relationships occurring between both natural-ecological (i.e. material) and social-semiotic dimensions (i.e. meanings, values, moral judgments). Transactions entail exchanges among diverse actors, assets and resources that play a major role in the sustainability and resilience of our environment. A particularity of these transactions is that they are not fungible, as changes in one dimension are related to changes in another dimension. As such, the concept of transaction has implications for SHCI in terms of how we think about the delineation and understanding of the design space. We see it as an arena wherein diverse compromises are made as a result of the multiple and multifarious dilemmas [15] designers are confronted with, when dealing with sustainability issues. ACM Classification Keywords H.5.m. Sustainable interaction design, Miscellaneous. INTRODUCTION How do we build systems that are in line with current understandings of Earth’s finite natural resources? How do we design them to catalyze and sustain social change? These and other questions have been introduced and developed within the Sustainable HCI (SHCI) literature [3,17,8,18,14,10,20,24]. This pioneering body of research work has drawn on theoretical and applied developments that can be related to third wave HCI [5]. They have on the one hand, touched upon key issues related to our everyday practices and culture and, on the other hand, they have overflowed third wave HCI boundaries as they are not only focused on the cultural, emotional, pragmatic or historical Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21301 37 situated within an ecosystem of ideals aspired to by society, such as sustainability, freedom, justice, peace, truth. None of these concepts exist in an absolute state; rather the incremental enactment of re-balancing the inherent inequity is what we focus on. To address the above, we consider alternative ways of framing the wicked problem of sustainability. One that moves away from the established belief within HCI, in technological quick-fixes. We adopt a critical lens to challenge the belief that by only changing behavior at an individual level, one can do away with unsustainability. This thinking, we argue, is flawed for many reasons, but mostly because of the wickedness of the sustainability problem. By analyzing the case of Fairphone, we illustrate how it is possible to imagine and design change at a broader level of community engagement, when it comes to concerns of fairness and sustainability. FROM THEORY TO PRAXIS: THE FAIRPHONE ROADMAP Based on the theoretical discussion above, we set out to examine what such an alternative might look like in practice. Fairphone started off as an awareness raising campaign in 2010, mainly focused on conflict minerals within the context of the smartphone industry. In the absence of a real alternative to point to, Fairphone emerged as a social enterprise in 2013, as the outcome of a crowdfunded campaign designed to produce a truly ‘fair’ phone. While the notion of fairness here was co-constructed by the founding members and the community of users, the four main action points that the movement was built around, were: mining, manufacturing, design and life-cycle. We chose this case as it provides us with a window into two worlds simultaneously: that of a social enterprise setting out to engineer and sustain a movement (based on changing relationships and practices within the domain of technology design and consumption), and, that of a technical artifact designed to embody the life-cycle approach - built on the “fairware” principles (open hardware and software, conflict free and fair in terms of workers rights, circular economy – a “cradle to cradle” approach). CRITICAL DESIGN ALTERNATIVES From the data analyzed, the following four themes emerged. From a logic of Volume to a logic of Fairness Design alternatives in some cases are based on the principle that change arises when technologies provide opportunities for individuals to live differently [22]. Sustainability in this context is a byproduct of the artifact itself. Other design choices frame change-making as arising from technology providing opportunities for community debate [9], in which the direction for change is set by communities themselves. Within the context of Fairphone, we observed both these drivers for change. Framing itself as a social enterprise, Fairphone has at the outset opted for a mandate that is more rooted in the ethics of sustainability than in any desire to be industry leaders in the smartphone domain. This has translated into a set of compromises or trade-offs. As a young start up, it has found itself in the position where it is unable to diversify its product range, or maximize much needed profits in order to grow; but instead keep a steady focus on the goal at hand which is to promote fairness and quality control of the existing artifact. In the words of the CEO of the organization, “Our mission is to create momentum to design this future. We started by making a phone to uncover production systems, solve problems and use transparency to invite debate about what’s truly fair. We believe that these actions will motivate the entire industry to act more responsibly.” Two years down the line, “This is where we still are. Our business model hasn’t changed, nor has our product focus – we will still concentrate on phones, and won’t branch out into other consumer electronics like laptops or tablets.” While this model has so far worked well for Fairphone, we are skeptical about the transferability and sustainability of this approach for other young start ups operating within a highly competitive market (without the same safety-nets). Another reservation we have pertains to the measurable impacts emerging from Empirical Design This paper draws on data collected at two levels. Primary: of semi-structured interviews conducted iteratively, with impact development and product strategy staff at Fairphone. Secondary: data collected via the website, blogs and online documentation, as well as the critical voices emerging from the wider community of users and supporters. The latter consists of early adopters, experts, users, designers and partner organisations. We also draw on social media data from Twitter (#wearefairphone) and Facebook. The analysis of the data was conducted using a procedure known as explication de texte, or close reading, an analytical method that originated in the humanities [19]. We took into consideration texts (i.e. interview transcripts, expert reports, blog entries, social media excerpts and forum debates) as our unit of analysis, from which we arrived at conceptual threads. They constitute the findings of our study, which we will unpack in the following section. Performing Fairness Defining the landscape of fairness is a tricky task, given that there is neither a single accepted definition, nor an absolute state of fairness per se. As a concept it is construed both in terms of procedural and distributive fairness [11]. It is the latter, equity-based logic for action, which concerns us the most within this inquiry, with respect to responsibilities and resource distribution. The global resource challenge shifts from being a matter of “living within limits” to one of “living in balance” between social and environmental boundaries [12]. Fairness is 38 this choice to prioritize fairness over profits or growth. In particular we wonder how this translate into tangible changes within the established industry of consumer electronics, where it remains to be seen if the model of Fairphone serves as an inspiration for change or a niche alternative which leaves no dent in their operations. From Technology Relationships Quick-Fixes to term. This links to our understanding of critical alternatives, in that it provides a mechanism to challenge the status quo with regard to consumption choices and locked-in design (i.e. proprietary systems that engender waste and unsustainability in their wake). During our interviews with the team, we were provided with the following view on both the evolution of the artifact and the movement: “Today Fairphone is not for everyone. The early adopters or aware consumers are the ones who are part of the community, because either they are interested in conflict minerals, or workers rights, or environmental impact. However by joining the movement, they gain a context to this awareness and see the bigger-linked picture of the life cycle approach we adopt. We are essentially providing them with the tools to be curious with.” The cocreation then takes place, we argue, not just at the level of the product design, but also in terms of education and capacity building on the part of the wider Fairphone community. The trade offs of opening up both the technical design of the phone and the future evolution of it to the wider user community, are manifold. Sustainability in hardware choices translated into decisions such as the modularity of the phone, its openness to friendly hacks and the conflict free nature of materials used, all make costly demands on resources, that are otherwise spent on user experience (e.g. larger screens, lighter phones, cheaper and faster processors, better cameras etc.) By adopting a lifecycle view, and interlinking design as well as consumption decisions made closer to home, with impacts in far away and often disconnected places, the very discourse of innovation and sustainability are being reexamined and reworked. Changing Just as DNA comes in pairs, we were informed in our discussion with the Head of Impact at Fairphone, that the DNA of this social enterprise consists of two strands: on the one hand, it aims to change the relationship between users and how they consume technology, on the other hand, it aims to use technology as an innovative tool to alleviate societal problems that emerge from its supply chain. From our critical standpoint, we see two key challenges standing in the way of realizing the goal of changing relationships between users and their technology fixes. That of scalability and the slow pace of change within the sustainability context. With regard to the former, a small outfit such as Fairphone (i.e. 30 fulltime staff in Amsterdam) simply (to put it in their own words) “can not afford to move the entire supply chain of production from China to Europe.” This alludes to a set of compromises that shift their roadmap and milestones to impacts more graspable, small scale and localized in the early stages of development. With regard to the latter challenge, in an industry and market such as that of smart phones and more broadly consumer electronics, the rate of change in product development feeds into an expectation of heightened novelty seeking. One member of the team, working on product design, commented that it was not just the consumer, with whom the relationship was transforming, but also with industrial traders (such as for example telecom providers and plastic suppliers) who were coming to view Fairphone as an experiment they increasingly sympathized with. It provided them proof that an alternative could exist. Within this environment, we ask, what does it entail to engage in a slow deliberate march towards the attainment of sustainability goals, most of which are not immediately apparent (i.e. changing worker relationships, acceptance of standards and regulation with regard to conflict free mining, recycle-reuse impacts)? Opening up the ‘black box’ of design One startling difference we found between the model adopted by Fairphone and mainstream consumer electronics manufacturers, concerns the openness in design and the holistic life-cycle view. Be it manifest in the urban mining workshops organized routinely or the e-waste reduction efforts in Ghana, the attempt here is one of creating conditions for debates around sustainability. Urban mining emerges from this context as a way to change the existing imbalance, by the extraction of minerals from existing products. The aim is to dismantle gadgets that have reached their end-of-life to uncover what’s within. The idea is that once users have unscrewed the back cover of their technological artifact (in this case a smart phone), and identified the components, the urban mining workshop would aim to unravel some of the phone’s hidden stories. From pollution and extremely dangerous working conditions to child labor, one can learn that a number of mining-related practices desperately require improvement. The approach adopted by mainstream players within this context, is to hide this inconvenient body of knowledge behind, sealed, glued and proprietary locked in devices, where the design serves as an impenetrable casing which keeps all unpalatable, guilt- From Locked-in Design to Collaborative Co-creation With regard to the actual technical artifact at hand, Fairphone has positioned itself along a roadmap towards ‘Fairware’, which is a concept expressed in both long and short term ideas. Thinking short term, Fairphone aims at being open source to allow lead users to optimize it, as well as modular, so others can repair/replace parts to use the phone longer. In the long term, “Cradle2Cradle” design would enable reuse scenarios, where old modules could be used to upgrade other devices, which would allow circulating across product cycles. This would considerably cut down on waste and forced obsolescence in the long 39 inspiring footprints of our consumption behaviours, neatly out of sight (and hence out of mind). The critical alternative being offered by Fairphone here is one of opening the design processes. 8. DiSalvo, C., Sengers, P., and Brynjarsdòttir. Mapping the landscape of sustainable HCI, In Proc. CHI’10. ACM (2010), 1975-1984. CONCLUSION 10. Fairphone. www.fairphone.com Far from being a flawless and self-contained movement, Fairphone is in its early stages of evolution, experiencing the teething pains expected of any initiative aiming to challenge the status quo. We see it as a step towards opening the discourse of critical alternatives within sustainable HCI. In this paper we have presented critical alternatives, both at the theoretical level (with the socioecological approach as a lens) and at an applied level (via the illustrative lens of the Fairphone case). In doing so we have contributed to a more holistic understanding of how social value laden enterprises along with open technological design can shape sustainable relationships between our environment and us. 11. Haidt, J. (2012). The righteous mind. Why good people are divided by politics and religion. Random House. 9. DiSalvo, C. Adversarial design. MIT Press, 2012 12. Hajer, M.; Nilsson, M.; Raworth, K.; Bakker, P.; Berkhout,F.; de Boer, Y., Rockström, J.; Ludwig, K. and Kok, M. Beyond Cockpit-ism: Four Insights to Enhance the Transformative Potential of the Sustainable Development Goals. Sustainability 2015, 7, 1651-1660. 13. Dourish, P. HCI and Environmental Sustainability: The Politics of Design and the Design of Politics. In Proc. DIS’10, ACM Press (2010),1-10. 14. Håkansson, M. and Sengers, P. Beyond Being Green: Simply Living Families and ICT. In Proc. CHI ´13. ACM (2013). 2725-2734. REFERENCES 1. Aoki, P.M.; Honicky, R.J.; Mainwaring, A.; Myers, C.; Paulos, E.; Subramanian, S. and Woodruff, A. A Vehicle for Research: Using Street Sweepers to Explore the Landscape of Env. Community Action. In Proc. CHI’09 ACM (2009), 375-384. 15. Håkansson, M. and Sengers, P. (2014). No easy compromise:Sustainability and the dilemmas and dynamics of change. Proc. of DIS’14. ACM (2014).1025-1034. 16. Lejano, R. and Stokols, D. Social Ecology, Sustainability and Economics. Ecological Economics, 89 (2013), 1-6. 2. Bellotti, V., Caroll, J.M. and Kyungsik, H. Random acts of kindness: The intelligent and context-aware future of reciprocal altruism and community collaboration. In Proc. of CTS. IEEE (2013), 1-12. 17. Nardi, B. The Role of Human Computation in Sustainability, or, Social Progress is Made of Fossil Fuels. In Handbook of Human Computation, New York, Springer, 2013. 3. Blevis, E., Makice, K., Odom, W., Roedl, D., Beck, C., Blevis, S., and Ashok, A.: Luxury and new luxury, quality and equality. In Proc. DPPI ’07, ACM (2007), 296–311. 18. Pargman, D. and Raghavan, B. Rethinking sustainability in computing: from buzzword to nonnegotiable limits. In Proc. NordiCHI ’14. ACM (2014), 638-647. 4. Busse, D.; Blevis, E., Beckwith, R.; Bardzell, S.; Sengers, P.; Tomlinson, B.; Nathan, B. Social Sustainability: An HCI Agenda. In Proc. CHI EA’12. ACM (2010), 1151-1154. 19. Richards, I.A. (1930). Practical Criticism. A study of Literary Judgment. Kegan Paul & Co.Ltd. London. 5. Bödker, S. When second wave HCI meets third wave challenges. In Proc. NordiCHI ‘06, ACM Press (2006), 1-8. 20. Silberman, S. and Tomlinson, B. Toward an Ecological Sensibility. In Proc.CHI ´10 EA. ACM (2010), 34693474. 6. Cerratto-Pargman, T. A European strand of Sustainable HCI? Workshop paper NordiCHI ‘14. (2014). 21. Stokols, D., Lejano, R. and Hipp, J. Enhancing the Resilience of Human–Environment Systems: a Social Ecological Perspective. Ecology and Society 18, 1 (2013), 1-12. http://www.researchgate.net/publication/276206198_Understa nding_Limits_from_a_Social_Ecology_Perspective 7. Cerratto Pargman, T. and Joshi, S. (2015). Understanding limits from a social ecology perspective. LIMITS 2015, Irvine, CA. USA 22. Woodruff, A., Hasbrouck, J., and Augustin, S. A bright green perspective on sustainable choices. In Proc. CHI ’08. ACM (2008).313-322. http://www.limits2015.org/papers/limits2015-cerrattopargman.pdf 40 Why Play? Examining the Roles of Play in ICTD Pedro Ferreira Mobile Life @ KTH Stockholm, Sweden i@pedropaf.com munication Technologies for Development (ICTD or ICT4D). The focus on the urgency to satisfy a pre-defined set of socio-economic needs relegates play as something poor people cannot afford to engage with. Generally speaking, there has not been a lack of excitement around the advent of the information age. ICTs were, early on, framed around their potential to improve and optimize business processes and efficiency [9, 27, 48]. In the late 80’s, Nobel Laureate Robert Solow challenged this enthusiasm with a famous quip: “You can see the computer age everywhere but in the productivity statistics” [78], echoing the concern that such explanations were insufficient to account for the overwhelming success and excitement around ICTs. While history has certainly justified the role of ICTs in promoting the advancement of a variety of socio-economic goals, this remains a narrow and insufficient way to account for and understand the mass personal adoption of these technologies. Within ICTD, this type of focus, is especially predominant and fails to account for some of the main aspects of technology use: i.e. entertainment, leisure or games. ABSTRACT The role of technology in socio-economic development is at the heart of ICTD (ICTs for development). Yet, as with much Human Centered technology research, playful interactions with technology are predominantly framed around their instrumental roles, such as education, rather than their intrinsic value. This obscures playful activities and undermines play as a basic freedom. Within ICTD an apparent conflict is reinforced, opposing socio-economic goals with play, often dismissed as trivial or unaffordable. Recently a slow emergence of studies around play has led us to propose a framing of it as a capability, according to Amartya Sen, recognizing and examining its instrumental, constructive, and constitutive roles. We discuss how play unleashes a more honest and fair approach within ICTD, but most importantly, we argue how it is essentially a basic human need, not antithetical to others. We propose ways for the recognition and legitimization of the play activity in ICTD. Author Keywords Play; ICTD; ICT4D; capabilities; freedom; games; entertainment Today, digital interactions, particularly in the form of smartphones, pervade Western and developing worlds alike. Mobile subscriptions have nearly matched the earth’s population, and 3 million people now have access to internet [80]. And while leisure and ludic interactions with technology are prominent, they lack the acceptance within academic work around technology use. It is now roughly 10 years since Bødker described third-wave HCI (Human Computer Interaction) in which understandings of technology must move beyond the office and into everyday life [8]. It is also close to 10 years since Gaver first published his work on ludic engagement [28], making play a legitimate goal in and of itself, pointing to the richness that play around technology brings into people’s lives. ICTD, in its more consumer focused facet, understanding and harnessing the potential of personal ICTs such as mobile phones, in what Heeks named ICT4D 2.0 [33], is also bordering on 10 years. This is the right time to review and understand how play has been treated within the community, and point to future directions for a more honest understanding of ICTD work, as well as the importance of recognizing individual and local agency through play. We need an understanding of digital technologies where play, gaming and entertainment are recognized legitimate activities, resisting the ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION Whenever we talk about ICT use, it is hard not to think about its gaming and entertainment potential. It is certainly self-evident that these aspects have been a major driver behind the success and mass adoption of ICTs worldwide (see for instance Sey and Ortoleva for a short review [75]). While apparently evident, we fail to see the acknowledgment of those activities as ends in themselves, intrinsically valuable and even as basic freedoms we enjoy as humans. Recognizing playfulness as it unfolds in everyday life becomes all the more challenging in Information and ComCopyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21264 41 temptation of marginalizing these forms of interaction, or legitimizing them only insofar they permit the achievement of other, external, goals. This is key for understanding the importance and desirability of digital technologies in people’s lives. “Inside the play-ground an absolute and peculiar order reigns. Here we come across another, very positive feature of play: it creates order, is order. Into an imperfect world and into the confusion of life it brings a temporary, a limited perfection.” [40] (p.10) We begin with a broad understanding of play, as an academic concern, so as to derive a common, everyday understanding of the term to be used in analyzing its role in ICTD work. We then build on Nussbaum’s notion of play as a basic capability [59] framing it under Sen’s notions of instrumental, constructive and constitutive roles [71] and drawing implications for understanding previous work in ICTD. Our aim is to acknowledge play as an activity people have a reason to value in itself, not contradictory with other ICTD goals. We discuss ways to reframe our thinking about play in ICTD, away from more paternalistic discourses, to repurpose the play concept as a valuable resource. For Huizinga, play is one of the most important and foundational activities humans (and animals) engage in. When Huizinga writes “not serious” it is to place it in contrast with the many requirements of everyday life. Play is of primary importance, particularly through its “limited perfection” and “order”, which stands in stark contrast with other everyday notions of play that emphasize its triviality. More recent scholars, like Brian Sutton-Smith, have emphasized the importance of play as an indispensable mechanism for survival [82]. And in their in-depth analysis of the different types of play existing in humans and animals alike, Brown & Vaughn focus on the importance of play for a normal development of the brain, and a transformative force, capable of improving people’s lives dramatically and vital for survival: “the opposite of play is not work, it is depression” [11]. WHAT DO WE MEAN WHEN WE TALK ABOUT PLAY? Play is an elusive concept, it is discussed from many different perspectives, and, as pointed out by sociologist Thomas Henricks, it suffers from the cantoning of disciplines around the concept [35]. Once more prominent within the domain of psychology and highly influential child development studies by Jean Piaget [61], it has captivated interest from a variety of disciplines, such as anthropology [30, 49] or sociology [35]. Henricks argues for revisiting the classical authors, less constrained by the tranches of academia, in order to get a more unified understanding of play. We proceed to a brief overview of play, to help us build an understanding, useful for thinking of play within ICTD. One of the concepts more closely tied to play is that of games. Huizinga in fact refers to play, mainly (but not only) as an action one does with games. If play is a challenging concept to unpack, game is certainly not easier. Roger Caillois categorizes play within four different types [13], building on the Greek terms: (1) Agon, referring to games of a competitive nature, (2) Alea, pertaining to games of chance, such as casino games, (3) Mimesis, referring to role playing and (4) Ilinx, which encompass aspects of extreme sports that provide adrenaline rushes and even taking hallucinogenic substances or riding rollercoasters. Games are one of the most extensively studied forms of play, both in academia in general [13, 40, 81], as well as around technology more specifically [42, 52, 69, 83]. We discuss games throughout, in the common, everyday understanding of the term, encompassing all these different forms. But it is worth emphasizing that not all play comes in the form of games. The origins and importance of play Dutch Cultural Historian Johann Huizinga, in his seminal work Homo Ludens [40], provides us with the first academic text focused exclusively on play. For Huizinga, play is not an element within culture, but rather a foundational aspect that precedes culture itself. In his words (the bold is ours for emphasis on foundational aspects): "Summing up the formal characteristic of play, we might call it a free activity standing quite consciously outside 'ordinary' life as being 'not serious' but at the same time absorbing the player intensely and utterly. It is an activity connected with no material interest, and no profit can be gained by it. It proceeds within its own proper boundaries of time and space according to fixed rules and in an orderly manner." [40] Free play For artist Allan Kaprow, the distinction between playing and gaming is actually an important one: “This critical difference between gaming and playing cannot be ignored. Both involve free fantasy and apparent spontaneity, both may have clear structures, both may (but needn’t) require special skills that enhance the playing. Play, however, offers satisfaction, not in some stated practical outcome, some immediate accomplishment, but rather in continuous participation as its own end. Taking sides, victory, and defeat, all irrelevant in play, are the chief requisites of game. In play one is carefree; in a game one is anxious about winning” [44] Huizinga lays down the foundations for what most have picked on to discuss play, namely: (1) that it is free, bounded by no other material interest than the play itself; (2) that it is explicitly non-serious, standing outside ordinary life; (3) that it absorbs the player; and (4) that it has boundaries in time, space as well as rules. For Huizinga, and for most authors who have discussed the concept, as we will see, these are important concepts that demarcate play: For Kaprow, even goals within the game itself (what Suits calls prelusory goals [81]), such as winning the game, a 42 goal, for Huizinga already separated from everyday life defined only through the willingness of players to accept those goals for the sake of the game playing itself. Kaprow is interested in liberating the player further, and is only willing to consider play in the form of free play, an exploration free from worries and goals whether internal or external to itself. It is not just an escape from ordinary life, but it is an escape from other forms of stress or anxiety. gaming, such as interpersonal communication and social networking. Studies such as Wyche’s around Facebook use in Kenya [86], amongst others [19, 22, 39] show the importance people ascribe to these uses, which should not be necessarily striking as they are otherwise prevalent amongst users in the Western world as well. How and why we use play While there is value in these categorizations and discussions, we will use play as encompassing most of the categories above. What is important here is an understanding of activities, which share some, or all of the following aspects: Most work regarding the importance of free play comes from psychology, and more specifically child psychology. This interest originates more prominently in Jean Piaget’s early work on children’s stages of development and the role of play as the motor through which children explore and derive their understanding of the world, as well as an important tool in education [61]. While Piaget’s focus was centered around child play, and not adult play for instance, his concept of play remains close to the one of Kaprow’s ideas of liberation through freer exploration and engagement with the world, and as a fundamental mechanism for development and learning [31]. Absorption, flow, autotelic and personal engagements An important aspect for understanding the play experience, and the desire to partake in it, is the often-mentioned absorption of the player, since all the way back to Huizinga’s characterization [40]. This aspect has again been more extensively studied within psychology, namely in reversal theory. Reversal theory tries to understand motivation and draw distinctions between the concept of paratelic, denoting the kinds of activities which are pursued as means to an end, and autotelic, denoting those engaged in for the enjoyment in the process itself [2]. Autotelic engagements denote moments in which people are absorbed in activities, escaping worldly, everyday realities. This has been discussed by Csikszentmihalyi’s through his influential work on flow [17], denoting that state in which one is absorbed into an intrinsically rewarding activity, pursuing it intensely for its own sake. While Csikszentmihalyi describes it as a personal feature, it is often taken, and easily translatable as a feature of the activities themselves: • They are engaged in freely, meaning its something people value and engage in willingly, outside of external pressures to do so. • They stand outside of ordinary life, in that they provide escape from the everyday routines and hardships. This is related to them being bounded in space and time, implying a demarcation from other activities, even when these are not always evident. • They are pursued for their own sake, meaning that they possess an autotelic or non-instrumental value, rather than serving an external purpose. This is often coupled with a state of flow, or intense absorption. These do not represent a strict set of necessary conditions to classify something as play, but rather guidelines to understand activities regularly placed under play, such as games, entertainment, fun or leisure. We will use these rather loosely, and at times include activities which may seem contrary to some of these principles, but which should still reasonably be understood as essentially playful. Other ways of classifying the same phenomena could be under unserious or non-instrumental, but using either of these, and other terminologies, ultimately undermines the focus of this work. For instance talking about these activities as noninstrumental implies a degree of uselessness, contrary to our main argument that these are the basis for leading a good life. Equally, to call them unserious would be, deny the serious of play itself, and its constitutive role as an essential freedom, as philosopher Kurt Riezler expressed: “An autotelic person needs few material possessions and little entertainment, comfort, power, or fame because so much of what he or she does is already rewarding. Because such persons experience flow in work, in family life, when interacting with people, when eating, even when alone with nothing to do, they are less dependent on the external rewards that keep others motivated to go on with a life composed of routines. They are more autonomous and independent because they cannot be as easily manipulated with threats or rewards from the outside. At the same time, they are more involved with everything around them because they are fully immersed in the current of life.” [17]. “Man’s playing is his greatest victory over his dependence and finiteness [...]. He can also play, i.e., detach himself from things and their demands and replace the world of given conditions with a playworld of his own mood, set himself rules and goals and thus defy the world of blind necessities, meaningless things, and stupid demands. Religious rites, festivities, social codes, language, music, art – even science – all contain at least an element of play. You can almost say that culture is play. In a self-made order of rules man enjoys his activities as the maker and master of his own means and ends. Take art. Man, liberated from what you call the real world or ordinary life, plays with a world of rhythms, sounds, words, colors, lines, enjoying the It is also building on these notions of intrinsically valued engagements, that we build our understanding of play. These may manifest themselves in very distinct ways than 43 triumph of his freedom. This is his seriousness. There is no ‘merely’.” [67] ment started with the carving of a literal Hole in The Wall between Mitra’s workplace, NIIT and an adjoining slum in New Delhi, known as Kalkaji [53], where a computer was put to use freely. This experiment proved to be a widely mediatized success story of free learning and playing by the children in the slum, who with only “minimal guidance” derived multiple useful interactions from the computer, achieving a number of measurable educational outcomes [53, 54]. This has then been scaled to hundreds locations and has reached thousands of children in India and Africa [37]. Part of the project’s success is the emphasis on how entertaining and motivating content contributed to the successful educational experiment. Play as capability Capability theory is a welfare economics concept introduced by Nobel Laureate Amartya Sen, and it stands at the root of arguably the most widely used measure of achievement of human needs, the Human Development Index (HDI) [1]. The HDI is closely tied to the Millenium Development Goals (MDG) [62], which constitute the basis established by the United Nations (UN) to guide development work in general, and as a consequence, ICTD efforts. The development of capabilities is, according to Sen, based on the idea that development should focus on enhancing people’s abilities to choose the lives they have reasons to value. One of the most notable efforts in this area since then has been led by Matthew Kam in designing for, and studying, the impact of digital gaming, particularly on the mobile, in English language education [42]. Long term trials have reported significant successes [43]. Studies in ICTD have praised the potential for games and leisurely activities in supporting educational goals, from literacy goals, to mathematics [5], with games being early on lauded as a good educational tool through its engaging function [29]. Play and games for education are arguably the most represented within ICTD understandings of play and games, and fall under the rubric of serious games [52] which is also an important field within technology studies at large. Martha Nussbaum, who worked with Sen in the development of capability theory, and is arguably the main authority on the subject, has written briefly about play as a capability. Nussbaum briefly describes play as “Being able to laugh, to play, to enjoy recreational activities.” [58] Nussbaum however, has not published as extensively on the topic of play as capability, with some notable exceptions discussing how some, in particular women, may lack opportunities for play and some of the consequences: “Often burdened with the “double day” of taxing employment and full responsibility for housework and child care, they [women] lack opportunities for play and the cultivation of their imaginative and cognitive faculties. All these factors take their toll on emotional well-being: women have fewer opportunities than men to live free from fear and to enjoy rewarding types of love—especially when, as is often the case, they are married without choice in childhood and have no recourse from a bad marriage.” [57] Job training Education is not the only field in which games and entertainment can have a positive, instrumental, role. Polly is a telephone based voice manipulation and forwarding system deployed in Pakistan by Raza et. al. [66]. It has had measurable impact during its large scale and long-term deployment with nearly half a million interactions and 85.000 registered users at the time, and showing considerable tendency in growth. The authors used entertainment to draw people to the service where other development services could be provided. They wanted to ask the questions: It is under this frame that we understand the importance of play, not as a superfluous activity, but as an essential capability towards living a good life; engaging in and exerting the freedoms that one has reasons to value. Building on Sen’s terminology, we will look at the role of play as instrumental, i.e. as a means-to-an-end, its constructive role, i.e. in building a consistent and honest interaction between researchers, designers and users. Finally, we arrive at our main goal here: how we can understand its constitutive role, i.e. as an unapologetic end-in-itself. “(1) Is it possible to virally spread awareness of, and train people in, speech-based services, in a largely low-literate population, using entertainment as a motivation? (2) Is it possible to leverage the power of entertainment to reach a large number of people with other speech-based services?” [65] Their evaluation turned out largely positive with the high rates of adoption and job creation, as well as plentiful evidence that users were extracting entertainment and leisurely value from the service, beyond the development offer [65]. PLAY AS INSTRUMENTAL Play, and games in particular, have been discussed in ITCD mainly given their instrumental roles. We will now look at the work done in this area, which represents the main contributions to the understanding of play within ICTD, drawing on examples from education, job searching and health. Non-prescribed uses Within health applications in ICTD, Schwartz and colleagues worked with the deployment of mobile devices to health workers in India. These devices were intended for purposes of collecting health data in eight different projects [70]. They note how workers also used these devices for Within education Perhaps the most famous and widely covered experiment with technology and education in development is Sugata Mitra’s Hole in the Wall (HiW) experiment. This experi- 44 more personal goals, beyond the intentions of deployment, calling these ‘non-prescribed uses’ [70]. The authors reflect on how explicitly allowing for these non-prescribed uses, helps to have a more honest engagement with the workers themselves. These did not feel the need to hide such behaviors, and it even worked as an extra motivation to participate in the project itself. The argument that Sen makes is for how democracy, as a capability, contributes to an ongoing dialogue, allowing a constant reorganization of priorities involving all levels of governance, rather than strictly thought out and imposed top-down. Development and aid theorist William Easterly, in his influential book “The White Man’s Burden” [20], explains how the poorest, which are often the ones targeted by aid and development interventions, are not engaged in this valuation process. Easterly argues that top-down aid and development goals are essentially anti-democratic, since the ones at the bottom do not dispose of any feedback or accountability mechanisms to change priorities at the top, such as voting or effective lobbying. In this example, we extend the notion of play to include much of the personal use of technologies more broadly, particularly in a work context. It is about respecting people’s desire to use these technological possibilities for their own purposes, the freedom to use them for stepping out of everyday work tasks, creating a playground beyond the instrumental value of the tasks they were given the devices to accomplish. Rather than fighting these uses, the project leaders figured that simply allowing for them created a better engagement with the health workers, leading to increased efficiency for the overall project. We do not have a solution for tackling the complexity and breadth of aid, development, or more specifically, ICTD initiatives with regards to their democratic framing and accountability. We argue that, within ICTD, the downplaying of playful engagements, which are, as we saw, prominent, and have the characteristics of freedom and voluntariness, harms communication and acknowledgment of people’s desires and value judgments, and, as a consequence, increase communication problems within the different layers of ICTD work. We will now look at how these problems have emerged, and have started being discussed within ICTD, as well as implications for the evaluation of ICTD projects and dynamics more broadly. We have seen how games and entertainment can bring about positive benefits as educational strategies. We have seen how they can help in engaging people with technological interventions, as well as providing motivation for users participating in other projects, by not restricting some of the non-prescribed uses. While we find this type of instrumentalization of play a rather narrow view on the phenomena, it is nonetheless one that should be respected, since it helps steer away from a vision where play may harm or detract from development goals. Echoing Sen’s discussion on democracy and development not being contradictory goals (as is sometimes assumed) and the idea of the “presumed existence of a deep conflict […] [when] studies show how both are important and have no conflict” [71] (p.149-150). Like Sen, we discuss this to help alleviate some of the concerns which emerge around play, and which we will see in more detail. But this should not detract from the importance of play in other ways, beyond and regardless of whether it achieves other goals or not. ICTD and play as failure In what is one of the first pieces in ICTD to reflect on these aspects, Ratan and Bailur provide us with examples of ICTD projects, which were classified as failures, despite, on a closer look, having brought significant value to people’s lives. One such project is a community radio and telecentre known as OurVoices, an UNESCO funded initiative which installed a significant amount of ICTs in a village called Bhairavi for the purpose of disseminating what was considered “relevant information”, and to examine ICT’s impact on reducing poverty and enhancing development. While initially branded a success in improving different metrics from women’s health to job creation, it quickly fell under scrutiny: PLAY AS CONSTRUCTIVE One aspect, which is not often mentioned when discussing, play, but with important consequences, is what Sen describes as (specifically referring to democratic rights) the “constructive role in the genesis of values and priorities” [71]. That is the right for people and communities to openly discuss and reorganize what values are important, and what goals to prioritize. This needs to be a constant, deliberative process that involving the affected, or targeted, communities. As Sen puts it: “We were told by the villagers that the radio set medium had been phased out soon after implementation (one of the reasons given was that the villagers started taking the radios out to their fields and listened to FM radio instead of OurVoices). During research listeners dismantled one of the radio loudspeakers in protest and used it to accompany the procession of the statue for a religious festival. The NGO’s reaction to this was that the people were ignorant and uninterested in their own “development”. They removed all the cabling, and set up the loudspeaker in another village” [64] “The need to discuss the valuation of diverse capabilities in terms of public priorities is, I have argued, an asset, forcing us to make clear what the value judgments are in a field where value judgments cannot be – and should not be – avoided [...] The work of public valuation cannot be replaced by some cunningly clever assumption.” [71] (p.110) The authors proceed to detail a number of consequences that ensued from this situation and causing significant distress to the local population. Ratan and Bailur discus yet 45 another project, Hole in the Office (HiO), which dealt with the introduction of PCs in a community, to aid with job searching [60]. Discontinuation of HiO occurred after it was determined that it was being used 25 times more for entertainment and games than for its intended function, thus satisfying no development goal: This is an issue, which affects ICTD’s understanding of ICT adoption dynamics and motivations. To illustrate this mismatch, Heeks discusses a survey conducted in Tanzania reporting that fewer than 15% of mobile owners believed that the benefits of owning mobiles justified the costs [56]. To this, Heeks asks: “Um . . . so if you believe that guys, why on earth do you own a mobile?” [34] Heeks is suggesting that a richer story is waiting to be told. A story that was not reported, either because there was no way to get at the rationale for the actual reasons leading people to make that investment, or because it is the kind of information that ICTD, with its narrower focus on pre-defined development goals, is in a way not always prepared to assimilate. “Yet, across the board the perception was that the PC was a “good” thing and that free access should not be discontinued” [64]. What these examples tell us is what Sey and Ortoleva have called “negative perceptions” of play within ICTD [75]. These perceptions lead to situations in which behaviors, that are not intended to occur around the specific ICTD intervention, are dismissed or censored even if they are highly valued by the communities. This has implications for people’s agency, defined by Sen: Ferreira and Höök have discussed these tensions in their work around mobile phone adoption in Rah Island, Vanuatu [24]. They focus on tensions not just between ICTD and the communities, but within the communities themselves, leading to obstacles in reporting and getting a deeper understanding as to people’s motivations, and indeed rightful desires, to acquire and engage with modern ICTs. These tensions should not be fed within ICTD dynamics, if we want an honest engagement with the recipients of these development projects, and an understanding as to what motivates the spending of significant amounts of resources in acquiring these, as Heeks explains: “I am using the term ‘agent’ [...] as someone who acts and brings about change, and whose achievements can be judged in terms of her own values and objectives, whether or not we assess them in terms of some external criteria as well” [71]. This highlights tensions between communities and individual’s values and goals, and external criteria for assessment, which, as we will now see, bears some damaging consequences for ICTD dynamics. “The significant amounts being spent by the poor on mobiles indicate that phones have a significant value to the poor […] we [the ICTD community] have long known […] that ‘poverty’ is not just about money and, hence, that poverty interventions and tools can usefully target more than just financial benefits” [34] Implications for evaluation The issue of downplaying the role of play within ICTD is not just one of overlooking community and individual agencies, as problematic as that is, but has implications for the evaluation of projects as well. Ramesh, the project manager of OurVoices explains why it is so hard to get an understanding of what is actually happening in ICTD projects such as OurVoices: This tension between perceived value and the financial efforts made by the poor to acquire these technologies is present in other settings, as pointed out by Song, where according to him people can spend more than 50% of their income on personal access to ICTs despite this representing an effort some would consider excessive [79]. The question is whether it is reasonable to simply assume irrationality in such behaviors, and subsequently risking censoring them, or whether there is something to learn from these uses. “It is hard to know if people are really listening. In a survey, if we …ask whether they watch TV or listen to us, they say yes. […] The minute they see us, they tell us what we want to hear. They say yes, yes we listened. They feel guilty for choosing entertainment over development, like something which is good for them.” [64] If we believe the latter to be the case, then we need to address the dynamics that exist within ICTD, and development work at large, between funding agencies, practitioners and recipients, that are partly responsible for undermining the appreciation of these more complex and irreducible, yet entirely legitimate forms of engaging with technology. The tension between what “insiders” may legitimately desire and what they perceive “outsiders” want to hear has been documented in other places [16, 36, 55] and may at times prevent even a basic understanding of the situation. This has been documented in developing work [16, 55], and has been more generally known to be an already existing characteristic of trials [10], which should not be encouraged, if we want to elicit a rich understanding of the communities as well as the ICTD projects themselves. These concerns are echoed by Kuriyan and Toyama, remarking that “what rural villagers want and what we think they need are frequently different” [46]. Summary Capabilities, according to Sen, should be: “Deliberative in order to accommodated ongoing discussions regarding priorities and new information, rather than taking any predefined metric (seducing as it may be) as an ultimate indicator of well-being, poverty or freedom of different kinds” [71] 46 If we are to return to Schwartz and colleague’s recognition and acceptance of non-prescribed uses [70], we see how it not only helped achieve the project’s goals, but also generated a more open and honest interaction between the different stakeholders. This in turn also helps avoid entirely undesirable feelings of guilt, as brought up by Ramesh from OurVoices, stemming from people using technology in other ways than their instructed use. These, in turn, may hide, what would otherwise be important aspects of technology use, which should be taken into consideration when trying to understand ICT usage and even, to some extent, explicitly catered to. The intentions and desires behind technology adoption, which often involves heavy financial sacrifices such as those Heeks [34] or Smyth and colleagues [77] have documented, remain underappreciated, given the failure to account for and respect people’s legitimate desires. And this is the most important aspect to keep in mind: not only this tension affects ICTD goals by undermining the relationships between the different parties, but, on a more fundamental level, they are highly prescriptive of images that some ICTD projects entertain about their participants, as Ramesh explains: field where this has been most emphasized is within psychology where from Piaget’s early work [61], the importance to stimulate and encourage play amongst children is an uncontroversial topic within the field. More recently psychologist Peter Gray has published extensive work on the topic, summing up the views of most cognitive psychology on the issue, arguing that, play is not only fundamental, and an important tool for learning, but also something to be unleashed [31]. The Convention on the Rights of the Child has long stated in its article 31: “States Parties recognize the right of the child to rest and leisure, to engage in play and recreational activities appropriate to the age of the child and to participate freely in cultural life and the arts.” [4]. While acknowledging the right of play for children is undoubtedly an important step, even if it is far from being a reality throughout the globe, it should not detract from the fact that play is important regardless of age. The desire to play We have seen an emergence of a, yet sparse, body of work which has began documenting playful behaviors around technology in ICTD. We have seen how South African mobile users have appropriated MXit, an instant messaging system, for the purposes of gaming [83], and other examples abound of such appropriations within the developing world [15, 75, 77]. Social networks and interpersonal communication have also been source of enthusiasm [70, 83, 86]. Larger surveys show users in the developing world engaging with gaming in internet cafes and other points of access [45, 74], and more generally many have discussed the significant amounts of resources that users, even within resource constrained situations, are willing to place in entertainment and gaming [26, 34, 63]. “We might be giving a programme about an agricultural scheme, which the government might have for him, which might significantly increase his yield, but he’s not interested in listening to it, because it’s boring. You know, he wants to watch the movie. That’s the competition we’ve got, the challenge we have to overcome.” [64] As a final note, the constructive role of play is, in a way, instrumental, since we frame it as a means towards achieving other goals (such as better evaluation and feedback loops). But because it can be a broad means towards negotiating the goals and values of development themselves, rather than achieving them directly, it is important to, as Sen does, view this role under a different light, more tied to the idea of freedom itself, rather than a narrower conception of social and economic progress. This desire for play, entertainment and personal uses in the developing world should not be striking, as this accounts for a significant portion of technological use outside the developing world as well. It is up to ICTD to acknowledge and study what people do, and want to do with technology, rather than ascribing a priori intentions. Here HCI can provide some guidance, where studies have emerged of technology use, studied and understood, beyond instrumental framings, and rather as an appreciation of legitimate technology use: for instance Juhlin and Weilenmann’s work on hunting [41] or Jeffrey and Shaowen Bardzell’s work on sex toys [6]. Both of these focus on the activities themselves, helping support the growing interest in those domains. However, we also find also some obscurity around these issues in HCI as well, as Genevieve Bell pointed out in her keynote at CHI 2010 [7]. Some similar resistances exist in appreciating activities for their own sake, with activities such as sports, sex or religion remaining largely underrepresented in research in comparison to their prominence within actual ICT use. PLAY AS CONSTITUTIVE We have devoted some time here to explain the importance of play from an instrumental and constructive viewpoints, partly in order to alleviate some concerns that may stem from an often – and wrongly – assumed contradiction between the triviality of play and the seriousness of other goals, such as development ones. Our main point in this work however, and what we will argue now in more detail, is that play is intrinsically important: its constitutive role as an integral part of the human experience. This is regardless of whether or not it helps achieve other goals, such as the ones discussed above. We have seen how most academic thought overwhelmingly speaks of play as a fundamental and innate aspect of life. From Huizinga’s initial framing of play, as the basis from which culture emerges [40], to Nussbaum’s classification of play as a basic capability in humans [59]. Arguably the We have seen how play occurs even around the simplest digital interactions [24], and how people engage in extensive work to creatively appropriate technologies for their 47 own desired uses, regardless of how limited the opportunities for play seem to be. But we argue that people should not have to work so hard to derive pleasure and other value from technologies which many of us take for granted, and that we should complement and assist, rather than resist, the motivations and desires that many communities are already very visibly displaying around ICT adoption. adopt ICTs, regardless of their socio-economic situation, than perpetuating play’s relative obscurity within ICTD work. A failure to understand motivations behind ICT use is likely to result in a failure to provide real benefits to those we target through our studies and interventions. DISCUSSION We have discussed the importance of looking at play as a capability within ICTD work, and analyzed its instrumental, constructive and constitutive roles, positioning relevant work according to this framework. In order to help alleviate some of the issues around the acknowledgment of play within ICTD, we propose some broader topics for discussion: (1) the dynamics of paternalism and a priori prescription of goals existing in ICTD, (2) how play has a narrative and rhetoric role within ICTD, and development more generally, which influences the way it is discussed, prioritized or accepted. We end by (3) discussing how this is not just an issue of designing, or deploying, playful technologies in ICTD, but rather a broader discussion that needs to take place in order to reform the term of play as, essentially, a quality of life and freedom issue. Taking play seriously There are two further examples we would like to bring up as part of emphasizing the non-triviality of play (although the play activities may themselves be trivial). In severely resource-constrained situations, play does not become a luxury one cannot afford. In many, rather severe, cases play becomes all the more important as a means to escape current hardships, particularly through the aspects of escape and order that play brings to life. It is in this way we would like to discuss these two important examples. The first comes from the Great Depression era, in the United States, where a growth in resources allocated to leisurely activities grew, rather than diminish, during the period between 1890 and 1940 [25], across socio-economic backgrounds. The second, and one of the most striking examples of this is given to us by George Eisen in “Children and Play in the Holocaust”, where Eisen tries to reconcile, and come to terms with, the horror of the situation and the importance of play among children [21]. In his work it is clear how easily one may completely ignore the importance and role of play in those situations. Eisen describes play as it occurred, dissecting its role within that context, without underappreciating or downgrading the urgency and deep precariousness of the situations those children found themselves in. It was, according to Eisen, precisely play, which allowed children to preserve a level of humanity throughout their plight. Paternalism and prescription Many have denounced the paternalistic dynamics that occur within development work [15, 32, 64]. Anthropologists like Arturo Escobar have discussed the making of concepts like “development” and “third” world, as narrow conceptions of the people described, as passive receivers of external knowledge which is good for them [23]. Manzo and colleagues discuss these dymanics: “What political economy wanted, in short, was to take the poor, inefficient Third World decision maker by the hand, lead him to the development candy store, and show him how to get the best buy for his meager pennies.” [50]. Ratan and Bailur also discuss this, through their recounting of Helena Norberg-Hodge’s [36] ethnographic fieldwork and her experience of coming back to Ladakh, India ten years after having been told by a local that there was no existing poverty in that location, only to find that same person begging tourists for help and emphasizing her situation of poverty . While tensions around the appreciation of play may appear particularly acute in ICTD settings, they have resonances in different moral tales present in the Western world as well. From Aesop’s fable of the grasshopper and the ant [14], in which the grasshopper finds its demise, after spending all summer singing and not preparing adequately for winter, while the ant thrives in its dutifulness (Suits dedicates a whole book defending the grasshopper’s lifestyle as the only way to lead a good life [81]), to for instance ideas about work ethics as a trait of cultures or religions (such as Max Weber’s treatise on protestant work ethics [84]). Within popular and academic thought, multiple traditions exist that reinforce these strict prioritizations between activities, most famously embodied by Mazlow’s pyramid of needs. What is at stake here is not just the dynamics that are created between the recipients of aid, donors and development practitioners, but more broadly the construction of these settings, as Escobar argues [23]. The interactions that development work generate can yield possibly detrimental dynamics between recipients of development work, funding organizations and development practitioners, and as Smyth and colleagues write, we must question: “Are needs really more urgent than desires? Who defines which is which? Do researchers exaggerate the urgency of ‘needs’ due to their own biases and preconceptions?” [77] These, and other, moral tales shape the way we conceive of play, work, dutifulness and so on, and are likely to permeate ICTD and academic work. However, we need to question these assumptions by showing how notions of play and work, for instance, are not contradictory. We are often better served appreciating the ways in which people choose to Perhaps here we can learn from a deeper understanding of local needs and aspirations to inform the nature of ICTD interventions. To gain this deeper understanding, we may 48 borrow even more from traditions of anthropology and sociology, as suggested by Burrell & Toyama [12]. This may allows us to step away from some of the more top-down, pre-defined metrics (as suggested by Burrel & Toyama [12], Heeks [33] and Van Dijk [18], to name a few). But this is not only a discussion between top down vs. bottom up approaches for development, as embodied most notably by Jeffrey Sachs [68] and William Easterly [20]. Both of these approaches have their own benefits and limitations. The important aspect to keep in mind is that, to be effective, both will require constant loops of feedback and accountability to avoid reinforcing some of these representational and paternalistic dynamics. researchers are only a part of the structures that determine and conduct development and ICTD work. The same concerns of justifying technology use that Ramesh, from OurVoices, brought up, may not be so dissimilar to the ways in which ICTD practitioners and researchers must justify their own work, as Adam Smith famously remarked: “The interest of the dealers, however, in any particular branch of trade or manufactures, is always in some respects different from, and even opposite to, that of the public.” [76] This is not intended to ascribe malice or ill intentions onto ICTD institutions, researchers, practitioners and other parties involved, from whom we assume an a priori deep commitment to their work. This is rather a reminder that dynamics which emerge from these institutional arrangements may at times ignore those with less voice to lobby for their interests, as Easterly discusses [20]. By not acknowledging playful behaviors around technology we may be engaging in a poor representation of people, and as a consequence, assuming priorities without deliberation. This may appear particularly prominent in development work, where the main focus is on a stricter view of socioeconomic needs, as already established in the MDG. Feelings of shame or guilt, such as those brought up by Ramesh from OurVoices, further marginalize the understanding of these practices and undermine the informational base on which ICTD work is conducted, i.e. the capacity to understand what the values and needs are for the users. Insights on steering clear from paternalistic modes of development thinking are not necessarily a novelty introduced in this work. But we argue that the acknowledgment of play, as a legitimate activity valued regardless of socio-economic situation, challenging us to observe and respect these behaviors, can be a key element to help unlock some of the vices that development work can generate. These are part of larger structures and dynamics, which can be difficult to escape. These are engrained, as we saw in academic thought, religious morals, theories around work ethics and even in children’s fables. Play can mean different things, few of which deserve the ill treatment and depreciation that it often gets. We suggest a repurposing of the concept of play at all levels, to acknowledge these fundamental human forms of engagement, enriching academic thought more generally, and ICTD interventions more specifically. Play as a narrative/rhetoric device We have defined play according to a series of criteria that are not all necessary to classify something as play or not. It is rather a matter of family resemblances [85] between different activities which we reasonably qualify under play, rather than a perfectly defined concept. The ways in which we choose to frame and acknowledge certain activities as play, or not, has important repercussions. One repercussion might be that donors are unwilling to fund technologies for entertainment if that is how they are framed rather than to achieve socio-economic outcomes for instance. As project manger Ramesh (from the OurVoices project) mentions when queried as to the discrepancies between the intended goals of the project, people less than serious activities and the difficulty at getting at these: Not (just) a design issue but a larger debate There are certainly better and worse ways to design play experiences with and around digital technologies, such lessons appear more prominently within game studies [69]. One important issue outside the gaming world, and which has inspired play thinking in HCI is Sengers’ and Gaver’s proposition for open interpretation [73] and Höök’s proposition for designing open surfaces [38], as a way to allow for a freer engagement with technology. “We can either approach community radio as what the community wants. If you make it that way, it will be music only. But at [the donor agency] we can’t justify all this equipment to play music all day. There has to be a development angle.” [64] Within an ICTD context, some such as ourselves have suggested looking at HCI, as a more mature, design oriented, discipline aimed at understanding people and technology, for design recommendations for play in ICTD, such as openness [15, 51]. While this certainly has some value, it is also important to remember that many play experiences, like games, are often extensively and carefully crafted to provide enhanced experiences of their own. In that sense openness is not a panacea to play, and both kinds of design strategies have a place when providing people with interesting and engaging playful experiences. The implications of framing something as important or urgent can result in different levels of attention paid, and funding awarded, to different projects and communities. Smyth and colleagues ask: “Do researchers exaggerate the urgency of ‘needs’ due to their own biases and preconceptions?” [77] While certainly biases and preconceptions from researchers play a part in how projects are conducted and evaluated, 49 Regardless of strategy, design alone can hardly challenge the existing institutional, moral and ethical structures, discussed earlier, that obfuscate play activities. These contribute to a perception, as Sutton-Smith describes, of other activities, as being “wiser” [82]. The work that must be done, is not just the technical work of designing these experiences, as important as that is, but a questioning of priorities and a constant renegotiating of these priorities with the people who are supposed to benefit from these technologies. It involves an ongoing dialogue, while steering away from a priori moral imperatives, which do not even seem to match with the ways in which people value digital technologies, both within ICTD contexts and at large. [2] [3] [4] [5] [6] Openness in designing systems can certainly be one of the mechanisms, amongst other ways of crafting and designing for play, with which to provide positive and desirable experiences with digital technologies. But we argue that openness as a value may be more important in crafting the ICTD projects and interventions themselves rather than the specific technologies, so that these can accommodate new insights, encouraging mutual deliberation and learning [47]. [7] [8] By resisting a priori goals, shifting the value of the term play and focusing more broadly on the interventions, rather than mainly on the systems’ designs, we hope to help alleviate some of the tensions around this topic, thinking more broadly about ICTD work and involving those who the work will affect, in more open deliberations over goals and priorities. Play is as an important capability to enhance, it should not be a source of guilt, but rather representing a large portion of what we value in technology and life. [9] CONCLUSION [12] [10] [11] We have seen how play, as a capability, helps us appreciate its different roles (instrumental, constructive and constitutive), showing how there is no inherent conflict between play and development, and how the acknowledgment of play is fundamental for a more honest and rigorous engagement with ICTD work. A dialogue around play will always exist within societies and academia, setting rules and deliberating over its appropriateness, this deliberation process should be encouraged, not restricted. But most importantly, this framing allows us to place the focus on its constitutive aspect, valuing play regardless of other development goals – in the same way we should ensure gender equal opportunities and democratic rights despite any other social and economic rationale, as essential freedoms, allowing people to deliberate on, and live, their lives in the ways they have reason to value and see fit – unapologetically focusing on a better quality of life. [13] [14] [15] [16] [17] [18] [19] REFERENCES [1] Anand, S. and Sen, A. 1994. Human Development Index: methodology and measurement. Human Development Report Office (HDRO), United Nations Development Programme (UNDP). [20] 50 Apter, M.J. 1989. Reversal theory: motivation, emotion, and personality. Routledge. Arora, P. 2012. The leisure divide: can the “Third World”come out to play? Information Development. 28, 2 (2012), 93–101. Assembly, U.G. 1989. Convention on the Rights of the Child. United Nations, Treaty Series. 1577, 3 (1989). Banerjee, A. et al. 2005. Remedying education: Evidence from two randomized experiments in India. National Bureau of Economic Research. Bardzell, J. and Bardzell, S. 2011. Pleasure is your birthright: digitally enabled designer sex toys as a case of third-wave HCI. Proceedings of the 2011 annual conference on Human factors in computing systems (New York, NY, USA, 2011), 257–266. Bell, G. 2010. Messy Futures: culture, technology and research. Bødker, S. 2006. When second wave HCI meets third wave challenges. NordiCHI ’06: Proceedings of the 4th Nordic conference on Human-computer interaction (New York, NY, USA, 2006), 1–8. Boyd, D.F. and Krasnow, H.S. 1963. Economic Evaluation of Management Information Systems. IBM Syst. J. 2, 1 (Mar. 1963), 2–23. Brown, B. et al. 2011. Into the wild: challenges and opportunities for field trial methods. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (2011), 1657–1666. Brown, S. and Vaughan, C. 2010. Play: How It Shapes the Brain, Opens the Imagination, and Invigorates the Soul. Scribe Publications. Burrell, J. and Toyama, K. 2009. What constitutes good ICTD research? Information Technologies & International Development. 5, 3 (2009), pp–82. Caillois, R. 1961. Man, play, and games. University of Illinois Press. Calder, A. 2001. The Fables of La Fontaine: Wisdom brought down to earth. Librairie Droz. Chirumamilla, P. and Pal, J. 2013. Play and power: a ludic design proposal for ICTD. Proceedings of the Sixth International Conference on Information and Communication Technologies and Development: Full Papers-Volume 1 (2013), 25–33. Crewe, E. and Harrison, E. 1998. Whose development. An ethnography of aid. (1998), 23–65. Csíkszentmihályi, M. 2008. Flow: The Psychology of Optimal Experience. HarperCollins. Van Dijk, J.A. 2006. Digital divide research, achievements and shortcomings. Poetics. 34, 4 (2006), 221–235. Donner, J. 2009. Blurring livelihoods and lives: The social uses of mobile phones and socioeconomic development. innovations. 4, 1 (2009), 91–101. Easterly, W.R. 2006. The White Man’s Burden: Why the West’s Efforts to Aid the Rest Have Done so Much Ill and so Little Good. Penguin Press. [21] Eisen, G. 1990. Children and play in the Holocaust: Games among the shadows. Univ of Massachusetts Press. [22] Ellwood-Clayton, B. 2006. All we need is love—and a mobile phone: Texting in the Philippines. Cultural Space and Public Sphere in Asia, Seoul, Korea. (2006). [23] Escobar, A. 2011. Encountering development: The making and unmaking of the Third World. Princeton University Press. [24] Ferreira, P. and Höök, K. 2012. Appreciating pleiplei around mobiles: Playfulness in Rah Island. Proceedings of the 2012 annual conference on Human factors in computing systems (New York, NY, USA, 2012). [25] Fischer, C.S. 1994. Changes in leisure activities, 1890-1940. Journal of Social History. (1994), 453– 475. [26] Galarneau, L. 2014. Global Gaming Stats: Who’s Playing What, and Why. Big Fish Games. (2014), 1– 5. [27] Gallagher, C.A. 1974. Perceptions of the value of a management information system. Academy of Management Journal. 17, 1 (1974), 46–55. [28] Gaver, W.W. et al. 2004. The drift table: designing for ludic engagement. CHI ’04 extended abstracts on Human factors in computing systems (New York, NY, USA, 2004), 885–900. [29] Gee, J.P. 2003. What video games have to teach us about learning and literacy. Computers in Entertainment (CIE). 1, 1 (2003), 20–20. [30] Geertz, C. 1991. Deep Play. Rethinking Popular Culture: Contemporary Perspectives in Cultural Studies. (1991), 239. [31] Gray, P. 2013. Free to learn: Why unleashing the instinct to play will make our children happier, more self-reliant, and better students for life. Basic Books. [32] Gumucio-Dagron, A. 2001. Making waves: Stories of participatory communication for social change. Rockefeller Foundation New York. [33] Heeks, R. 2008. ICT4D 2.0: The Next Phase of Applying ICT for International Development. Computer. 41, 6 (2008), 26–33. [34] Heeks, R. 2008. Mobiles for Impoverishment? ICTs For Development. [35] Henricks, T.S. 2006. Play reconsidered: Sociological perspectives on human expression. University of Illinois Press. [36] Hodge, H.N. 2013. Ancient futures: learning from Ladakh. Random House. [37] Hole-in-the-Wall: 2011. http://www.hole-in-thewall.com/insight.html. Accessed: 2015-01-24. [38] Höök, K. 2006. Designing familiar open surfaces. Proceedings of the 4th Nordic conference on Humancomputer interaction: changing roles (New York, NY, USA, 2006), 242–251. [39] Horst, H.A. and Miller, D. 2006. The cell phone: an anthropology of communication. Berg. [40] Huizinga, J. 1939. Homo Ludens: A Study of the Play-Element in Culture. Routledge. [41] Juhlin, O. and Weilenmann, A. 2008. Hunting for fun: solitude and attentiveness in collaboration. Proceedings of the 2008 ACM conference on Computer supported cooperative work (New York, NY, USA, 2008), 57–66. [42] Kam, M. et al. 2008. Designing e-learning games for rural children in India: a format for balancing learning with fun. Proceedings of the 7th ACM conference on Designing interactive systems (2008), 58–67. [43] Kam, M. et al. 2009. Improving literacy in rural India: Cellphone games in an after-school program. Information and Communication Technologies and Development (ICTD), 2009 International Conference on (2009), 139–149. [44] Kaprow, A. 2003. Essays on the Blurring of Art and Life. Univ of California Press. [45] Kolko, B. et al. 2014. The value of non-instrumental computer use: Skills acquisition, self-confidence, and community-based technology training. TASCHA. [46] Kuriyan, R. et al. 2008. Information and communication technologies for development: The bottom of the pyramid model in practice. The Information Society. 24, 2 (2008), 93–104. [47] Loudon, M. and Rivett, U. 2014. Enacting Openness in ICT4D Research. Open Development: Networked Innovations in International Development. (2014), 53. [48] Lucas, H.C. 1973. User reactions and the management of information services. Management Informatics. 2, 4 (1973), 165–172. [49] Malaby, T.M. 2007. Beyond play a new approach to games. Games and culture. 2, 2 (2007), 95–113. [50] Manzo, K. 1991. Modernist discourse and the crisis of development theory. Studies in comparative international development. 26, 2 (1991), 3–36. [51] Marsden, G. 2009. UNDER DEVELOPMENT Electronic tablecloths and the developing world. Looking ahead. 16, 2 (2009). [52] Michael, D.R. and Chen, S.L. 2005. Serious games: Games that educate, train, and inform. Muska & Lipman/Premier-Trade. [53] Mitra, S. et al. 2005. Acquisition of computing literacy on shared public computers: Children and the“ hole in the wall.” Australasian Journal of Educational Technology. 21, 3 (2005), 407. [54] Mitra, S. 2003. Minimally invasive education: a progress report on the “hole-in-the-wall” experiments. British Journal of Educational Technology. 34, 3 (2003), 367–371. [55] Mosse, D. 2001. People’s knowledge’, participation and patronage: Operations and representations in rural development. Participation: The new tyranny. (2001), 16–35. 51 [56] Mpogole, H. et al. 2008. Mobile Phones and Poverty Alleviation: A Survey Study in Rural Tanzania. Proceedings of 1st International Conference on M4D Mobile Communication Technology for Development (Karlstad University, Sweden, 2008), 69–79. [57] Nussbaum, M. 2000. Women’s capabilities and social justice. Journal of Human Development. 1, 2 (2000), 219–247. [58] Nussbaum, M.C. 2011. Creating capabilities. Harvard University Press. [59] Nussbaum, M.C. et al. 1993. The quality of life. Clarendon Press Oxford. [60] Pawar, U.S. et al. 2008. An „Office Hole-in-theWall‟ Exploration. Microsoft Research Technical Report. [61] Piaget, J. and Cook, M.T. 1952. The origins of intelligence in children. (1952). [62] Poverty, E. 2015. Millennium development goals. United Nations. Available online: http://www. un. org/millenniumgoals/(accessed on 23 August 2011). (2015). [63] Radha, G. 2013. Mobile Gaming in Emerging Markets - Some Insights. TreSensa. [64] Ratan, A.L. and Bailur, S. 2007. Welfare, agency and “ICT for Development.” Proceedings of the International Conference on Information and Communication Technologies and Development (Bangalore, 2007). [65] Raza, A.A. et al. 2013. Job opportunities through entertainment: Virally spread speech-based services for low-literate users. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (2013), 2803–2812. [66] Raza, A.A. et al. 2012. Viral entertainment as a vehicle for disseminating speech-based services to lowliterate users. Proceedings of the Fifth International Conference on Information and Communication Technologies and Development (2012), 350–359. [67] Riezler, K. 1941. Play and Seriousness. The Journal of Philosophy. 38, 19 (1941), 505–517. [68] Sachs, J. 2005. The End of Poverty: Economic Possibilities for Our Time. Penguin Press. [69] Salen, K. and Zimmerman, E. 2004. Rules of play: game design fundamentals. MIT Press. [70] Schwartz, A. et al. 2013. Balancing burden and benefit: non-prescribed use of employer-issued mobile devices. Proceedings of the Sixth International Conference on Information and Communications Tech- [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] 52 nologies and Development: Notes-Volume 2 (2013), 140–143. Sen, A. 1999. Development as freedom. Oxford University Press. Sen, A. and others 1993. Capability and well-being. na. Sengers, P. and Gaver, B. 2006. Staying open to interpretation: engaging multiple meanings in design and evaluation. Proceedings of the 6th conference on Designing Interactive systems (2006), 99–108. Sey, A. et al. 2013. Connecting people for development: Why public access ICTs matter (eBook). (2013). Sey, A. and Ortoleva, P. 2014. All Work and No Play? Judging the Uses of Mobile Phones in Developing Countries. Information Technologies & International Development. 10, 3 (2014), pp–1. Smith, A. and Nicholson, J.S. 1887. An Inquiry Into the Nature and Causes of the Wealth of Nations... T. Nelson and Sons. Smyth, T.N. et al. 2010. Where there’s a will there’s a way: mobile media sharing in urban india. Proceedings of the 28th international conference on Human factors in computing systems (New York, NY, USA, 2010), 753–762. Solow, R.M. 1987. We’d better watch out. New York Times Book Review. 36, (1987). Song, S. 2009. Nathan and the Mobile Operators. Many Possibilities. Statistics, I.T.U. 2014. International Telecommunication Union. ITU ICT Statistics. Retrieved February. Suits, B. 1978. The grasshopper: Games, life and utopia. Broadview Press. Sutton-Smith, B. 2009. The ambiguity of play. Harvard University Press. Walton, M. and Pallitt, N. 2012. “Grand Theft South Africa”: games, literacy and inequality in consumer childhoods. Language and Education. 26, 4 (2012), 347–361. Weber, M. 1905. The Protestant Ethic and the Spirit of Capitalism: and other writings. Penguin. Wittgenstein, L. 1953. Philosophical investigations. John Wiley & Sons. Wyche, S.P. et al. 2013. Facebook is a luxury: An exploratory study of social media use in rural Kenya. Proceedings of the 2013 conference on Computer supported cooperative work (2013), 33–44. Double Binds and Double Blinds: Evaluation Tactics in Critically Oriented HCI Vera Khovanskaya Information Science Cornell University vdk9@cornell.edu Eric P. S. Baumer Communication and Information Science Cornell University ericpsb@cornell.edu Phoebe Sengers Information Science and Science & Technology Studies Cornell University pjs54@cornell.edu of AI research and nine years later named the most important paper in the conference that year. It was not, however, influential in the way that at least one of the authors intended. For Philip Agre, the paper represented not only a breakthrough in the technical development of AI, but also a critical reframing of AI as a research program. As later worked out in Computation and Human Experience, for Agre the primary goal of this work had been to critically analyze the limitations of the philosophical conceptualizations undergirding technical work in AI and to offer a new conceptual alternative, embodied in technical language that engineers could understand. This approach Agre termed critical technical practice. Yet, to his frustration, while AI enthusiastically picked up on the technical work, the fundamental critical reconceptualization of human activity embodied in the technology was not picked up; eventually the work was reabsorbed into what Agre saw as business as usual in AI [44]. ABSTRACT Critically oriented researchers within Human-Computer Interaction (HCI) have fruitfully intersected design and critical analysis to engage users and designers in reflection on underlying values, assumptions and dominant practices in technology. To successfully integrate this work within the HCI community, critically oriented researchers have tactically engaged with dominant practices within HCI in the design and evaluation of their work. This paper draws attention to the ways that tactical engagement with aspects of HCI evaluation methodology shapes and bears consequences for critically oriented research. We reflect on three of our own experiences evaluating critically oriented designs and trace challenges that we faced to the ways that sensibilities about generalizable knowledge are manifested in HCI evaluation methodology. Drawing from our own experiences, as well as other influential critically oriented design projects in HCI, we articulate some of the trade-offs involved in consciously adopting or not adopting certain normative aspects of HCI evaluation. We argue that some forms of this engagement can hamstring researchers from pursuing their intended research goals and have consequences beyond specific research projects to affect the normative discourse in the field as a whole. While critical technical practice as a concept did not have the lasting influence on AI that Agre had hoped, in recent years the idea of deeply coupling technical design work with critical reflection on its conceptualization has been picked up in a neighboring field, Human-Computer Interaction (HCI). For example, within HCI, Dourish cites Agre's notion of critical technical practice as a major inspiration for his landmark work Where the Action Is [16, p 4]. Sengers et al.'s articulation of reflective design [41] expands critical technical practice to be used not only to work toward the resolution of specific technical impasses, but as a core method throughout all phases of technology’s design and use. The engagement of critical practice within technology design specifically and HCI more broadly has also been spurred through research drawing on other critical traditions. For example, Bardzell, Bardzell, and Blythe integrate criticism and critical theory into design [5,6,7,10]; DiSalvo and Hirsch draw on critical and political traditions in the arts to frame new opportunities and roles for design [15,23]; and Pierce and Löwgren use philosophy of technology and design theory to articulate and execute critical approaches to interaction design [35,36,30]. In short, there is a foment of critical approaches in HCI which aim to push forward HCI research while simultaneously troubling some of its foundational assumptions. Thus, in contrast to Agre's impact within AI, in HCI critically Author Keywords Critical technical practice; critically oriented HCI, evaluation ACM Classification Keywords H.5.m. Information Interfaces and Presentation (e.g. HCI): Miscellaneous INTRODUCTION In 1987, two graduate students at MIT published a paper describing a new way to construct intelligent programs in the major annual Artificial Intelligence (AI) conference [2]. This paper became widely influential, spawning a new area Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21266 53 oriented research has clearly achieved lift-off as a vibrant research program. What does tactical engagement with HCI look like? We will ground our ongoing discussion in this section in three examples of 'working systems' which have been influential in the field and inspirational for our own practice. We chose these particular exemplars as instances where authors begin to “lift the curtain” to reveal some of the tensions in evaluating critical HCI projects. In this paper, we take the emergence of a wider body of research integrating critical reflection with technology design as an opportunity to better understand the dynamics and problematics of tactically engaging with HCI in order to expand, reconceive, or otherwise critically reform the field. We argue that the challenges that Agre faced in marrying critical reflection with an ongoing technical discourse that otherwise tends to bracket such questions have not simply disappeared but remain relevant to understanding how researchers can successfully tactically engage HCI. Moreover, HCI adds a significant new complexity to this challenge because of the split audience of HCI work. HCI researchers must engage not only other scholars, but also end users or other participants, who are an obligatory passage point for commonly accepted definitions of what makes a system or approach 'work' in HCI. The first system, Affector, is an interactive video installation developed by Phoebe Sengers, Simeon Warner, and Kirsten Boehner [12,40,42]. The system is comprised of a video window between the offices of two friends that enables them to communicate their emotions by systematically filtering the video feed according to sensor readings. Affector is presented as a challenge to dominant threads in computer-mediated communication toward greater realism and accuracy in representation of affect by not directly representing emotion in the system and instead emphasizing openness to interpretation. Methodologically, it challenges HCI notions of objectivity and scientific observation by using an 'autobiographical design' method in which the same people are designers, users, and evaluators, working together to iteratively design and develop the system in reaction to their personal experiences. Within this set of considerations, we aim to make three main contributions. First, we articulate how the need for tactically engaging with the lingua franca in HCI shapes the nature of our interactions with the participant in ways that can hamstring at least some of researchers' critical intentions. Second, through examining unpublished backstories of three of our own projects, we demonstrate how this happens concretely on the ground, leading to some undesirable practical consequences in relationships with participants despite our largely being able to position the projects as successful from an HCI research point of view. Finally, based on our own work and related published projects, we identify three specific ways in which what counts as 'working' in mainstream HCI discourse manifests in critically oriented HCI discourse in ways that can inadvertently lead to particular forms of breakdown in the relationship between designers and publics. The second system, Local Barometer, is one of three “threshold devices”, poetic intermediaries between home and the environment, developed by Gaver’s research studio at Goldsmiths [21, 33]. The Local Barometer is intended to provide people with a new sense of the “sociocultural texture” around their house. The barometer itself is composed of six small screens which display scrolling images and texts from classified advertisements sourced from the internet. The local wind conditions, as measured by the device, determine which advertisements appear on the screens: the harder the wind blows, the greater the geographic distance from which the ads “travel.” Gaver et al. position this work to resist a dominant ethos in technology design towards accuracy and efficiency. Instead, they stress the value of multiple, unresolved narratives in understanding the meanings of technology, including play, ambiguity and interpretation in design [42]. Thus, the evaluation of this system involved collecting and juxtaposing multiple narratives of use, rather than emphasizing a single correct meaning or use of the system. TACTICALLY ENGAGING HCI The problematic we explore in this paper is associated with work in HCI that proposes significant alterations to core foundational assumptions of HCI and embodies these alternatives in working systems or methods. Some of this work is explicitly inspired by critical technical practice, some of it comes from other critical traditions, and some is not explicitly framed as 'critical' but still aims to open significant new spaces for HCI by inverting dominant assumptions of what makes for good design or research. In this section, we will describe how tactical engagement with HCI works in several well-known projects and articulate three key attributes of these projects that support effective tactical engagements with HCI. We will argue that these three key attributes lead to a core problematic for critically oriented HCI work: that the need to articulate systems as 'working' within HCI’s discursive norms significantly troubles our relationships with participants. The third system is an "ultra low cost sensing system" developed by Kuznetsov, Hudson, and Paulos to support air quality activists in detecting particulate pollution [25]. The “lo-tech” sensors can be assembled inexpensively from common paper materials. The authors describe the system and its deployment in a local environmental community, highlighting a series of tradeoffs between expensive and inexpensive sensing systems, usability and generalizability, and the role of inexpensive sensors for encouraging reflection and community action. Through this system, the authors aim to alter the political economy of environmental sensing by making it possible for non-experts to collect, 54 measure, and reflect on local air quality, rather than requiring them to partner with scientists who would be in charge of the data gathering and interpretation. Kuznetsov et al. test the system in collaboration with a local activist group. This project is part of a longer-term design research program on participatory approaches to environmental data gathering, which has highlighted political issues in the design of environmental sensing systems [4,25,26,27,28]. they assigned to their third collaborator the role of an “external evaluator” to provoke some of the reflection and moderate discussions about the system. In other work, Sengers legitimated these practices by arguing that approaches such as Affector's which address multiple interpretations are still “definable and testable” by HCI standards [42, p.107]. For Gaver’s team, there is a similarly strong commitment to presenting an alternative to dominant technology design methods. This work is not framed as 'critical' - Gaver et al. see this work rather as a positively and practically oriented design practice. Nevertheless, it directly challenges some HCI orthodoxies, particularly traditional goals of “usefulness and usability,” and aims to open HCI eyes to the possibilities of designing for open-ended playfulness (vs tight functionality) and leveraging ambiguity (vs. establishing clear answers) [18]. Based on these examples, we now describe three key attributes of these systems and others like them that support effective, critical engagement with more mainstream HCI. Later, we examine how these strategies interact to shape what it is possible and not possible to say through constructing working systems in HCI. Engaging the Lingua Franca One attribute emerging from these examples is that they make their critical interventions by challenging some aspects of the normative discourse while simultaneously engaging, and to a degree upholding, its other elements. Even while each approach highlights the ways it upends one or more significant aspects of standard HCI practice, each also supports the viability of its critical move by making sometimes implicit and sometimes explicit appeals to HCI's established “lingua franca”—the common forms of communication of an intellectual field. In this section, we explore how these three examples do so by looking at what aspects of HCI they queer and what they hold steady. This challenge is embodied not only in the goals designed for in these systems, but also in their evaluation. In contradistinction to approaches that aim to establish definitively how a system is taken up in use, the evaluation of the Local Barometer, as well as other domestic technologies such as the Drift Table and the Key Table, focuses on capturing the intervention’s multiply interpretable assessments [42]. Gaver and his team study families who are tasked to “live with” the technology by using ethnographic techniques. They also solicit the views of semi-independent commentators, such as journalists and documentary filmmakers, who, like the participants, are not informed of any designerly intentions for the deployment. These methodologies can read as strange to an HCI audience more attuned to laboratory-based 'user tests' conducted with behavioral social-science methodologies, and are probably explicitly intending to be provocatively different. These moves are not naive; as with Affector, Gaver's explicit intention is to co-opt the documentary method while recouping it for a different epistemological orientation. For his team, the double-blind techniques are a way of capturing the multiply interpretable, to avoid narrowing the scope of the interpretation and to “open up” the design space as much as possible. With Affector, the authors were not only arguing for a new, non-representational way of looking at affect. They also argued for shifting the evaluation of design practice in HCI from ascertaining whether a design is a success or failure by predetermined metrics to narrating how the system “came to be known or lived as a success or failure” [12, p. 12:24]. They do so by recognizing “multiple, perhaps conflicting interpretations” [42 p. 101] of the design through iterative and thoughtful analysis not only of the system design, but also of the evaluation criteria. As the authors point out, one of the “problematics” [12 p.12:19] their work revealed was that knowledge gained through these approaches “is not always compatible with what is recognized as knowledge production in HCI” [12 p.12:19]. In the low-tech sensors case, the challenges to HCI orthodoxy have to do with questions about the framing of the relationship between researchers and participants. Particularly in Aoki et al.'s ethnographic work into the challenges of environmental sensing [4], on which the lowtech sensors work explicitly draws, a key issue is seen in the way activist groups defy usual HCI understandings of fixed, well-defined user populations who are positively oriented to technology and to research. That study identified activist groups’ skepticism, directed towards scientists seeking to understand air quality, around who has access to the data collected and who is in charge of its interpretation. This finding lead directly to Kuznetsov et al.’s choice to design air quality sensors that can be easily The evaluation of Affector was explicitly motivated in part by the need to counter perception by the broader HCI community that allowing for multiple possible interpretations rather than a single, correct answer would lead to an “anything goes” mentality. One stated goal was to show that multiply-interpretable systems could still be rigorously evaluated [42]. In an effort to make room for their new approaches and to habilitate them to HCI sensibilities, the authors appropriated and reframed existing HCI methods. In their writing, the authors explicitly aimed to use these methods coupled to a different epistemological stance. For example, they used a diary for the two users to track their emotions and interactions with the system, and 55 interpreted by activists themselves, thus removing scientists from the equation. The design is evaluated through a community workshop in which activists try out the sensors and the researchers interpret the results. The narration of the results blends issues and factors from a standard, apolitical HCI with those that come from the political stance of the researchers, ranging from discussions of ease of use to reflections on the degree to which the sensors allow for end users to feel a sense of ownership and understanding of the data. At the same time, the evaluation is couched in terms drawn from standard HCI methodologies, particularly the form of the focus group. For example, the interpretations provided are the researchers', the participants are anonymized and referred to through generic, interchangeable labels, and the argument is substantiated through empirical evidence in the form of direct quotations from participants. more recognizably interpretation [11]. behavioral-scientific forms of Still, the risks of being re-assimilated seem attenuated in HCI compared to AI because of the presence of an established community of researchers schooled in critically oriented sensibilities, which likely supports more radical methodological transformations than would be possible if papers were only ever reviewed by researchers committed to mainstream methodologies. But while the risks described by Agre in AI may be less dire in contemporary HCI, the adoption of critical engagement with technology design as part of HCI practice also introduces a significant new element, in the form of users. Constructing what ‘works’ A second attribute emerging from these examples is that they aim to make critical interventions in HCI through embodying them in ‘working systems’, or, to put it more precisely, systems that can be narrated to the HCI community as 'working.' The critical arguments gain rhetorical force in the community by demonstrating that one can challenge significant norms of HCI practice and still end up with a system that is practical, usable, functional, or otherwise clearly 'good' by standards that are at least somewhat recognizable from mainstream HCI. In all three examples, then, we see critical interventions couched as challenging HCI norms, coupled to aspects that buttress the validity of the challenge with appeal to terms and sensibilities that resonate, or are thought to resonate, with a more mainstream HCI community. Such strategies are sometimes framed as concessions or compromises required to get a work published [29]. That framing, however, does not do justice to the creative challenge constituted by the commitment to engage the mainstream discourse for change. Each of these examples remixes the mainstream and the subversive in a different way in order to establish new grounds. By identifying this dynamic, we wish, in part, to call it out as an explicit and challenging research design problem worthy of sustained attention within the critically oriented HCI community. This 'work ethic' [3] is also a central feature of Agre's critical technical practice in AI; his systems were intended to give rhetorical force to his philosophical critique by demonstrating that alternatives generated from that critique can be embodied in a working system. In [3], Agre argues that the notions of what it means for a system to 'work' shape the claims it is possible (or not) to make through them. Within HCI, particularly in the last 10-15 years, claims to a 'working system' have been rooted nearly universally, as the previous section might suggest, in processes of 'evaluation', where evaluation almost invariably involves tests with human participants In order to make the claim that a critical intervention 'works' as a design principle, researchers frequently choose to engage in 'evaluation,' i.e., muster arguments that involve claims based on empirical evidence of how participants understand or interact with a system. Sometimes, as with the low-tech sensors, critical interventions in HCI are embodied in challenging assumptions or norms about use in the design of a system, while those systems are evaluated using fairly standard strategies such as usability tests or focus groups. Other times, as with Affector and Local Barometer, researchers alter the evaluation methodologies themselves in ways they consider to better align with the critical goals of the project. At the same time, we need to be aware of the consequences and costs of the specific ways in which the dance between the mainstream and subversion takes place. Agre's experience in AI, described in the introduction, serves as a case in point. In [3 pp. 152-153] Agre describes why it is difficult for critical alternatives to truly redirect a technical field. One of these reasons is that "it is difficult to apply [a] method [embodying a critical alternative] to any real cases without inventing a lot of additional methods as well, since any worthwhile system will require the application of several interlocking methods, and use of the existing methods may distort the novel method back toward the traditional mechanisms and the traditional ways of talking about them." There is thus a risk that the necessary business of leveraging traditional HCI methodologies and sensibilities may limit our ability to shape true critical alternatives. Further, accommodating mainstream sensibilities may make it possible for critical alternatives to be picked up, but at the same time invite re-appropriation into the status quo. Boehner et al., for example, argue that Gaver's methodology of 'cultural probes,' intended as a radical reframing of HCI techniques for assessing use contexts that undermined claims to objectivity, became quickly reframed and "improved upon" by assimilating it to In considering these choices in a critically oriented design project, we note that, since the validity of the design argument hinges on a convincing evaluation, there is likely more room to maneuver tactically in the design of a system than in its evaluation. Thus, the question of evaluation is a 56 lynchpin for understanding how critically engaged projects gain a rhetorical foothold in HCI. At the same time, 'evaluation' must not be framed too narrowly. As we saw with the Affector, Local Barometer, and Lo-tech Sensing examples unpacked previously, the negotiation of "what works" is present not only during the literal “doing” of the evaluation, but throughout the construction and narration of design and evaluation. The Problematic So far in this argument, we have identified three key attributes of projects that embody critical challenges to mainstream HCI in working systems. First, researchers must necessarily engage with the lingua franca to make their critical challenges legible and defensible in the field; they do so by holding some aspects of mainstream methodology steady while violating others. Second, in order to argue for the value of their critical challenge, they frequently engage in some form of 'evaluation' which is based on narrating the results of empirical interactions with human participants. Third, part of the critical project in HCI involves queering the relationships among designers, technology, and users as commonly imagined in the field. Reframing relationships with users Given that participants are a central resource for knowledge construction in HCI, a third attribute that emerges from these examples is that in critically oriented HCI not only the technology itself but also relationships between designers, technologies, and their users become material to be critically reframed. All three systems share an interest in challenging traditional relationships between technology interventions and human participants by presenting alternative relationships in the form of working systems. Affector, for example, is grounded in a critique of how standard affective computing frames the relationship between human emotional experience and its mechanic representation; it also is used to argue that richer notions of emotional experience can be brought into design by violating common notions of objectivity in HCI through folding designers' experiences directly into design. Gaver et al. use the Local Barometer to argue that forms of human experience normally marginalized in design are legitimate design material. Further, the system is based on reframing the relationship between designers and users from designers attempting to control user experience to enabling more open-ended forms of engagement. Low-tech sensors are designed for a politically engaged audience that implicitly challenges traditional HCI conceptions of the discipline as scientific and therefore politically neutral. Explicitly, the sensor project aims to reframe relationships between scientists and activists by putting its users in control of sensing and interpretation and cutting the scientists out of the loop. A major resulting problematic, and one we have experienced in our own work, is that the act of making critically oriented design interventions legible to the HCI community—i.e. tactically engaging with the “lingua franca”—shapes the nature of interactions with participants in ways that can undermine the critical goals of the project. Keeping in mind that legitimate evaluation is part of the HCI definition of “working,” it follows that discursive norms in the HCI community will have an impact on how critically researchers engage with the human participants in their work (i.e. their “users”) in order to be able to speak convincingly to the broader HCI community. These simultaneous conversations with the twin audiences of scholars and participants can result in complicated double binds. These double binds arise because critically engaged projects necessarily find themselves grappling with a legacy of evaluation in HCI in order to negotiate what it means for their own systems to “work” in accordance with its discursive norms. And what makes an evaluation 'convincing' in HCI is strongly influenced by its inheritance from the intellectual traditions of psychology, cognitive science, and human factors. This is the complex and somewhat contradictory discursive world in which critically oriented researchers – often coming from humanities, arts, or design traditions to some degree at odds with behavioral social-science epistemologies – must maneuver to make our arguments hold. More generally, and in contradistinction to the case in AI (at least at the time Agre was working), critical engagement by HCI researchers is oriented not only to critical reflection on the conceptual commitments of the discipline, but also to altering how we imagine and perform relationships between researchers and users. Sometimes this is embodied in a commitment that not only researchers, but also end users, participants, or publics should critically reflect on technology and its relationship to human life [e.g. 41]. This interest in considering and responding to the problematic politics of designers and users deeply resonates with and is partially inspired by the long legacy of participatory design. But, unlike participatory design, in which these reframed relationships are widely considered an inviolable aspect of what it takes to create knowledge, in HCI the commitment to developing alternative relationships to users comes under substantial stress. The next section explains why. In the rest of this paper, we aim to lay out concretely how the way we construct legitimate claims in evaluation shapes the relationships it is possible for us to have with participants. Our goal in this analysis is to look at specific ways that tactical engagement takes place, and the consequences of trade-offs made in that engagement. We will argue that our relationships with users become contorted in particular ways through specific strategies for legitimation inherited from mainstream HCI. These trade-offs and their consequences are often hard to see, particularly in published papers. This is because those papers are necessarily framing interactions with users in ways that will support claims to legitimacy. The contortions and problematics are outside that frame and perhaps even to 57 some degree outside researchers' conscious understanding. Next, we will explore the nature of and substantiate this problematic by looking at problems that emerged in three examples from our own work. We will lift the veil from our previously published papers to discuss issues in our relationships with participants that troubled us at the time, but seemed to fall outside the frames of what we could easily discuss in the published work. Later in the paper, we will integrate these experiences with published works to identify discursive tropes which shape engagement with audiences and account for design and evaluation outcomes. assess the biases of different sources and sort out opinion from factual news reporting and identify reporterly bias. While instances of frame reflection that were in line with Schön and Rein's [39] description of frame reflection did occur, they were mentioned more rarely. The first set of tensions arises from participants' insistence on the value of neutral, factual news reporting. Indeed, during controlled experiments and focus group studies that involved explicitly mentioning the concept of framing [37], participants often interpreted “framing” as “spin” or “bias.” The system was seen as a tool for identifying and avoiding such manipulations of objective news reportage. This usage, however, runs contrary to literature on framing, which suggests that there is no such thing as an un-framed fact [17,13]. This distinction between biased and “straight” news, between opinion and fact, led participants to use the system not in a reflective mode but in an evaluative one. Thus, the accounts that participants often gave were in some ways at odds with the frame reflection and the conceptual goals for the paper. TENSIONS, TRADE-OFFS, AND DOUBLE B(L)INDS The following three research projects have been published in some capacity in HCI venues and have involved at least one of the authors. Some of these examples explicitly draw from critical technical practice and critical design methodologies, while others are not explicitly “critical” and instead engage in subversion through other forms of unusual HCI practices. What is held in common in these three accounts is the presence of the three attributes described above: engaging with HCI’s lingua franca and with human participants, while simultaneously challenging status quo relationships between researchers, users, and technological systems through “working” alternatives. By examining our own work through these internal accounts we can investigate the ramifications of the interactions between attributes more fully than by reading only the published portions of this critically oriented work. Second, we had hoped the system would help participants identify frames at work, consider the assumptions on which those frames are based, and look at the ramifications of those frames for the perception. However, as prerequisite to engaging in frame reflection, participants would have needed to accept that framing operated in the way described by these theories. That is, in order to engage in frame reflection, our participants would have needed to adopt a meta-perspective that (1) multiple perspectives on (i.e., framings of) reality exist and (2) those different perspectives (framings), generally speaking, are each equally valid. Once this meta-perspective is adopted, a tool such as the one we designed could potentially become quite useful. However, without such a meta-perspective, the visualizations provided by the tool become open to (mis)interpretations that align with a different fundamental take on the world and the nature of (objective) reality. Designing for Frame Reflection One of the authors was involved in designing and implementing an interactive visualization tool that leverages computational linguistic analysis to present patterns of language in political blogs and news sources. In contrast to traditional data visualization, the design sought not to provide an overview of what is being said but rather how things are being said, i.e., to encourage attention to and reflection on how issues are “framed” [17,13]. In addition to lab studies and focus groups [37], this system was deployed in a field study, during which the tool was used for at least 8 weeks by regular readers of political news coverage during the 2012 U.S. election campaign [8]. The evaluation sought to assess the tool's capacity for supporting frame reflection [39], the ways that users integrated tool use with their existing reading practices, and broader issues in how participants interpreted the computational analysis and visualization. So as not to bias their responses, participants were not initially told that the primary intended purpose of the system was to foster frame reflection. In retrospect, we experienced a researchers’ double bind: if we told participants about framing, frame reflection, and multiple perspectives, then any evidence of frame reflection we saw in their accounts would be more likely attributable to those statements than to their experiences with the tool per se. However, when we kept such information from them, as was the case here, then participants were left with little scaffolding for the kind of thinking the tool was meant to engender. This double bind did not result in the research becoming entirely incapacitated, as some participants were able to come to an understanding of framing, but the process was hamstrung by our desire to evaluate the validity of the system for supporting frame reflection. The field study identified a variety of ways that participants used the frame reflection tool, but particularly relevant here is that participants often attempted to use the system as a ‘bias detector.’ Participants would examine specific patterns of language visualized using the tool in an effort to Challenging Self-Optimization In this project, two of the authors were involved in the development and deployment of a reflective, critically oriented personal informatics tool [24]. This tool was 58 designed to inspire reflection not on peoples’ own behaviors, as personal informatics systems usually do, but instead, on the infrastructures underlying the gathering and presentation of personal data and the narrow modes of engagement that traditional personal informatics systems promote. We encouraged this critical reflection by gathering users’ web browsing data and displaying it using different design strategies: for example, we hoped to encourage reverse engineering of data gathering algorithms from our participants by using purposeful malfunction in the visualization. Because we were arguing for the role of critical design in challenging the status quo of selfoptimization narratives in personal informatics, and because we wanted to show that our design strategies could potentially be employed toward similar means in other personal-data contexts, our evaluation needed to convincingly illustrate the ways that the designs themselves provoked the specific sorts of reflection we intended. Probing Community Values This research was part of a larger effort to explore issues broadly related to sustainability [43] at a local farmers’ market using cultural probes [9]. As part of this work, two of the authors sought to develop a methodology for doing “community probes,” applying a similar approach and sensibility as cultural probes, but as a means of fostering conversations among community members rather than between community members and designers. Responses to a cultural-probe-style diary were used to inspire a series of such community probes in the form of speculative design proposals. For example, inspired by themes of stress and chaos in the diary responses, one design suggested a series of ropes and poles that automatically reconfigured themselves to allow for optimal foot traffic among the crowded market stalls. These proposals were intentionally provocative, as we sought to incite reactions from participants pertaining to what they liked and disliked about the market. In order to evaluate our system, we recruited participants for a study about reflecting on web browsing data and about unexpected approaches to visualizing personal information. While this was not directly deceptive, it did not directly call out the critical agenda in our work. During the interviews with our participants, we found ourselves in a complicated tension. Participants picked up on the strangeness of our system’s design but were also trying to use the system as one would use a traditional personal informatics tool. Often, this interpretation of the tool would happen in the context of “critical” conversations where participants speculated on the norms, limitations, and assumptions in personal informatics systems. While some participants expressed skepticism toward personal informatics systems and the self-optimizing values they espouse, it is undeniable, too, that some of those same participants were optimistic about tracking and imagined using the tool to “improve” their web browsing habits. These designs were displayed on large posters at the market along with markers and post-it notes for visitors to leave comments. Adhering to the “aesthetic control” of cultural probes [18], the posters were not branded with the research lab or university where the authors worked. One of the central challenges that presented itself in the deployment of these cultural probes was that despite the satirical nature, or in some cases technical infeasibility, of the proposals, some market visitors thought not only that they were serious proposals for changes, but that these changes were being proposed by the market administration. When several market visitors complained about the proposals to the administration, we were asked to amend the posters with a disclaimer that the designs were in no way affiliated with or endorsed by the official market administration. We later followed up with the management, asking about conducting a follow-up where they could be more involved in the design process, but we were told that the market administration was not interested in running any more “surveys” at this time; in retrospect, these conversations illustrated not only a substantial misunderstanding of our research intervention, but also an antagonistic relationship between us and the public we wanted to engage. We discovered a complex negotiation between the participants and the interviewers that evolved in response to users’ conflicted interpretations of the system. We wanted to get responses from participants during the interview without revealing our own motivations, but simultaneously felt tempted to explain our critical stance in order to have an informed conversation about the limitations of personal informatics. The way this played out in practice was that the former behavior manifested during the “formal” interview, while the latter discussion occurred after the interviewer began “debriefing” the participant about the goals of the study. This ad-hoc negotiation with traditional HCI protocols allowed researchers to pick up on threads previous articulated by participants earlier in the interview while maintaining the “hands-off” ethos of not biasing our participants (at least not initially). The negotiation also demonstrates the ways that our research goals shaped the ways we engaged with the participants in our study. We saw similar tensions stemming from tactical non-disclosure manifest more dramatically in the following project. Under some interpretations of cultural probes, we could read the overwhelmingly negative response to the community probe, both from the market attendees and the market administration, as a response and legitimate result of the probe: the refusal of even infeasible technologies points to community values around sustainability that oppose optimization narratives in favor of community gathering, revealing a community-oriented sustainability practice. We are very aware, however, of the interpersonal tensions that manifested when we came to the Farmer’s Market and attempted to provoke critical responses from the public while temporarily masking our affiliations with the local university. While we were able to gather experiences and community responses to the probes we set up, the severed 59 relationship with the market administration points to the ways our relationships suffered as a consequence of the levels and types of information withheld. participants in psychological experiments. Bardzell et al. connect this struggle to the complexities of critical design, where the goal is to deal with uncomfortable topics (in their case, sex and gender) in deliberately provocative ways. THINGS GO AWRY In our cases, we see traces of participant unease as to the nature of their relationship with researchers even in work that did not explicitly involve “critical design,” such as the speculative sketches we installed at the farmers’ market being (mis)interpreted as official proposals. This relationship becomes even more complex when the orientation of the participant toward the researcher is explicitly suspicious or even adversarial, as was the case with the activist communities in the Aoki et al. Street Sweepers fieldwork [4]. There, activists took on a generally oppositional stance toward academic researchers, who the activists saw as people who enter into activist communities for the benefit of their own research and disappear, along with the data, when this research is complete. Again, in navigating this complex relationship that comes as a unique consequence of employing subversive tactics in HCI, either in embodying alternatives in working systems or methods, we see noticeable frictions that arise when researchers negotiate different aspects of HCI’s lingua franca at different times in response to the specifics of their situation, We initially found it difficult to articulate our experiences beyond a simple run-down of cases where we had “gone wrong” in doing critically oriented work, or a general indictment of empirical values unconsciously permeated into critically oriented HCI. Eventually, we realized that the evaluation practices researchers take to legitimate their research as working systems in HCI are deliberate, tactical engagements with the lingua franca, practices which in turn take a role in negotiating discursive norms in the community. For this reason, it is important not to treat our own tensions as isolated cases of research malpractice, but as actions situated in the context of the critically oriented community. Using our own experience as a guide, we looked for traces of the same problematic in other critically oriented. We felt connections between the tensions in our work and those documented by Gaver et al. in “Anatomy of a Failure,” where the research team reflects on challenges in the evaluation of a sensor-based system, the Home Health Monitor, that was designed to be deliberately ambiguous in order to encourage interpretation and appropriation in domestic settings [20]. Similarly to our own methods, and in keeping with the studio’s practices, the researchers did not tell the participants in their study about their own intentions to avoid biasing the participants’ interpretations of the system. As the researchers describe, this initial lack of transparency contributed to a lingering attitude of uncertainty and suspicion that persisted even as researchers attempted to clarify their intentions through modifications in the system’s design. In the paper, the researchers discuss the ways in which withholding information and leaving the design open to interpretation made the system almost completely inscrutable and, simultaneously, almost completely uninteresting. We found points of resonance in their analysis in the ways that non-disclosure shaped the relationship between researchers and study participants. THREE DISCURSIVE TROPES We have seen that in all the cases above, both our own and others from the literature, researchers are holding constant or appealing to some elements of more orthodox evaluation practices while challenging, bending, or exploring alternatives to others. How do we actually articulate what these practices are, to better understand what is being held constant and what is being changed? Here, we draw on values sourced from controlled scientific experimentation to highlight how HCI’s historical undercurrents of cognitive psychology and computer science shape elements of the current lingua franca in evaluation. For each of these three values, or discursive tropes, we describe how they can be employed by critically oriented researchers to creatively maneuver subversive research toward conventional notions of legibility. In practice, these three values are not clearly separable and are in fact bound up with one another in intricate ways. For the purposes of analysis, we tease them apart, but it should be noted that in practice they rarely separate cleanly. Our intention is to provide conceptual language with which to better understand both the nature and the ramifications of tactical engagement with discursive norms of evaluation in HCI. Engagement between design researchers and their participants, and the ways that the relationship is made ever more complex by explicitly critical agendas, is also addressed in the work of Bardzell et al. [5]. In this work, the researchers deploy critical designs meant to provoke critical reflection on gender and divisions of domestic labor. This paper recognizes their participants’ struggle to participate (and be “good participants”) in the research. Like with Home Health Monitor, Bardzell et al. address aspects of participant suspicion through a desire to know what “the study is really about.” Part of this struggle in subject participation has to do with the idea that relationships between researchers and their participants do not start from a “blank slate.” One couple interviewed by Bardzell et. al. was particularly skeptical about the hidden motivations of the researchers stemming from their previous experiences in graduate school as Demand Characteristics Psychologists and other social scientists have long acknowledged that the substrate they study—human beings—differs in important ways from that studied in the physical sciences. Namely, experimental subjects often reason, consciously or unconsciously, about the purpose of the experiment in which they are participating, not for 60 nefarious ends but often rather for compliance. Many experimental volunteers desire to be “good subjects” and see their participation as a contribution to the furthering of science [34]. The subject, then, also has a stake in the experiment turning out well. Thus, “as far as the subject is able, [s/]he will behave in an experimental context in a manner designed [...] to validate the experimental hypothesis” [34 p. 778, emphasis original]. Thus, an experimental result may not actually be due to the experimental manipulation itself but to the willing compliance of the subjects. Indeed, the subject may leverage a variety of cues – including the study description, the informed consent forms, the demeanor of the experimenter, the study procedures themselves - to reason (again, either consciously or unconsciously) about the purpose of the experiment. This “totality of cues which convey an experimental hypothesis to the subject [are called] demand characteristics” [34 p. 779]. We see maneuvers toward the reduction of demand characteristics across other critically oriented work in HCI, notably the techniques described earlier by Gaver et al. where researchers gather multiple perspectives on design work by informing neither the participants nor the independently hired film crew of their design intentions. Again, it is important to stress what while these negotiations appear ‘scientific’—and Gaver even compares the practices of evaluating Local Barometer to experimental evaluations done during his graduate work in Cognitive Science—the efforts to reduce demand characteristics are employed deliberately, under different (non-empirical) epistemic methodologies, to serve alternative ends [33 p. 140]. While this strategy can be rhetorically leveraged to tactically engage with HCI audiences, it can also create tension between researchers and study participants and impede researchers’ ability to pursue study goals. Representative Sampling Each of the above case studies demonstrates different attempts to reduce demand characteristics. For example, the work applying critical design to personal informatics was described to participants as dealing with personal data. We did not tell subjects that the design was “critical” in nature or that it was intended to prompt reflection on the value commitments that underlie personal informatics. In the frame reflection study, participants were told that the tool was about understanding political language, but framing was not explicitly mentioned. In the farmers’ market work, we did not want to tell participants that we were interested in sustainability so that they would not focus on a single aspect of their experiences at the market. This tactic followed from the advice of Gaver et al. [18] about ensuring that the speculative proposals had no trappings or accouterments of a university research lab, and other examples of researchers being deliberately ambiguous in the presentation of their design. Traditional evaluation techniques in HCI in their significant conceptual borrowing from empirical practices in experimental psychology and demographic sociology often center on questions of whether the participants in a study comprise a “representative sample”. Representative sampling is a statistical technique based on the idea that while it is rarely practical or even possible to get information from every person in a targeted population, the qualities of a population as a whole can be closely paralleled by a much smaller, properly selected segment. By selecting for some key variables such as gender, age, technical expertise, or socioeconomic status, and controlling for others, the representative sample becomes an effective ‘stand-in’ for other parts of that population. Traditionally in HCI, representative sampling emerges as a way to learn about potential users of a technology [38]. However, as HCI accrues interdisciplinary practices into its standard methodologies (e.g. interpretive and ethnographic ways of knowing,) the concern over representative sampling itself came to represent HCI’s internal tensions. As described in [33 p. 17], an ethnographer may refer to “sampling” a population to account and interpret their experiences and practices, and sometimes seek to make broader statements based on their interpretations without having their population “stand for” a broader statistical phenomenon in the same ways that controlled demographic or experimental approaches might attempt to do. In each case, though, these decisions, intended rhetorically to minimize demand characteristics, ended up affecting, and in some cases limiting, the potential of our designs to engage participants in the very types of critical thinking and reflection we sought to engender. With the critical personal informatics study, participants expressed critical ideas about normative values in personal informatics in conjunction with literal interpretations of how the tool could be used to optimize behavior, because they suspected that this was our research goal. With the frame reflection system, participants were left a bit baffled as to how they might use the tool. Since frame reflection is admittedly a bit outside the normal approach to political news coverage, participants fell back on familiar forms of identifying and comparing partisan bias. At the farmers’ market, participants reacted most dramatically, thinking that these were official proposals rather than reflecting on broader themes about the market to which the speculative designs sought to draw attention. The question of representative sampling was salient at several moments of our research that are often not accounted for in traditional research accounts. For example, when writing on the critical personal informatics work went through peer review, one issue raised by reviewers was that our sample was not representative of the population as a whole, and that the results that we got might be specific to the participants we recruited. While we leveraged a rhetorical appeal to this same trope by arguing that our “sample” was representative of the audience to whom 61 personal informatics applications are traditionally marketed (i.e. upper middle-class, tech-savvy, college educated), in retrospect, this interaction was an example of the way that critically oriented researchers negotiate and enforce HCI’s evaluation norms. Representative sampling negotiations are also visible in the Affector research, where the authors point to concerns about “auto-biographical blinders” that might emerge from designers testing technology on themselves, and employed the use of a “third party” evaluator in response to this concern [42 p. 12:22]. than to the evaluation protocol. It is relevant to note here that the effort to evaluate towards an imaginary where prototypes exist in the world and are used by people without the designers’ intervention is not unique to critical projects in HCI, and that most (if not all!) HCI deployments employ elements of speculation. However, doing so tended to limit our ability to achieve some of the very goals our studies were intended to show we had accomplished, such as the double bind that occurred in the frame reflection example. Again, we see similar attempts to engage stimulus-response causality in others’ work: the low-tech sensors research argues that the design of the sensors afforded particular types of usage. For example, by explicitly not attaching instructions about how the sensors worked, and then citing how all of the participants said the sensors were easy to use, the research team rhetorically convey that the devices were usable by traditional HCI standards. The research team was able to appeal to this trope of stimulus-response causality while also engaging participants in reflective and interpretive practices around environmental sensing, which illustrates the methodological creativity of critically oriented researchers to negotiate different values into the evaluation of their systems. As mentioned previously, the idea that research participants in critically oriented work can be used empirically to justify broader claims is also echoed by the very ways that we invoke scientific language of lingua franca to refer to participants in our writing (e. g. “P5” to refer to specific participants.) This is of course not to say that researchers literally dehumanize their participants into data, but rather that engaging with empirical epistemologies can be used rhetorically to make legible and render valuable to HCI some of the intellectual contributions of critically oriented researchers. Stimulus-Response Causality Experimental controls attempt to isolate various factors from one another in order to identify causal stimulusresponse effects. To establish causal links, researchers often look for differences between an experimental condition and a control condition, where the two conditions are identical in every way with the exception of a single experimental manipulation: the stimulus. Thus, any differences observed between the two conditions must result directly from the stimulus. Generally, evaluations of critically oriented HCI are not situations of experimental control. Researchers rarely compare a control and experimental group or explicitly attempt to demonstrate causal relationships between some manipulated stimulus and some independent response. That said, many of the examples discussed above, both from our work and from others, evidence an interest in showing a type of causal relationship between a stimulus (i.e., design intervention) and a response (i.e., critical thinking or reflection). SYNTHESIS AND DISCUSSION Broadly speaking, engagement with these tropes can legitimate new research directions while also shaping the nature of system evaluation and particularly the modes of engagement between researchers and participants. As we discuss in the next sections, these ramifications are not solely limited to individual research programs, and also bear important consequences for the critically oriented community as a whole. Engagement and Tempered Radicals In previous sections we described how, when compared to Agre’s work in AI, the challenge of simultaneously engaging two audiences emerges as a quality unique to HCI. However, negotiating relationships between different communities while challenging the status quo is a practice that also finds parallels in other fields. We have found a helpful parallel between ideas espoused in critical technical practice and a concept from organizational science of the “Tempered Radical” [31]. Tempered radicals are individuals who identify with and belong to a certain organization while simultaneously being committed to a cause, community, and ideology that is fundamentally different from and at times at odds with the dominant values of the individual's organization. Tempered radicals seek to challenge the status quo by building up legitimacy within their organization and identifying strategic opportunities to negotiate change within their institution. For example, based on the project’s motivation, we had a vested interest in demonstrating that the frame reflection tool itself, rather than any element of our study protocol, could effectively foster reflection on the framing of political issues. Similarly, we sought to show that the critical personal informatics themselves, and the generalizable strategies we used to design them, and not our interview questions, provoked people to consider the value commitments on which personal tracking technologies hinge (for example, to show that deliberate malfunction provoked critical reflection on infrastructure through reverse engineering.) In spanning these boundaries, the tempered radical adopts an “ambivalent” stance that leads to tension. However, as Meyerson and Scully describe [31], the tempered radical can also be remarkably well positioned to assess where and These evaluations were designed so that observed responses could be attributable to the system designs per se, rather 62 at which times “small wins” can be successfully enacted. Over the course of our analysis, we have been continuously impressed at the ways in which researchers have negotiated traditional HCI rhetoric and values to make room for and legitimate new research programs. Though this work is not without risks of “loss” or re-absorption into mainstream practices (e.g., in the case of [11]), we feel that critically oriented research practice can be used, gainfully, to challenge status quo values and practices in the design and evaluation of technology. we offer a generalizable stance or lens that critically oriented researchers can use to articulate their own tactical engagements with HCI’s discursive norms. We believe that such conversations will help critically oriented researchers work with each other to acknowledge and innovate evaluation methods in HCI. ACKNOWLEDGEMENTS We would like to thank the NSF under Grants Nos. IIS1110932 and IIS-1217685; the Intel Science & Technology Center for Social Computing; Melissa Mazmanian and the Doctoral Research and Writing Seminar at UC Irvine; and the anonymous reviewers for their wonderful feedback. Politics of Evaluation An essential component for the success of the “tempered radical” is their affiliation with other like-minded members of their organization. Previously, we have described critically oriented researchers as people engaging simultaneously with two audiences, their participants and the broader HCI academic community. However, methodological decisions made by critically oriented researchers also impact a third group of people, namely other critically oriented researchers in HCI. Here we channel an argument made by Cohn et al. that methods used in designing and evaluating have more general consequences for discursive and practical action because methods enable specific discourse and forms of knowledge production [14]. In other words, just as the HCI community enforces a set of discursive norms, so do critically oriented sub-communities within HCI. In this sense, critically oriented researchers should be conscious of their role in normalizing practices for future researchers who wish to engage in related forms of subversion. We recognize that mainstream HCI publications may not always be the most appropriate venues for these discussions (although it has certainly been possible to write about tensions while couching them in lingua franca rhetoric of success, e.g. [20]). We do believe that conversations among critically oriented researchers lead to helpful research contributions (e.g. [20, 5, 32]) and hope that our analysis can be used to articulate the trade-offs that critically oriented researchers make when engaging with HCI practices. REFERENCES 1. Agre, P. Computation and human Cambridge University Press, 1997. Experience. 2. Agre, P., Chapman, D. 1987. Pengi: an implementation of a theory of activity. In Proceedings of the sixth National conference on Artificial intelligence - Volume 1 (AAAI'87), Vol. 1. AAAI Press 268-272. 3. Agre, P. "Toward a critical technical practice: Lessons learned in trying to reform AI." Bridging the Great Divide: Social Science, Technical Systems, and Cooperative Work, Mahwah, NJ: Erlbaum (1997): 131157. 4. Aoki, P. M., Honicky, R. J., Mainwaring, A., Myers, C., Paulos, E., Subramanian, S., & Woodruff, A. "A vehicle for research: using street sweepers to explore the landscape of environmental community action." In Proc. CHI ‘09 5. Bardzell, S., Bardzell, J., Forlizzi, J., Zimmerman, J., and Antanitis, J. "Critical design and critical theory: the challenge of designing for provocation." In Proc. DIS ‘12 6. Bardzell, J., Bardzell, S, and Stolterman, E. 2014. Reading critical designs: supporting reasoned interpretations of critical design. In Proc. CHI ‘14 7. Bardzell, J., and Bardzell, S. 2013. What is "critical" about critical design?. In Proc. CHI ‘13 CONCLUSION By looking back on our research and by thinking through our analysis of the values we engage and their trade-offs, it becomes possible to imagine potential future trajectories for our projects. For instance, in the frame reflection example, we could change the conditions of the demand characteristic reduction to allow sharing information about framing that we previously withheld from participants. Instead, we could re-negotiate the “response” portion of the stimulus response to ask not whether the system makes people reflect on framing but rather to evaluate the ways in which participants then conceptually engage with ideas of framing using the system. More broadly, what we hope to contribute through this analysis is not a generalizable method or framework for critically oriented evaluation or engagement with the traditional HCI lingua franca. Rather, 8. Baumer, E. P. S., Cipriani, C., Davis, M., He, G., Kang, J., Jeffrey-Wilensky, J., Lee, J., Zupnick, J., and Gay, G. K. (2014). Broadening Exposure, Questioning Opinions, and Reading Patterns with Reflext: a Computational Support for Frame Reflection. Journal of Information Technology & Politics, 11(1), 45–63 9. Baumer, E. P. S., Halpern, M., Khovanskaya, V., & Gay, G. Probing the Market: Using Cultural Probes to Inform Design for Sustainable Food Practices at a Farmers’ Market. In J. H. Choi, M. Foth & G. Hearns (Eds.), Eat, Cook, Grow: Human-Computer Interaction with Human- Food Interaction. MIT Press, 2014. 63 10. Blythe, M., Overbeeke, K., Monk, A., Wright, P. (eds). Funology: From Usability to Enjoyment. Kluwer Academic Publishers, 2003. 28. Kuznetsov, S., Odom, W., Moulder, V., DiSalvo, C., Hirsch, T., Wakkary, R., & Paulos, E. HCI, politics and the city: engaging with urban grassroots movements for reflection and action. In CHI EA '11. 11. Boehner, K., Vertesi, J., Sengers, P., & Dourish, P "How HCI interprets the probes." In Proc. CHI ‘07 29. Lieberman, H. The Tyranny of Evaluation, http://web.media.mit.edu/~lieber/Misc/TyrannyEvaluation.html 12. Boehner, K., Sengers, P., and Warner, S. 2008. Interfaces with the ineffable: Meeting aesthetic experience on its own terms. ACM Interact. 15, 3, Article 12 (December 2008), 29 pages. 30. Löwgren, J. 2013. Annotated portfolios and other forms of intermediate-level knowledge. interactions 20, 1 (January 2013), 30-34. 13. Chong, D., & Druckman, J. N. (2007). Framing Theory. Annual Review of Political Science, 10(1), 103–126. 31. Meyerson, D., and Scully, M. "Crossroads tempered radicalism and the politics of ambivalence and change." Organization Science 6.5 (1995): 585-600. 14. Cohn, M., Sim, S., and Dourish, P. 2010. Design methods as discourse on practice. In Proc. GROUP ‘10 32. Michael, M. (2012). “What are we busy doing?” Engaging the idiot. Science, Technology & Human Values, 37(5), 528-554. 15. DiSalvo, C. Adversarial Design. The MIT Press, 2012. 16. Dourish, P. Where the action is: the foundations of embodied interaction. MIT press, 2004. 33. Olson, J., and Kellogg, W. Ways of Knowing in HCI. Springer, New York, NY, 2014. 17. Entman, R. M. (1993). Framing: Toward Clarification of a Fractured Paradigm. Journal of Communication, 43(4), 51–58 34. Orne, M. T. (1962). On the social psychology of the psychological experiment: with particular reference to demand characteristics and their implications. American Psychologist, 17(11), 776–783. 18. Gaver, W. W., Boucher, A., Pennington, S., & Walker, B. (2004). Cultural probes and the value of uncertainty. interactions, 11(5), 53-56. 35. Pierce, J. 2012. Undesigning technology: considering the negation of design by design. In Proc. CHI ‘12 19. Gaver, William W., Jacob Beaver, and Steve Benford. "Ambiguity as a resource for design." In Proc. CHI ‘03. 36. Pierce, J. and Paulos, E. “Counterfunctional things: exploring possibilities in designing digital limitations.” In Proc. DIS ‘14 20. Gaver, W., Bowers, J., Kerridge, T., Boucher, A., & Jarvis, N. "Anatomy of a failure: how we knew when our design went wrong, and what we learned from it." In Proc. CHI ‘09. 37. Polletta, F., Pierski, N., Baumer, E. P. S., Celaya, C., & Gay, G. (2014). A “Peopled” Strategy of Frame Reflection. In Annual Meeting of the American Sociological Association (ASA). San Francisco. 21. Gaver, W., Boucher, A., Law, A., Pennington, S., Bowers, J., Beaver, J., Humble, J., Kerridge, T., Villar, N., and Wilkie, A. “Threshold devices: looking out from the home”. In Proc. CHI ‘08 38. Preece, J., and D. Maloney-Krichmar. "The humancomputer interaction handbook." (2003): 596-620. 22. Goffman, E. (1974). Frame Analysis. Cambridge, MA: Harvard University Press. 39. Schön, D. A., & Rein, M. (1994). Frame Reflection: Toward the Resolution of Intractable Policy Controversies. New York: Basic Books. 23. Hirsch, T. 2010. Water wars: designing a civic game about water scarcity. In Proc. DIS ‘10 40. Sengers, P., Boehner, K., Mateas M., and Geri Gay. 2008. The disenchantment of affect. Personal Ubiquitous Comput. 12, 5 (June 2008), 347-358. 24. Khovanskaya, V., Baumer, E.P.S., Cosley, D., Voida, S. & Gay, G. “Everybody knows what you’re doing: A critical design approach to personal informatics.” In Proc. CHI ‘13 41. Sengers, P., Boehner, K., David, S., & Kaye, J. J. "Reflective design." In Proc. CC ‘05 25. Kuznetsov, S., Hudson, S., and Paulos, E. "A low-tech sensing system for particulate pollution." In Proc. TEI ‘14 42. Sengers, P., and Gaver, B. “Staying open to interpretation: engaging multiple meanings in design and evaluation.” In Proc. DIS ‘06 26. Kuznetsov, S., Davis, G. N., Paulos, E., Gross, M. D., & Cheung, J. C. 2011. Red balloon, green balloon, sensors in the sky. In Proc. UBICOMP ‘11 43. Silberman, M., Blevis, E., Huang, E., Nardi, B. A., Nathan, L. P., Busse, D. Preist, C., and Mann, S."What have we learned?: a SIGCHI HCI & sustainability community workshop." In CHI EA'14. 27. Kuznetsov, S., Davis, G., Cheung, J., & Paulos, E. “Ceci n'est pas une pipe bombe: authoring urban landscapes with air quality sensors.” In Proc. CHI ‘11 44. Suchman, L. Human-machine reconfigurations: Plans and situated actions. Cambridge University Press, 200 64 Note to Self: Stop Calling Interfaces “Natural” Lone Koefoed Hansen & Peter Dalsgaard Department of Information Studies and Digital Design, Participatory Information Technology, Aarhus University Helsingforsgade 14, Aarhus N, Denmark koefoed@cavi.au.dk, dalsgaard@cavi.au.dk (NUI) and a verb, “natural interaction”. When viewed from the point of view of interaction design, this is rather strange, given that the experiential qualities of technologies are very often hard to describe; in order to understand and design meaningful interfaces, we need all the help we can get from the words we have at our disposal. Since we in the HCI community evaluate, design, or research relations between a user and a technological setup, we must be better at reminding ourselves that we also need to develop the way that we talk about and characterize the nuances of this interaction. ABSTRACT The term “natural” is employed to describe a wide range of novel interactive products and systems, ranging from gesture-based interaction to brain-computer interfaces and in marketing as well as in research. However, this terminology is problematic. It establishes an untenable dichotomy between forms of interaction that are natural and those that are not; it draws upon the positive connotations of the term and conflates the language of research with marketing lingo, often without a clear explanation of why novel interfaces can be considered natural; and it obscures the examination of the details of interaction that ought to be the concern of HCI researchers. We are primarily concerned with identifying the problem, but also propose two steps to remedy it: recognising that the terminology we employ in research has consequences, and unfolding and articulating in more detail the qualities of interfaces that we have hitherto labelled “natural”. In this paper, we continue the brief statement in [6] and argue that the use of terms such as “natural” is problematic because the terminology highlights qualities that it does not help us understand and explain adequately, obscuring important aspects at the same time. To frame it in the spirit of the Critical Alternatives conference, we will first offer a critique and then outline alternatives. Author Keywords “NATURAL” USER INTERFACES Natural user interfaces; criticism; terminology. We specifically wish to criticize the way the term natural is employed to describe user interfaces—an issue that we see not only in marketing, but also in our own work and interaction with colleagues in the field. Even if [13] attempted to dismiss the term in 2010, we nevertheless see university courses on designing for NUIs, workshops at academic conferences with NUI in the title, as well as an increasing number of publications in the ACM Digital Library that explicitly mention NUI (from 19 publications in 2007 to 85 in 2014), of which only few (e.g. [14]) seek to discuss and clarify the use hereof. At present, a wide array of interfaces bears the label. The most common are touch screens, gesture-based interaction, and speech recognition, but the term also encompasses stereo 3D and haptic interfaces [12]. It even extends to specific types of interactions with wellknown instruments, such as the trackpad on a MacBook Air that can be adjusted to offer users a “natural scroll direction” (see Figure 1). ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION The words we use to describe technology matter. They carry with them connotations, establish expectations about the role and use of an object or an interface, and foreground particular qualities, while obscuring others. In many areas of interaction design and HCI, especially the ones pertaining to technical aspects of interaction, rich terminologies are continuously being developed in order to address the intricacies of a particular field. However, in other areas, often those pertaining to the experience and perceived qualities of interfaces, the terminology is disturbingly unnuanced. One instance of this is “natural”, a catch-all term that has seen widespread use in industry as well as in academia, both using it as a noun, “natural user interface” The above list shows that the “naturalness” of an interface is neither defined nor delimited by the underlying technologies, nor does it refer to something that exists in, or emerges from, nature as it is clearly fully man-made and artificial. Rather, the common conception of a natural user interface refers to the experience of interacting with or through the interface, as Microsoft Research Labs’ description states “People already use gesture and speech to interact with their PCs and devices; such natural ways to interact with Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21316 65 meaning of a particular word through time in order to understand its connections to other words. Naturalized representations An important part of Foucault’s discourse theory is that a discourse is practically invisible in the period in which it exists—it is often naturalized. In a Foucauldian sense, the dominant discourse of a given period (which can extend over hundreds of years) makes it hard to recognize that the way the world is framed and experienced could be in any other way than it is. Though operating on a smaller scale, French semiotician Roland Barthes’ concept of the naturalized connotation is similar; it describes when a connotation, i.e. a representation or understanding of something, is mistaken for a denotation, i.e. a fact, a given, a natural occurence [1]. What Barthes saw was that this communication strategy is used widely in politics and advertising, and he argued that a central aspect of being a modern citizen is to be a skilled sign-reader, i.e. to learn how to spot and challenge attempts of deception. Figure 1: Scroll Direction Tutorial for Apple Trackpad. with technologies make it easier to learn how to operate them” [9]. On Wikipedia, a NUI is described as “a user interface that is effectively invisible, and remains invisible as the user continuously learns increasingly complex interactions. The word natural is used because most computer interfaces use artificial control devices whose operation has to be learned.” In this definition we have at least three different concepts of what “natural” might mean: it is unnoticable (“effectively invisible”) since it does not involve a physical (“artificial”) input device; it is a walk-up-and-use interface (it doesn’t have to be learned); and becoming a super-user is easy. This definition clearly echoes Weiser’s vision for the computer for the 21st Century, which would “disappear into the background” [17]; indeed Weiser himself employed metaphors from nature, stating that using computers would become as refreshing “as taking a walk in the woods”. These strategies of ‘deception’ are not always on purpose, though, as becomes clear when we turn to feminist theories and critiques where making naturalizations visible in order to change them is a central goal. With language and discourse analyses as “weapon,” the goal is to make visible how particular ideas of gender, minorities, etc. are part of power structures that are deeply embedded in a given culture. “We must be wary of the suggestion that anything is natural,” is how feminist philosopher Haslanger summarizes this stand, and although she specifically argues that not everything is a cultural construction, language is not one of those things [6]. Language is understood as an encoder of meanings and norms, for instance when the word “man” both means a person of the male sex and the human race in general. Since terminology can be a barrier to accurate communication, the argument is that words must be chosen very carefully, allowing for greater nuances in our understanding of the world and the possibilities we have for changing it. See [15] for an excellent overview of the intersection of feminist critique and language. WHY IS IT PROBLEMATIC TO CALL INTERFACES “NATURAL”? None of the above readings are intended by those that use the term, obviously, but this only points to a core problem and to our purpose with this paper: words are not only descriptive but also formative—they are important because they help constitute the way we perceive (our possibilities to act in) the world. Central to many language theories is the idea that the world is discursively constructed—we understand the world as well as our possibilities herein also through the way we put it into words. Lakoff’s cognitive linguistics is an example of this; in [9] he and Johnson famously highlight how metaphors shape our lived experience of the world. Likewise, Foucault’s concept of “discourses” places language as an important part of power structures that change over time, enabling certain developments of society whilst making others less likely [4]. On a micro as well as on a macro level all highlight that it is important to question the terminologies we use. This is done with critique, here understood both in the Frankfurt School tradition of using critical theory to investigate hidden structures, and in the etymological tradition of tracking the In summary, these examples of discourse critique from other fields than HCI show that a constant focus on representation and language in general and on terminology in particular is necessary if we don’t want the words we use to describe the world to prevent us from exploring its potentials fully. Following this, we argue below that the term “natural”, when referring to user interfaces, is problematic in at least three different respects, namely in regards to the unclear meaning of the term, the issues that it foregrounds, and even more so the issues that it obscures. What does “natural” mean? The first and arguably most obvious critique to raise is that the meaning of the word natural is unclear and imprecise. Considering the definitions outlined earlier, it is difficult to decipher exactly what the common denominator for natural 66 user interfaces is, unless we move to a very abstract level: NUIs require no instruments, they are easy to learn, and they disappear into the background. These are dubious claims for most NUIs. Furthermore, “natural” is being employed in ways that contradict this very broad definition, as exemplified by the “natural scroll direction”, which clearly indicates that instruments do exist and that there are certain “natural” ways of using them, although, ironically, this way needs to be demonstrated in a tutorial. the realm of interaction design and HCI is unclear and imprecise at best, and self-contradictory at worst. What does “natural” foreground, what does it obscure, and why does it matter? Employing discourse analysis to this field, it is clear that the term “natural” is not neutral; rather, it bears with it primarily positive and desirable connotations. Weiser’s walk in the woods is “refreshing”, the Leap Motion “senses how you naturally move your hands [so that you can do] things you never dreamed possible”, [10] and NUIs are “more intuitive, engaging and captivating” [12:1]. From a pragmatic point of view, the desire to label interfaces as natural is understandable. Owing to the imprecise definition of the term, which easily accommodates new technologies as they emerge, this is a non-committal way of imbuing a novel and unfamiliar product with positive associations. Especially in a marketing discourse, “natural” handily uses the naturalization trick that Barthes talked about, marking objects as inescapably good; who wants something unnatural if you can get the natural? However, this approach is unsuitable in a research discourse, where—following Barthes—a central aspect of being a highly skilled and reflective researcher is to be a highly skilled reader and critic of signs, given that words shape the way we give agency to the world. In many NUIs, instruments are quite clearly still in play: a gesture based interface such as the Leap Motion controller may obviate the need for a mouse, but it is still clearly present as an instrument that requires the user to maneuver the hands in a specific manner within a constrained interaction zone; likewise, speech recognition may obviate the need for a keyboard, but it still needs to be activated and deactivated in a specific manner, e.g. by picking up a phone and uttering particular activation commands, be it “OK Google” or the anthropomorphic counterpart “Siri”. In continuation, these technologies have not eliminated the need for learning how to interact, but rather introduced new interaction modes for us to learn. We need to learn how to do micro-gestures that can be sensed with a reasonable level of precision by the Leap Motion controller; we need to learn how to speak in a monotonous, robotic voice (preferably without accents) to minimize speech recognition errors. In short, we need to learn how to use our bodies as instruments in new ways to accommodate the new “natural” interfaces. Thus, neither the claim of getting rid of instruments nor the claim of not having to learn them seems to hold true. Furthermore, the broad use of the term natural user interface makes “user” a generic concept, implying that some users are more (un)natural than others—as when a Kinect makes it impossible to operate a natural interface when missing an arm or sitting in a wheelchair. By using as biased and imprecise a term as “natural”, there is a risk of conflating these two very different discourses in HCI, which makes it particularly pertinent for researchers to strive for terminological precision—a prerequisite for the development of a research field. Precision prompts designers to go beyond the surface of a phenomenon, examining and articulating in greater detail how and why it unfolds as it does. Further, precision makes arguments presented in a research community contestable and opens them to joint elaboration, discussion, and development—an essential part of a research field. Looking beyond these murky definitions of what makes an interface natural, there also appears to be an underlying assumption that the use of instruments is unnatural, or at least that some instruments are more natural than others. These assumptions, too, are problematic. Firstly, they run counter to the study of human evolution [as discussed in e.g. 5], which indicates that the use of instruments is a crucial aspect of human nature. Secondly, the distinction between more or less natural interfaces is questionable and potentially untenable; e.g. touch and gesture interfaces are referred to as natural user interfaces, but it is unclear how a four-finger swipe is a more natural way of switching to a new desktop space than moving a mouse pointer to a hotspot corner; or how it is more natural to wave your arms in front of a camera mounted on a tv screen in order to calibrate camera tracking so that you can control an in-game cartoon avatar than using a game controller? Thirdly, they ignore the body of work in HCI and beyond that addresses how the use of instruments becomes internalized into human action [e.g. 3]. In summary, the definition of natural in While this may appear to be terminological nit-picking, we argue that the words we use matter, even if we are more interested in developing technologies than in discussing semantics. Words have real consequences for how we understand the interfaces we make; for how well we can convey what they do and why to the others; and for how we as a field of both researchers and practitioners can work together to further develop them. The problem with using a term such as natural is that it emphasizes purportedly positive qualities of an interface, but it does not help us understand if, how, and why the interface works. If we are interested in anything beyond promoting a novel interface—and as HCI researchers and practitioners, we ought to be—then we need a better vocabulary to understand, discuss, criticise, and develop these interfaces. IMPLICATIONS Changing a discourse is difficult, but that should not keep us from proposing alternative courses of action, especially in a field that transforms as rapidly as HCI. We consider 67 two intertwined steps towards addressing the over-reliance on the term “natural”. The first step is to raise awareness that this is in fact problematic. We have laid out a series of criticisms here, from theoretically grounded arguments to more practically oriented ones. While we consider it highly relevant to discuss this topic in the context of the Critical Alternatives conference, we need to bring these criticisms to a wider HCI audience, including venues that might be less amenable to them, if we are to have any hope of having an impact. For this to be constructive, it must be combined with the second step, namely to develop a more refined vocabulary for articulating and addressing what is currently covered by the umbrella term “natural”. novel interfaces; on the opposite, we are advocating the development of a discourse that is rich enough to address the intricacies of these interfaces. REFERENCES 1. Barthes, R. 1993 [1957], Mythologies. Vintage, London. 2. Blackler, A.L. & Hurtienne, J. 2007, Towards a unified view of intuitive interaction: definitions, models and tools across the world. MMI-Interaktiv, 13. pp. 36-54. 3. Bødker, S. Through the interface: A human activity approach to user interface design. Hillsdale, NJ, Lawrence Erlbaum, 1991. 4. Foucault, M. 2001 [1966] The order of things: an archaeology of the human sciences, Routledge. The types of interfaces currently labelled natural are being developed and adopted at a rapid pace, and we need to also develop and adapt a language for interfaces able to match the complexity of them. We propose that this can be inspired and informed by recent contributions seeking to unfold the notion of “intuitive interfaces”. As with natural, intuitive is a term laden with positive connotations and attached to many an interface, but also one that in itself does not say much about the qualities of an interface and its use. However, recent contributions, including [2;8;16] have examined in more depth what intuitive interaction might mean, and in turn what can make an interface intuitive. At an overarching level, [2] examines properties how prior cultural, sensori-motor, and acquired knowledge influence what is perceived as intuitive; in an attempt to develop tools for examining and refining specific interfaces, [16] has developed a questionnaire for studying if and how users of a given interface conceive of it as being intuitive; and [8] examines what intuitive interaction means in installations in public spaces, drawing among others on the work of [2]. These contributions highlight complementary approaches to developing a more extensive vocabulary of interaction in ways that are theoretically well founded as well as of value to HCI practitioners. since a richer understanding of the notion of intuitive interaction can in turn help developers create better products. The study of the so-called natural user interfaces is in need of similar contributions. 5. Gibson, K.R. & Ingold, T. (eds.) 1993, Tools, Language, and Cognition in Human Evolution, Cambridge University Press. 6. Hansen, L.K. 2014, What's in a word?: Why natural isn't objectively better, Interactions, 21, 1, pp. 22-23. 7. Haslanger, S. 2000, Feminism in metaphysics: Negotiating the natural., in Hornsby & Fricker (eds.), The Cambridge companion to feminism in philosophy, Cambridge University Press. 8. Hespanhol, L. & Tomitsch, M. 2015, Strategies for Intuitive Interaction in Public Urban Spaces, Interacting with Computers 2015. 9. Lakoff, G. and Johnson, M. 1980, Metaphors we live by. University of Chicago Press, Chicago. 10. Leap Motion. “Leap Motion, Product”. http://leapmotion.com/product, accessed 2015-02-16. 11. Microsoft Research “NUI: Natural User Interface”, http://research.microsoft.com/en-us/collaboration/ focus/nui/default.aspx, accessed 2015-02-16 12. Murphy, S. 2013, White Paper: Design Considerations for a Natural User Interface (NUI). Texas Instruments. 13. Norman, D. 2010, Natural user interfaces are not natural. interactions 17, 3. CONCLUSION 14. O'hara, K., Harper, R., Mentis, H., Sellen, A. and Taylor, A. 2013, On the naturalness of touchless: Putting the “interaction” back into NUI. ACM Trans. Comput.Hum. Interact. 20, 1, Article 5. Academic fields outside of HCI have long-standing traditions of discussing the terminology applied within the field itself. Barthes is famous for pointing out how many of the things we say and do are ideologies in camouflage, altough often unintendedly so. This is akin to Luhmann’s notion of the blind spot, as well as Marxist and feminist cultural critiques, all focusing on how the way we talk about things shapes our perception of and actions in the world. In this paper, we pursue a similar line of critique in order to examine the notion of “natural” user interfaces. We have argued that the term “natural” does not at all suffice for articulating and understanding these types of interfaces; the use of the term is unnuanced and marred by imprecision and contradictions. Our objective is not to critique the development of 15. Saul, J. 2012, ‘Feminist Philosophy of Language’, in Zalta, E. N. (ed.) The Stanford Encyclopedia of Philosophy (Winter 2012 Edition),. 16. Ullrich, D. & Diefenbach, S. 2010, INTUI. Exploring the Facets of Intuitive Interaction. In J. Ziegler & A. Schmidt (Eds.) Mensch & Computer 2010 (pp. 251260). München: Oldenbourg. 17. Weiser, M. 1991, The computer for the 21st century. Scientific American, Sept. 1991, pp. 94-104. 68 Gaza Everywhere: exploring the applicability of a rhetorical lens in HCI Omar Sosa-Tzec Indiana University 901 E. 10th Street Bloomington, IN 47408 omarsosa@indiana.edu Erik Stolterman Indiana University 901 E. 10th Street Bloomington, IN 47408 estolter@indiana.edu Martin A. Siegel Indiana University 901 E. 10th Street Bloomington, IN 47408 msiegel@indiana.edu An enthymeme is an argument that provides the claim or conclusion to the audience, but leaves one premise unstated, thus it is also known as a truncated syllogism. It then becomes the task of the audience to fill in that premise [7, 13]. In the case of a visual enthymeme, the observer detects or interprets a claim or conclusion from the observation of the composition, whose elements serve as the stated premise, and fills in the unstated premise through reasoning derived from such observation [13]. ABSTRACT By examining application software as a type of rhetorical artifact, it is possible to highlight its social, ethical and moral implications. In this paper, we explore one possibility for such a lens: application software functioning as a visual enthymeme. To explore the applicability of that concept in HCI, we analyze one web application as a first step. In our analysis, we observe that interaction and usability are two features that support an application in functioning as a visual enthymeme. Also, online sharing could help the user take the role of the arguer. Our analysis allows us to outline the elements of a user-centric persuasive experience and shows promise for further explorations regarding the applicability of rhetoric in HCI. In this paper we argue that rhetoric is an appropriate and useful lens in HCI. To explore the applicability of this lens, we focused on one possibility: application software functioning as a visual enthymeme. In particular, we focus on applications with a GUI. As a first step of this exploration, we chose to analyze a web application known as “Gaza Everywhere” which compares the Strip of Gaza with other territories in the world [16]. During news coverage regarding the Gaza Strip in the summer of 2014, relevant tweets, and an online publication from The Independent [3] suggested that Gaza Everywhere could function as a visual enthymeme. The application’s interface contains demographic statistics about the Gaza Strip. It also contains a Google Maps widget that overlaps the Gaza Strip with any territory selected by the user. Moreover, it allows the user to adjust the position and orientation of the Gaza Strip in the map. Tweets about the application are also shown in the interface. Through the analysis of the application’s interface, interaction, and usage, we observe that Gaza Everywhere illustrates a case of visual enthymeme. Also, we notice that interaction and usability are two features through which the user can fill in the unstated premise. Moreover, online sharing provides a means for the user to take the role of someone making an argument. Author Keywords Visual Enthymeme; Rhetoric; Persuasive Technology; Experience Design; Design Theory ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): User Interfaces – Graphical User Interfaces (GUI), Screen Design, Theory and Methods INTRODUCTION Rhetoric is a critical human ability that permeates all forms of human communication [7,10]. Rhetoric can mobilize people; it can influence their perception of reality and truth [7]. Many people are exposed to symbolic, persuasive, and visual artifacts every day [8,12,13]. Some of these artifacts are intended to make a point, increasing awareness about a particular situation, and thus change the beliefs, attitudes or values of people. For example, the 1996 advertisement of United Colors of Benetton showed three hearts, labeled “White, Black, Yellow,” intended to point out that people are equal regardless of their skin color.1 This is an example of a visual enthymeme, a form of argument aimed at persuading an audience, rather than finding truth. These visual enthymemes can be found in printed advertisements, televised political campaigns, and documentary photographs. This paper is structured as follows. First, we present our case study. Second, we discuss some observations derived from the analysis of Gaza Everywhere. Third, we emphasize possible ways in which a rhetorical lens could be advantageous in the analysis and design of software. CASE STUDY Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM Object of analysis The interface of Gaza Everywhere (GE) is mostly composed of a Google Maps widget. It also includes a Twitter widget and some demographic information about the Gaza Strip. At the beginning of the interaction, the map shows a 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark 1 Advertisement available at http://goo.gl/NIR224 DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21314 69 translucent red shape of the Gaza Strip overlaying its territory (Fig. 1). As a result of a search, the map places the translucent shape on the territory of interest. At any time during the interaction, the user can zoom and drag the map, as well as drag and rotate the translucent shape. Also, the software allows the user to share the current state of the map online or take a screenshot of it. The user can access the source code through GitHub or embed the application in a webpage. According to A. Nassri, author of GE, the web application is intended to “help visualize Gaza’s relative size in comparison with other territories in the world.” Most of the tweets shown in the interface’s Twitter widget align with this description. Yet, some of them express a political opinion, including a note from i100, an online publication from The Independent [3]. In that note, images derived from the interaction with GE are used to support the note’s claim, “Gaza Everywhere app highlights true scale of humanitarian crisis”. The web application is embedded at the end of the note, so the user can explore the claim herself. Figure 1. Initial state of GE. Screenshot from the WWW. crease her awareness while doing visual comparisons with GE. Unlike other traditional forms of visual enthymeme, GE allows the user to play with the composition at will. That means altering the GUI as a result of interaction with the application. Traditional forms of visual enthymeme tend to be static [12,13]. However, the visual comparisons in GE can be done at any time and place as long as the user has access to the Internet and a device with a web browser. Those visual comparisons could elicit experiential knowledge, particularly the territories in which the user has lived. In that regard, GE could elicit sensations and emotions as a consequence of interaction, supporting not only the user experience but also the detected or interpreted claim. Thus, interaction and experiential knowledge help the user to fill in the unstated premise in order to support the detected or interpreted claim. Nevertheless, the user’s awareness is affected after each interaction and by other sources of information. Consequently, the user might revisit the detected or interpreted claim, which makes the persuasive effect of GE evolve with the user. Analysis and Observations The impact of GE relies on its GUI. The physicality of the interface depends on the device being used to interact with the web application (e.g., smartphone). However, the device has no major effect on the visual information or the interaction (i.e., search, interaction with the map and sharing online). GE doesn’t employ audio in its composition. However, the visual information conveyed in the map is inaccurate since Google employs a variation of the Mercator projection. In this sense, the GE’s intent could be regarded as illustrative, rather than affirmative. GE’s intent is not oriented to provide a territorial truth. The user might be familiar with Google Maps, which makes the interaction with GE simple. Looking for a particular territory and adjusting the map is a task that should not represent a major usability issue. GE’s usability supports filling in the unstated premise. Later, it is relatively easy for the user to share an image or link of the current map. Another user that observes that map follows the same dynamics as the first user: detect or interpret a claim and validate it based on her awareness about the Gaza Strip and the territory shown on the map. In a similar fashion, access to the source code allows a knowledgeable user to modify GE in order to reflect her vision of the web application and thus affect other users. These characteristics give GE a versatility that is not commonly seen in traditional forms of visual enthymemes. Regardless of the stated intent of GE, each user has a different awareness regarding the conflict in the Gaza Strip, which affects the perception of the application’s intent. For example, some users might see GE as a means to emphasize political issues. The note from The Independent’s i100 is an exemplar of this situation. Gaza Everywhere gives room for several interpretations that are related to that awareness. The interpretations correspond to claims that can be only supported through the combination of reasons derived from the user’s awareness and reasons provided through the interface’s information. These characteristics allow GE to be classified as a visual enthymeme [13]. Regarding the interface, the detected or interpreted claim is supported by the overlap shown by the map in juxtaposition with the demographic information shown on the interface. Yet, such a claim is only accessible after interaction with the web application. The initial state of GE represents a visual identity wherein Gaza is equal to Gaza since the translucent shape and the territory shown on the map are the same, including their dimension and orientation. Nonetheless, the shape becomes a metaphorical visual unit of measurement after interaction with the web application. Later, the capability of manipulating the shape helps the user in- DISCUSSION The case of Gaza Everywhere exemplifies how application software can work as a visual enthymeme. An enthymeme, either linguistic or visual, is rooted in probability; it relies on the beliefs, presuppositions, and experiences of the audience [7]. Consequently, software functioning as a visual enthymeme not only relies on its design in order to be persuasive; it also relies on the user’s beliefs, presuppositions, 70 topoi, the lines of reasoning for inventing persuasive arguments [7], nor rhetorical figures for the composition of software functioning as a visual enthymeme. Thus, we point out the opportunity to explore the interrelation and applicability of rhetoric to the design of application software and the so-called user experience. Below, we highlight some implications of adopting a rhetorical lens in HCI and call for further research efforts in this direction. A rhetorical lens could help analyze or generate a design of application software. Efforts related to analyzing a design would contribute to the existing body of knowledge in HCI focused on the persuasive and phenomenological aspects of technology [e.g., 2,9,19 and Fallman in 24]. Such an analysis could reveal aspects that help the designer understand the dynamics and intricacies of a persuasive user experience (e.g., Fig. 2). A rhetorical lens could aid the designer to criticize and reflect upon the design of application software, especially in the comprehension of denotations and connotations conveyed through the design, and thus the meaning that emerges during the user experience [6,15, 22]. Figure 2. Elements of the user experience when interacting with software functioning as a visual enthymeme. and experiences in order to leverage its persuasive character. In that sense, the user is placed as the central element of a persuasive experience. Based on our analysis, placing the user in that position causes the persuasive character of the application software to be conditioned by: 1) the user’s computer and information literacy; 2) the user’s beliefs, presuppositions, and experiences, especially those related to the situation for which the application was intended; 3) the effect of the particular characteristics found in the context of use, including the influence of the discourse [10] about the situation for which the application software was intended and the rhetorical exchange with other people (Fig. 2). Efforts related to generating a design can contribute to many aspects of HCI. Here, we will emphasize how it contributes to the existing body of knowledge in HCI focused on design pedagogy [11,23]. A rhetorical lens would allow the designer to understand the design of application software as a composition intended to address a (rhetorical) situation that needs to be solved or improved [6,7,14,20]. Additionally, it provides vocabulary for the designer to characterize the participants involved in the user experience and their interrelations [e.g., 1,6,18 and Christensen & Hasle in 24]. This might support reflection-in-action [21] during the design process due to the designer’s awareness of the current rhetorical situation and the people involved, as well as the possible impact of the design approach used in composing the solution [e.g., 14, 20,22]. Moreover, none of these conditions are static. After each interaction with the application software, the user gains knowledge and a certain rhetorical agency to influence the discourse related to the situation for which the software was intended. That influence occurs at least at a personal level. In terms of persuasive technology, this means inheriting an autogenous persuasive intent [9]. However, GE illustrates that the technical features of an application (e.g., online sharing), as well as its usability could expand the scope of that influence. One user could persuade other users to interact with the application. In terms of persuasive technology, the user of software functioning as a visual enthymeme could become an element of the exogenous persuasive intents [9] to other potential users. Moreover, the impact of such an influence could affect the designer and hence future designs. As we learned from GE, software working as a visual enthymeme outlines a user-centric persuasive experience, a set of complex, evolving relations of people, technology, and discourse (Fig. 2). Considering the design of application software as a composition encourages the exploration of tropes and schemes, the so-called figures of speech or rhetorical figures, as a generative tool for the designer. In graphic design, such an exploration has helped teach designers to conceptualize before execution, and to be aware about the social, moral and political dimensions of design [8]. An understanding about rhetorical figures could help the designer not only better comprehend the concept of metaphor and metonymy, probably the most used rhetorical figures in HCI, but also to notice and conceptualize interfaces and interactions beyond these two figures. OUTLINING THE APPLICABILITY OF RHETORIC IN HCI Throughout this paper, we have illustrated how application software can function as a visual enthymeme. The case of GE is pertinent to HCI because it illustrates that design is inherently persuasive [19] and functions as an argument [5,6], regardless of the designer’s intent. Certainly, the case of GE provides a glimpse of the implications of application software functioning as a visual enthymeme. Yet, it doesn’t completely illustrate the generative applicability of the concept. For instance, this analysis does not address the use of Besides understanding the rhetoricity of HCI, rhetorical knowledge and experience could expand the designer’s set of competency [17]. The designer could apply rhetorical knowledge in order to convey a design argument before peers, the client, and other stakeholders during the design process. Consequently, the designer could strengthen her ability to reflect-on-action [21]. Rhetoric entails persuasion, either in the form of communication that brings about 71 change in people [4] or in identification with common ideas, attitudes or material possessions [10]. This could help the designer achieve an empathetic state of alignment not only with the client but also with the user. Knowing about rhetoric reinforces the following goals: that the agency of all the people involved in the process should be managed, trust created, and a common understanding achieved [17]. 9. Fogg, B.J. (1998). Persuasive Computers: Perspective and Research Directions. In Proc. CHI ’98. ACM Press/Addison-Wesley (1998), 225-232 10. Foss, S.K., Foss, K.A., & Trapp, R. Contemporary Perspectives on Rhetoric. 2nd Edition. Waveland Press, 1991. 11. Getto, G., Potts, L., & Salvo, M.J. Teaching UX: Designing Programs to Train the Next Generation of UX Experts. SIGDOC ’13. ACM Press (2013), 65-69. CONCLUSION In this paper, we took a first step exploring one way through which a rhetorical lens can be applied to the analysis of software. We analyzed Gaza Everywhere, which illustrates the possibility of application software functioning as a visual enthymeme. This analysis allowed us to outline the set of complex, evolving relations involved in such a case, introducing a schema of user-centric persuasive experience. We consider that a rhetorical lens could bring awareness of the social, moral and ethical implications not only about a design but also about the user experience. Providing such a lens to designers could help them comprehend or become sensitive to the pervasiveness of everyday rhetoric, intentional and unintentional. The user experience could be interpreted as a rhetorical interface between one person, other people, and artifacts. Through this interface discourses are shaped – discourses that go back to the designers and influence the way and with whom they identify. 12. Handa, C. Visual Rhetoric in a Digital World. Bedford/St. Martin Press, 2004. 13. Hill, C.A., & Helmers, M. (Eds.) Defining Visual Rhetorics. Lawrence Erlbaum Associates, 2004. 14. Hullman, J., & Diakopoulos, N. Visualization Rhetoric: Framing Effects in Narrative Visualization. T. on Visualization and Computer Graphics 7, 12 (2011) 22312240 15. Kannabiran, G., & Graves Petersen, M. Politics at the interface: A Foucauldian power analysis. In Proc. NordiCHI 2010. ACM Press (2010), 695-698. 16. Nassri, A. Gaza Everywhere (2014, August 8). URL: http://ahmadnassri.github.io/gaza-everywhere/ 17. Nelson, H.G., & Stolerman, E. The Design Way. MIT Press, 2012 ACKNOWLEDGMENT This work is supported in part by the National Science Foundation (NSF) Grant Award no. 1115532. Opinions expressed are those of the authors and do not necessarily reflect the views of the entire research team or the NSF. 18. Price, J. A Rhetoric of Objects. SIGDOC ’01. ACM Press (2001), 147-151 . 19. Redström, J. Persuasive Design: Fringes and Foundations. W. IJsselsteijn et al. (Eds.) PERSUASIVE 2006, LNCS 3962 (2006), 112-122. Springer. REFERENCES 1. Arvola, M. The Mediated Action Sheets: A Framework for the Fuzzy Front-End of Interaction and Service Design. Proc. European Academy of Design Conference, (2013). 20. Rosinki, P., & Squire, M. Strange Bedfellows: HumanComputer Interaction, Interface Design and Composition Pedagogy. Computers and Composition 26, 3 (2009), 149-163. Elsevier. 2. Bardzell, J. Interaction Criticism: An introduction to the practice. Interacting with Computers 23, 6 (2011), 604621. Elsevier. 21. Schön, D.A. The Reflective Practitioner. How Professionals Think In Action. Basic Books, 1984. 22. Sosa-Tzec, O., & Siegel, M.A. Rhetorical Evaluation of User Interfaces. NordiCHI ’14. ACM Press (2014), 175178. 3. Barlett, E. Gaza Everywhere app highlights true scale of humanitarian crisis. i100 from The Independent (August 14, 2014). Shortened URL: http://goo.gl/b3t2lV 23. Torning, K., & Oinas-Kukkonen, H. Persuasive System Design: State of the Art and Future Directions. Persuasive ’09, ACM Press (2009). 4. Bostrom, R.N. Persuasion. Pretince-Hall, 1983. 5. Buchanan, R. Declaration by Design: Rhetoric, Argument, and Demonstration in Design Practice. Design Issues 2,1 (1985) 4-22. 24. Kord. Y., IJsselsteijn, W., Midden, C., Eggen B., Fogg, B.J. (Eds.) Persuasive Technology. 2nd International Conference on Persuasive Technology, PERSUASIVE 2007. Revised Selected Papers. Springer (2007). 6. Carnegie, T.A.M. Interface as Exordium: The Rhetoric of Interactivity. Computers and Composition 26, 3 (2009) 164-173. Elsevier. 7. Covino, W.A., & Jolliffe, D.A. Rhetoric: concepts, definitions, boundaries. Allyn and Bacon, 1995. 8. Ehses, H., & Lupton, E. Rhetorical Handbook. An Illustrated Manual for Graphic Designers. Design Papers 5 (1988), 1-39. 72 Human-computer interaction as science Stuart Reeves Mixed Reality Lab School of Computer Science University of Nottingham, UK stuart@tropic.org.uk between Newell, Card, Carroll and Campbell around the deployment of cognitive psychology for designing user interfaces, and the prospects of developing “a science of human-computer interaction” [43, 12, 44]. Since then there have been sporadic expressions—a tendency if you will— towards cultivating some element of ‘scientific disciplinarity’ for HCI. This may be seen in the form of panels and workshops on matters like scientific replication [58, 59] or interaction science [32] that have surfaced at the ACM CHI conference in the last few years. Most recently Liu et al. [36] and Kostakos [35] have argued that HCI is a poor scientific discipline when measured against other bone fide examples (such as those of the natural sciences or disciplines with ‘science’ in their title). In this analysis HCI is found devoid of central motor themes that are taken as a signature of thoroughbred scientific disciplines, thus representing a presumed failure of the HCI programme. Echoing the calls of Greenberg and Thimbleby in 1992 [27], work is thus required to make HCI “more scientific” [35]. ABSTRACT The human-computer interaction (HCI) has had a long and troublesome relationship to the role of ‘science’. HCI’s status as an academic object in terms of coherence and adequacy is often in question—leading to desires for establishing a true scientific discipline. In this paper I explore formative cognitive science influences on HCI, through the impact of early work on the design of input devices. The paper discusses a core idea that I argue has animated much HCI research since: the notion of scientific design spaces. In evaluating this concept, I disassemble the broader ‘picture of science’ in HCI and its role in constructing a disciplinary order for the increasingly diverse and overlapping research communities that contribute in some way to what we call ‘HCI’. In concluding I explore notions of rigour and debates around how we might reassess HCI’s disciplinarity. Author Keywords Science; disciplinarity; cognitive science. ACM Classification Keywords In exploring these complex debates, this paper addresses a range of cognate concerns in HCI: ‘science’, ‘disciplinarity’ and ‘design’. The argument I present in this paper contends that the status anxiety over HCI as an academic object has its origins in the early formulation of HCI’s research practice. This practice blended the application of cognitivist orientations to scientific reasoning with Simon’s view of design [56], in order to establish a particular research idea—what I refer to as the scientific design space. This guides both what human-computer interactions are, and how we investigate them. This idea, I argue, has configured how many HCI researchers relate to interactive artefacts in their work practices and thus shaping HCI’s disciplinary circumstances and discussions. H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION Human-computer interaction (HCI) is represented by a large and growing research community concerned with the study and design of interactive technologies. It rapidly emerged from the research labs of the 1970s, matching the lifetime of the Århus Decennial conferences. Perhaps characteristic of all ‘new’ research communities, anxieties have been expressed over its status as an academic object from the very beginning (indeed, as an early career researcher I myself have often felt a similar confusion about what HCI is as an academic object). Many of these anxieties centre around disciplinary shape and how that shape relates to ‘science’—the topic of this paper. It is not the intention of this paper to suggest that cognitivist scientific reasoning is the only orientation to reasoning present in HCI research. It is also not within the scope of this paper to fully map out the landscape of different forms of reasoning in HCI (e.g., the ‘designerly’), nor to evaluate the claims of different approaches compared with their achievements. Neither is it the intention to imply that disciplinary anxiety is solvable. Instead, this paper focusses upon cognitivist scientific reasoning and its expression through the scientific design space, arguing that this has been an important and persistent force in the broader logics of significant portions of HCI’s programme. This is despite Discussion about the role of science both in and of HCI can be traced to various formative exchanges in the early 1980s Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21296 73 repeated suggestions of transformative intellectual changes that HCI as a whole may have gone through. Law?), and no obvious shared commitment to a certain set of problems or ways of approaching them [36]. This paper firstly unpacks the debates that I broadly subsume into questions over the status of HCI as an academic object and its relationship to ‘science’; through this I detail two important anxieties that have animated this debate. I then relate these to a core concept—the scientific design space—that I argue emerged in HCI’s formative years at the confluence of cognitive science and interface engineering challenges. In this I revisit HCI’s relationship to cognitive science, and the introduction of a particular ‘picture of science’, tracking subsequent influence of this on attempts at crafting HCI’s disciplinary architecture. Finally, the paper’s discussion returns to evaluate these concepts. Debate about this incoherence problem has two poles of discussion: descriptions of ‘how things are’ (which often include explanations), and prescriptions of ‘how things should be’. A key way of describing present incoherence to HCI is the accumulation of diversity in its approaches over time. As Rogers states, “HCI keeps recasting its net ever wider, which has the effect of subsuming [other splinter fields such as CSCW and ubicomp]” [50 p. 3]. As a matter of this accumulation, HCI researchers have tended to ‘bolt on’ new approaches repeatedly. This accumulation has involved accommodating new epistemological perspectives and the disciplinary objects that come with them. This is perhaps best illustrated by theory and theorising in HCI, which has been accorded importance given its association as a signature of established disciplines. Beginning with a particular cohort of theories drawn from cognitive science and psychology, ‘theory’ in HCI has increasingly been adopted from diverse origin disciplines [50, 29]. This has introduced a great diversity of technical senses in which ‘theory’ is meant (see [50 pp. 16-17] for a descriptive account). Question marks are raised over what qualifies, what a theory is useful for, how they are to be organised, what the relationships between them actually are, how new theories should (or can) be developed, amalgamated, divided, or simply decided as ‘good’ or ‘bad’, relevant or out of scope. As Bederson and Shneiderman admit prior to an attempt to define theory in information visualisation and HCI, “a brief investigation of the language of theories reveals its confusion and inconsistent application across disciplines, countries, and cultures” [2]. THE STATUS OF HUMAN-COMPUTER INTERACTION An external view of HCI’s disciplinary status could assume that it is secure. For instance, in CHI 2007’s opening address, Stuart Feldman (then president of the ACM) described the HCI research as “absolutely adherent to the classic scientific method” [1]. But the picture from within HCI seems radically different. By reviewing broad discussions around HCI’s disciplinarity in this section, I intend to sketch a background for subsequently addressing the specifics of ‘science’ in HCI. Questions over HCI’s disciplinarity emerged early in its development. In 1987 ergonomist and HCI pioneer Brian Shackel asked during his INTERACT conference keynote whether “HCI was a discipline, or merely a meeting between other disciplines” [15]; a couple of years later, Long, Dowell, Carroll and others discussed what kind of discipline HCI might be described as [38, 11]. Although Carroll characterises this and his exchanges with Newell and Card as the “theory crisis” of the mid-1980s [11 p. 4], one only need glance at a standard textbook to notice that HCI seems still to be routinely presented by an ambiguous constellation of overlapping disciplinary descriptors (e.g., interaction design, user experience, etc.). The term itself is also problematised, and HCI can be taken to perhaps subsume or compose these various related descriptors. For example, one position adopted by a key textbook— Interaction Design: Beyond Human-Computer Interaction [51]—formulates HCI as a contributing academic discipline to a broader field of interaction design. With theory also comes attendant epistemological commitments (which may or may not be honoured, of course) and other associated objects. For instance, these are the methods (e.g., experimental design, experience sampling, anthropological ethnography) and corresponding instruments for administering these methods (e.g., NASA TLX, social network analysis metrics [23]). This “remarkable expansion” [50 p. xi] tends to be taken sometimes as indications of success (in the form of rich emerging discipline) but perhaps more significantly as a signifier of problems, such that—as Carroll states—“an ironic downside of the inclusive multidisciplinarity of HCI is fragmentation” [10]. Similar views are expressed perhaps most commonly within the program committees of the SIG CHI conference, and in other public fora e.g., Interactions magazine [29]. An absence of uniformity of theory, method and instrument is unsettling when compared to formal accounts of how disciplines should be, particularly the disciplines with coveted scientific status. The choice thus seems stark; Rogers raises the question of prescribing disciplinary order: i.e., whether to “stem the tide and impose some order and rules or let the field continue to expand in an unruly fashion” [50 p. 2]. The dichotomy Here I sketch out two features of HCI’s disciplinary anxieties: incoherence and inadequacy. Later on I argue that invocations of ‘science’ are often attempts to remedy these perceived problems in HCI. Incoherence In essence, the core of the incoherence problem lies in the idea that HCI seemingly has few ‘secured’ propositions that researchers generally agree upon (except, perhaps Fitts’s 74 presented is now a familiar one where the opposite of disciplinary prescription is “unruly”. specifically, when considering the prospects of HCI’s position amidst ‘the disciplines’, this debate is often configured as a disciplinary contrast between the ‘hard’ and ‘soft’ sciences, and correspondingly, between the rigorous and the less disciplined, between the quantitative and the qualitative, between maturity and immaturity, and between the ideal-scientific and the striving-scientific. Bederson and Shneiderman present a paradigmatic expression of this in an attempt to explain the inadequacies: “Mature scientific domains, such as physics and chemistry, are more likely to have rigorous quantitative laws and formulas, whereas newer disciplines, such as sociology, or psychology, are more likely to have qualitative frameworks and models” [2] (also see Card’s comments on this [55]). Such comparisons, we might note, are of course performed under the assumption of an essential disciplinary comparability, in spite of the questions over the relevance of this endeavour—a point I return to later. What of the response to this in HCI? One possibility is to redescribe HCI in such a way that creates some semblance of order. This includes attempts to rationalise the existing range of work that occupies the HCI space, perhaps most visibly represented in discussions around ‘turns’ [50] ‘waves’ [5] and ‘paradigms’ [30]. For instance, Rogers offers four key turns: design (early 1990s), culture (late 2000s), ‘the wild’ (mid 1990s) and embodiment (early 2000s) [50]. Another way is to prescribe standardisation; and where standardisation comes, so do calls for ‘scientific’ ways of establishing order. These calls are largely intended, I think, to strengthen HCI’s disciplinary coherence. For example, Whittaker et al. argue that HCI’s “radical invention model” works against “the development of a ‘science’ of HCI”, with their proposed solution being the development of standardised sets of “reference tasks” to support “cumulative research” [57]. This view is consonant with programmatic statements arguing for developing a more science-like approach in HCI through practices like routine replication [27, 26, 58, 59, 31] or other prescribed forms of evaluation [45]. Relatedly, Liu et al. have also argued for the need of prescriptive standards of order via the development of what they term as shared “motor themes” [36]. THE SCIENTIFIC DESIGN SPACE AND HCI In order to understand the twin anxieties of incoherence and inadequacy, I think it helps to return to HCI’s formation. By examining how some key features of present-day HCI emerged from the research labs of North America in particular, this section discusses the intellectual origins of an orienting concept: the scientific design space. As part of this I am interested in the relationship to both the ‘picture of science’ in HCI as well as ideas of architecting HCI as a (possibly) scientific discipline. In closing, this section then turns to assess some of the intellectual foundations of this concept, based in Simon’s perspective of design and science. Inadequacy The second and interrelated expression of anxiety is that of HCI’s (intellectual) inadequacy when positioned as an academic discipline against a roster of other, better established disciplines. In 2002 panel, positions on this were represented in a panel of key HCI figures (Shneiderman, Card, Norman, Tremaine and Waldrop) discussing 20 years of the CHI conference [55]. Shneiderman argued the importance of “gain[ing] widespread respect in the scientific communities”, Norman commented on HCI as a “second-class citizen”, while Card reflected on his aspirations for HCI to graduate to “something somewhere in the lower half of the engineering disciplines, maybe civil engineering”. The situation seems to be unchanged since 2002: in a foreword to a recent (2012) handbook for HCI, Shneiderman reflects that “HCI researchers and professionals fought to gain recognition, and often still have to justify HCI’s value with academic colleagues or corporate managers” [54]. Designing the mouse Prior to broad recognition of HCI as a distinctive, nameable research activity—and naturally prior to debate about its status even as a ‘discipline’—the design of the humancomputer interface was primarily approached by pioneering research labs as a construction problem involving the provision of some possible control of a computer system. Many early fundamental interaction techniques were initially conceived of in this way (e.g., direct manipulation, the mouse, windowing environments [42, 41])—i.e., as primordially technological engineering endeavours that aimed to produce task-functional and efficient user interfaces. Verification of the usability of such interfaces tended to be applied only after design decisions had been made and implemented [10 p. 2]. Challenges were mounted to this software engineering focussed approach by nascent HCI. From this emerged a pairing between design work and cognitive scientific work [9]. Norman described this as “cognitive engineering” [46], although he retained a separation between the different roles of design and cognitive science. Yet, as we will see, the situation of this pairing in its formation was ambiguous in terms of whether it also is a conflation of the two. These concerns are political ones—they are motivated by the ways in which HCI is seen to be judged by others (academics, research funders, etc.). As part of this it seems standard practice to perform disciplinary comparisons between HCI and disciplines that are formally labelled as ‘sciences’, including physics (which remains the favoured model of scientific purity in positivistic philosophy of science), chemistry or biology [50 p. xii, 31]. More 75 The practical foundations of the scientific design space (i.e., a science of design [9], not the scientific study of design [14]) erupted from research within Xerox PARC, perhaps illustrated most emblematically by Stuart Card and his pioneering work on the computer mouse. In the book Designing Interactions [41]—documenting the history of the design of interactive systems and devices using firsthand accounts from its key players—Moggridge states of the computer mouse: “Stu [Card] was assigned to help with the experiments that allowed [mouse developers Doug Englebart and Bill English] to understand the underlying science of input devices”. Card’s aim was, in his own words, to develop a “supporting science” that would undergird the design activity of emerging interface technologies like the mouse or the desktop metaphor based graphical user interface. While earlier human factors and software engineering influenced work in HCI had been concerned with the use of ergonomic theory, its role in relation to design was generally verificationist [9, 10 p. 2], i.e., not used in predictive ways that serviced design. Basic atheoretic trial-and-error engineering approaches were typically being used at the time: in optimising the mouse’s design Card states that “the usual kind of A-versus-B experiments between devices” were no longer sufficient since “the problem with them was that [English, the designer] did not really know why one [mouse design] was better than the other” [41 p. 44]). hand-eye coordination than the device itself. In addition to its explanatory power, Card found that applying cognitive scientific concepts to conceptualise and shape the design space could also be generative; it could offer different design possibilities through prediction. For example, it was found that by designing a mouse-like input device that incorporated “putting fingers together you can maybe double the bandwidth [of input precision]”. This resulted in some clear advice for designers: to “put your transducer in an area that is covered by a large volume of motor cortex” [41 p. 45]—i.e., direct the design towards the capacities of the cognitive motor processor and its relationship to the rest of the cognitive subsystem. From this example it becomes clear how the cognitivist orientation was applied not only to the user, but also in shaping the idea of a scientific approach to the design space itself. Card, Mackinlay and Robertson later summarised this approach thus: “Human performance studies provide the means by which design points in the space can be tested. We can use this basic approach as a means for systematizing knowledge about human interface technology, including the integration of theoretical, human performance, and artifact design efforts.” [6]. The scientific design space approach and HCI’s disciplinary architecture The scientific design space of Card and others offers a consistent and strong visionary prescription for how research in HCI can proceed. That Card and colleagues have had a large influence on much of HCI’s development in uncontroversial, but I argue that some aspects of this influence on key forms of reasoning in HCI have been overlooked. In this section, I explore its role in the broader endeavour of describing and prescribing HCI’s disciplinarity. In this sense, I am interested in how the scientific design space offers solutions to some of the disciplinary anxieties of HCI. Card instead saw a role for “hardening” the “soft sciences of the human-computer interface” [43] through applying cognitive science as a way of explaining why interfaces failed or succeeded and thus predictively guiding design work. Cognitive science, offering a representational theory of the mind, could be deployed in order to construct a correct “theory of the mouse” [41 p. 44] based on theories of interoperating mental structures that were being developed in cognitive science research. Although initially drawing upon the non-cognitive behaviourist model of Fitts’s Law and Langoff’s work on this, Card’s work integrated this into a fuller assembly of cognitive units (e.g., perceptual processors, motor processors, memory stores, etc.). Cognitive psychology, with its mappings between human action and cognitive units, offered explanations for how Card might formally rationalise a ‘design space’ of the mouse so as to guide the mouse designers’ work along the right pathway according to the predictions of cognitive science. In this sense, cognitive science was employed to ‘tame’ the apparent ‘irrationality’ of design work. In 1983’s The Psychology of Human-Computer Interaction, Newell, Card and Moran sought to develop “a scientific foundation for an applied psychology concerned with the human users of interactive computer systems” [8]. Card would later point to the development of the mouse as an “ideal model of how the supporting science could work” [41 p. 45]. Although this hinted at a separation between design space exploration and cognitive science, the nature of that relationship remained largely unspecified. Yet, drawing out the implications of this “ideal model” meant extending the scientific design space of the mouse to a scientific design space approach for human-computer interactions more generally. The first step was to consider input devices as a whole [6], but linked to a broader prescriptive programme of “discovering the structure of the design space and its consequences” [7]. At times Card’s novel application of cognitive psychology challenged assumptions about what was actually important for design decisions being made for input devices like the mouse. For example, using this approach Card found that the differences between mouse designs was more to do with 76 ‘lower’ levels of organization than cognition; cognition, in turn, is a lower level of organization than social interaction.” [10]. Shneiderman also adopts this hierarchical descriptive model of relationships between more fundamental and ‘higher’ sciences. For instance he distinguishes between “microHCI”, where researchers “design and build innovative interfaces and deliver validated guidelines for use across the range of desktop, Web, mobile, and ubiquitous devices”, and “macro-HCI”, which deals with “expanding areas, such as affective experience, aesthetics, motivation, social participation, trust, empathy, responsibility, and privacy” [53]. Macro and micro HCI have “healthy overlaps” yet have different “metrics and evaluation methods”. Finally, Rogers presents a description of HCI that seems oriented by a similar hierarchical sensibility. In HCI Theory: Classical, Modern and Contemporary, HCI’s (scientific?) disciplinary structure is a logical, hierarchical arrangement of “paradigms, theories, models, frameworks and approaches” that vary in scale, “level of rigor, abstraction and purpose” [50 p. 4]. In this scheme paradigms are the large scale “shared assumptions, concepts, values and practices” of a research community, while a theory indicates an empirically “well-substantiated” explanatory device that resides within a particular perspective or theoretical tradition (and is presumably associated with a particular paradigm). Beneath this sit models, which are predictive tools for designers, of which Fitts’ Law is the most familiar. Figure 1: Architecting the discipline—positioning HCI within a scientific order (figure reproduced from [43]). A clearer scientific disciplinarity to HCI was also developed subsequent statements. In 1985 Newell and Card presented a vision for the role of psychological science in HCI. This work offered a descriptive account of a layered model in which temporal bands (between decades and milliseconds) map to different action units, associated memory and cognitive capacities, and the relevant theoretical bands which apply [43] (see Figure 1). Within this schema HCI’s phenomena of interest sit largely within the psychological and bounded rationality bands, happening to coincide precisely with the concerns of cognitive science of the time and ordering the rest of the space in terms of various related sciences. Unpacking the idea of the scientific design space Returning again to the core idea of the scientific exploration of the design space, here I want to unpack its orienting ideas. In doing so, I think we can better understand the lines of reasoning being deployed in its pursuit. The significance of this layered model should not be underestimated. It offers a unified, reductive organisation that is similar to positivistic philosophy of science where higher order sciences are reducible to lower ones (ultimately physics) [20]. In this scheme, it is social and organisation science, various ‘levels’ of the cognitive sciences, neurosciences, and biological sciences, that fill in this order. Critically, this model for HCI is also cumulative: thus, latter developments such as Information Foraging Theory [48]—an influential theory that explains the information search and collection behaviours of web users (for example)—fit within the ‘rational’ band, yet build upon lower, more foundational bands from which information foraging “gains its power and generality from a mathematical formalization of foraging and from a computational theory, ACT-R, of the mind” [32]. The central idea of design spaces and their systematic and empirical investigation as a scientific matter seems to have a strong resonance with the work of Herbert Simon. Simon was frequent collaborator with Newell, whom Card had been a student of. Simon’s influential book, The Sciences of the Artificial [56], is important for this concept in that that not only lays out a programme for a scientific approach to design but also discusses this within the context of Simon’s prior work around solution-searching within bounded rationality (e.g., ‘satisficing’). In essence Simon argues that the phenomena of the design activity itself demands a scientific approach of its very own: a new “science of design” [56 ch. 5], but one that is unshackled from the tendency to employ methods from the natural sciences. Simon’s conception of this new science of design shares a link to the formulations of Card, Newell and others that I described earlier. Crucially, both these approaches conceptualise design as an optimisation problem. As an optimisation problem, the design of “the artificial” (here meaning human-constructed objects such as As a guiding notion of hierarchical disciplinary order, this work seems to have been influential in sparking attempts at descriptions of HCI. Firstly, Carroll, while in disagreement with Newell and Card’s presentation of the role of psychology in HCI [12], elsewhere concurs to some extent with this organisation in terms of “level of description. Thus, perception and motor skill are typically thought of as 77 interactive digital artefacts) may be rendered as dimensions, enumerated, rationalised, and essentially made docile. This step gives rise to the application of a spatial metaphor of design: i.e., the idea of a design space that is “parametrically described” and populated with “device designs as points” [7]. Card positions cognitive science as a scientific way to “structure the design space so that the movement through that design space was much more rapid” [41 p. 45]. Simon offers a more generalised version of this: design spaces not only of artefacts but also economic, organisational and social systems. This view is broad and encompassing, and it is perhaps a conclusion drawn from a fundamentally cognitive conception of the mind. Hence Simon states that “the proper study of mankind is the science of design” [56 p. 138]. For Simon a scientific approach to designing such things involves constructing all design problems as computational spaces—ones that are amenable to formalisation and therefore computational search, optimisation and so on. And it is this notion which I argue undergirds the design space concept as it has found its way into forming a scientific approach to HCI research. This influence was initially made possible through the importance of cognitive science in HCI’s formation. Conceptually the prospect of transforming design problems into reductive computational spaces (to be addressed by scientific methods) was facilitated by the central implications of a ‘strong’ cognitive science position. Specifically this position offers an isomorphism between the computer and the human (i.e., as a cognitive object, with various input / output modalities). There are two key conflations that follow: firstly that the human is computational, or can be described with computational concepts; and secondly that the designer and therefore design itself is computational. This underlying idea in HCI has enabled the adoption process of the design space model. The practical expression of these ideas may be readily found the research outputs of many HCI conference venues like CHI and UIST, but also in those of related communities such as Ubicomp. Following Card and colleagues, it has become the approach of choice for work that evaluates novel input or output devices, but also for innovative interaction techniques for established interface forms (e.g., GUIs, touch and gestural interfaces, etc.). The ‘ideal expression’ of the scientific design space is often conducted under the broader glossed label of psychology. It characteristically involves engaging in task-oriented interface evaluations via a hypothesis-driven experimental format—being often classed as ‘usability evaluation’ [26]. In building hypotheses and delivering their results, classic features of cognitive science theory are recruited, for example cognitive objects like memory, task load and so on. This might also include methods where the rationale is grounded in cognitive science reasoning, such as ‘think aloud’ techniques. Hypothesis testing enables an organised and systematic traversal of the design space particular to the class of device and interface under investigation [6, 7]. As part of this, cumulative, replicable, generalisable and theoretically-informed findings are delivered both to HCI as a whole but also potentially back to cognitive science as instances of applied research. It is also possible that attempts to form novel cognitive theory specific to interactive systems may result, for example Information Foraging theory presents one well-known instance of this (also note its relationship to the ACT-R cognitive architecture [48]). Returning to the disciplinarity question once again, the scientific design space idea in its broader construction from Simon seems to offer both security for the status of HCI (in terms of academic adequacy and coherence), and also answers some of the questions presented earlier regarding the status of HCI as a discipline—whether engineering, craft or science [38]. The idea effectively responds by reformulating design problems in ways that let them be “reduced to declarative logic” [16 p. 176], meaning that “design [becomes] a system with regular, discoverable laws” [24 p. 26]. We will return to this topic of ‘design’ in the closing of the discussion section. DISCUSSION Having firstly described HCI’s disciplinary anxieties, and then covered formative concepts around scientific design spaces, the discussion now broadly explores the relations between them. I will initially do this through picking apart the intellectual coherence of the design space approach. First I argue that this idea has been very important in HCI research, and moreover, still is particularly in the evaluation of novel input devices. Then I wish to discuss the conceptual problems of the ‘scientific design space’, and in doing so must necessarily also turn to tackle the more general issue of the role of ‘science’ in HCI. Finally, I turn to review the introduction of ‘designerly’ perspectives in HCI and examine how this relates to scientific design spaces. As an adoption phenomenon in HCI, this approach seems very well-established. But complaints have emerged about the trappings of this approach being prioritised at the expense of “rigorous science” [26]. For instance, Greenberg and Buxton articulate this trend as one of “weak science” in HCI that is more concerned with the production of one-off “existence proofs” of designs than systematic rigour that can “put bounds” on the wider design space under test [26]. In comparison, Card, Mackinley and Robertson emphasise the importance of uncovering design space structure [7], and through this taking different device designs as inputs or “points in a parametrically described design space” [7]—in The adoption of scientific design spaces in HCI Via Card and others, it seems that Simon’s way of conceptualising the design of artefacts has been a significant contributor to HCI’s ‘DNA’. Yet this impact of Simon on HCI is only occasionally acknowledged [9]. 78 other words, “systematizing” [6]. In contrast, the prioritisation of existence proofs has meant theory-free meandering explorations of an unspecified design space with little clear epistemic justification for taking a design space approach in the first place—or to put it another way, the methods of this approach end up deciding the problems to be tackled [26]. The scientific design space is something of a ‘muscle memory’ for HCI’s relationship to interactive devices and systems. three forms: 1. design as a (cognitive) scientific activity; 2. the application of (cognitive) science in design—i.e., retaining some notional separation between these activities; 3. the development of scientific understandings of design itself (see [14 ch. 7]). The question for the scientific design space in HCI is based around some blurring between the first two forms. In some ways the initial programme of Card’s recounted in Designing Interactions appears different to Simon’s, yet in others seems to have a similar end result. At first glance Card seems to position his work as offering a “supporting science” that does not replace design (i.e., the application of cognitive science to design). He reminds us that it still requires “good designers to actually do the design” [41 p. 45]. Yet, I think the picture is somewhat more complex than this. In later work, such as morphological design space analysis [7], extant designs become ‘inputs’ to a design space which itself generates the parameters along which “good designers” must work. In this way designers must engage with the predictive authority of the (cognitive) scientific shaping of the design space—so as to “alter the designer’s tools for thought” [44]. (Card, Mackinley and Robertson’s story of the “headmouse” user is also worth reflecting upon in this regard [7 p. 120].) I must note, of course, that there are broader issues at play here in HCI beyond the intellectual framework of design spaces. For instance, publishing cultures that reward quantity and the ‘least publishable unit’ will also tend to tolerate existence proofs. This is likely encouraged or at least further enabled by the idea of ‘scientific’ cumulative progress in HCI. Design spaces and scientific method Now that I have discussed the adoption of the scientific design space mode of HCI research, I wish to tackle the very idea itself in two ways. In this section here I will argue that this mode of work borrows from the natural sciences in spite of Simon’s original articulation. In the section after this I will link the discussion to much more general issues around the notion of ‘science’ itself in HCI’s discourse. HCI’s ‘picture of science’ The first point is that curiously, even given his call for genuinely new sciences of the artificial that were specifically not derivative of the natural sciences, the position Simon outlines nevertheless relies upon this strategy anyway. It seems that this is a product of the way in which Simon conceives of and specifies design as an activity—and how this could help knowledge about design move on from what he saw as “intellectually soft, intuitive, informal, and cook-booky” ways of conceiving of it at the time [56 p. 112]. At the core of the design space notion—I believe—is a desire to bring some aspect of (or perhaps all of) HCI closer towards a scientific disciplinarity. Doing so offers the promise of addressing anxieties over incoherence and inadequacy. Yet in order to better understand this I argue that we must start to unpack what is meant by ‘science’ and how it is used in HCI’s discourse. I also wish to bracket off ‘science’ and ‘science talk’ more generally and look at what is done with it. After examining this, I then move on to contrast these deployments with understandings of scientific practice from philosophy of science and empirical studies of scientists’ work—contrasts that reveal a problematic dissonance between the two. As Ehn expresses it, Simon performs something of a ‘trick’ which “poses the problem of design of the artificial in such a way that we can apply the methods of [formal] logic, mathematics, statistics, etc., just as we do in the natural sciences” (original emphasis) [16 p. 175]. In other words, because of Simon’s conceptualisation of design activities in terms of spatial, computable design spaces that are essentially reduced to search problems, the deck becomes stacked in such a way that ‘textbook’ understandings of the methods of the natural sciences just happen to turn out to be the relevant choice. And, following Simon, we find the science of the design space in HCI also relies upon perceived methods of the natural sciences. Interestingly, this perception is itself second-hand: it is actually that of cognitive science’s version of the methods of the natural sciences not the actual practices per se. Perhaps the first debates over what kind of science might be relevant to HCI can be found in the Newell, Card, Carroll and Campbell exchanges of the early 1980s [43, 12, 44]. Here, Carroll and Campbell took issue over Newell and Card’s characterisation of “hard science” (“quantitative or otherwise technical”) and “soft, qualitative science” [43] in their arguments about the possible role of science (cognitive psychology) in HCI. As Carroll and Campbell disputed, this was a “stereotype of hard science” and psychology itself; a false dichotomy around science was created by Newell and Card in order to support a positivist programme [12]. Since then, discourse on ‘science’ in HCI research broadly has featured a web of conflicting deployments of the term. These deployments can offer both descriptions of HCI as scientific and prescriptions about ensuring that HCI is scientific (in some way). It is possible to offer two The second point here is the lack of clarity in this notion around design and science as activities (I shall come to unpack ‘science’ and ‘design’ in the sections following this one). Specifically we could draw a distinction between 79 contrasting poles of this so as to illustrate the difficulties. At one end ‘science’ can be deployed in its loosest meaning as a synonym of rigour. For instance, I would suggest that Carroll often uses this formulation (see [11]); it is an invocation that seems to avoid scientism or positivist assumptions. At the other end of the scale we see ‘science’ being used to denote a specific set of ‘scientific qualities’ that are seen as gold standards of being scientific. Elsewhere I have summarised what is typically meant here in terms of three linked concepts [49]: 1. accumulation— science’s work is that of cumulative progress (e.g., [44, 57]); 2. replication—science’s work gains rigour from being reproducible by ‘anyone’ [28]; and 3. generalisation—science’s cumulative work builds toward transcendent knowledge. As part of this, various other descriptions of what it means to be “more scientific” [35] are pointed to: an adherence to the scientific method, empiricism and the “generation of testable hypotheses” [32]. In this sense ‘science’ is produced as a description of a general knowledge production procedure that offers standardisation and guarantees against ‘bias’. In this account ‘science’ represents a self-correcting incremental process of knowledge accumulation: “the developing science will gradually replace poor approximations with better ones” [34]. Other studies of scientific work in astronomy, physics, chemistry and biology practice unpack the ‘lived order’ of these practices, describing how natural phenomena must be socially ‘shaped’ and transformed into scientific “Galilean objects” for presentation in academic papers and so on [21, 39]. Following this, I think that HCI discourse around ‘science’ broadly tends to make a common mistake in expecting formal accounts and material practices to be interchangeable, leading to various confusions (elsewhere this necessary dissonance has even lead to accusations of fraud [40]). For instance, the replication of results as a necessity in principle for HCI has been described as “a cornerstone of scientific progress” [59]. This notion trades on the idea that formal accounts should provide an “adequate instruction manual for replication work” [28]. Yet when we turn to studies of the natural sciences, from which the principle is ostensibly extracted, we find that firstly most scientific results do not get replicated and secondly that where it is used, replication is a specifically motivated, pragmatic action for particular contested, relevant cases [13]. Not only is there the potential for confusion between formal accounts and material practices, but studies of the natural sciences also question the notion of the homogeneous way in which ‘science’ broadly is conceptualised, further problematising the term as we find it in HCI. This use tends to presuppose the existence of a coherent specific set of nameable enumerable procedures that make up ‘the scientific method’; procedures that are also held to be representative of a standardised approach to science in general. Philosophy of the sciences instead argues that there are sciences, plural, each sui generis, with no uniform approach or necessary promise of ultimate unity [20]. The notion of ‘the scientific method’ is itself questionable. Drawing attention to problems of an induction-based model of scientific progress, Feyerabend has instead argued that there exists “anarchistic methodology” and counterinduction in scientific practice that is not visible in formal accounts [18]. This is not that ‘anything goes’ methodologically, but rather that adherence to formal accounts of method alone cannot explain how science progresses. As Bittner puts it, the natural sciences have “tended to acquire arcane bodies of technique and information in which a person of average talent can partake only after an arduous and protracted preparation of the kind ordinarily associated with making a life commitment to a vocation” [3]. Hence it becomes potentially distorting to compress such a diverse set of practices into a singular but unspecifiable method of ‘science’ so as to draw out principles for being “more scientific” [35]. There are of course many other ways of talking about ‘science’ which have found their way into HCI [49]. The most obvious would be in vernacular usage—where ‘being scientific’ is used in place of ‘being professional’, ‘being reasonable’, ‘being careful’ or ‘being scholarly’ and so on. Often these forms may be political, for example, using ‘science’ as a rhetorical strategy to assert epistemic or moral authority (‘science’ as good work or a transcendent truth). It could also be used as a way of categorising research as scientific and non-scientific. Or it can be an aspirational label that requests peer recognition—for example, computer science rather than informatics [49]. It seems, then, that the overall ‘picture of science’ in HCI is confused. I wish to now to examine this picture with that presented by empirical studies of the natural sciences. The first general point that I address to the broad HCI description / prescription of ‘science’ is the difference between formal accounts of science (e.g., scientific papers) and the material practices that are carried out to produce them. In ethnomethodological accounts of scientific practice the relationship between the two, and their essential inseparability, is explored. Livingston, for instance, describes the ad hoc and profoundly local achievement of performing everyday practical laboratory tasks (e.g., determining chemical soil composition). He makes the observation that “From an external point of view, the procedures themselves determined the results. From the point of view of those working in the lab, our joint experiences helped us find what those procedures were in practice and how we could make them work.” [37 p. 155]. HCI’s relationship to design In this section I want to discuss another key element of the design space besides its ‘scientific’ sensibilities—‘design’. Of course, design has always been a concern for HCI 80 research. As Card, Newell and Moran stated, “design is where the action is in the human-computer interface” [8 p. 11]. The question is what kind of design—the deployment of the term ‘design’ is itself somewhat like ‘science’, i.e., a potential source of great confusion [61]. Regardless of the conceptual challenges I raised, Card and colleagues nevertheless very clearly articulated a strong sense of design following Simon. Yet I have argued this adoption at the same time provides an intellectual foundation to some of HCI’s recurrent concerns about developing a scientific disciplinarity. The formalised rigour of the scientific design space has meant this particular conceptualisation of design has flourished in HCI; this conceptualisation is grounded in background(ed) and unreflective scientific framing devices. It is for this reason that many novel interactive systems and technologies are still evaluated as inputs to the design space model. although there have been distinctions made between design practice and critical design [61]. Further, as they are expressed within HCI, designerly perspectives have similar rehearsals of arguments around matters of disciplinary order, anxieties [19], and the relevance of scientific disciplinarity to design as an activity. For example, Zimmerman, Stolterman and Forlizzi specifically call for research through design to “develop protocols, descriptions, and guidelines for its processes, procedures, and activities” and find its own sense of “reliability, repeatability, and validity” [62]. Gaver characterises this as designerly approaches succumbing to HCI’s tendency towards scientism [22]. It seems that the debate around these designerly perspectives is thus no less susceptible to HCI’s orienting conversations around (scientific) disciplinarity that I have discussed in this paper. However, perhaps because of the struggle for recognition in HCI, designerly perspectives do tend to consistently emphasise the importance of assessing and valuing designerly research correctly—the argument being that the products of this work do not always fit with how HCI values research. This is highlighted by Gaver (“appropriate ways to pursue our research on its own terms” [22]), Wolf et al. (“its own form of rigor” [60]), Fallman and Stolterman (“rigor and relevance have to be defined and measured in relation to what the intention and outcome of the activity is” [19]), and Zimmerman, Stolterman and Forlizzi (“a research approach that has its own logic and rigor” [62]). Yet this argument about rigour has not been generalised to HCI as a whole and remains stuck within the frame of the designerly perspective attempting to gain legitimacy within the HCI community—in concluding I want argue for that generalisation. Since at least the late 1990s and certainly the early 2000s, however, HCI has seen the development of its own subcommunity of researchers concerned with what I will gloss here as ‘designerly’ perspectives (a gloss that I will later problematise). As they appear in HCI, designerly perspectives work with a range of terms like ‘design research’, ‘research through design’, etc.; to highlight a few examples I discuss here: [17, 60, 61, 19, 62, 22]. This relatively small but distinct subcommunity is one of the few places in HCI that actually foregrounds Simon’s conception of design—which is suggested by some to be a “conservative account” [17]. In this account, Simon effectively democratises and deskills design by arguing that “[e]veryone designs who devises courses of action aimed at changing existing situations into preferred ones” [56 p. 111]. This logically flows from his bounded rationality view of design as search. Yet within the context of this designerly perspective, it has been argued that traditions along the lines of the scientific design space model tend to mask wider discussion about the nature of the design activity itself; as Fallman argues, “Design is thus a well-established and widespread approach in HCI research, but one which tends to become concealed under conservative covers of theory dependence, fieldwork data, user testing, and rigorous evaluations” [17]. In short, the absence of designerly alternatives to Simon has meant design is “at best limiting and at worst flawed” in its usage in HCI [60]. CONCLUSION In this paper I have sought to examine disciplinary anxieties in HCI through picking apart its ongoing relationship to ‘science’. This has meant identifying the idea of the scientific design space—an approach conceiving of designed artefacts as scientific objects, influenced by formative early applications of cognitive science to input devices. This significant approach to design seems to have subsequently configured much HCI research discourse, leading to discussions around scientific qualities of accumulation, replication, and generalisation. Yet, as I have tried to show, matters of science and scientific disciplinarity in this perspective are somewhat problematic conceptually and far from settled. Further, in spite of announcements over HCI’s various ‘turns’ and successive ‘waves’ [5] of development, or even ‘paradigms’ [30], I have contested that some of the key assumptions of HCI have been quite resilient to such apparent changes, even with the introduction of designerly perspectives to HCI that challenge Simon’s conceptualisation of design—indeed, there we also find similar debates played out around (design) science and (design) disciplinarity. Yet, following the pattern of eclectic importations of new literatures to HCI, these debates around designerly perspectives typically (perhaps necessarily) offer a somewhat simplified presentation. As a point of contrast, Johansson-Sköldberg detects five interrelated but different discourses of “designerly thinking”: as creation of artefacts (Simon), as reflexive practice (Schön), as problem-solving (Buchanan), as sense-making (Cross), or as meaningcreation (Krippendorff) [33]. Within HCI accounts of design, the literature tends to brush over such nuances, 81 In concluding I wish to suggest two implications. Firstly, that HCI researchers should—with some caveats—stop worrying about ‘being scientific’ or engaging in ‘science talk’ and instead concern themselves with working in terms of appropriate forms of rigour. Secondly, that HCI should—again, with some caveats—stop worrying about disciplinary order or ‘being a discipline’ and instead engage with the idea of being interdisciplinary and all the potential difficulties and reconfigurations that requires. must arise from and be appropriate for the actual problem or research question under consideration” [26]. There are other purposes with which ‘science’ terms may be put to in HCI that might be necessary—albeit as a double-edged sword. For example, in political purposes, such as rhetorical, or persuasive uses, ‘science’ may have importance for communicating how HCI fits within research funding structures that adhere to the normative descriptive / prescriptive forms that this paper has questioned. The danger here, of course, is the apparent cynicism of this approach. From science to rigour The first point turns on ‘science’ in HCI. To summarise, the paper has presented the role that ‘science’ plays in descriptions of HCI, e.g., accounts of HCI research as having scientific qualities. Secondly the paper has highlighted the use of ‘science’ in building prescriptions for HCI research, e.g., programmes by which HCI research can be conducted as a scientific discipline of design. These descriptions and prescriptions pertain to HCI research practice. Yet, in both cases I have tried to show that these are problematic when we consider how these articulations compare with what the model—the natural sciences—looks like as a set of lived practices. I have argued that the model being invoked by HCI—i.e., formal accounts—and the everyday material practices of natural scientists do not match up (for good reason), meaning that the case for employing said formal accounts of ‘science’ as a descriptor of or prescription for HCI seems weak and potentially confusing. At the most charitable we might say that ‘science’ could be used as a synonym for rigour, albeit a highly loaded one. From discipline to interdiscipline The second concluding point turns rethinking disciplinarity in HCI and concerns about constructing a rationalised disciplinary order. While the importance of multidisciplinarity has long been identified in HCI (e.g., [10]), what I argue for here is underlined by Rogers’s characterisation of HCI (for good or bad) as an “eclectic interdiscipline” [50 p. xi]. In other words, the difficulties of assembling HCI into some disciplinary order may be the natural state, the key characteristic of HCI. Reflecting upon this, Blackwell has suggested that HCI could be best conceived of as a catalytic interdiscipline between disciplines, rather than a discipline that engages in the development and maintenance of a stable body of knowledge [4]. If HCI is to be a rigorous interdiscipline then it will require working more explicitly at the interface of disciplines. We will need more reviews of and reflections upon the landscape of different forms of reasoning in HCI and through this better ways of managing how potentially competing disciplinary perspectives meet together. This paper has touched only one part of the landscape, but there are many more. In abandoning the formal-scientific, I want to emphasise the notion of appropriate rigour. This is the idea that rigour in HCI must be commensurate with the specific intellectual origins of the work; e.g., this may be (cognitive, social, etc.) psychology, anthropology, software engineering, or, more recently, the designerly disciplines. This runs counter to the desire for hierarchical disciplinary orderings, standardisation, or other forms of positivist reductionism in HCI that I have discussed in this paper. Firstly, such orderings will necessarily foreground contradictory accounts of rigour, and secondly, invocations of ‘science’ will tend to replace focus on seeking the relevant frame of rigour. Instead, appropriate rigour is achieved “not through the methods by which data is collected, but through the ways in which the data can be kept true [...] during the analysis” [52]. At the same time it should be noted that there are dangers here too: being an interdiscipline can mean that HCI research diffuses into contributing disciplines and ‘credit’ is never recognised for HCI. This suggests that in addition to reconceptualising HCI as an interdiscipline, we must think of new and perhaps radical ways to characterise HCI as it is presented to the outside world. ACKNOWLEDGMENTS This work is supported by EPSRC (EP/K025848/1). Elements of this paper developed through conversations with Bob Anderson, to whom I am grateful. Thanks also to those who conversed with me on the paper and / or the topic: Susanne Bødker, Andy Crabtree, Christian Greiffenhagen, Sara Ljungblad, Lone Koefoed Hansen, Alex Taylor, Annika Waern, and anonymous reviewers. To highlight the notion of appropriate rigour I point to the “damaged merchandise” controversy of the late 1990s, where the reliability and validity of well known usability methods in prominent studies were critiqued [25]. Perhaps one of the reasons why this critique produced a significant response [47] was that it took the studies to task using the intellectual framework that had been implicitly ‘bought into’ (cognitive psychology). As Greenberg and Buxton argue, “the choice of evaluation methodology—if any— EPSRC Research Data Statement: All data accompanying this publication are directly available within the publication. 82 17. Fallman, D. 2003. Design-oriented human-computer interaction. In Proc. CHI ‘03. ACM, New York, NY, USA, 225-232. REFERENCES 1. Bartneck, C. 2008. What Is Good? – A Comparison Between The Quality Criteria Used In Design And Science. In Proc. CHI ‘08 Extended Abstracts. ACM, New York, NY, USA, 2485-2492. 18. Fallman, D., Stolterman, E. 2010. Establishing criteria of rigour and relevance in interaction design research. Digital Creativity 21, 4, Routledge, 2010, 265–272. 2. Bedersen, B., Shneiderman, B. 2003. The Craft of Information Visualization, chapter 8, p. 350, Elsevier, 2003. 19. Feyerabend, P. K. 2010. Against Method. 4th ed., New York, NY: Verso Books. 3. Bittner, E. 1973. Objectivity and realism in sociology. In Psathas, G. (ed.), Phenomenological Sociology. Chichister: Wiley, 109-25. 20. Fodor, J. A. 1974. Special sciences (or: The disunity of science as a working hypothesis). Synthese 28, 2, Springer, 1974, 97-115. 4. Blackwell, A. F. 2015. HCI as an Inter-Discipline. In Proc. CHI ‘15 Extended Abstracts. ACM, New York, NY, USA, 503-516. 21. Garfinkel, H., Lynch, M., Livingston, E. 1981. The Work of a Discovering Science Construed with Materials from the Optically Discovered Pulsar. Philosophy of the Social Sciences 11 (June 1981), Sage, 131-158. 5. Bødker, S. 2006. When second wave HCI meets third wave challenges. In Proc. of NordiCHI ‘06. ACM Press, 2006. 22. Gaver, W. 2012. What should we expect from research through design?. In Proc. CHI ‘12. ACM, New York, NY, USA, 937-946. 6. Card, S. K., Mackinlay, J. D., Robertson, G. G. 1990. The design space of input devices. In Proc. CHI ‘90. ACM, New York, NY, USA, 117-124. 23. Golbeck, J. 2013. Analyzing the Social Web. Elsevier, 2013. 7. Card, S. K., Mackinlay, J. D., Robertson, G. G. 1991. A morphological analysis of the design space of input devices. ACM Trans. Inf. Syst. 9, 2 (April 1991), 99122. 24. Goodman, E. 2013. Delivering Design: Performance and Materiality in Professional Interaction Design. PhD thesis, University of California, Berkeley, 2013. 8. Card, S. K., Newell, A., Moran, T. P. 1983. The Psychology of Human-Computer Interaction. L. Erlbaum Assoc. Inc., Hillsdale, NJ, USA. 25. Gray, W. D., Salzman, M. C. 1998. Damaged merchandise? A review of experiments that compare usability evaluation methods. Hum.-Comput. Interact. 13, 3 (September 1998), 203-261. 9. Carroll, J. M. 1997. Human–computer interaction: Psychology as a science of design. International Journal of Human-Computer Studies, Volume 46, Issue 4, April 1997, Pages 501-522. 26. Greenberg, S., Buxton, B. 2008. Usability evaluation considered harmful (some of the time). In Proc. CHI ‘08. ACM, New York, NY, USA, 2008. 10. Carroll, J. M. (ed.) 2003. HCI Models, Theories, and Frameworks: Towards a Multidisciplinary Science, Elsevier, 2003. 27. Greenberg, S., Thimbleby, H. 1992. The weak science of human-computer interaction. In Proc. CHI ‘92 Research Symposium on HCI, Monterey, California, May. 11. Carroll, J. M. 2010. Conceptualizing a possible discipline of human-computer interaction. Interact. Comput. 22, 1 (January 2010), 3-12. 28. Greiffenhagen, C., Reeves, S. 2013. Is replication important for HCI? In Workshop on replication in HCI (RepliCHI), CHI ‘13. 12. Carroll, J. M., Campbell, R. L. 1986. Softening up Hard Science: reply to Newell and Card. Human–Computer Interaction, 2(3):227-249, Taylor and Francis, 1986. 29. Grudin, J. 2015. Theory weary. interactions, blogpost (posted 14/03/14), retrieved 15th June, 2015, from: http://interactions.acm.org/blog/view/theory-weary 13. Collins, H. M. 1975. The seven sexes: A study in the sociology of a phenomenon, or the replication of experiments in physics. Sociology, 9(2):205-224, 1975. 30. Harrison, S. Tatar, D., Sengers, P. 2007. The Three Paradigms of HCI. In Proc. alt.chi, CHI ‘07, San Jose, CA, May 2007. 14. Cross, N. 2007. Designerly ways of knowing. Birkhäuser GmbH, Oct. 2007. 31. Hornbæk, K., Sander, S. S., Bargas-Avila, J. A., Simonsen, J. G. 2014. Is once enough?: on the extent and content of replications in human-computer interaction. In Proc. CHI ‘14. ACM, New York, NY, USA, 3523-3532. 15. Dix, A. 2010. Human-computer interaction: A stable discipline, a nascent science, and the growth of the long tail. Interact. Comput. 22, 1 (January 2010), 13-27. 16. Ehn, P. 1990. Work-Oriented Design of Computer Artifacts. L. Erlbaum Assoc. Inc., Hillsdale, NJ, USA. 32. Howes, A., Cowan, B. R., Janssen, C. P., Cox, A. L., Cairns, P., Hornof, A. J., Payne, S. J., Pirolli, P. 2014. 83 Interaction Science Spotlight, CHI ‘14. ACM, New York, NY, USA, 2014. 48. Pirolli, P. 2007. Information foraging theory: Adaptive interaction with information, vol. 2. Oxford University Press, USA, 2007. 33. Johansson-Sköldberg, U., Woodilla, J., Çetinkaya, M. 2013. Design Thinking: Past, Present and Possible Futures. Creativity and Innovation Management. 22, 2 (June 2013), 121–146. 49. Reeves, S. 2015. Locating the “Big Hole” in HCI Research. Interactions 22, 4 (July 2015). 50. Rogers, Y. 2012. HCI Theory: Classical, Modern and Contemporary, Synthesis Lectures on Human-Centered Informatics 5, 2, Morgan & Claypool, 2012, 1-129. 34. John, B. E., Newell, A. 1989. Cumulating the science of HCI: From S-R compatibility to transcription typing. In Proc. CHI ‘89. ACM, New York, NY, USA, 109-114. 51. Rogers, Y., Sharp, H., Preece, J. 2011. Interaction Design: Beyond Human—Computer Interaction, 3rd Edition. Wiley, 2011. 35. Kostakos, V. The big hole in HCI research. Interactions 22, 2 (February 2015), 48–51. 36. Liu, Y., Goncalves, J., Ferreira, D., Xiao, B., Hosio, S., Kostakos, V. 2014. CHI 1994-2013: mapping two decades of intellectual progress through co-word analysis. In Proc. CHI ‘14. ACM, New York, NY, USA, 2014. 52. Rooksby, J. 2014. Can plans and situated actions be replicated? In Proc. CSCW ‘14. ACM, New York, NY, USA, 603-614. 53. Shneiderman, B. 2011. Claiming success, charting the future: micro-HCI and macro-HCI. Interactions 18, 5 (September 2011), 10-11. 37. Livingston, E. 2008. Ethnographies of Reason. Ashgate, 2008. 54. Shneiderman, B. 2012. The Expanding Impact of Human-Computer Interaction. Foreword to The Handbook of Human-Computer Interaction, Jacko, J. A. (ed), CRC Press, May 2012. 38. Long, J., Dowell, J. 1989. Conceptions of the discipline of HCI: craft, applied science, and engineering. In Proc. 5th conference of the BCS-HCI Group on People and computers V, Sutcliffe, A., Macaulay, L. (eds.). Cambridge University Press, 9-32. 39. Martin, A., Lynch, M. 2009. Counting Things and People: The Practices and Politics of Counting. Social Problems 56, 2 (May 2009), 243-266. 55. Shneiderman, B., Card, S., Norman, D. A., Tremaine, M., and Waldrop, M. M. 2002. CHI@20: fighting our way from marginality to power. In Proc. CHI ‘02 Extended Abstracts. ACM, New York, NY, USA, 688691. 40. Medawar, P. B. 1963. Is the scientific paper a fraud? The Listener 70 (12 September), 377–378. 56. Simon, H. A. 1996. Sciences of the Artificial. 3rd ed., MIT Press, Sept. 1996. 41. Moggridge, B. 2006. Designing Interactions. MIT Press, Oct. 2006. 57. Whittaker, S., Terveen, L., Nardi, B. A. 2000. Let’s stop pushing the envelope and start addressing it: a reference task agenda for HCI. Hum.-Comput. Interact. 15, 2 (September 2000), 75-106. 42. Myers, B. A. 1998. A brief history of human-computer interaction technology. interactions 5, 2 (March 1998), 44-54. 58. Wilson, M. L., Mackay, W. 2011. RepliCHI—We do not value replication of HCI research: discuss (Panel). In Proc. CHI ‘11 Extended Abstracts. ACM, New York, NY, USA. 43. Newell, A., Card, S. K. 1985. The prospects for psychological science in human-computer interaction. Hum.-Comput. Interact. 1, 3 (September 1985), 209-242. 59. Wilson, M. L., Resnick, P., Coyle, D., Chi, E. H. 2013. RepliCHI: The workshop. In CHI ‘13 Extended Abstracts. ACM, New York, NY, USA, 3159-3162. 44. Newell, A., Card, S. 1986. Straightening Out Softening Up: Response to Carroll and Campbell. Hum.-Comput. Interact. 2, 3 (1986), 251-267. 60. Wolf, T. V., Rode, J. A., Sussman, J., Kellogg, W. A. 2006. Dispelling “design” as the black art of CHI. In Proc. CHI ‘06. ACM, New York, NY, USA, 521530. 45. Newman, W. M. 1997. Better or just different? On the benefits of designing interactive systems in terms of critical parameters. In Proc. DIS ‘97. ACM, New York, NY, USA, 239-245. 61. Zimmerman, J., Forlizzi, J., Evenson, S. 2007. Research through design as a method for interaction design research in HCI. In Proc. CHI ‘07. ACM, New York, NY, USA, 493-502. 46. Norman, D. A. 1986. Cognitive engineering. In D. A. Norman & S. W. Draper (Eds.), User centered system design: New perspectives on human-computer interaction. Hillsdale, NJ: Erlbaum Associates. 62. Zimmerman, J., Stolterman, E., Forlizzi, J. 2010. An analysis and critique of Research through Design: towards a formalization of a research approach. In Proc. DIS ‘10. ACM, New York, NY, USA, 310-319. 47. Olson, G. M., Moran, T. P. 1998. Commentary on “Damaged merchandise?”. Hum.-Comput. Interact. 13, 3 (September 1998), 263-323. 84 Designed in Shenzhen: Shanzhai Manufacturing and Maker Entrepreneurs Silvia Lindtner University of Michigan Ann Arbor, MI-48104 lindtner@umich.edu Anna Greenspan NYU Shanghai Shanghai, China-200122 ag158@nyu.edu David Li Hacked Matter Shanghai, China-200050 david.li@hackedmatter.com design [3, 6, 7, 27, 43]. Drawing on this work, this paper argues that contemporary processes of technology design necessarily include the place and culture of production. ABSTRACT We draw from long-term research in Shenzhen, a manufacturing hub in the South of China, to critically examine the role of participation in the contemporary discourse around maker culture. In lowering the barriers of technological production, “making” is being envisioned as a new site of entrepreneurship, economic growth and innovation. Our research shows how the city of Shenzhen is figuring as a key site in implementing this vision. In this paper, we explore the “making of Shenzhen” as the “Silicon Valley for hardware.” We examine, in particular, how maker-entrepreneurs are drawn to processes of design and open sharing central to the manufacturing culture of Shenzhen, challenging conceptual binaries of design as a creative process versus manufacturing as its numb execution. Drawing from the legacy of participatory design and critical computing, the paper examines the social, material, and economic conditions that underlie the growing relationship between contemporary maker culture and the concomitant remake of Shenzhen. Today, PD’s call for involvement of users into the design process is not only accepted in popular design approaches such as human-centered design, but has also morphed into a business strategy. Bannon and Ehn, for instance, document the ways in which corporations promote the view that users and designers co-create value [3]. They illustrate the expansion of “a managerial version of user driven design” rooted in “market-oriented business models removed from PD concerns” [3]. Closed company innovation has increasingly given way to “open innovation” models, where creativity, knowledge and expertise of users are leveraged for company profit. PD’s call for critical intervention is further complicated by the recent flurry of devices and tools ranging from social media apps to smart devices (or Internet of Things), whose value depends on the participation of users. While companies like Facebook mine behavior data online to sell it back to its users in the form of ads, newer companies like Misfit see the value of smart wearables in the sensitive data their users generate and share by wearing the device while sleeping, walking, driving, working, exercising, etc. Advocates of the “maker movement” also celebrate a new formulation of user participation. By providing the tools, machines and platforms that enable people to make their own technologies, “makers” hope to turn passive consumers into active participants not only in technological design, but also in economic processes and civic matters (for prior work see e.g. [2, 18, 25, 31, 33, 39, 40, 45]. Open hardware platforms like the Arduino and affordable CNC machines like the 3D printer are envisioned to enable otherwise passive consumers to produce their own devices, tools, and eventually machines. Author Keywords Maker culture, industrial production, manufacturing, participation, open source, DIY, China, , Shanzhai. ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION Critical scholarship of computing has long been committed to questioning the apparently strict separation between production and consumption, design and use. One of the most widely known and impactful approaches has been participatory design (PD). With roots in the Scandinavian labor movement in the 1970s, PD emerged alongside outsourcing, automation and the introduction of Information Technology into the workplace. PD sought to intervene in these processes, promoting the view that the user and the larger social context and surrounding material culture should be central to considerations and processes of This contemporary promotion of “participatory production” [3] has critical gaps, as a return to the original concerns central to participatory design makes clear. Although design has been opened up to include and benefit from the participation of users (as elaborated above), the question of who is considered a legitimate participant in the design process has remained fairly limited. In particular, there is often an unspoken separation between what happens in the design studio, makerspace, hardware incubator or home office (the site of ideation, co-creation, appropriation, and Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21265 85 day-to-day use) and what happens on the factory floor (the site of manufacture, production, and wage labor). The “human” in human-centered design, the “participant” in participatory design, and the “maker” who advocates the “democraticization” of production concentrates on the designer-user/producer-consumer relationship, but rarely on the relationship to the factory worker, producer, mechanical engineer, and so on. This is particularly ironic considering PD’s original concern to intervene in processes of outsourcing, deskilling of labor, and the re-organization of work [3, 6, 7]. The central argument of what follows is that ‘participation’ in the design process does not only include the social context of the end user, but also, crucially, the material, socio-economic and cultural context of production. This paper demonstrates this by focusing on the manufacturing hub of Shenzhen, China, as a crucial agent in much of the design and creation of contemporary technology. models of technological production in which design and innovation are seen to emerge predominantly from global epicenters in the West (e.g. Silicon Valley) [2, 3, 13, 23, 25, 38]. Our work builds upon this research, by taking seriously manufacturing as site of expertise, design and creative work. We draw from long-term ethnographic research with factories, makers, and hardware start-ups in Shenzhen, a global hub of electronic manufacturing located in Southern China. In this paper, we analyze the social, technological, and economic processes of manufacturing in Shenzhen, rooted in a culture of tinkering and open source production that has evolved in the shadows of global outsourcing and large-scale contract manufacturing. We demonstrate that a growing number of maker entrepreneurs have begun to intersect with this manufacturing ecosystem, experimenting with modes of design, production, and collaboration. Examining these intensifying collaborations enables a deeper and nuanced conceptualization of both design and of the ongoing transformation of Shenzhen. “Making” is often celebrated as a method that might revitalize industrial production in Western knowledge economies, e.g. [1]. In reality, this is not a straight-forward or easy process. Many hardware start-ups face difficulties in transitioning from hobby to professional making and manufacturing [16, 49]. A number of businesses have tried to capitalize on these difficulties by providing makerentrepreneurs with an access to manufacturing in China. Take, for instance, Highway1, a hardware incubator in San Francisco, which promises start-ups a smooth transition into mass manufacturing without having to spend substantial amounts of time at their China-based manufacturing sites. Here, engaging with manufacturing expertise is rendered a problem space and an inconvenient hurdle for designers, makers and start-ups. Implicit in this approach is a widespread conception of technology production, which splits manufacturing and design along geographical lines; in which technology is conceived and designed in the West, and then manufactured in low-wage regions with loose regulatory environments. The evidence of this idea of design is emblazoned on the iPhone: “Designed by Apple in California. Assembled in China.” Designers, here, are understood as the agents, with their ideas being executed elsewhere. In its most extreme formulation this division corresponds to a Cartesian inspired ‘mind-body dualism’ in which an active rational mind in the West guides a passive, inert body in the so-called developing world. Shenzhen & the maker movement In the last years, there has been a growing interest in the potential impact of a so-called “maker” approach to technological innovation, education, and economic growth [29]. “Making” is thought to enable a move from tinkering and play, to prototyping and entrepreneurship and, finally, to help revive industries and sites of manufacturing lost due to histories of outsourcing. Making is drawing investment from governments, venture capitalists, and corporations around the world. While the US government promotes digital fabrication and making as a way to return to the “made in America” brand (with the White House hosting its own Maker Faire) [33, 36], the European Union has introduced formal policies aimed at rebuilding manufacturing capacities and know-how in order to sustain their knowledge economies [15]. Large international corporations have also started to invest. In 2013, Intel introduced the Arduino compatible Galileo board; an “Intel inside” microcontroller platform aimed at branding Intel as a champion of the maker approach. Our work challenges the dominant narratives of maker culture by critically investigating the relationship between making, designing, and manufacturing. We argue for a return to one of the most fundamental concerns of PD, i.e. to foreground the expertise, tacit and situated knowledge of everyday work practice [43, 46]. Our focus is on the ways in which the city of Shenzhen has emerged as a central player in the broader imaginary as making shifts from hobby to entrepreneurial practice. Shenzhen figures in the global maker imaginary as a “maker’s dream city” or “the Silicon Valley for hardware,” where visions of technological futures get implemented today. Until recently, few technology researchers and people in the broader IT media sector have paid much attention to Shenzhen. This began to change, when a growing number of “makers” In this paper, we build on prior work that has begun challenge simplistic binaries of design-production, examining processes and cultures of design, making, and repair in regions outside of the United States and Europe [3, 23, 25, 39]. Drawing from research with mobile repair workers in rural Namibia, for example, Jackson et al. [25] focus on mundane sites of repair, breakdown and reuse as important, but often neglected sites of design. In engaging with these often overlooked places, commonly thought of as technologically, economically and socially “behind,” scholars have argued for an approach that challenges 86 traveled to the coastal metropolis to turn their ideas into end-consumer products. Well-known examples of these made-in-China devices are the virtual reality goggles Oculus Rift and the Pebble smart watch. In 2012, one of the first hardware incubator programs, HAXLR8R (now renamed as HAX), opened its offices in Shenzhen. Other investment programs such as Highway1, Bolt, and Dragon Innovation followed suit. Shenzhen draws not only makers and hardware start-ups, but also large corporations such as Intel, Texas Instruments, Huawei, and more. Intel, for instance, has invested 100 million USD in what the company calls the “China Technology Ecosystem” in Shenzhen [22]. Since 2013, the MIT Media Lab has organized tours for its students through Shenzhen’s electronic markets and factories. In a recent blog post Joi Ito, head of the Media Lab, records his impressions, describing local factories as “willing and able to design and try all kinds of new processes to produce things that have never been manufactured before” [24]. participant observation in five hackerspaces and at over thirty maker-related events such as Maker Faires, Maker Carnivals, Hackathons, Barcamps, and Arduino workshops across the cities of Shanghai, Beijing, and Shenzhen, as well as several months of ethnographic fieldwork at a hardware incubator in Shenzhen, following the day-to-day workings of ten start-ups and their journeys of moving from idea into production. Participant observation at hackerspaces included joining daily affairs such as prototyping, space management, member meet-ups, open houses, and the organization of workshops. The research at the hardware incubator included daily observations at the office space as well as accompanying start-ups during sourcing, prototyping, and manufacturing. Between 2012 and 2014, we made numerous trips to Shenzhen to focus on the history and culture of the region’s local manufacturing industry. We hosted a series of handson workshops and intensive research trips in Shanghai and Shenzhen (in total 5 over the duration of two years). These events enabled us to bring together an interdisciplinary network of 120 scholars, makers and industry partners from China, the United States, South-East Asia, and Europe concerned with “making.” Backgrounds of our participants spanned the fields of HCI, the arts, design, engineering, manufacturing, science fiction writing, and philosophy. How did Shenzhen, once known as a site of cheap and low quality production, become the place to be for contemporary hardware innovation? How have design processes, such as those that Ito speaks of, developed and fed into the culture of manufacturing that has emerged in the city over the past three decades? Who is considered a legitimate participant and what sites of expertise and design are rendered invisible? Throughout these events, we collated hundreds of hours of video and audio material of interviews, field visits, panel discussions, hands-on workshops and discussion sessions. In total, we conducted over 150 formal interviews with relevant stakeholders including makers, members and founders of hacker and maker spaces, organizers of maker related events, factory workers, owners, and managers, government officials and policy makers, employees in design firms and large IT corporations who were invested in making and manufacturing, artists and urban planners, entrepreneurs and investors. As common in ethnographic research, we prepared sets of interview questions, which we expanded and modified as we went along and identified emergent themes and new questions. We combined discourse analysis, situational analysis [11], and research through design [5, 51]. Although we have interviewed people from a wide range of backgrounds, for the purposes of this paper, we draw on a subset of our interviews, which were conducted with people active in Shenzhen’s manufacturing industry as well as those from the global maker scene who are intersecting with manufacturing. As many of our interviewees are public figures, we refer them, when they spoke in a public context (e.g. at workshops, conferences, Maker Faires, etc.), by their real names. We anonymized all informal conversations and interviewees who preferred not to be named. The findings presented in this paper challenge the common binary of “made in China” versus “designed in California” that inherently associates the West with creativity and innovation and China with low quality production. We argue that what we witness in Shenzhen today has an important impact on the relationship between making, manufacturing and design. This paper contributes by shedding light on a situated practice of design, prototyping and ideation that emerges from within manufacturing. The paper, thus, provides new insights into histories and cultures of professional design and making that have emerged outside of more familiar IT hubs such as Silicon Valley [41, 47]. Our aim is to foster an engagement with mundane sites of contemporary industrial production – like Shenzhen – in order to advance a critical inquiry of design, maker production, global processes of technology work and labor, and participation. METHODS & APPROACH We draw from long-term research about technology production in China in order to examine the cultural and technological processes that unfold at the intersection of design and manufacturing. This includes in-depth ethnographic research conducted over 5 years, hands-on participation in maker and manufacturing projects, and the hosting of a series of interdisciplinary workshops and conferences that brought together scholars and practitioners concerned with making and manufacturing. Ethnographic research conducted by the first author included long-term Our research team comes from a mixed background including interaction design, HCI, cultural anthropology, China studies, urban studies, philosophy, entrepreneurship, and physical computing. This has proven to be effective in 87 allowing an in-depth engagement with both the technological and social practices of making and manufacturing. All of us speak Mandarin Chinese (one of us is a native speaker and the other two have received formal language training for more than 5 years). Interviews were conducted in both English and Chinese. All formal interviews were professionally translated and transcribed. As contract manufacturers grew in size, and began catering predominantly to large brands, a network of entrepreneurs saw an opportunity to establish themselves in the gaps of the global economy. A dense web of manufacturing businesses emerged in Shenzhen, catering towards less well-known or no-name clients with smaller quantities, who were not of interest to the larger players. This less formal manufacturing ecosystem (known as shanzhai 山 寨 in Chinese) is comprised of a horizontal web of component producers, traders, design solution houses, vendors, and assembly lines. They operate through an informal social network and a culture of sharing that has much in common with the global maker movement (though largely motivated by necessity rather than countercultural ideals). We now turn, in greater detail, to this local manufacturing culture. SHENZHEN: FROM OUTSOURCING TO SHANZHAI Shenzhen is a young city; the build up of its urban landscape dates back only 30 years ago, when a series of village collectives began to be transformed into one of the world’s largest manufacturing hubs, e.g. [14, 34]. This was in part enabled by the implementation of a government policy that declared Shenzhen a Special Economic Zone (SEZ) [19, 30]. In 1979, when the SEZ policy went into effect, Shenzhen had a population of under 50 000, by 2010 it had morphed into a metropolis of over 10 million people1. Shanzhai 山寨 Shanzhai translates into English as mountain stronghold or mountain fortress, and connotes an informal, outlaw tradition. The term has been in use in China for a long time and features most prominently in folk stories like the Shuihuzhuan (water margins) that tells the adventures of 108 rebels, who hide in the mountains and fight the establishment. Building on this common narrative, Jeffrey, describes shanzhai as the story of “outlaws who have gone away to the mountains, doing things within their own rules. There's an element of criminality about shanzhai, just the way that Robin Hood is a bit of an outlaw. But it's really about autonomy, independence, and very progressive survival techniques.“ [26]. The growth of Shenzhen coincided with, and was propelled by, an outsourcing boom, which, to quote Lüthje et al., “emerged from the massive restructuring of the US information technology industry that began in the 1980s” [30]. Throughout this period, companies in the US and Europe moved their manufacturing facilities into low-cost regions of the so-called developing world. Shenzhen constituted a particularly attractive site; as an SEZ the barriers of entry for foreign corporations were significantly lowered, with a range of incentives including tax reductions, affordable rents and investments aimed at integrating science and industry with trade. The outsourcing of factories and manufacturing clusters radically reshaped the high tech districts of the United States. As a result, by the 1990s, with the rise of the “new economy,” the IT industry was “no longer dominated by vertically integrated giant corporations such as IBM but rather was shaped along horizontal lines of specialized suppliers of key components such as computer chips, software, hardware disk drives, and graphic cards” [30]. Scholars speculate that the term was first applied to manufacturing in the 1950s to describe small-scale familyrun factories in Hong Kong that produced cheap, low quality household items, in order to “mark their position outside the official economic order” [19]. They produced counterfeit products of well-known retail brands such as Gucci and Nike, and sold them in markets that would not buy the expensive originals. As electronic manufacturing migrated to Shenzhen the informal network of shanzhai manufacturing found a perfect product in the mobile phone. Shanzhai production includes not only copycat versions of the latest iPhone, but also new creations and innovations of phone design and functionality (see Figure 1). With the gradual upgrade of technological and organizational skills in former low-cost assembly locations, a process of vertical re-integration began to take place. By the late 1990s, Taiwanese ODMs (original design manufacturing) such as Acer, HTC, Asus and Foxconn, which designed the manufactured product on behalf of their brand-name customers, started to develop substantial intellectual property rights on their own [30]. One particularly famous example is the ODM HTC that entered the market with its own branded cell phone. This shift began challenge the global leadership of established hightech economies. Within China, shanzhai devices are catered towards lowincome migrant populations that could not afford more expensive branded products. Shanzhai phones have a strong global market, targeting low-income populations in India, Africa, and Latin America [20, 48]. As the shanzhai ecosystem matures, we are beginning to see the development of branded phones. Xiaomi (小米 to take but one example, is an affordable smart phone that comes with a chic design and makes use of sophisticated branding techniques. Although it grew by leveraging the shanzhai industry, Xiaomi is rarely associated with it. 1 We can’t do justice here to the complexity of Shenzhen’s history and direct the reader to work by Mary Ann O’Donnell, Juan Du, Winnie Wong, Josephine Ho, Carolyn Cartier, and others [9, 14, 19, 34, 35, 50]. 88 hardware hacking in the West is celebrated as enabler of future innovation, the open manufacturing mechanism of shanzhai is often denounced as holding China back on its modernization path due to its lack of principles and norms such as the international copyright law [19]. In the next section we describe in greater detail the particularities of shanzhai’s open production. OPEN MANUFACTURING: GONGBAN & GONGMO During our research in Shenzhen, we met and interviewed many different players in shanzhai production ranging from component producers, vendors, traders, assembly, and design solution houses. One consistent element that we found to be at the core of shanzhai was the production of so-called “public boards,” called gongban ( 公 版 ) in Chinese; production-ready boards designed for endconsumer electronics as well as industry applications. Gongban are typically produced in independent design houses that link the component producers (e.g. a chip manufacturer) and the factories that assemble the different parts into phones, tablets, smart watches, medical devices, and so on. Figure 1. Examples of four shanzhai phones (from left to right): phone shaped as apple, phones shaped after children's toy and Chinese alcohol brand, phone that also functions as flashlight and radio. Photos taken by authors, 2012-2014. Rather it has become widely accepted as a national phone brand that many Chinese are proud of. While some people associate shanzhai with stealing and low quality goods [48], there is a growing endorsement of shanzhai as a prime example of Chinese grassroots creativity that has innovated an open source approach to manufacturing. One strong proponent is Bunnie Huang, who gained widespread recognition when he hacked the Xbox in 2003. In a series of blog posts, Huang details the workings of shanzhai as a unique “innovation ecosystem [that developed] with little Western influence, thanks to political, language, and cultural isolation” [21]. Huang here refers to a highly efficient manufacturing ecosystem that rests on principles of open sharing that are different from, but also compatible with, more familiar open sharing cultures. During our research, we followed closely the process of one of the region’s largest distributers and their internal design house that produces about 130 gongban per year. The design house does not sell any of these reference boards, but rather gives them out to potential customers for free, alongside a list of components that go into making the board as well as the design schematics. The company makes money by selling the components that go into the boards. As such, it is in their interest to support as many companies as possible to come up with creative “skins” and “shells” (called gongmo in Chinese) that are compatible with their boards. Their customers, then, take a gongban of their liking as is or build on top of it. The boards are designed so that the same board can go into many different casings: e.g. one board can make many different smart watches or many differently designed mobile phones. Since 2010, years before Pebble Watch or the Apple Watch made news, thirty some companies in Shenzhen were shipping their own smart watches based on this open production mechanism (see Figure 2). Shanzhai is neither straightforward counterculture nor prosystem. As a multi-billion USD industry, it is deeply embedded in contemporary modes of capitalist production. At the same time, with its roots in and ongoing practices of piracy and open sharing, shanzhai challenges any inherent link made between technological innovation and the tools, instruments, and value systems of proprietary, corporate research and development. As Jeffrey and Shaowen Bardzell argue, analysis that upholds the strict boundaries between critical design and affirmative design; resistance culture and capitalist culture is often too simplistic [4]. The gongban public board functions like an advanced version of an open source hardware platform such as the Arduino, yet differs in that it constitutes a bridge into manufacturing. “We call this shanzhai in Shenzhen. It’s a mass production artwork,” explained Larry Ma (anonymized), the head of the aforementioned distributor’s design house. To Larry Ma, there is no question that shanzhai is different from simple copycat. “First, shanzhai needs creativity: it is something only a person with a quick reaction who knows the industry chain very well can do. Shanzhai producers are acutely aware of the global market economy, and have developed incisive and canny strategies to negotiate, subvert, criticize, ironize, and profit from it [19]. The early and affordable shanzhai versions of the smart phone, for instance, were designed for customer segments that could not afford the expensive and branded phones on the market. Shanzhai disrupted who gets to decide over new markets, customers, and how tech business was to be done. In other words, issues of concern in critical and reflective design practice – such as “passivity,” “reinforcing the status-quo,” “illusion of choice” [4] – are as salient in shanzhai production as they are in conceptual design. It is particularly ironic, then, that while open 89 and important decisions with regards to investment, release dates, and collaboration partners are often made over informal dinner meet-ups and weekend gatherings. These social connections are central to getting business done in Shenzhen, as we discuss in greater detail in the next section. Many of our interlocutors saw themselves as belonging to a grassroots community and maintained that it was the mutual support of Shenzhen’s open manufacturing culture that enabled their competitive advantage. MAKING IT IN SHENZHEN Shenzhen’s population comes from elsewhere. More than 95% of the city’s population is migrants. Shenzhen’s technology sector grew from the intersection of two early flows. The first were technological entrepreneurs from Taiwan, involved in the early chip industry, who sought to take advantage of China’s economic opening and it’s initial experiments with SEZ’s. This stream of capital cross-fed into a giant internal movement throughout the Mainland, in which a vast ‘floating population’, freed from the controls of the command economy, poured into the coastal cities looking for work. This dynamic is still very much at work today. In the summer of 2014, Foxconn was reported to be recruiting 100 000 workers to build the iPhone 6. Figure 2. Gongban (public/open board) and Gongmo (public/open casing) of a smart watch, Shenzhen, China. Photo taken by authors, April 2014. Shanzhai makers are asking themselves what the normal people will need next… It is very important that you are very familiar with the upstream and downstream industry chain. And there is a kind of hunger. These three elements together make it an art work… it’s about being hungry for the future.” Larry Ma’s R&D unit is one of many corporate entities in the shanzhai ecosystem that have grown over the years into substantial businesses. This growth has occurred outside the traditional IP regime, using an open manufacturing ecosystem rooted in open reference boards, and a culture in which the bill of materials (a list of all the materials that goes into making a particular device, something that a company like Apple keeps strictly closed) is shared. This open culture of production has enabled local chip manufacturers such as Allwinner and Rockchip to compete with renowned international corporations like Intel. At the crux of this manufacturing process is their speed to market, driven by what Larry Ma describes as “hunger.” In the shanzhai ecosystem, ideation, prototyping and design happen alongside the manufacturing process. Products are designed in relation to the demands of a fast changing market. Rather than spending months or years deliberating over the next big hit, shanzhai builds on existing platforms and processes, iterating in small steps. In this way, shanzhai brings new products to the market with remarkable speed. In Shenzhen, cellphones can go from conceptual designs to production-ready in 29 days. Products are market-tested directly by throwing small batches of several thousand pieces of a given product into the market. If there is demand and they sell quickly, more will be produced. There is a commitment to never building from scratch (an approach that is shared by the open source community). Prototyping and consumer testing occur rapidly and alongside the manufacturing iteration process, rather than occurring beforehand (where it is commonly placed in Westerncentric design models). It is not only the promise of a better income, but the hopes for a different future that motivate hundreds of thousands of migrant workers every year to seek employment in Shenzhen, often far away from their home towns and families, sending back remittances. Though, as is widely reported, there is an issue of sweatshop labor in Shenzhen, many of the people we met during our research promote Shenzhen as full of opportunities, a dream city, a place where “you can make it” in China today. Violet Su, for instance, worked her way up from a part-time job to personal assistant to Seeed Studio2’s CEO “Shenzhen is a good place to live,” she says. “If you go to another city, people treat you like outsiders. But here everyone belongs. It’s like as if everyone was born here. When I first came to Shenzhen I really liked one of the city’s slogan that decorated the bus: ‘When you come to Shenzhen, you are a local person.’” Many who enter the shanzhai ecosystem do not come from privileged socio-economic backgrounds. Take, for instance, Ye Wang (anonymized), the manager of a shanzhai tablet company. Wang is one of the few, who “made it.” His company has revenue of several million USD a year, shipping tablets to South America, Eastern Europe, Russia, and the United States. Wang originally came to Shenzhen at the urging of a relative who was working at the Chinese car manufacturer BYD (Build Your Dream) and who helped 2 Seeed Studio is a Chinese hardware facilitator that sells open hardware products and educational kits, and connects makers driven to move from prototyping into production with Shenzhen’s manufacturing ecosystem. www.seeedstudio.com A particular social dynamic is crucial to this design in manufacturing process. Personal and business lives blend, 90 Wang to get a corporate scholarship that funded his college education. After college, Wang entered what he calls “the shanzhai community.” He made a name for himself by leading a development team that produced one of the first copycat versions of the Apple iPad. The localized, slightly altered version of the tablet was introduced into the Chinese market before Apple had officially released the iPad in the United States. This did not go by unnoticed by bigger players in the shanzhai ecosystem. Wang explained how, once one has gained trust and made a name for themselves, it is easy to find partners who are willing to freely share resources: “Shenzhen is working just like this. You can understand it as crowdfunding. It works differently from crowdfunding via online social networking … you must be firmly settled in the industry, be recognized, have a good personality… Everybody in the industry chain gives you things for free, all the materials, and only when you have sold your product, you do the bills [and pay back].” drawn to the city’s abundance of materials and the production processes located here. For many of these newcomers the first stop in Shenzhen are the markets of Huaqiangbei ) , a 15-by-15-city block area, filled with large department store buildings. Each mall contains a labyrinth of stalls spread over several floors (see Figure 3). Malls specialize in everything from basic components such as LEDs, resistors, buttons, capacitors, wires, and boards to products such as laptops, phones, security cameras, etc. For makers, the markets provide immediate access to tools, components and expertise. Ian Lesnet from Dangerous Prototypes, a company that sells open hardware kits, describes the lure of Huaqiangbei and Shenzhen as a whole: “The wonderful thing about Shenzhen is that we have both horizontal and vertical integration. In Huaqiangbei, you can buy components. Go a little bit further out, people sell circuit boards. A little bit further out, there are people who manufacture things and attach components to circuit boards. So you can actually have something built. And a little further out there are people who make product cases. A little further out you have garages with large-format printers who make labels for your products and a little further out they recycle it back down again. So you can build something, design it entirely, have it manufactured, sell it, and then break it down its components and recycle it back into the center of the markets. You have all the skills and all the people who can do that and they are all here in one place. And that’s what’s really enticing about Shenzhen.” Wang, here, describes an important funding mechanism that enables people who lack the financial resources to nevertheless receive support from within the larger shanzhai network. People become part of this social network by participating in both informal face-to-face gatherings (over dinner, lunch, at the manufacturing site) and networking via mobile social media platforms such as Wechat (www.wechat.cn). Much of the offline activity takes place over alcohol-infused meals, KTV bars and massage parlors, establishments that are frequented by a largely male clientele, (all of which speaks to a strong gender hierarchy that infuses shanzhai culture). People in shanzhai think of themselves as driven and hard working, committed to improving their standard of living and to make money. Many considered the level of entrepreneurial possibilities unique to Shenzhen: “there is no other place like this in China. Here you find a lot of opportunities, you can become yourself, you can realize your dream, you can make a story out of your life.” “Living in Shenzhen is like living in a city-size techshop,” echoes Zach Smith, one of the co-founders of the 3D printer Makerbot. Smith first came to Shenzhen when Makerbot started to collaborate with a local manufacturing business. Since then he has spent many years working and living in the city and has learnt to adapt to what he calls Shenzhen’s “native design language.” “If you come to Shenzhen, you are going to take your American design language and you Shanzhai production is fast and nimble mostly due to this unique social fabric through which decisions about new products, design and pricing are made collaboratively. This process entails people to be “on 24/7.” Every personal interaction, no matter if offline or online, is also about furthering a collective goal: the expansion and spread of business opportunities, the discovery of niche markets and the distillation of new mechanisms that will generate additional sales. In this way shanzhai production culture is not dissimilar from Silicon Valley, with its male dominated management and entrepreneurial leadership, hard-driven work ethic and peer pressure, all of which forms a closeknit community of informal socializing and information sharing [41]. MAKERS IN SHENZHEN Figure 3. Huaqiangbei markets (upper left to bottom right): USB sticks shaped as plastic figurines, stacks of wires, assortment of magnets, department store building view from the top. In the last few years, Shenzhen has begun to draw yet another wave of migrants – mobile elites such as tech entrepreneurs, hackers, makers, geeks and artists, who are 91 through bodily reactions to the size of a button or the feel of a knob, as Ian Lesnet elaborates: “When you design electronics, it's not just an engineering problem. It's a design process. Being able to just walk into Huaqiangbei, touch buttons, push them, be like, ‘Oh, this one is weak. This one is strong.’ Choosing things. Holding things. Get this amount of knowledge that you don't get sitting at a computer sitting somewhere else in the world” (see Figure 4). Many agreed that this tacit and embedded learning had become central to their design process and was something they learned only after they had arrived in Shenzhen. “In school, they don't teach you DFM, design for manufacturing, at all,” says Antonio Belmontes from Helios Bikes, “the factory helps us bring our ideas down to design for manufacturing. They also help you save money. Especially when you approach them during the design process.” are going to have to translate it,” Smith explains. “If you are out here you can start to learn that local design language, and start using it in your own designs… It helps you make designs that are easier to manufacture, because you are not substituting a bunch of stuff… People out here can build their designs in this native way. As you go and meet with manufacturers you understand their design process, how they want to build things, or what they are capable of building. This changes the way you want to do your design, because as a designer, if you are a good designer, you are going to try and adapt to the techniques instead of making the techniques adapt to you.” What Smith describes here was something many of the makers we interviewed experienced; transforming their designs through interactions with factories, engineering processes, machines and materials. Manufacturers and makers work together to prototype, test materials and functionality, continuously altering everything from the shapes of product casings to PCB design (Printed Circuit Board). Together, they iterate and shape the design of the final product through a process that typically spans several months of frequent often-weekly meetings. Take, for instance, maker entrepreneur Amanda Williams, one of the few women active in the scene. She has been working closely with several different manufacturing units in Shenzhen during the process of designing an interactive lamp. Williams reflects on these collaborations as follows: “sometimes you find out from a factory that this won't work or that won't work, or you can't use this size because you need a certain amount of wall thickness or this material's gonna break… working with the factories, we understand how to modify our design, in order to make it better for mass manufacturing.” What draws tech entrepreneurs, makers and designers to Shenzhen is that phases of ideation, design, market testing, and industrial production evolve together in an iterative process (as opposed to design practices in which ideation and prototyping are thought of as phases that proceed and then guide processes of execution). What emerges is a tactile and deeply embodied design practice that requires close connections with both materials and the local skillsets that many describe as a highly professionalized form of making in action. John Seely Brown, former director of Xerox Park, during a visit to Shenzhen, reflected upon this process by speaking of tacit versus explicit knowledge. “What you are really doing,” he said speaking of hardware production in Shenzhen, “is modulating a conversation between your tools and the materials your are working on for some end result. And you are overseeing that dance in its own right.” Makers working in Shenzhen are brought closer to the tactility that lies at the heart of hardware design. In molding their visions whilst enmeshed – rather than removed from – the context of manufacturing, their designs become tuned to the materiality of the hardware, modulating their visions SEEED STUDIO & THE 2014 SHENZHEN MAKER FAIRE Much of what we see with regards to maker entrepreneurialism in Shenzhen today goes back to the early efforts of Seeed Studio, a Chinese hardware facilitator that connects Shenzhen’s world of manufacturing with the global maker scene. Seeed Studio was founded in 2008 by the then 26-year old Eric Pan ( and grew quickly from a two-people start-up into a successful business that now has more than 10 Million USD annual revenue and over 200 employees. Seeed Studio sells hardware kits, microcontroller platforms, and custom-made printed circuit boards to makers. It also provides highly personalized services. One of Seeed Studio’s core businesses is to enable maker start-ups to move from an idea to mass production by identifying what Eric Pan calls “pain points”—moments of transition, where a company lacks the knowledge of how to scale up. Seeed Studio products have gained reputation worldwide. They are offered for purchase online, on makerspecific platforms, and in mainstream retailers in the US. When HAXLR8R opened its doors as one of the first hardware incubator programs in Shenzhen in 2012 it was with the help and in the offices of Seeed Studio. Figure 4. in Huaqiangbei: makers getting a "feel" for different components. Photos by first author, 2013. 92 moving from making one thing to making hundreds of thousands or millions of things. While Dougherty emphasized processes of tinkering and play, Tong and Lin focused on the role of design in the professionalized manufacturing process, or as Lin put it: “the process of making just one thing is very different from continuous production. It requires cross-disciplinary work. Hardware is different from the Internet. You need to think about design from the beginning. Design is central to all steps of the process of manufacturing including differentiation, customization, standardization… You also need to design for future manufacturing, for the next assembly you need to think about this from the beginning of the design process.” Figure 5. "Innovate with China," product label by Seeed Studio. The Shenzhen Maker Faire was Dougherty’s first visit to China. When we interviewed Dougherty during his visit, he reflected on the differences of making in the US and in China. “It’s an indeterminate problem of ‘how do I get this made?” he said speaking of the difficulties many hobbyist and professional makers face in the US, “where should I go to find the parts?” Makerspaces address part of the issue, he further elaborated, but scaling up was almost impossible: “They don't necessarily have the context, skill sets or knowledge to make. Even, "What are the right things to make or not make at all?" Part of it is that American manufacturing is geared to large companies, and so those interfaces aren't there for a small company.” Dougherty, here, counters the overly euphoric narratives that view making as enabling an easy return to the “made in America” brand. “I see this as an information problem,” he says “you might find out while being here that if you manufacture it this way, you should have designed it differently.’” Eric Pan has become an influential voice of China’s maker scene eager to demonstrate that “made in China” can mean something more than just copycats and cheap labor. The first thing one reads, when entering the offices of Seeed Studio, is the tagline “innovate with China,” painted on a large mural wall. A pun on the “made in China” brand, it is also the label that adorns Seeed Studio products (see Figure 5). “When I came to the US in 2010, people there knew us and liked our products, but nobody wanted to believe that we are a Chinese company,” Pan recalls, “nobody had thought that cool and innovative products could come of China. That’s why, ever since, we have been using ‘innovate with China’ on our product labels to demonstrate that manufacturing in China can mean ‘partnership’ and innovation instead of cheap labor and low quality.” “Innovate with China” was also the slogan of China’s first featured Maker Faire that took place in April 2014, organized and hosted by Seeed Studio. The Maker Faire constituted an opportune moment for Seeed Studio to demonstrate its vision of China’s creative role in the world of making and manufacturing. People who attended the Maker Faire were well-known figures in the maker community, and included amongst others Dale Dougherty, founder of MAKE magazine, Chris Anderson, who authored the book Makers, Tom Igoe who co-founded Arduino, Jay Melican who carries the informal title “Intel’s Maker Czar,” Eri Gentry from BioCurious, Vincent Tong and Jack Lin from Foxconn. In Shenzhen, design on the factory floor is not unique to shanzhai, as those involved in the process know well. For instance, what transpired from a couple of visits to a big contract manufacturer (anonymized), even companies like Apple have their designers and engineers (just like maker entrepreneurs) work side by side with the designers and engineers at the factory, iterating together until the very last minute, when the product is frozen for release. This is in contrast to a common perception of Apple as the creator who outsources to the cheap labor provided by the manufacturer. DISCUSSION The talks and presentations at the Shenzhen Maker Faire were wrapped between two keynote speeches: Dale Dougherty, considered by many to be the founding father of the US maker movement, gave the opening speech, while Vincent Tong and Jack Lin , from Foxconn, gave the closing plenary. Dougherty, in his talk, focused on the creativity that lies in making one thing. He emphasized the culture of hobbyist creation and tinkering that went into the early stages of development of the first Apple computer, and described making as an adventure where the outcomes are uncertain. Tong and Lin, on the other hand, talked about the opportunities and challenges that lie in scaling up, “Apple computers are this really big example. Designed in California, made in Shenzhen. We pride ourselves on design and we don't have to do that other work. Remember the paperless office. Things would just be designed on computers and then made. It was almost like we didn't need that dirty world near us. It could be in China… But physical things have properties that speak to us intuitively that we cannot just analyze on a computer screen, no matter how much resolution we have. That's calling into question that split between designed here and made there. “ 93 representation of production [17]. This separates the designer and maker from the embedded and embodied practice of production and the tacit knowledge that is essential to cultures of production documented here. Our aim is to challenge a mythic structure of technology innovation in which the “creative” work of design is highlighted, while the work of manufacturing remains at arms length. In short, we follow Bannon and Ehn in arguing, alongside the tradition of design anthropology, that the insights “from an understanding of material culture” be “more directly fed on to the practices of participatory design” [3]. A rigorous participatory design practice not only includes a deep engagement with the social context of users, but also with the material and social conditions of contemporary production. (Dale Dougherty, Interview with the authors, April 2014) This paper sets out to question a prevailing myth of technological production in which design is separated from what Dougherty, here, calls the “dirty world” of manufacturing. It does so by focusing on the culture of open production and design that has developed in Shenzhen over the last 30 years. More specifically, our research has concentrated on how the ecosystem of shanzhai emerged alongside the more-well known processes of outsourcing and governmental policy that opened up the region to foreign investment. In doing so, our work challenges some of the prevalent discourses and practice around making and its engagement, however implicit, with participatory design. Central to the early efforts of participatory design, and critical scholarship of computing more broadly, has been an emphasis on the user and a desire to empower those who might have less say in technological production. Prominent figures of the maker movement have turned this call for individual empowerment into a powerful business strategy, e.g. [1]. Many maker kits and smart devices are marketed as educational in that they train their consumers to become producers themselves. Today, many users of digital fabrication tools and open hardware platforms are indeed producing a wide and rich variety of software code, electronic schematics, 3D designs, and so on. Committed to the culture and spirit of open source, many of these users also freely share their design contributions. Maker products, in this sense, function much like social media apps such as Facebook or virtual worlds like Second Life, in which the value of the product is significantly shaped by what people “make” with it [8]. While this certainly broadens the range and number of “participants” in the design of technology it is also subject to a growing critique of the “sharing economy,” in which, “the labor of users, fans, and audiences is being put to work by firms” [45]. Maker entrepreneurs who come to Shenzhen to turn visions of smart and networked devices into products are intersecting with these embedded and tactile processes of production. Indeed, it is the close proximity to the processes and materials of production that makes the city so enticing to makers. As we have shown in this paper, it is not just access to tools and machines, but a particular process of design that draws makers into Shenzhen; prototyping is part and parcel of fabrication, rather than preceding it; and testing and designing evolves through daily interactions with the workings of machines, materials, components, and tools. From the electronic markets and craftsman workshops to assembly lines and design solution houses, Shenzhen immerses technology designers in a mode of prototyping that is tied to the feel and touch of materials as well as the concrete processes of manufacturing. Many of the people we interviewed agreed that “being in it” was crucial to learning, understanding, and working with what they considered to be an open, informal and highly professionalized design practice. The goal of this paper has been to critically unpack contemporary maker discourse by examining the remake of Shenzhen. In so doing, we question the imaginary of Shenzhen as the “Silicon Valley for Hardware,” that has been fueled by promotional campaigns of hardware incubators and corporate investment in the region. These often linear stories of progress, which assume that Shenzhen is “catching up” with innovation centers like Silicon Valley, tend to be void of the intricacies of the region’s production processes described in this paper; from its history of outsourcing and piracy to the global scale of contemporary shanzhai production. We have shown that innovation, design and production are necessarily situated, evolving in close relation to particular histories of technological, economic and social development. In this, the paper follows the call to locate design [9, 23, 25, 46] so as to include the site of industrial production itself. Efforts in critical computing have long called upon researchers and designers to reflect upon “the values, attitudes, and ways of looking at the world that we are unconsciously building into our technologies” as well as the “values, practices and Moreover, digital fabrication tools such as the 3D printer or the CNC milling machine, which are envisioned to enable a broader audience to engage with processes of making, often keep the designer at arm’s length from the kind of tacit knowledge necessarily involved in the manufacturingcentered design process we have described in this paper. While digital fabrication tools provide techniques for rapid prototyping in a design studio, they do not engage one with the situated and embodied processes of manufacturing on a large scale. What becomes increasingly clear from our engagement with Shenzhen is that, to repeat Dougherty’s point stated above, “physical things have properties that speak to us intuitively that we cannot just analyze on a computer screen, no matter how much resolution we have.” Thus, whilst the promotion of a return to hands-on making is pervasive (“everyone is a maker”), many of the software applications aimed at bringing designers into the production of hardware have been oriented around creating an abstract 94 experiences that are unconsciously, but systematically left out” [42]. 8. Boellstorff, T. 2008. The coming of age in Second Life: an anthropologist explores the virtually human. Princeton University Press. Clearly this extends well beyond the common user-designer relationship. What values, norms and attitudes towards manufacturing and production do we consciously or unconsciously build not just into our designs, but also into our critical theories and practices? What new possibilities are opened up if we take seriously diverse and distributed cultures of production? Who is considered a legitimate participant in the “maker” revamp of industrial production? What expertise and work is rendered invisible as makers turn visions of networked objects into mass-produced artifacts? These questions recall the central concerns of early theorists of participatory design: a deep engagement with sites of production, labor, and manufacturing. 9. Cartier, C. 2002. Transnational Urbanism in the Reform-era Chinese city: Landscapes from Shenzhen. Urban Studies 39: 1513-1532. 10. Chan, A. Networking Peripheries: Technological Futures and the Myth of Digital Universalism. MIT Press (2014). 11. Clarke, A. 2005. Situational Analysis: Grounded Theory after the Postmodern Turn. SAGE Publications. 12. Dourish, P. 2001. Where the Action Is. The Foundations of Embodied Interaction. MIT Press. 13. Dourish, P. and Mainwaring, S.D., "Ubicomp’s Colonial Impulse," Proc. of UbiComp'12, Springer (2012), 133142. ACKNOWLEDGMENTS We would like to thank everyone who contributed to this research, particularly those at Seeed Studio, Chaihuo , DFRobot, and XinCheJian as well as all the makers, entrepreneurs, and shanzhai producers, who shared with us their time and insights. This research was in part funded by the National Science Foundation (under award #1321065), the Lieberthal-Rogel Center for Chinese Studies, and the Intel Science and Technology Center for Social Computing. 14. Du, J. Shenzhen: Urban Myth of a New Chinese City. Journal of Architectural Education, Volume 63, Issue 2, pp.65-66. 15. Factory in a day. http://www.factory-in-a-day.eu/ 16. Greenspan, A., Lindtner, S., Li, D. 2015. Atlantic 17. Hartman, B., Doorley, S., Klemmer, S. Hacking, Mashing, Gluing: Understanding Opportunistic Design. IEEE Journal of Pervasive Computing, 2008, Vol. 7, No. 3, pp. 46-54. REFERENCES 1. Anderson, C. 2012. Makers. The New Industrial Revolution. Crown Publishing Group, New York. 18. Hertz. G. 2013. Critical Making. http://conceptlab.com/criticalmaking/ Last accessed September 2014. 2. Ames, M., Bardzell, J., Bardzell, S., Lindtner, S., Mellis, D. and Rosner, D. Making Cultures: Empowerment, Participation, and Democracy - or Not? In Proc. Panel at CHI'14, ACM Press (2014).Anderson, C. Makers: the new industrial revolution. Random House, 2012. 19. Ho, J. 2010. Shanzhai: Economic/Cultural Production through the Cracks of Globalization. Crossroads: Cultural Studies Conference. 20. Houston, L. Inventive Infrastructure: An Exploration of Mobile Phone, Dissertation: Lancaster University, 2014. 3. Bannon, L. and Ehn, P. 2012. Design Matters in Participatory Design. In: J. Simonsen and T. Robertson (eds) Routledge Handbook of Participatory Design, pp. 37-63. 21. Huang, B. 2013. The $12 Gongkai Phone. http://www.bunniestudios.com/blog/?p=3040, last accessed September 2014. 4. Bardzell, J. and Bardzell, S. 2013. What is “Critical” about Critical Design. ACM Conf. Human Factors in Computing Systems CHI 2013 (Paris, France), pp. 3297-3306. 22. Intel Newsroom. 2014. Intel CEO outlines new computing opportunities, investments and collaborations with burgeoning China Technology Ecosystem. http://newsroom.intel.com/community/intel_newsroom/ blog/2014/04/01/ last accessed June 3, 2015. 5. Bardzell, J., Bardzell, S., and Hansen, L.K. 2015. Immodest Proposals: Research Through Design and Knowledge. Proc. of ACM Conference Human Factors in Computing Systems CHI'15 (Seoul, South Korea). 23. Irani, L., Vertesi, J., Dourish, P., Philip, K., and Grinter, R. 2010. Postcolonial Computing: A Lens on Design and Development. Proc. ACM Conf. Human Factors in Computing Systems CHI 2010 (Atlanta, GA), 13111320. 6. Beck, E. 2002. P for Political: Participation is Not Enough. Scandinavian journal of Information Systems, 14 (1), pp. 77-92. 7. Bødker, S. 1996. Creating Conditions for Participation – Conflicts and Resources in Systems Development. Human-Computer Interaction, (11:3), pp. 215-236. 24. Ito, J. 2014. Shenzhen trip report – visiting the world’s manufacturing ecosystem. LinkedIn Pulse. https://www.linkedin.com/pulse/article/2014081706093 6-1391 last accessed September 2014. 95 25. Jackson, S.J., Pompe, A. and Krieshok, G. Repair worlds: maintenance, repair, and ICT for development in rural Namibia. In Proc. the ACM 2012 conference on Computer Supported Cooperative Work (CSCW 2012), ACM Press (2012), 107-116. office/2012/03/09/remarks-president-manufacturingand-economy) 38. Philip, K., Irani, L., and Dourish, P. 2012. Postcolonial Computing: A Tactical Survey. Science, Technology, and Human Values. 37(1), 3-29. 26. Jefferey, L. 2014. Mining Innovation from an Unexpected Source: Lessons from the Shanzhai. Lecture at the Shenzhen Maker Faire, 2014. 39. Rosner, D. and Ames, M. 2014. Designing for repair?: infrastructures and materialities of breakdown. In Proc. the 17th ACM conference on Computer supported cooperative work and social computing, 319-331. 27. Kensing, F. and Blomberg, J. 1998. Participatory Design: Issues and Concerns. Computer Supported Cooperative Work (7:3-4), pp. 167-185. 40. Rosner, D. Making Citizens, Reassembling Devices: On Gender and the Development of Contemporary Public Sites of Repair in Northern California. Public Culture 26, 1 72 (2014), 51-77. 28. Norman, D. and Klemmer, D. 2014. State of Design: How Design Education must change. LinkedIn Pulse. https://www.linkedin.com/pulse/article/2014032510243 8-12181762. 41. Saxenian, A. Regional Advantage: Culture and Competition in Silicon Valley. Harvard University Press 1996. 29. Lindtner, S., Hertz, G., Dourish, P. 2014. Emerging Sites of HCI Innovation: Hackerspaces, Hardware Startups, Incubators. In Proc. of the ACM SIGCHI Conference on Human Factors in Computing Systems CHI'14 (Toronto, Canada), pp.439-448. 42. Sengers, P., Boehner, K., David, S., Kaye, J. 2005. Reflective Design. AARHUS’05, pp.49-58. 43. Shapiro, D. 2005. Participatory Design: the will to succeed. AARHUS’05, p.29-38. 30. Luethje, B., Huertgen, St., Pawlicki, P. and Stroll, M. 2013. From Silicon Valley to Shenzhen. Global Production and work in the IT industry. Rowman & Littlefield Publishers. 44. Sivek, S.C. "We Need a Showing of All Hands" Technological Utopianism in MAKE Magazine. Journal of Communication Inquiry 35, 3 (2011), 187-209. 45. Söderberg, J. 2013. Automating Amateurs in the 3D printing community: connecting the dots between ‘deskilling’ and ‘user-friendliness.’ Work organization, labour and globalization, vol 7, no. 1, pp. 124-139. 31. Mellis, D.A. and Buechley, L. Do-it-yourself cellphones: an investigation into the possibilities and limits of high-tech diy. In Proc. the 32nd annual ACM conference on Human factors in computing systems (CHI 2014), ACM Press (2014), 1723-1732. 46. Suchman, L. 2002. Located Accountabilities in technology production. Scandinavian Journal of Information Systems, 14 (2): 91 – 105. 32. Milian, M. The Sticky Situation that delayed the Pebble Smartwatch. Global Tech, Bloomberg, Sep 16, 2013. 47. Turner, F. 2006. From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Catalogue, and the Rise of Digital Utopianism. Chicago: University of Chicago Press. 33. Nguyen, Josef. MAKE Magazine and the Gendered Domestication of DIY Science. Currently under the review for: Perspectives on Science. 34. O’Donnell, M-A. 2010. What exactly is an urban village anyway? http://shenzhennoted.com/2010/03/19/whatexactly-is-an-urban-village-anyway/ last accessed June 1, 2015. 48. Wallis, C. and Qiu, J. 2012. Shanzhaiji and the Transformation of the Local Mediascape of Shenzhen. In Sun, W. and Chio, J. (eds). 2012. Mapping Media in China, pp. 109-125, London: United Kingdom: Routlege. 35. O’Donnell, M-A. 2011. Utopian Shenzhen 1978-1982. http://shenzhennoted.com/2011/08/27/utopianshenzhen-1978-1982/ last accessed June 1, 2015. 49. Williams, A. and Nadeau, B. 2014. Manufacturing for Makers: From Prototype to Product. Interactions XXI, Nov-Dec 16, p.64. 36. Office of the Press Secretary, the White House (2013). Remarks by the President in the State of the Union Address. (http://www.whitehouse.gov/the-pressoffice/2013/02/12/remarks-president-state-unionaddress). 50. Wong, W.W.Y. Van Gough on Demand: China and the Ready Made. University of Chicagor Press, 2014. 51. Zimmerman, J., Forlizzi, J., Evenson, S. 2007. Research through design as a method for interaction design in HCI. In Proc. of the SIGCHI Conference on Human Factors in Computing Systems, pp. 493-50. 37. Office of the Press Secretary, the White House (2012). Remarks by the President on manufacturing and the economy. (http://www.whitehouse.gov/the-press- 96 Material Speculation: Actual Artifacts for Critical Inquiry Ron Wakkary1,2, William Odom1, Sabrina Hauser1, Garnet Hertz3, Henry Lin1, Simon Fraser University, Surrey, British Columbia, Canada1 Eindhoven University of Technology, Eindhoven, Netherlands2 Emily Carr University of Art and Design, Vancouver, British Columbia, Canada3 {rwakkary, wodom, shauser, hwlin}@sfu.ca; r.l.wakkary@tue.nl; ghertz@ecuad.ca In considering the productive pairing of design and fiction to advance critical speculation, there is an opportunity to explore other forms of fiction informed practices that might nurture and expand interaction design research efforts. To date, fictional thinking in design has focused on science fiction and scenarios, and on conceptual artifacts like nonfunctioning prototypes, storytelling props, and fictional objects. The HCI community has paid less attention to other theories of fiction in addition to science fiction. Relatedly, HCI researchers have largely overlooked the role that actual and situated artifacts in the everyday can offer for speculative and critical inquiries in design. This shift in attention to actual and situated artifacts would reveal design artifacts and everyday settings to be sites for speculative and critical inquiry. ABSTRACT Speculative and fictional approaches have long been implemented in human-computer interaction and design techniques through scenarios, prototypes, forecasting, and envisionments. Recently, speculative and critical design approaches have reflectively explored and questioned possible, and preferable futures in HCI research. We propose a complementary concept – material speculation – that utilizes actual and situated design artifacts in the everyday as a site of critical inquiry. We see the literary theory of possible worlds and the related concept of the counterfactual as informative to this work. We present five examples of interaction design artifacts that can be viewed as material speculations. We conclude with a discussion of characteristics of material speculations and their implications for future design-oriented research. This paper introduces a complementary concept to design fiction that we call material speculation. This concept draws on the literary theory of possible worlds [cf. 48]. Material speculation emphasizes the material or mediating experience of specially designed artifacts in our everyday world by creating or reading what we refer to as counterfactual artifacts. Material speculation utilizes physical design artifacts to generate possibilities to reason upon. We offer material speculation as an approach to critical inquiries in design research. In plain fashion, for this paper we consider speculative inquiries that aim to generate progressive alternatives to be critical inquiries. Author Keywords Material Speculation, Speculative Design; Design Fiction. Critical Inquiry ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION Interaction design and human-computer interaction (HCI) have long borrowed from fiction in design techniques like scenarios, personas, enactments, and even prototyping. Speculative inquiries in design like futuring, forecasting, and envisionments have also deeply incorporated practices of fiction. Recently, design fiction has emerged as a uniquely productive approach to speculative inquiries. Most importantly, design fiction has extended the speculative aim of design–its future orientation–into more reflective realms that critically challenges assumptions we hold about design and technology. This is a valuable step in interaction design research toward offering approaches to more critical speculative inquiries. Our work builds on speculative and critical design, which can be seen as broad yet established approaches to design aimed at exploring and questioning possible, plausible, probable, and preferable futures [18, 19, 24]. Notions of speculative and critical approaches to design have a long history that extends across several disciplines and continue to be the subject of ongoing theorization and debate [4, 18, 41, 30, 1, 50, 27]. A primary goal of this paper is to contribute to the growing relevance and interest in a speculative and critical position on design in the HCI community. We do this through proposing material speculation as a conceptual framing for reading and creating design artifacts for critical inquiry. It is important to note that we do not propose material speculation as a means of classification or definition of artifact types or design approaches. In this sense, aspects of material speculation may well overlap with design fiction, or related notions like speculative design or critical design. Our aim is Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21299 97 not to develop a mutually exclusive term or to create a hierarchy among concepts. However, we believe there are unique benefits and outcomes in using the conceptual framing of material speculation to understand design artifacts and to create them. “Design Fiction is making things that tell stories. It’s like science-fiction in that the stories bring into focus certain matters-of-concern, such as how life is lived, questioning how technology is used and its implications, speculating about the course of events; all of the unique abilities of science-fiction to incite imagination-filling conversations about alternative futures.” Our contributions in this paper are multi-fold. We extend the critical and reflective speculation in interaction design and HCI research by articulating the concept of material speculation. This concept advances the notion that the material existence of specifically designed artifacts situated in the everyday represent a unique and productive approach to critical inquiry. This paper also contributes a theorized understanding of material speculation and, by extension, fiction-focused practices in design research through the reasoned adaption of possible worlds theory to interaction design research. More broadly, the paper can be seen as another step forward in supporting the need for more reflective and critical forms of knowledge production in HCI and interaction design. Importantly, we stress that this type of critical inquiry occurs through the conceptualizing and crafting of design artifacts to generate theoretical articulations and intellectual argumentation. For Dourish and Bell [17], science fiction provides a representation of a practice in which technical and material developments will be understood. It is not only that science fiction stories offer imaginary prototypes of things to be but also that science fiction creates environments in which these things are discussed, understood, and used. The fictional embedding of design and technology with fictional people and in practices brings to the fore the cultural questions of these futures and the roles of technologies. Dourish and Bell [17] argue that these cultural issues are inherent in our notions of design and technology. Science fiction reveals our prior cultural commitments before any implementation of design or technology. What emerges in their readings of science fiction is an “imaginative and speculative figuring of a world” in which new things and technologies will inhabit; and the bringing into focus of the “central role of sociological and cultural considerations” that are often obscured in our techno-centric reasoning of actual technologies [17]. What is evident is that science fiction affords an enhanced form of critical reasoning on technologies and design. BACKGROUND AND RELATED WORK The review of related work that follows is intended to establish two main points: 1) the potential of critical speculation in design, 2) the crafting and material strategies in speculative inquiries in design that serve as antecedents and inspirations for material speculation. Similarly, Reeves [42] sees design fictions as texts to be productively read and unpacked. Reeves argues a greater role should be given to fiction in the futuring activities of design that he refers to as envisioning. In his view, a more critical envisioning would disentangle the aspects of fiction from the less productive qualities of forecasting and extrapolations [42, p.1580] (see [10] as an exemplary approach to envisioning through fiction). Reeves specifically cites Bleecker’s method of design fiction that sets out the goals of not only reading but generating design fictions that express multiple futures and by that let go or challenge assumptions about the direction and breadth of progress. Bleecker and Reeves see in design fictions a design method that engages assumptions of the future as a means to derive critical understandings of the present. Speculation and fiction in design Fiction has had a long trajectory of use in interaction design research, particularly as a means to aid the process of interaction design and more recently as a mode of critical inquiry. Design as a discipline is concerned with change and preferred futures. As a result there is a natural orientation towards the future and the use of futuring activities in design. For example, the creation of personas fictional characters representing potential users – [15] and scenarios – narrated descriptions of future design details to inform design rationales – [14]. In parallel, prototyping has been a useful technique in design leading to final (future) designs through mockups and models. These practices of design can be seen to draw on fictional thinking or include fictional components. In design fiction, the use of fictional practices in design becomes more explicit and a source for investigating the speculative potential of design. Similarly, Bardzell and Bardzel [3] see how fiction and science fiction can reinvigorate visioning in HCI through what they conceptualize as cognitive speculation.1 They propose a form of speculative thinking that is grounded in the realities of current science but rely on imaginative extrapolation that is intellectually rigorous and reasoned. This methodological approach mobilizes science fiction to Design fictions relate to representations of the future from science fiction to design scenarios that detail “people, practice and technology” [8, p.133]. Discussions on technological futures have been well established within Science and Technology Studies (STS) [13, 52, 43], yet these discussions are relatively new to interaction design and HCI. The term design fiction arose in a presentation given by Julian Bleecker in 2008 [9]. Bleecker sees in the idea of science fiction a genre-methodology for design [9]: 1 It should be noted that our use of the name “material speculation” is largely coincidental to the name of Bardzell and Bardzell’s concept. 98 critically inquire and envision technological futures that foreground the lived experiences of the future world. speculate on industrial progress and consumer culture, and on the nature of design itself in these contexts. Dunne makes clear that the ‘functionality’ of these design artifacts is to act as materialized props through which alternative stories emerge that operate in the space between rationality and reality, where “people can participate in the story, exploring the boundaries between what is and what might be” [20, p. 67]. In other words, through leveraging the seductive language of design, design artifacts are crafted to provoke people to imagine it in use and the possible future that would manifest around it—to “become lost in a space between desire and determinism” [20, p. 67]. These deeply articulated discussions help to reveal the significance and potential of science fiction in design and critical inquiry. In essence, the practices of science fiction bring to design research the reasoning on multiple futures that challenge assumptions and the sociological, cultural, and political tendencies that underlies our representations and considerations of design and technology. In considering a design orientation to critical inquiry, cognitive speculation can be seen to have higher-level concerns that run parallel to our comparatively grounded concerns with material speculation. With respect to design fiction, its limitation is its emphasis on the creation or reading of fictional texts that are embedded with references to design and technologies. The artifact in design fictions is a mere reference, prop, or non-functioning prototype referred to as a diegetic prototype based on Kirby [29]. To paraphrase Kirby, a diegetic prototype is a technology or technological artifact that only exists in the fictional world but fully functions in that world. In opposition to that, our discussion of material speculation opens the critical functioning of alternative futures in design through the crafting of material artifacts that operate and exist in the actual world. This shift to materialized and crafted speculations draws on the work we generally refer to as speculative design. Dunne and Raby [18] develop this strategy further through their notion of physical fictions (p. 89), where design artifacts extend beyond being merely props for films never made, to being situated as things in exhibition spaces that “prescribe imaginings” and “generate fictional truths” (see p. 90). The aim of physical fiction design artifacts is to critically project different possibilities and to “challenge the ideas, values, and beliefs of our society embodied in material culture” (p. 90). Similar to Sengers and Gaver’s [49] aim to shift the site of meaning-making from the maker to the user, physical fictions aim to open up moments of suspended belief and, in doing so, shift our role from user to imaginer. However, similar to para-functional objects these are discursive objects—crafted interventions to create discussions. Dunne and Raby [18] refer to these as “intentional fictional objects” with no aim to be real: “physical fictions that celebrate and enjoy their status with little desire to become ‘real’” (p.89). For Dunne and Raby, the ideal dissemination for their research and experimentations is in the form of exhibitions in museums and galleries, which “function as spaces for critical reflection” [18, p. 140]. Crafting material speculations Speculative artifacts have played important and ongoing roles in design-oriented research in and outside of HCI. For example, Sengers and Gaver [49] unpack a range of speculative design artifacts that critically inquire into—and often complicate or unsettle—the relationship between functionality and user interpretation in interactive systems design. While these design artifacts were diverse and targeted various contexts, they are united in their aim to speculatively open up situations that subvert a single authoritative interpretation of a system in the service of provoking people to arrive at their own self-determined understanding of the meaning and ‘use’ of a system. Across these cases, ambiguity is leveraged as a resource to create embodied, functional systems to provoke dynamic, varied, and speculative interpretations of the design artifacts from the perspective of the user. Designers Auger and Loizeau (see www.augerloizeau.com) create speculative design projects that, similar to Dunne and Raby, take the form of installations within galleries. Drawing on a range of projects from the Royal College of Art, Auger [2] offers insights into the crafting of speculation in design. Borrowing from a range of techniques in the humanities and sciences, Auger reflects on important dimensions surfaced from these speculative design projects—from generating tension to conflict with engrained systems in our familiar everyday ecologies to carefully managing the uncanniness of the design artifact to provoke viewers to engage with the issue(s) it speculates on. These dimensions are important for constructing what Auger calls the perceptual bridge: “In effect, a design speculation requires a bridge to exist between the audience’s perception of their world and the fictional element of the concept” [2, p. 2]. Dunne’s earlier notion of para-functionality [20] predates the related work of Sengers and Gaver [49], where speculative design artifacts are intentionally crafted to encourage reflection on how technological devices and systems shape (and often constrain) people’s everyday lives, behaviors, and actions in the world. In articulating para-functionality, Dunne draws on a wide range of examples from furniture exhibitions to radical architecture proposals to satirical design projects to unpack how design artifacts can construct social fictions that critically The work discussed above reveals the role the materialized design artifact can play in critical inquiry and to shift the authority of the interpretation to the “imaginer” or “user”. A further lesson is that the material forms, along with the 99 concepts, mutually shape the inquiry and through this process become unique or specialized types of artifacts. In all cases the artifact serves as a bridge between our current world and an imagined critical alternative or transformed view of our world. advantageously occupy a creative space at the boundary between actual and possible worlds. We also elaborate on how counterfactual artifacts generate possible worlds through encounters with people. As a consequence of these features, material speculation acts as a form of critical inquiry. The intent and shaping in Dunne and Raby, and Auger and Loizeau are rhetorical strategies aimed at material artifacts as discursive interventions. As such, these designers situate their work in exhibitions arguing it occupies a critically reflective space between the real and the unreal. We argue through material speculation that the converse is equally insightful—that material speculations find a critical space of inquiry by occupying the actual or everyday world as opposed to a gallery space. Possible worlds theory Possible worlds is a philosophical concept developed in the latter twentieth century by the analytical school, including philosophers Saul Kripke and David Lewis [34, 48] and was later adopted by literary theorists [cf. 38, 22, 45]. Philosophically, possible worlds is an approach to the problem of counterfactual statements in modal logic. For example, Kripke asks what is the truth condition of the statement that Sherlock Holmes “does not exist, but in other states of affairs he would have existed” [31]; or this counterfactual statement by Ryan [48], “if a couple hundred more Florida voters had voted for Gore in 2000, the Iraq war would not have happened.” In modal logic, the question is how is each of these counterfactual statements interpreted to be true or false. The philosopher David Lewis who bridged analytical philosophy to literary theory [32] offered the position that propositions like counterfactual statements can be seen to be either true or false dependent on in which worlds the statement is true and which worlds the statement is false [34]. This allows for a reasoned argument to be made on the inevitability of the Iraq war if Al Gore was indeed elected president in 2000 despite the fact that he was not elected president. This allows for the fictional world of Sherlock Holmes to unfold such that any faltering of the detective’s deductive reasoning would be perceived as false or a negative development in the character’s intellect. In relation to the open-ended and leveraging of ambiguity in design in the work of Sengers and Gaver [49], work like the Prayer Companion [23] serves as an antecedent to material speculation. Through crafting and situating a very particular material and functional form in the everyday world it speculates and reasons in a highly critical fashion. The use of science fiction design has extended it into realms of critical inquiry that have productively opened new territory. Design fiction makes explicit the potential of fiction in combination with design to challenge and reconsider assumptions and underlying issues in order to more critically and reflectively consider next steps and advances in design and technology. Moving beyond fictional texts that embed references to design and technologies, crafted speculations are equally critical and have the capacity and dexterity to tackle broad topics for inquiry that may reflect back on design or focus beyond the field. Crafted speculations reveal the potential of shifting interpretation and meaning making to users and audiences of design through an openness and provocation embodied in the design artifacts. In the next section, we begin our descriptions of material speculation and how it contributes to this body of work. Counterfactuals are central to the theory of possible worlds. By virtue of contradicting one world (e.g., the world in which Al Gore lost the presidential election), they elicit and open up another possible world (e.g., a world in which Al Gore won the presidential election). Lewis describes counterfactuals as similar to if…then operators that create conditional modes in which possible worlds may exist [32, 48]. MATERIAL SPECULATION In what follows, we begin with an introduction of the literary theory of possible worlds. We follow this with a description of five examples of interaction design artifacts that can be viewed as material speculations. Lastly, we describe and interpret characteristics of a material speculation. Possible worlds theory relies upon the ideal that reality is comprised of all that we can imagine and that it is composed of the “actual world” and all “possible worlds” [47]. Philosophically, there are different approaches to this idea however Lewis’ view tends to prevail and is most influential with respect to literary theory [47]. Lewis sees the actual world as the concrete physical universe over time. In most respects, possible worlds are materially similar to the actual world, however they are not connected in any way spatially or temporally [33]. Importantly, Lewis also views actual worlds as having no privilege over possible worlds, rather actual worlds are simply our world, the one we inhabit. The actual world is indexical. It merely refers to the world of the inhabitant or the one who is Manifestations of possible worlds Here we articulate how particular design artifacts can be seen to generate possible worlds. We draw on key concepts from possible worlds theory to support our idea of material speculation. These include the notion of actual versus possible worlds and the notion of the counterfactual. We discuss how design artifacts can be seen as counterfactual artifacts while still being material things. We argue that the material actuality of counterfactual artifacts enables them to 100 speaking within a given world. In this sense, all worlds like the actual world hold their own internal logic and autonomy. fiction a reader construes a possible world to conform as much as possible to his or her actual world. In other words, the reader departs from his or her perceived reality only when necessary. The obvious benefit for fictional authors is that there need be no accounting for the rising and setting of the sun; if its not described in the text, a reader can assume the daily rotation of the planet. In addition, critical differences can be focused upon such as in a reference to a winged horse—a reader can imagine the combination of a known horse with known wings and speculate on that difference between the possible and the actual. These aspects give a critical functioning to the boundary between the actual and the possible. Truth conditions of the possible are seen to be relevant to the actual or at least open to be speculated upon. Further, there is a set of relational propositions that are automatically considered such as: What kind of saddle might a winged horse have? Where do flying horses migrate and settle? Is there a whole new biological class between the classes of mammals and birds? The theorist Thomas Pavel [38] referred to this as the adoption of a new ontological perspective that gives possible worlds a degree of autonomy. Metaphorical transference of possible worlds to interaction design in material speculation Amongst literary theorists there is the question of the legitimacy of considering fictional worlds as possible worlds [e.g. 44]: Would analytical philosophers validate the idea of fictional worlds as possible worlds? Ryan is equally content with the notion of a metaphorical transference between disciplines [48]. She cites fellow theorists Lubomír Doležel to argue that even if considered a metaphorical transference, the validity of the application of possible worlds theory is its potential to identify unique features of fiction that other approaches do not [48]. It is in this spirit that we extend possible worlds theory to interaction design and HCI. Material speculation is the adaption of possible worlds theory to design research. As we have discussed, when considering possible worlds and counterfactuals, in philosophy or fiction we are concerned with either a statement of logic or a text. In design, we are concerned with a material thing. A counterfactual is a virtual or tangible artifact or system in design and HCI rather than statement or text. Hence we refer to it as a counterfactual artifact. The notion of an actual counterfactual is a departure from Lewis’ criterion that possible worlds have no spatial or temporal connections to the actual world— they are remote. Yet, here we view this departure more advantageously than negatively. Given that counterfactual artifacts sit on the actual side of the boundary between the actual and possible worlds this sense of a new ontological perspective is arguably more pressing. An actual counterfactual artifact not only opens up speculation on the artifact but on its conditions as well. When encountering a material speculation, potential reasoning would include not only ‘what is this artifact’ but also ‘what are the conditions for its existence’ (e.g., including the systemic, infrastructural, behavioral, ideological, political, economic, and moral). Material speculation probes the desirability of the truth condition of the proposition and the conditions bound to it. The creative boundary between the actual and the possible There is a productive and creative space at the boundary between the actual and possible worlds, or the real and the fictional. There are many examples from fiction in literary texts, theatre or film where authors intentionally blur the distinction between actual and possible worlds for its creative possibilities. Whole genres have emerged like mystery or interactive dinner theatres that directly involve audiences in the fictional world. Live action role-playing games, augmented reality games or alternate reality games actively transgress the boundary between the real and fictional. Janet Murray’s notion of the “fourth wall” in interactive media aimed to cross theatrical illusion and actuality [35]. In these cases, interactivity is the counterfactual action that crosses the divide: fictional characters are not supposed to interact with actual people or in the actual world. In material speculation, it is making the counterfactual into an actual artifact that crosses the divide between the actual and possible worlds since, as we discussed earlier (see Possible worlds theory), counterfactuals are not supposed to exist in the same time or place as the actual world. The counterfactual artifact as proposition and generation of possible worlds In material speculation we can see the counterfactual artifact as embodied propositions similar to propositions in counterfactual statements in analytical philosophy. It is helpful to think of the counterfactual artifacts as being if…then statements as we discussed earlier (see Possible worlds theory). In this sense, the counterfactual artifacts trigger possible world reasoning that extends beyond them. In other words, the possible world or fictional account is not embodied fully in the counterfactual artifact rather it is generated by interactors in the encounter or experience of the counterfactual artifact. It is not a limitation that the counterfactual artifact is of our actual world, rather it is this very actuality that provokes or catalyzes speculation by being at the boundary of the actual and the possible. In fiction, the discussion of where the possible world is situated is more complex. Since the influence of poststructuralist thinking on literary theory, namely in concepts of open work by Umberto Eco [21] and textuality by Roland Barthes [7], meaning and fiction are seen to be Ryan [48] referred to this potential in her principle of minimal departure in which she argues that in the case of 101 generated in the act of reading by readers and not solely by the author. The importance of this for material speculation is that those interpreting or reasoning upon the counterfactual artifacts also generate possible worlds in multiplicity. Eco viewed a literary text as “a machine for producing possible worlds” [21, p.246] and in this sense we view a counterfactual artifact as a “machine” for producing possible worlds. our aim is to illustrate how they exemplify counterfactual artifacts. We only hint at the multitude of possible worlds each may generate since this is part of the lived experiences of each example. Our accounts of possible lived worlds of these material speculations are not intended to be exhaustive since this would require a separate and more indepth treatment of each that is beyond the aims of this paper. Inaccessible Digital Camera [40]—The inaccessible digital camera is a digital camera that is made of concrete; all photos are stored locally inside of its concrete case. The only way for the owner to view the photos stored on the camera is to, in effect, break the camera and retrieve the memory card stored inside (see figure 1). The inaccessible camera is part of a larger set of ‘counterfunctional devices’ designed to explore how enforcing limitations in the design of interactive technologies can potentially open up new engaging possibilities and encounters. In a later project [39], elements of the inaccessible camera design were embodied in another counterfunctional camera, named the Obscura 1C, which again had a form comprised of cement that required its owner to break it to access the digital photos stored on a memory card inside. Participants were exposed to the capsule camera in a lab setting and, later in a different study, the Obscura 1C was handed out or sold to people via an online website (e.g., Craigslist.com). The transference of criticality to interaction design The adaption of possible worlds gives fiction an immense criticality. Fictional texts can speculatively yet critically inquire upon our world. As Ryan argues [48], fiction has the capacity of truth and falsity giving it more consequence than when perceived as artistic lies or fantasy. Fiction assumes a real world shared between actual and possible worlds giving it a perch for relevance and critical insights into our actual world. However, through counterfactuals, it does not mimic the actual world, rather readers and authors alike construct possible worlds different than the actual world leading to creative and reasoned speculations [48]. In interaction design, counterfactual artifacts can also be seen to gain a perch in this critical inquiry space of consequential propositions rather than matters of functionality or consumption. The critical nature afforded to fiction with recourse to possible worlds is that its embodied propositions can be accepted as truthful under certain conditions. With respect to the actual world, or our world, the choice can be to regard the propositions as false under the conditions of our actual world. Or it can be seen as a critical alternative, which is to change the conditions of our actual world to make the proposition truthful. This, in essence, is the model for material speculations in interaction design research as a mode of critical inquiry. The inaccessible camera can be seen as counterfactual in that it draws its owner into a familiar device and interaction—taking a photo with a camera. However, its form and composition depart into an alternative situation in which one must destroy the digital device recording one’s life experiences in order to access these digital records. In our contemporary world of constant availability and connectedness, these counterfunctional cameras project a critical stance on ‘functionality’—one based on inhibiting, restricting or removing common or expected features of a technology. To initiate consumption of one’s digital photographs, one must first encounter the discomfort of destruction. On a broader level, encounters with the inaccessible camera invite critical reflection on one’s own practices contributing to unchecked digital content production and the almost unnoticed or assumed eventual obsolescence and disposal of everyday digital devices. Summary In summary, we can see how possible worlds theory, enabled by the work of literary theories, can be applied to interaction design research to develop the notion of material speculation. The basic outlines of material speculation can be summarized as the manifestation of a counterfactual in a material artifact that we refer to as a counterfactual artifact. As a material thing it occupies the boundary between actual and possible worlds. The counterfactual artifact is also an embodied proposition that, when encountered, generates possible world accounts to reason on its existence. These two aspects combined afford material speculations a position in critically speculating on the actual world. Examples of material speculation In what follows we provide an overview of examples of interaction design artifacts that can be read as material speculations. We aim to emphasize their actual material existence situated in everyday settings. With our examples, we focus largely on the description of the design artifacts as Figure 1. The Inaccessible Digital Camera [40]. 102 Rudiment #1 [26]— Rudiment #1 is a small machine encased in wood and plastics that affixes magnetically to surfaces, such as a refrigerator door (see figure 2). This machine consists of two main parts or ‘modules’ that are connected by a flexible cable. One part moves across a magnetic surface with magnetic wheels when it’s narrowrange infrared detector senses peripheral movement. The speed and direction are also randomly changed each time these sensors detect close by movement. The second part is wired to the first part; it provides the first module with power and also signals further movement when its own wide-range infrared detector senses movement. Both parts are able to detect when an edge or obstacle is reached and are programmed to change direction if either are encountered. The authors’ aims for designing Rudiment #1 (and also its cousins Rudiment #2 and #3) were to speculatively explore how interactive machines might exhibit autonomy, and how this might be interpreted and speculated on by people living with them. Two households in the southeast region of the United Kingdom experienced rudiment #1 for roughly four weeks each. Figure 3. The table-non-table [54]. participants to more broadly consider their relationships to other machines and objects in the home. The Rudiments effectively struck the balance between offering relatively familiar materials and formal aesthetics, while operating in unfamiliar ways that opened up new possibilities for thinking about human relations to everyday computational things. Rudiment #1 itself, along with the speculations triggered by household members’ lived-with encounters, helped develop and advance a speculative space for moving beyond the design of “machines for our own good, to a possibility of interactive machines that might exhibit autonomy, but not as we know it” [26, p. 152]. From the beginning household members struggled to make sense of Rudiment #1 when they encountered it. Interactive technologies and machines commonly occupy our everyday environments, but the combination of an unclear ‘function’ or purpose paired with unfamiliar, yet resolved aesthetics prompted a range of speculations on what Rudiment #1 is and the nature of its intelligence and autonomy. These design qualities provoked and challenged household members to encounter a world in which machines may exhibit and enact a form of autonomy that is very different from what we know and understand in our world today. This was evident in household members’ initial use of ‘pet’ metaphors in attempts to describe their relations to the Rudiments. However, participants eventually migrated to focus on qualities of function and engagement to make sense of the autonomy exhibited by the Rudiments. In one case household members perceived that Rudiment #1 could be networked to other machines within their home, and in other cases ongoing encounters with the Rudiments led Table-non-table [37, 54]—The table-non-table is a slowly moving stack of paper supported by a motorized aluminum chassis (see figure 3). The paper is common stock (similar to photocopy paper). Each sheet measures 17.5 inches by 22.5 inches with a square die cut in the middle to allow it to stack around a solid aluminum square post that holds the sheets in place. There are close to 1000 stacked sheets of paper per table-non-table, which rest on the chassis about one half-inch from the floor. The movement of the table is in short durations (5-12 seconds) that occur once during a longer period of time (a random selection between 20 to 110 minutes). The table-non-table lived with one household for five months, became part of two households for six and three weeks respectively, and became part of two households in a preliminary deployment for several days in Vancouver, British Columbia. In some ways similar to Rudiment #1, the table-non-table provoked a range of speculations as participants attempted to make sense of its purpose and place within their homes. While initially its owners attributed anthropomorphic qualities to the table-non-table (e.g., perceiving it had the abilities to ‘hide’ or ‘pretend’) [54], over time different Figure 2. Rudiment #1 vertically affixed to a surface [26]. 103 relations emerged as encounters with it accumulated. The table-non-table became an artifact that was curiously computational and clearly required electricity, yet many of the ways participants used and related to it mirrored manipulations and reconfigurations more commonly associated with non-digital things. The flat surface of the table-non-table opened it up to being subtly drawn on, at times in unknowing ways as other objects were stacked on top and it slowly became just another thing in the background of domestic life. When its movement was noticed the owners often relocated it to different locations in the house or apartment as if trying to reveal different understandings of the artifact. In other cases, the subtle yet persistent movement of the table-non-table catalyzed emergent, creative interactions by people and their pets as a way of “resourcing” the table-non-table. For example, cats alternated between using it as a bed and viewing it as another entity with either caution or curiosity. In fact, one cat began to treat a heater appliance next to the table-nontable in a similar fashion as if similarly constituted objects were now alive. Both pets and people played with the sheets of paper from ripping them. People made drawings on the paper and turned them into large snowflakes [54]. The table-non-table can be seen as radically departing from how many people experience domestic technology on an everyday basis. In this way, people, pets and their material environments were reconfigured over and over again to try to incorporate the table-non-table alongside other domestic artifacts, spaces, and experiences over time. small opening in the panel to allow a photo to drop onto the central platform of the box. The Photobox’s behavior is enacted through an application, which runs on a laptop that wirelessly connects to the embedded printer via Bluetooth. At the start of each month, Photobox indexes its owner’s Flickr archive and randomly printed four or five photos that month. In similarly random fashion, it selects four (or five) photos and generates four (or five) timestamps that specify the print time and date for each photo; at print time, the matching photo was printed. The Photobox was created to speculatively explore how an interactive artifact could critically intervene in experiences of digital overload (i.e., the proliferation of digital photos) and, more generally, what might comprise a material instantiation of the slow technology design philosophy [25]—an interactive thing whose relationship with its owner could emerge and evolve over time. Three nearly identical Photoboxes were designed and implemented; three households in Pittsburgh, Pennsylvania, USA subsequently lived with a Photobox for fourteen months respectively. The Photobox combines a recognizable form—a wooden chest—and a common experience—viewing and engaging with digital photos—within a networked design artifact that provoked both familiar and alien experiences across all households as they encountered this design artifact over time. The Photobox can be seen as counterfactual in that well-worn wooden chests do not manifest material rendering of one’s online photo collection, at least in the world as we experience it today. Nonetheless, it was perfectly functional and developed a unique character and configuration within each of the three households that owned one. In addition to stimulating reflections from its owners about the memories it surfaced from deep within their digital photo archives, its unfamiliar (and uncontrollable) slow pacing paired with its classification among participants as ‘a technology’, triggered a range of reflections about participants relations to other technologies in the home and what values ought to constitute a domestic technology. The Photobox clearly departed from any kind of familiar combination of form, materials, and computational behavior that typically characterize domestic technologies. As a result, participants speculated on the nature of the artifact and technology in everyday life, though lived-with encounters over a long period of time. These encounters opened a productive space for framing future speculative design inquires. Photobox [36]—The Photobox is a domestic technology embodied in the form of a well-worn antique chest that prints four or five randomly selected photos from the owner’s Flickr collection at random intervals each month (see figure 4). The two main components of Photobox are an oak chest and a Bluetooth-enabled Polaroid Pogo printer (which makes 2x3 inch photos). All technological components are embedded in an upper panel in the chest in an effort to hide of ‘technological’ components from view. The printer is installed in an acrylic case that secured it to a Mediated Body [28]—Mediated Body is a symbiotic system consisting of a human (“the performer”) wearing custombuilt technology (“the Suit”). The system offers a play session for a single participant (i.e. a person that is not the performer). The role of the technology is to sense physical bare-skin connection between the performer and the participant, where the sensing yields analogue values that range from a few centimeters from actual touch to light touch to full contact (see figure 5). The values are converted into a relatively complex soundscape, which is Figure 4. The photobox largely took the form of a European oak chest. [36] 104 played back in the headphones that both the performer and the participant wear. Thus, from the participant’s point of view, the performer is a musical instrument that she can play by touching. However, due to the design of the system, the instrument can also play its player: When the performer touches the participant, the soundscape is affected in the same way. The headphones make the interactive soundscape a shared experience between performer and participant, and they also serve to limit surrounding sounds and thus make the experience more intimate and private for the two players. Further, the suit includes bright lights on the performer’s chest, which serve two purposes. First, the lights enhance the interactive properties of touch by changing color and pulse when a touch is sensed. Second, they broadcast some of the interaction dynamics of the ongoing session to the surrounding area. The mediated body was encountered at the week-long Burning Man Festival and in public spaces, such as the subway in Berlin (see Figure 5). engaged in making sense of this encounter in ways that differ considerably from the performer and participant (see Figure 5). The Mediated Body speculates on many issues pertaining to the mobile experience of new media, the cultivation and expression of personal space in public places, the human body as a technical interface, and the richness and tensions entangled across all of these themes. Characteristics of material speculation Based on the related works, adaption of possible worlds theory, and the accounts of material speculation examples, we summarize our conceptual framing with a series of characteristics. Material speculation is the coupling of counterfactual artifacts and possible worlds––Material speculation is the sum of the counterfactual artifact that is designed to exist in the everyday world to be encountered and the multitude of possible worlds it generates by those encounters. Counterfactual artifacts exist in the everyday world— The counterfactual nature of material speculations rely on the contradiction of the artifact not appearing to “fit the logic of things” in the everyday world yet undeniably existing in the actual world to be encountered. Counterfactual artifacts situated in the everydayness of our world offer a new ontological perspective that over time makes more visible assumptions, implications, and possible change. It is important for the depth and quality of the emergent possibilities that material speculations be a lived experience rather than simply an intellectual reflection. More diverse and deeper possibilities are generated through cohabitation, interactions, and constant encounters over time. While different in many ways in terms of its materials, form and interactivity, the Mediated Body leverages familiar interactions (e.g., touching another human) to venture into unfamiliar territory similar to all of the prior examples. It was evident that encounters with the Mediated Body not only continually reconfigured relations between the performer and participant, but also the evolving social and material ecology encompassing these interactions. It generated encounters in which issues of social conformity became peripheral for the performer and participant in favor of direct, intimate engagements in public spaces. However, these engagements extended beyond the two people directly involved in the interaction as those around them also Counterfactual artifacts are generators of possible worlds—Counterfactual artifacts in material speculations do not embody possible worlds, rather they act as propositions that, if considered, generate lived-with engagements with new possibilities encapsulated within possible worlds. As we discussed earlier, counterfactual artifacts are machines that generate possible worlds. These include the world(s) as imagined by the designers, world(s) imagined by those who encounter the counterfactual artifact, and most speculatively, the counterfactual artifact itself can be understood to imagine a world. Counterfactual artifacts are specially crafted— Counterfactual artifacts in material speculations are specially designed artifacts. They are crafted with the intent and purpose to inquire on new possibilities. This is not a straightforward practice; it requires expertise and design judgment to create an artifact that successfully contradicts and deviates from the world around it, yet is entertained as a viable proposition in our everyday world. As evident across our examples, counterfactual artifacts are carefully shaped and designed through materials, form and computation such that the artifact is balanced between “falsely” existing in the actual world while being “true” in a possible world. Figure 5. Top: Mediated Body [28] in use at the Burning Man festival. Bottom: The Mediated Body performed on a Berlin subway among many onlookers. 105 Material speculation is critical inquiry––Counterfactual artifacts by nature challenge the actual world since they are designed to occupy the boundary between the actual and the possible. The criticality of a material speculation can arise from the quantity of possible worlds it opens up or the quality in which it suggests fewer possible worlds. In either case this speaks to the nature of the critical space revealed. The precision and promise of the critical inquiry is mediated through the crafting of the counterfactual artifact, which can be shaped and directed toward critical or needed spaces of inquiry. the position that the things bare knowledge in distinct and complex ways. While there are important differences in their own epistemological commitments, emerging theoretical notions such as Bogost’s [11] carpentry— constructing artifacts that do philosophy, where their real and lived existence embodies intellectual argumentation— and Baird’s [6] thing knowledge—artifacts can embody and carry knowledge prior to our ability to theorize or reason through language—offer intriguing perspectives that can be seen both as critical and generative mechanisms framing how we do and how we unpack approaches like material speculation within the interaction design community. DISCUSSION In our discussion we review our contributions and the importance for HCI and interaction design research of the centrality of the actual and situated artifact. We also consider the potential future of speculative materiality as part of everyday practices. Speculative materiality of everyday practices If critical speculative research in design shifts the emphasis of interpretation to the design audience or users, might we expect this same audience to go beyond interpretation and appropriation to the active engagement in crafting speculation themselves? Similar concerns have been explored previously in considerations of the intersection of social practices and design fiction. For example, Tanenbaum et al. [53] looked at the practice of Steampunk with the lens of design fiction. They discussed how Steampunk enthusiasts manifest both making and fictional practice that serves as an exemplar of the meeting of social practices and speculative inquiries through design. The authors argue that this intersection of fiction with materiality through making verges on making a possible world actual. Despite the impossibility of the Steampunk world becoming real it forms an ongoing social practice of non-designers centered on making such impossibility a reality. Wakkary et al. [55] explore the relation between speculation and material practices in a different manner. These authors describe how speculative inquiries in sustainable design like hydroponic kitchen gardens of the future or vertical gardens enabled a ‘generative approach.’ These speculations are taken up within the existing competences and materials of Green-DIY enthusiasts and realized today through, for example, Ikea-hacks for the hydroponic kitchen and reclaimed truck pallets turned into vertical gardens. Here, the meeting of the speculative and the material takes an everyday turn that is distinct but related to our discussion of material speculative inquiries for design-oriented research. Actual and situated artifacts as knowledge A core goal of this paper is to complement the nascent and growing interest in design fiction in HCI and interaction design research through offering an alternative conceptual framing for critical and speculative inquiries in design. We aimed to expand on the criticality that design fiction brought to speculation in design research by motivating and developing the role that actual design artifacts can play in critical inquiries. Further, we sought to build on the traditions of craft and material work of speculative and critical design, as well as the shifting of responsibility for interpretation to users by including situatedness and lived with experience as sites for critical inquiry. A further contribution we made was to provide an in depth theoretical account to nurture the development of material speculation within HCI. We aimed to be as clear and transparent in our use of theory alongside the material speculation concept such that other researchers can not only refine and revise our concept, but also refine and revise our theoretical work, which may in turn lead to other insights. On a disciplinary level, fundamental to the concept of material speculation is the centrality of the actual and situated artifact as a producer of knowledge in interaction design research. This articulation can be seen to support the increasing turn within interaction design and HCI research to develop forms of knowledge production centered on the essential role of designed artifacts. This advances the notion of making ‘things’ as a site of inquiry that produces insights, theories, and argumentation that is unique to interaction design research and distinct from critical art practice and the humanities [5, 51, 30]. We see our work as contributing to nascent and growing interest in design artifacts as generators of knowledge (e.g. annotated portfolios [12], critical making [41] and adversarial design [16], among others). CONCLUSION This paper has motivated and articulated material speculation as a conceptual framing to further support critical and speculative inquires within HCI and interaction design research. In this, we have reviewed and synthesized a theoretical account of possible worlds theory and the counterfactual, described and interpreted a set of examples of material speculations, and proposed characteristics of material speculations. Importantly our aim is to not be prescriptive nor conclusive, rather we intend to provide a conceptual framing to inspire future generative work at intersection of critical and speculative inquires targeted at More broadly, our work parallels movements emerging outside of HCI and interaction design that critically advance 106 the everyday in HCI. We concluded with actual and situated artifacts as knowledge and speculative materiality of everyday practices as opportunity areas for framing future contributions of material speculations in HCI and design. In our future work, we aim to refine and expand these concepts, both materially and theoretically. As the HCI community continues to seek out critical alternatives for exploring the nature of interactive technology in everyday life, we hope material speculation can be seen as a complementary framing for supporting these initiatives and, more broadly, the need to recognize and develop ways of practicing more reflective forms of knowledge production. 10.Blythe, M. Research through design fiction: narrative in real and imaginary abstracts. In Proc. CHI 2014, ACM Press (2014), 703-712. 11.Bogost, I. Alien Phenomenology, or What It’s Like to Be a Thing. University of Minnesota Press, Minneapolis, MN, 2012. 12.Bowers, J. The Logic of Annotated Portfolios: Communicating the Value of ‘Research Through Design.’ In Proc. DIS 2012, ACM Press (2012), 68–77. 13.Brown, N., B. Rappert, and A. Webster (eds). Contested Futures: a sociology of prospective technoscience. Ashgate, Surrey, UK, 2000. ACKNOWLEDGMENTS 14.Carroll J. M. 1997. Scenario-based design. In Handbook of Human-Computer Interaction. Helander M., Landauer T.K., and Prabhu P. (eds). Elsevier, Amsterdam, NL, 383-406 The Social Sciences and Humanities Research Council of Canada (SSHRC), Natural Sciences and Engineering Research Council of Canada (NSERC), Banting Postdoctoral Fellowships, and Canada Research Chairs supported this research. We thank the many authors who gave us permission to reprint their work and images. 15.Cooper, A. The inmates are running the asylum. Macmillan Publishing Company Inc., Indianapolis, IN, 1999. REFERENCES 16.DiSalvo, C. Adversarial Design. The MIT Press, Cambridge, MA, 2012. 1. Agre, P.E. 1997. Toward a critical technical practice: lessons learned in trying to reform AI. In Bridging the Great Divide: Social Science, Technical Systems, and Cooperative Work, Bowker, G. et al. (eds). Erlbaum, Mahwah, NJ, 131-157. 17.Dourish, P., and Bell, G. Resistance is futile: reading science fiction alongside ubiquitous computing. Personal and Ubiquitous Computing, 18, 4 (2014), 769-778. 2. Auger, J. Speculative design: crafting the speculation. Digital Creativity, 24, 1 (2013), 11-35. 18.Dunne, A., and Raby, F. Speculative everything: design, fiction, and social dreaming. MIT Press, Cambridge, MA, 2013. 3. Bardzell, J. and Bardzell, S. A great and troubling beauty: cognitive speculation and ubiquitous computing. Personal and ubiquitous computing, 18, 4 (2014), 779-794. 19.Dunne, A., and Raby, F. Design noir: The secret life of electronic objects. Springer Press, New York, NY, 2001. 4. Bardzell, J. and Bardzell, S. What is "critical" about critical design? In Proc. CHI 2013, ACM Press (2013), 3297-3306. 20.Dunne, A. Hertzian tales: Electronic products, aesthetic experience, and critical design. MIT Press, Cambridge, MA, 1999. 5. Bardzell, S., Bardzell, J., Forlizzi, J., Zimmerman, J. and Antanitis, J. Critical design and critical theory: the challenge of designing for provocation. In Proc. DIS 2012, ACM Press (2012), 288-297. 21.Eco, U. and Robey, D. The Open Work. Harvard University Press, Cambridge, MA, 1989. 22.Eco, Umberto. The Role of the Reader: Explorations in the Semiotics of Texts. Bloomington: Indiana University Press, Bloomington, IN, 1984. 6. Baird, D. Thing Knowledge: A Philosophy of Scientific Instruments. University of California Press, Berkeley, CA, 2004. 23.Gaver, W., Blythe, M., Boucher, A., Jarvis, N., Bowers, J., and Wright, P. The prayer companion: openness and specificity, materiality and spirituality. In Proc. CHI 2010, ACM Press (2010), 2055-2064. 7. Barthes, R. and Howard, R. S/Z: An Essay. Hill and Wang, New York, NY, 1975. 8. Bell, G., and Dourish, P. Yesterday’s tomorrows: notes on ubiquitous computing’s dominant vision. Personal and Ubiquitous Computing, 11, 2 (2007), 133-143. 24.Gaver, B., and Martin, H. Alternatives: exploring information appliances through conceptual design proposals. In Proc. CHI 2000, ACM Press (2000), 209216. 9. Bleecker, J. Design fiction: A short essay on design, science, fact and fiction. 2009. Retrieved June 16, 2015 from http://blog.nearfuturelaboratory.com/2009/03/17/designfiction-a-short-essay-on-design-science-fact-and-fiction/ 25.Hallnäs, L. and Redström, J. Slow Technology – Designing for Reflection. Personal and Ubiquitous Computing, 5, 3 (2001), 201–212. 107 26.Helmes, J., Taylor, A.S., Cao, X., Höök, K., Schmitt, P., and Villar, N. Rudiments 1, 2 & 3: design speculations on autonomy. In Proc. TEI 2011, ACM Press (2011), 145-152. 40.Pierce, J., and Paulos, E. Counterfunctional things: exploring possibilities in designing digital limitations. In Proc. CHI 2014, ACM Press (2014), 375-384. 41.Ratto, M. Critical Making: Conceptual and Material Studies in Technology and Social Life. The Information Society: An International Journal, 27, 4, (2011), 252260. 27.Hertz, G. Critical Making: Manifestos. Telharmonium, Hollywood, CA, 2012. 28.Hobye, M., and Löwgren, J. Touching a stranger: Designing for engaging experience in embodied interaction. International Journal of Design, 5, 3 (2011), 31-48. 42.Reeves, S. Envisioning ubiquitous computing. In Proc. CHI 2012, ACM Press (2012), 1573-1582. 43.Retzinger, J. P. Speculative visions and imaginary meals. Cultural Studies 22, 3–4, (2008), 369–390. 29.Kirby, D. The Future Is Now: Diegetic Prototypes and the Role of Popular Films in Generating Real-World Technological Development. Social Studies of Science 40, 1 (2010), 41-70. 44.Ronen, R. Possible worlds in literary theory (Vol. 7). Cambridge University Press, Cambridge, UK, 1994. 45.Ryan, M.-L. The Modal Structure of Narrative Universes. Poetics Today 6, 4, (1985), 717–56. 30.Koskinen, I., Zimmerman, J., Binder, T., Redstrom, J., & Wensveen, S. Design research through practice: From the lab, field, and showroom. Elsevier, Amsterdam, NL, 2011. 46.Ryan, M.-L. Possible Worlds, Artificial Intelligence and Narrative Theory. University of Indiana Press, Bloomington, IN, 1991. 31.Kripke, S.A. Semantical Considerations on Modal Logic. Acta Philosophica Fennica 16, 1963 (1963), 83– 94. 47.Ryan, M.-L. Possible-Worlds Theory. In Routledge Encyclopedia of Narrative Theory. Herman, D et al (eds). Routledge, London, UK, 2010, 446–450. 32.Lewis, D. Truth in fiction. American Philosophical Quarterly, 15, 1 (1978), 37-46. 48.Ryan, M.-L. Possible Worlds. In the living handbook of narratology. Hühn, Peter et al. (eds.), 2012. Retrieved June 16, 2015 from http://www.lhn.unihamburg.de/article/possible-worlds 33.Lewis, D. K. On the plurality of worlds. Blackwell, Oxford, UK, 1986. 34.Menzel, C., Possible Worlds. In The Stanford Encyclopedia of Philosophy (Spring 2015 Edition), Zalta, E. (ed.), 2015. Retrieved June 16, 2015 from http://plato.stanford.edu/archives/spr2015/entries/possib le-worlds/ 49.Sengers, P., and Gaver, B. Staying open to interpretation: engaging multiple meanings in design and evaluation. In Proc. DIS 2006, ACM Press (2006), 99-108. 50.Sengers, P., Boehner, K., David, S., and Kaye, J. J. Reflective design. In Proc. CC 2005, ACM Press (2005), ACM Press, 49-58. 35.Murray, J. H. Hamlet on the holodeck: The future of narrative in cyberspace. Simon and Schuster, New York, NY, 1997. 51.Stolterman, E. and Wiberg, M. Concept-Driven Interaction Design Research. Human-Computer Interaction 25, 2 (2010), 95–118. 36.Odom, W. T., Sellen, A. J., Banks, R., Kirk, D. S., Regan, T., Selby, M., Forlizzi, J., and Zimmerman, J. Designing for slowness, anticipation and re-visitation: a long term field study of the photobox. In Proc. CHI 2014, ACM Press (2014), 1961-1970. 52.Sturken, M., Thomas, D., And Ball-Rokeach, S. Technological Visions: Hopes and Fears That Shape New Technologies. Temple University Press, Philadelphia, PA, 2004. 37.Odom, W. T. and Wakkary, R., Intersecting with Unaware Objects. In Proc. C&C 2015, ACM Press (2015), in press 53.Tanenbaum, J., Tanenbaum, K., and Wakkary, R. Steampunk as design fiction. In Proc. CHI 2012, ACM Press (2012), 1583-1592. 38.Pavel, T. Possible Worlds in Literary Semantics. Journal of Aesthetics and Art Criticism 34, 2, (1975), 165–76. 54.Wakkary, R., Desjardins, A., Hauser, S. Unselfconcious Interaction: A Conceptual Construct. Interacting with Computers, (2015), (in press). 39.Pierce J., and Paulos, E. Making multiple uses of the obscura 1C digital camera: reflecting on the design, production, packaging and distribution of a counterfunctional device. In Proc. CHI 2015, ACM Press (2014), 2103-2112. 55.Wakkary, R., Desjardins, A., Hauser, S., and Maestri, L. A sustainable design fiction: Green practices. ACM Transactions on Computer-Human Interaction 20, 4, (2013). Article No. 23 108 Charismatic Technology Morgan G. Ames Intel Science and Technology Center for Social Computing University of California, Irvine charisma@morganya.org relation to the charisma of past technologies. This historical perspective highlights the ideological commonalities between all of these charismatic objects. It also suggests that far from being a new phenomenon, charismatic technologies have been captivating their users and admirers for decades, and will likely continue to do so for decades to come. We will see that charismatic technologies help establish and reinforce the ideological underpinnings of the status quo through utopian promises [39] – promises that persist even when the technology does not deliver. ABSTRACT To explain the uncanny holding power that some technologies seem to have, this paper presents a theory of charisma as attached to technology. It uses the One Laptop per Child project as a case study for exploring the features, benefits, and pitfalls of charisma. It then contextualizes OLPC’s charismatic power in the historical arc of other charismatic technologies, highlighting the enduring nature of charisma and the common themes on which the charisma of a century of technological progress rests. In closing, it discusses how scholars and practitioners in human-computer interaction might use the concept of charismatic technology in their own work. The goal of this paper is to expose the ideological stakes that buttress charismatic technologies. Those who create, study, or work with technology ignore the origins of charisma at their own peril – at the risk of always being blinded by the next best thing, with little concept of the larger cultural context that technology operates within and little hope for long-term change. Recognizing and critically examining charisma can help us understand the effects it can have and then, if we choose, to counter them. However, it is also important to acknowledge that charisma can smooth away uncertainties and help us handle contradictions and obstacles. As such, the purpose of this paper is not to ‘prove’ charisma ‘wrong,’ because its rightness or wrongness is beside the point. As we will see, what matters is whether a technology’s charisma is still alive. Author Keywords Charisma, childhood, education, history of technology, ideology, One Laptop per Child, religion, science and technology studies, technological determinism, utopianism. ACM Classification Keywords K.4.0. Computers in Society: general. INTRODUCTION Scholars have noted the holding power that some technologies seem to have – a power that goes beyond mere form or function to stimulate devotion, yearning, even fanaticism [2,39,44,62]. While Apple products, especially iPhones, are often held up as the most common example of this holding power [11,31,37,40,52,56], it exists in various forms for many technologies, from sports cars to strollers. This paper provides a framework for understanding how charisma operates in relation to technologies, how it might be identified, and what is at stake when we are drawn in. It provides tools for identifying charismatic technologies and teasing out the implications of this charisma, from the hardware and software of the object itself to the ensembles, agendas, and trajectories of the globalized organizations around it [18], and back down to the ways that those same groups shift, contest, or perpetuate the object’s charisma. This borrows from Actor-Network Theory the idea that nonhuman “actors” have agency in technosocial discourses [33], and from Value-Sensitive Design a normative examination of the ways in which the myriad values influence design and use [19]. This analysis adds to these theories a detailed case study of the role that charismatic authority plays in the design and use of technologies, digging beneath professed values to identify the ideological underpinnings upon which values, and charisma, rest. This paper describes this holding power as charisma. Applying Weber’s theory of charismatic authority [68] to objects, it presents a case study of a technology that was highly charismatic to its makers and supporters (and remains so to a devoted core): the One Laptop per Child (OLPC) project’s “XO” laptop. With about two and a half million in use globally, OLPC’s green-and-white XO remains a focal point for discourses about children, technology, and education, even a decade after its 2005 debut. This analysis explores the roots of the laptop’s charisma and the important role that charisma played in OLPC’s heady early days. It then reflects on the charismatic elements present in this project in Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21199 109 network of relations [33]. ANT, however, tends to neglect the beliefs that underlie these networks, especially if these beliefs do not take material form [34:2]. Thus, while ANT provides tools for analyzing the ‘scripts’ that designers build into technologies [1], it falls short of providing the means to account for how ideological frameworks animate or inhabit these products – a gap charisma can fill. METHODS AND SOCIAL-HISTORICAL ORIENTATION This paper draws on archival research, interviews, and ethnographic observations conducted between 2008 and 2015. This includes an investigation of the forty-year development of the ideas behind One Laptop per Child (OLPC) through a review of the project’s mailing lists, wikis, and discussion boards; the publication history of its founders; and interviews with some developers. The author also conducted seven months of fieldwork of an OLPC project in Latin America (see [5,6,53]), but this data is not directly included here. Analysis followed an iterative, inductive approach common in anthropology and cultural studies, combining the themes that emerge ground-up from a thorough understanding of participants’ worldviews with a critical interpretation of these themes as ‘texts’ able to expand or contest broader theoretical questions [10]. Distinct from fetishism’s focus on form, which stipulates a fixation on the materiality of the (presumably) passive object itself as a source of power, a charismatic object derives its power experientially and symbolically through the possibility or promise of action: what is important is not what the object is but what it promises to do. Thus, the material form of a charismatic technology is less important than how it invokes the imagination. As sociologist Donald McIntosh [36] explains, “charisma is not so much a quality as an experience. The charismatic object or person is experienced as possessed by and transmitting an uncanny and compelling force.” A charismatic technology’s promises are likewise uncannily compelling, evoking feelings of awe, transcendence, and connection to a greater purpose [39,44]. This paper contextualizes the patterns noted in OLPC’s rhetoric and design within the broader arc of technological development, as told by historians of technology. The combination of historical and contemporary data lends itself to reaching beyond the often bounded scope of qualitative research to answer more long-ranging questions about the trajectory of technological development and use. Charisma moreover implies a persistence of this compelling force even when an object’s actions do not match its promises – hence the magical element of charisma. This is where a charismatic technology’s link to religious experience is especially strong, as a system built on faith that is maintained and strengthened outside of (or even counter to) the auspices of ‘evidence’ [2]. This is also one of the places that the consequences of charisma are most visible, as a technology’s devotees maintain their devotion even in the face of contradictions. THEORIZING CHARISMATIC TECHNOLOGIES To explain the holding power that OLPC’s laptop has had on technologists and others around the world, I develop the idea of a charismatic technology. This section defines charisma, outlines the salient features of charismatic technologies, and details the connection between charisma and related concepts from social theory including fetishism, religion, technological determinism, and ideology. This potential for charisma to override ‘rational’ thought is something that is not lost on marketers. While they may not call it as much, their promotions often tap into the charisma that certain technologies have, and journalists often echo it as well. In fact, religion scholar William Stahl examines the popular discourses about technology and concludes that “our language about technology is implicitly religious” [56:3]. Other scholars have also documented the connections between technology and religion, highlighting the prevalence of religious language in branding discourse [37,40,52,56], revealing parallels between religious and engineering practice [2], or showing how specific technologies are rhetorically connected to divinity and redemption in use [11,28,31,39]. Thus, while charismatic objects construct and reinforce their own charisma, media are often implicated in amplifying it, as we will see below. Charisma as a sociological construct was theorized by Max Weber to describe the exceptional, even magical, authority that religious leaders seem to have over followers. In contrast to (though sometimes subsumed by) other types of authority, such as traditional or legal/rational, charismatic authority is legitimized by followers through a belief that a leader has extraordinary, even divine, powers that are not available to ordinary people [68]. Though charisma is generally used to describe the power of humans, not objects, it has been applied to nonhumans as well. Maria Stavrinaki, for instance, describes the Bauhaus African Chair as a ‘charismatic object’ within the Bauhaus community [57]. Relatedly, Anna Tsing [59] discusses how the idea of globalization has been charismatic to some academics, who uncritically naturalize or even reinforce globalist agendas by characterizing globalization as universal and inevitable. Tsing’s model of charisma – a destabilizing force that both elicits excitement and produces material effects in the world (even if these effects differ from those that were promised) – is at play here as well. In their often utopian promises of action, charismatic technologies are deceptive: they make both technological adoption and social change appear straightforward instead of a difficult process fraught with choices and politics. This gives charismatic technologies a determinist spirit, where technological progress appears natural, even inevitable. This naturalizing element can lead us to underestimate the sustained commitment needed for technological adoption. In By treating an object as charismatic, these approaches, and this paper as well, utilize perspectives from actor-network theory (ANT), which subject both human and non-human ‘actors’ to the same analytical lens in a mutually-constituted 110 building out the railroad in the mid-nineteenth century, for instance, Nye shows that the charisma of the locomotive led to the U.S. paying an enormous price in resources and lives in an attempt to realize the utopian promises of rail [44]. By the same token, charisma’s naturalizing force can make critique and debate appear unnatural. operates within and little hope for long-term change. Recognizing and critically examining charisma might, as Mosco explains, “help us to loosen the powerful grip of myths about the future on the present” [39:15]. Still, it is also important to recognize that charisma plays an important, even indispensable, role in our lives. Charisma – whether from leaders or technologies – can provide direction and conviction, smoothing away uncertainties and helping us handle contradictions and adversities. Rob Kling asserts that faith in technology can play a major role in cultural cohesion: “During the 1930s,” he explains, “this almost blind faith in the power of the machine to make the world a better place helped hold a badly shattered nation together” [30:48]. As such, the purpose of this paper is not to ‘prove’ charismatic technologies ‘wrong,’ because matters here is whether charisma is still ideologically resonant. However, charisma contains an irony. A charismatic technology may promise to transform its users’ sociotechnical existence for the better, but it is, at heart, fundamentally conservative. Just as charismatic leaders confirm and amplify their audiences’ existing worldviews to cultivate their appeal [36], a charismatic technology’s appeal is built on existing systems of meaning-making and largely confirms the value of existing stereotypes, institutions, and power relations. This unchallenging familiarity is what makes a charismatic technology alluring to its target audience: even as it promises certain benefits, it simultaneously confirms that the worldview of its audience is already ‘right’ and that, moreover, they are even savvier to have this technology bolster it. THE CHARISMA OF ONE LAPTOP PER CHILD OLPC’s laptop, called the ‘XO,’ was the first of its kind to combine a rugged design, an open-source educational software suite, and full (though purposefully underpowered, in an attempt to prolong battery life) computer functionality, with the goal of overhauling education across the Global South. The project is a culmination of over forty years of work at the MIT Media Lab and its predecessors, particularly the intellectual legacies of MIT professors Seymour Papert and Nicholas Negroponte. While many were involved in OLPC, Negroponte and Papert together were largely responsible for establishing the direction of the project. This worldview that charisma reinforces is what social theorists refer to as an ideology: a framework of norms that shape thoughts and actions. Cultural theorist Stuart Hall describes an ideology as a “system for coding reality” that “becomes autonomous in relation to the consciousness or intention of its agents” [22:71]. We live with many ideologies, reinforced not just by charisma but across our socio-political landscape: neoliberal economics, patriarchal social structures, and Judeo-Christian ethics are among the many dominant ideologies in the United States, for instance. Both Negroponte and Papert are themselves charismatic, and both used it to build the charisma of the OLPC project and the XO laptop. While Negroponte has been the public face of the project, glibly flinging XOs across stages at world summits to demonstrate their ruggedness [64] and talking about “helicopter deployments” of laptops to remote areas [66], Papert was the project’s intellectual father. His whole career focused on the idea of computers for children, leading to the development of LOGO, Turtle Graphics, Lego Mindstorms, and, finally, One Laptop per Child. Though the results of these projects have been lackluster at best [4,6,51,55,67], Papert is still often considered a central figure in education and design (e.g. [8,20]), and his books remain foundational to the curriculum at the MIT Media Lab [2]. Hall notes that ‘ideology’ has been useful in social theory as “a way of representing the order of things which endowed its limited perspectives with that natural or divine inevitability which makes them appear universal, natural and coterminous with ‘reality’ itself” [22:65]. What is important in this definition is the way that ideology fades into the background: by one metaphor commonly used in anthropology, ideologies are as invisible to most people as we imagine water is to a fish [15:123]. In Hall’s words, an ideology “works” because it “represses any recognition of the contingency of the historical conditions on which all social relations depend. It represents them, instead, as outside of history: unchangeable, inevitable and natural” [22:76]. By referencing and reinforcing existing ideological norms, charismatic technologies by extension can also appear ‘unchangeable, inevitable and natural.’ While the charisma of these two men were important to OLPC, I also want to emphasize that OLPC’s XO laptop itself included a host of charismatic features that were eagerly discussed in the tech community and media alike, even when some never actually existed and others did not work in practice. In this way, the machine itself embodied and performed its charisma, and the discussion around the machine amplified and perpetuated these promises. Thus, charismatic technologies help establish and reinforce the ideological underpinnings of the status quo. They do so through promises that may persist among true believers even when the technology does not deliver. The task of this paper, then, is to “break the spell of the present” [56], exposing the ideological stakes that underpin charisma. Technologists ignore the origins of charisma at their own peril – at the risk of always being blinded by the next best thing, with little concept of the larger cultural context that technology One of the most charismatic features of the initial proposal for the laptop was a hand crank for manually charging it. Though this feature never existed in a working prototype and all laptops in use were charged via standard AC adapter [53], journalists still mention the hand crank or make claims based 111 And yet, none of these niggling realities mattered. For several years after its debut and even today in the popular press, the project and its charismatic green machine were the darlings of the technology world. High-profile technology companies including Google, Sun, Red Hat, and others donated millions of dollars and hundreds of hours of developer labor to the project from 2005 to 2008. The open-source community, wooed by the promise of a generation of children raised on free software, continue to enthusiastically contribute to the development of OLPC’s custom-built learning platform, Sugar. Unsubstantiated and quite likely apocryphal stories about children teaching parents to read, of impromptu laptopenabled classrooms under trees, or of the laptop’s screen being the only light source in a village (never mind how the laptop itself was recharged) were echoed enthusiastically across the media and tech worlds. Even today, with ample evidence that OLPC has not lived up to these promises, they are still repeated, and XO laptops have been added to the collections of major art museums. These hyperbolic claims themselves, whether based on design features or more general claims about how the laptop would change the world, became part of the machine’s charisma. The project’s leaders, like many who lead development projects, had the utopian, if colonialist [6,17,27], desire to transform the world – not only for what they believed would be for the better but, as we will see, in their own image. Negroponte infamously said that the project and laptop “is probably the only hope. I don’t want to place too much on OLPC, but if I really had to look at how to eliminate poverty, create peace, and work on the environment, I can't think of a better way to do it” [63]. Though Negroponte made an easy target by often saying outrageous things that sometimes left even others on the project shaking their heads or backpedaling, many on the project echoed variations on this particular theme. Figure 1. The XO in standard laptop configuration, showing the first-generation Sugar interface (promotional picture). on its presumed existence. Two antennae “ears” on either side of the screen, which also acted as latches and covers for ports, were added to visually replace it (Figure 1). They also served to anthropomorphize the laptop, as did the person-like “XO” name/logo which, when turned on its side, was meant to look like a child with arms flung wide in excitement. The XO included a “view source” keyboard button, a feature hailed as revolutionary even though web browsers had long included a similar function. Moreover, the button often did not work and I never saw it used in months of observations. The laptop’s hardware was meant to be similarly accessible, and OLPC leaders boasted about the laptop being particularly easy to repair. However, compromises with manufacturers and cost-cutting made this promise unattainable as well, and projects were plagued with breakage [4,53]. This raises the question, though, of why these worlds so enthusiastically accepted these claims. Here is where the charismatic authority of the XO laptop becomes important. Again, my goal here is not to ‘prove’ OLPC’s charisma ‘wrong’: highlighting the project’s continued charm despite its many failures is simply a method for demonstrating that charisma is indeed at play. In the next three sections, I will show how this charisma was based on specific ideas about childhood, school, and computers. As I describe in more detail in [3], OLPC hoped to replace a school experience that Papert claimed had ‘no intrinsic value’ [49:24] with a computer to encourage children to learn to think mathematically [48]. This was aimed at elementary school students who were assumed to be precocious, scientificallyinclined, and oppositional. Not coincidentally, this target audience matched developers’ conceptions of their own childhoods. In the following sections, I will briefly describe each of these and why they were important to the charisma of OLPC’s XO laptop. Then I will show that these same themes undergird a number of other historically charismatic technologies, demonstrating charisma’s conservatism. The laptop had an innovative screen, invented by fellow MIT Media Lab professor and founding OLPC contributor Mary Lou Jepsen, that could be swiveled and flattened, and its back light turned off, like an e-book – though I found that the screen was also the second-most-common component to break and the second-most-expensive to replace [53]. The first-generation XO included a mesh network so that laptops could connect to one another directly without an access point, though in practice the underpowered laptops would grind to a halt if more than a few were connected and the feature was dropped from later versions of the laptop’s software [53,65]. Even the original name of the project – the “hundred-dollar laptop” – was never achieved: the lowest price was $188 [6,53]. This in part was because OLPC’s goal of having hundreds of millions of laptops in use across the Global South also never came to fruition – instead, about 2.5 million have been sold, most of them in Latin America [6]. 112 The charisma of childhood group that has managed to hold onto the magic of childhood: independent thinkers, lifelong learners, and, not coincidentally, many in the hacker community. They generalized from experiences with largely white, middleclass American youth – or from their own idiosyncratic childhoods – that all, most, or at least the most ‘intellectually interesting’ [49:44–50] children are innately drawn to tinkering with computers and electronics, or in Papert’s words, ‘thinking like a machine’ [48]. One Laptop per Child explicitly stated and implicitly built into their laptop the idea that children are born curious and only need a small impetus (such as a computer or an electronics kit) to keep that curiosity alive and growing. OLPC moreover specified the kinds of learning that children are naturally inclined to do: engineering-oriented tinkering. In this section, we will briefly consider the cultural narratives that Papert, Negroponte, and others draw on about what childhood should mean and what constitutes a good one, narratives that have become deeply rooted in middle-class American culture and reflect American cultural values such as individualism and (certain kinds of) creativity. The anti-charisma of school The One Laptop Per Child project frames itself and its laptop as a radical break from an unchanging global educational tradition [7,43,48–50]. In contrast to the naturally curious state of children, OLPC leaders paint school as a stultifying rote experience that has not changed in over a century [42,43,48,49]. In his writing, Papert villanizes what he calls ‘instructionism,’ or lecture-heavy, curriculum-based education. Papert argues that instructionism creates people ‘schooled’ to think in limited ways and seek validation from others out of ‘yearners,’ the innate, creative, playful state of children, focused on thinking independently, not caring what others think, and seeking answers via many routes for questions in which they are personally interested [49:1]. This ideology of childhood, though seemingly universal, is historically, geographically, and socioeconomically situated. Childhood came to be understood as a distinct developmental state in nineteenth-century Europe and America [12,26,38,70,71], an ideological shift that was justified with Romantic-era ideals of childhood as a noble state closer to nature. These ideologies of childhood spread to mainstream middle-class parenting culture by the mid-twentieth century [26,32,70], when the idea of nurturing childhood creativity, play, and individualism through consumerism (via large numbers of aspirational toys) gained popularity among middle-class white American families [45,46]. Papert is unequivocal about his disdain for school, calling the classroom ‘an artificial and inefficient learning environment’ [48:8–9] with ‘no intrinsic value,’ its purpose molding children out of their natural state and into a more socially ‘desirable’ form [49:24]. With the exception of a few brave teachers who fight against the establishment [49:3], Papert’s description of a monolithic ‘School’ (always capitalized) is of an institution unchanged for over a century, ‘out of touch with contemporary life,’ and shamelessly ‘impos[ing] a single way of knowing on everyone’ [49:1–6]. Around this time, toy manufacturers began to reinforce the idea that engineering was a space for natural masculine creativity through construction toys aimed at boys [16,45,46], even though those patterns of play were relatively new and far from universal [12]. Alongside these shifts, ‘healthy rebellion’ also became an accepted part of American boy culture [38,46]. Far from the more ideologically intimidating rebellion that threatens to actually change the status quo, the rebellion that is sanctioned in boyhood is often tolerated as ‘boys will be boys’ or even encouraged as free-thinking individualism. From Mark Twain’s Tom Sawyer to today, popular culture has linked these relatively harmless forms of rebellion against school and society with creative confidence, driven by ‘naturally’ oppositional masculine sensibilities. Papert is not alone in the expressing scorn for education. Many of OLPC’s contributors, whether affiliated with MIT (including MIT professors) or the open-source software community, describe similar sentiments (e.g. [9,25,41]). Even though not all people in the OLPC community actually rejected school (and many, in fact, excelled), they published and shared narratives about how boring, stifling, and unfulfilling classroom education was. The charisma of OLPC’s laptop references this ideology of childhood. As I show in [3], though the XO was ostensibly for children of any gender, its design echoed tropes of boyhood specifically: its mesh network embodied a resistance to authority and its programming-focused applications echoed the engineering toys that companies had been marketing to boys for almost a century before [46]. Papert’s and Negroponte’s writing also often praised iconoclastic or free-thinking children – generally boys – who took to computers (and to their experiments) easily. In the process, they glossed over the many complexities of childhood – not to mention how things like household instability or food insecurity affect learning – by universalizing children as natural ‘yearners’ [49]. These narratives resonated in the technology community as well as across American culture more broadly, where it is common, and even encouraged, to disparage public education and recount tales of terrible teachers (while excellent ones are often forgotten, and the role of school as a social leveler or cultural enricher are similarly unmentioned). In this way, the anti-charisma of school has become a common cultural trope, and we will critically examine it later in this paper. For OLPC, it served two purposes. First, it aligned the project with this broader backlash against public school. Second, it provided a rhetorical foil to ideologies of childhood – an opportunity to reinforce the importance of individualism, (technically-inclined) play, and rebellion important to the idea of childhood OLPC relies on. These actors also referenced these ideologies of childhood in discussing themselves. In his writing, for example, Papert describes himself and others like him as part of a rarefied 113 The charisma of computers childhoods to other children around the world [3]. Papert, for one, explicitly credits his own formative experiences with computers as inspiration for the project. “I realized that children might be able to enjoy the same advantages” with computers as himself and other MIT hackers, Papert explains in his book The Children’s Machine – “a thought that changed my life” [49:13]. As an alternative to school, Papert proposes giving each child a ‘Knowledge Machine,’ a clear forerunner to OLPC’s XO laptop. Indeed, many who have watched a child with a touchscreen or a videogame have marveled at how children seem to be naturally enamored with technology. Stories abound of precocious young children, especially boys, who seem to take to electronics fearlessly and naturally [41,48,49]. Both Papert and Negroponte rhapsodize about the ‘holding power’ that computers can have. Papert says, In conversations with OLPC developers and those who identify as ‘hackers’ across Silicon Valley and in Boston, I often encountered the story of ‘teaching myself to program.’ These actors describe learning about computers as something they did on their own, driven by feelings of passion and freedom, and independent of any formal instruction [24,25,41,58] – just as OLPC’s intended beneficiaries are meant to do. Shunning school in favor of this libertarian learning model, these people saw their excitement kindled and powers developed from their personal relationship with computers, with hours spent tinkering with Commodore 64s or Apple IIs, chatting with others in BBS’s or Usenet groups. The computer is the Proteus of machines. Its essence is its universality, its power to simulate. Because it can take on a thousand forms and can serve a thousand functions, it can appeal to a thousand tastes. [48:viii] The stories OLPC tells about the potentials of its charismatic laptop roll into one package many of the promises connected to computers and cyberspace: of newfound freedoms and potentials for computer-based self-governance, of the inversion of traditional social institutions (putting, of course, the computer-savvy like themselves at the top), of the “flattening” of bureaucracies, of the end of geographic inequity, of a “reboot” of history [30,39,44,54,60,61,69]. Even though computerization has largely entrenched existing power structures, these ideals of the computer age live on, especially among technologists whose cocooned middle-class existence may not bring them in contact with those whose lives and livelihoods have not benefited from these technologies or who have been actively excluded from the wealth in the technology world. But as I dug deeper into how this self-learning worked, I found that in all cases they benefited from many oftenunacknowledged resources. This included a stable home environment that supported creative (even rebellious) play, middle-class resources and cultural expectations, and often (though not always) a father who was a computer programmer or engineer. Moreover, many also acknowledged, sometimes readily, that they were not typical among their peers in their youth, and were often shunned for their unusual interests or obsessions. Even among peers who also had computers at home they were often unique in their interest in learning to ‘think like a machine,’ while many others with the same of access were not nearly as captivated by computers. The connection between rebellion and computing cultures is also well-established, particularly as countercultural norms of the 1960s were embraced by early cyberculture communities [60,61]. Narratives about the kinds of ‘all in good fun’ rebellion that computers could enable were popularized with the establishment of the ‘hacker’ identity in the 1980s (e.g. [21,35]). This imbued cyberspace with metaphors of the Wild West, of Manifest Destiny, of a new frontier of radical individualism and ecstatic self-fulfillment [60,61,69]. It also encouraged a libertarian sensibility, where each actor is considered responsible for their own actions, education, and livelihoods – conveniently ignoring the massive infrastructures that make the cyberspace ‘frontier’ possible. However, neither Negroponte nor Papert discuss the possibility that they, and their students at MIT, enjoyed privileged and idiosyncratic childhoods, nor do they dwell on the sociotechnical infrastructure that enabled that privilege. If anything, Papert’s accounts of employing his theories in classrooms or other settings (e.g. see [48,49]) reinforce notions of exceptionalism by focusing exclusively on the few engaged children, those rare emblems of ‘success’ who appear to prove his theories, and ignoring the rest. These ideals of computers and cyberspace motivated OLPC’s commitment to provide computers to children around the world, and helped the project resonate with many, from journalists to governments, similarly enamored with technology. Imbued with this infinite potential, laptops could take priority over teachers, healthcare, even food and water. And they took on yet greater power in utilizing ideas of childhood, school, and computers together, as we see next. OLPC’s idea of the self-taught learner who disdains school for computers also discounts the critical role that various institutions – peers, families, schools, communities, and more – play in shaping a child’s educational motivation and technological practices. Instead, Papert and other OLPC developers essentialize the child-learner and make the child and the laptop the primary agents in this technosocial assemblage, favoring technological determinism – all it takes is the right kind of computer to keep kids as ‘yearners’ – over the complicated social processes involved in constructing and negotiating childhood. The nexus of childhood, computers, and learning: the charisma of the self-taught hacker These three ideologies come together in One Laptop per Child through the narrative of the self-taught hacker and developers’ desire to bring the experiences of their own This reliance on personal experience underscores an ideological slippage between the ideas taken up by XO 114 laptop designers and children’s activities on the ground [5,6]. OLPC aimed to provide access to pedagogical materials with the assumption that children’s interests would take care of the rest, but did not account for their own idiosyncratic childhoods and ultimately reinforced existing socioeconomic and gender inequalities [5,6,53]. Their reliance on personal experience also enabled their naturalized images of childhood to justify a particular set of normative social objectives. Locating these slippages, and the tenacity to which supporters held to their ideals even when confronted with them [2], provides a powerful method of identifying charisma and uncovering its possible consequences. progress and Western expansion, and a guarantor of economic development for all it reached [44].1 Even when railroads cost vastly more time, money, and human lives to construct than anticipated, and even when they did not bring endless prosperity and worldliness to their termini, the locomotive’s charisma persisted for some time, in the same way that subsequent charismas did over a century later as highways were laid down alongside or over rail lines and as airports replaced train stations as the newest symbols of progress and modernity [39,44,54]. Indeed, these same utopian discourses were applied to many other technological advances over the next century – many of which we now take for granted – including the steamboat, canals, bridges, dams, skyscrapers, the telegraph, electricity, the telephone, the radio, the automobile, television, cable television, and more [16,39,44,54,62]. Even the airplane was hailed as a “winged gospel” that would “foster democracy, equality, and freedom” – despite its terrifying wartime role [69]. PUTTING OLPC’S CHARISMA IN HISTORICAL CONTEXT As we have seen, OLPC’s charisma has the ability to evoke awe, a feeling of spiritual transcendence, and utopian visions for the kind of world that its laptops might make possible. In particular, the XO rolled into one package many of the promises also connected to computers and cyberspace, to childhood and play, and to learning outside of school. A key component of this charisma was its ability to evoke the nostalgic stories that many in the technology world tell about their own childhood experiences with computers. The project has the utopian, if colonialist [17,27], desire to remake the world in its own image. While it may be easy to discount these past charismatic technologies given the perspective and tarnish of time, they contain two lessons, one about the enduring importance of charisma to the modern cultural imagination, and the other about its limits. In particular, there is a striking parallel between the charisma of computers that OLPC draws on and these earlier charismatic technologies. Historian Howard Segal notes that the rhetoric of the power of the Internet to spread democracy “was identical to what thousands of Americans and Europeans had said since the nation’s founding about transportation and communications systems, from canals to railroads to clipper ships to steamboats, and from telegraphs to telephones to radios” [54:170]. Vincent Mosco also notes these similarities in a call for both more empathy for the past and more skepticism of the present. “We look with amusement, and with some condescension,” he writes, “at nineteenth-century predictions that the railroad would bring peace to Europe, that steam power would eliminate the need for manual labor, and that electricity would bounce messages off the clouds, but there certainly have been more recent variations on this theme” [39:22]. However, OLPC’s XO laptop is but the latest in a long line of charismatic technologies. Historians of technology show that from the railroad in the mid-nineteenth century to the Internet today, many new technologies have been hailed as groundbreaking, history-shattering, and life-redefining. “Since the earliest days of the Industrial Revolution,” Langdon Winner writes, “people have looked to the latest, most impressive technology to bring individual and collective redemption. The specific kinds of hardware linked to these fantasies have changed over the years … [b]ut the basic conceit is always the same: new technology will bring universal wealth, enhanced freedom, revitalized politics, satisfying community, and personal fulfillment” [69:12–13]. Here, I briefly describe the history of charismatic technologies as a way to locate and bound OLPC’s charisma. In short, a historical perspective helps (in William Stahl’s words) “break the spell of the present” [56]. It demonstrates that today’s charismatic technologies are neither natural nor inevitable, but are ideologically conservative: even as they promise revolution, they repeat the charisma of past technologies and ultimately reinforce the status quo. This, in turn, allows us to better identify new charismatic technologies and to understand charisma’s consequences. The next section examines a charismatic technology that provides these lessons for One Laptop per Child: the radio. Historian David Nye identifies the first time feelings of redemption were linked to a new technology in his description of the locomotive of the mid-nineteenth century. From the inauguration of America’s first rails in Baltimore in 1828 to the completion of the transcontinental line at Utah’s Promontory Point in 1869, Nye describes a nation gripped with railroad mania. The railroad was implicated in dozens of hyperbolic claims, including the “annihilation of space and time” [29] with its sustained speed, the demise of manual labor with steam, a liberation of humankind from provincialism via increased contact and communication, the triumph of reason and human ingenuity, an ‘engine’ of 1 Nye also notes the flip side of the sublime: a small group decried the locomotive as a “device of Satan” because of its ‘unholy’ speeds [44]. While a discussion of dystopianism is beyond the scope of this paper, it is the flip side of utopianism: though ruled by fear instead of hope, it is still beholden to the same ideologies [30,39,69]. 115 Lessons from the charismatic radio dreams. These “Radio Boys,” as Mosco identifies them, were initially the “heroes” of the Radio Age, “youngsters who lent romance and spirit to the time by building radios, setting up transmitters, and creating networks” [39:2]. But as compelling as their messages may be to anyone sympathetic to contemporary hacker culture, the dreams of the Radio Boys did not prevail. As radio became more commercialized, the connection between technological tinkering and utopian thinking largely receded into niche communities such as HAM radio [23] until reappearing, with a new set of actors but some remarkably similar practices, around the networked personal computer [35,58,61] – the actors that Papert says inspired his work on designing a ‘children’s machine.’ Of all the charismatic technologies of the past, the one that has the strongest resonances with the charisma of computers, education, and childhood – charismas that OLPC relies on – is the radio. Aside from an increasingly marginalized culture of HAM radio operators, it can be hard to imagine radio in the U.S. (today so often a commercialized audio wasteland of top-40 songs on repeat, with a few public stations limping from one pledge drive to the next) as an intensely charismatic technology. But radio took 1920’s America by storm, capturing the country’s collective imagination with promises blending technological miracles and manifest destiny. Historian Susan Douglas explains that radio, as envisioned in 1924, “was going to provide culture and education to the masses, eliminate politicians’ ability to incite passions in a mob, bring people closer to government proceedings, and produce a national culture that would transcend regional and local jealousies” [16:20]. Commentators described the replacement of telegraph wires with radio waves with psychic metaphors and compared it to magic [16:41]. Historians have argued that remarkable parallels between this group and the ‘hackers’ who designed personal computers, the Internet, and OLPC’s XO laptop are no accident. The charisma of radio, computers, the Internet, locomotives, and more draws on the same set of utopian stories about technology, youth, masculinity, rebellion, and selfdetermination. For instance, the individualist strains in these earlier technologies found voice in the cyber-libertarianism of recent decades [61]. The new frontiers of the imagination that the railroad opened in mid-nineteenth century America are rhetorically echoed in the new frontiers of radio in the 1920s and cyberspace in the 1980s, spaces of radical individualism and ecstatic self-fulfillment [60,61,69]. The charismatic appeal of tinkering that 1920s ‘Radio Boys’ championed found voice again in computer hacking, in projects like OLPC, and most recently in maker culture [3]. Thus, while the technologies that these charismas are attached to may shift with time, the charisma lives on. Many of the amateur enthusiasts and educators who pioneered radio were especially excited by the medium’s apparent ability to transcend political and economic controls, enabling virtual communities and informed populism – all hopes that have been echoed more recently about computers, the Internet, and OLPC. After all, Mosco observes wryly, “How could any material force get in the way of invisible messages traveling through the ether?” [39:27] However, many things can and did get in the way of radio’s utopian dreams. Douglas notes that until the Internet, the radio was the main medium where struggles between commercialism and anti-commercialism played out, in a cyclic pattern, over decades [16:16]. When nascent radio stations and other businesses realized how much money could be made selling advertising time on the radio, they advocated for commerce in the ‘ether.’ Governments, too, wanted to control this new technology, and most either took complete control or shared bandwidth with industry, sidelining educators, amateur operators, and other early enthusiasts with little leverage to realize their own hopes. Even by the 1930s, a mere decade after radio’s most charismatic days, “radio was no longer the stuff of democratic visions” [39:27] – much like OLPC has receded into the fog of history for many former supporters a decade after its debut, or the Internet, now decades old, is gradually co-opted for commercial interests. What could be problematic about the feelings that these charismatic technologies can evoke, at least when still relatively new? After all, many of these technologies did, in time, transform the technosocial worlds in which they existed. However, Vincent Mosco argues that we actually prevent these technologies from having their full effect as long as we remain enthralled by their charisma. It was not until they recede into the ‘mundane’ and we understand how they could fit into the messy realities of daily life, rather than making us somehow transcend it, that they have the potential to become a strong social force [39:2,6,19]. If this is the case, then perhaps it is just now, with the spell of OLPC broken, that the XO laptop can start to show its lasting effects among those using it. Time, and independent analysis [6], will tell. Lessons from technologies in educational reform Still, the small group of dedicated enthusiasts who had pioneered the medium – almost all men – continued to use radio to “rebel against the technological and programming status quo in the industry” [16:15], flaunting intellectual property, government rules, and big business in the process [39:2]. This group continued to conceptualize their pastime of tinkering, listening, and programming the radio as a “subversive” activity [16:15] and maintained their equally subversive hopes for the future of the wireless medium. And it is understandable that they would be loath to give up their Education is of course not immune to the draw of charismatic technologies. In fact, educational reform has been a target of techno-utopian discourse long before the One Laptop per Child project was announced, even before Papert began his career of studying children and computers. Charismatic technologies from radio to the Internet have been hailed as saviors for an educational system seemingly perpetually on the brink of failure [62]. The hyperbolic promises of OLPC echo these ongoing efforts to combine the twin promises of charismatic technology and school reform to reach for utopia. 116 Because learning was the central goal of OLPC, examining the history of charisma in this context lets us contextualize the project and identify other charismatic technologies that pervade education. Moreover, a survey of the history of school reform shows that much of the anti-charisma of school described above is based more in stereotypes than fact, reinforced by a century of technological failures in education. Finally, we will see that education reforms like OLPC are often compelled to use charismatic technologies to promise utopia in order to secure attention and funding, which then sets them up for failure and short-lived projects. As reformers then shift to the next charismatic technology, charisma will continue to impede real, if incremental, change. communication technologies including early motion pictures, radio, television, and computers were (and, in the latter case, still are) often championed as fast cure-alls for educational woes by innovators who were often not involved in public schools themselves. “Impatient with the glacial pace of incremental reform, free of institutional memories of past shooting star reforms” that left no effect in day-to-day schooling, and “sometimes hoping for quick profits as well as a quick fix,” Tyack and Cuban explain, these reformers “promised to reinvent education” with technology [62:111]. Like the more general utopian visionaries of technology discussed in previous sections, many of these innovators over-promised the scope and ease of change and lacked a nuanced understanding of the day-to-day social, cultural, and institutional roles of the actors most directly involved in the worlds they wanted to transform. When the messy, expensive, time-consuming realities of using technology in the classroom inevitably clashed with hyperbolic promises, disillusioned innovators, along with the media and the general public, would often blame schools and especially teachers for not solving problems with technological adoption that were, in reality, beyond their reach. Then, research on the effects on the new technology in the classroom would start to roll in, showing that the technology was, as Tyack and Cuban describe, “generally no more effective than traditional instruction and sometimes less” [62:121–122]. Meanwhile, attention and resources had been diverted from more complicated, expensive, or politically charged social or educational reforms that did not promise a quick fix [62:3] and were thus less ‘charismatic.’ Over at least the last century and a half, American schools have been simultaneously held up as the foundation for reforming society and under constant pressure to reform themselves. Starting in the 1830s, upper-class activists such as Horace Mann evangelized the idea of ‘common schools’ as a savior from moral ruin, detailing the horrors that could result (and, in some cases, were resulting, for instance via unregulated child labor) if the country did not adopt universal public schooling [14:8,62:1, 16, 141–142]. Far from the ‘factory worker mills’ framed in some more recent educational reform literature (see [48,49] for examples from Papert), public schools and subsequent innovations in schooling such as age-organized instruction (and sometimes the abandonment of it), curriculum-based instruction (and sometimes the abandonment of it), cheap paper and textbooks, kindergarten, subject-separated instruction, middle school, school lunch, desegregation, televised instruction, and even standardized tests were all framed by socially and politically elite reformers for more than a century and a half as a cure-all for assimilating immigrants and marginal populations, ending child labor, modeling good citizenship, and instilling morality in the next generation [13,14,62]. This is not to say that reform or technological adoption in education is impossible. Though the broader institutional structures of schools and many of the strategies teachers use to reach students appear to be unchanged, there have been many successful innovations that have altered the daily lives of students and teachers, sometimes dramatically. Still, even more realistic reforms well-grounded in the realities of students and schools sometimes have difficulty gaining broad popular support outside of the school unless they add a charismatic gloss of rapid, revolutionary change. The panacean promise of education was gradually shifted in the late twentieth century from moral to economic. In introducing “Great Society” educational reforms meant to eliminate poverty in the 1960s, President Johnson avowed that “the answer to all our national problems comes down to a single word: education” [62:2]. Two decades later, as perceptions of the United States’ waning influence prompted a sense of panic about reforming education to achieve economic competitiveness, President Reagan’s National Commission on Educational Excellence and Reform took up the specter of impending dystopia in an alarmist report titled A Nation at Risk: The Imperative For Educational Reform that to this day motivates U.S. educational reform [13,14,62]. This charismatic pressure can put even open-eyed educational reformers in a catch-22 [6]. They must promise dramatic results to gain the social and financial support for reforms, and then they must either admit to not achieving their goals, or pretend that they did achieve them. Either way, funders will declare that the project is finished and withdraw financial support, and then researchers and other observers will begin to note the discrepancies between reformers’ promises and their own observations. Thus, projects that rely on charismatic technologies are often short-lived, cut off before charisma recedes into the background and the technology can become part of everyday classroom experience. This catch-22 has dogged educational reform for well over a century, and as the educational technology community moves on to the next charismatic technologies Throughout these efforts, reformers often implicated the technologies most charismatic at the time as integral to achieving their visions. Some technological innovations, like chalkboards, inexpensive paper, overhead projectors, and computer-assisted standardized testing, have, in fact, had large and lasting effects on schools. But some of the most charismatic technologies have not. Many media and 117 (whether they be MOOCs or makerspaces), it will continue to hamper the possibility of real, if incremental, change. childhood with universalizing concepts like ‘yearners’ and ‘schoolers,’ Papert and others in OLPC drew on cultural narratives about what childhood should mean and what constitutes a good one, narratives that have become deeply rooted in American culture and reflect American cultural values such as individualism and (certain kinds of) creativity. These actors have generalized from their experiences with largely white, middle-class American youth – or from their own idiosyncratic childhoods – that all, most, or at least the most ‘intellectually interesting’ [49:44–50] children are innately drawn to tinkering with computers and electronics, or in Papert’s words, ‘thinking like a machine’ [48]. What is the alternative to this catch-22 of charismatic education reform? Incremental reforms, what Tyack and Cuban call “tinkering,” are more effective in the long-term, even if they are not charismatic. “It may be fashionable to decry such change as piecemeal and inadequate, but over long periods of time such revision of practice, adapted to local contexts, can substantially improve schools,” they explain. “Tinkering is one way of preserving what is valuable and reworking what is not” [62:5]. CONCLUSION The case of OLPC shows us why it is dangerous to ignore the origins of charisma: one risks being perpetually blinded by the newest charismatic technology as a result. Indeed, those who pin reform efforts on charismatic technologies are often caught in a catch-22 where their projects are cut short whether they register success or not, because the promises of charisma are ultimately unattainable. Through an analysis of One Laptop per Child and a survey of past charismatic technologies, this paper exposes the ideological stakes that underpin charisma – the ability for technologies (or, as originally theorized, people [68]) to evoke feelings of awe, transcendence, and connection to a greater purpose. It shows how the promises that charismatic technologies make are ideologically charged, how they can be identified, and what is at stake when we are drawn in. While it may be easy to discount examples from the past given the perspective and tarnish of time, taking a historical perspective on charismatic technologies show us how conservative charisma actually is – the same kinds of promises have been made over and over, with different technologies – and also how unattainable its promises are. Moreover, as long as we are enthralled by charisma we might actually prevent these technologies from becoming part of the messy reality of our lives, rather than helping us transcend it. We must remember that charisma is ultimately a conservative social force. Even when charismatic technologies promise to quickly and painlessly transform our lives for the better, they appeal precisely because they echo existing stereotypes, confirm the value of existing power relations, and reinforce existing ideologies. Meanwhile, they may divert attention and resources from more complicated, expensive, or politically charged reforms that do not promise a quick fix and are thus less ‘charismatic.’ Examining charisma can help us understand its effects and, through understanding, counter them. This analysis is meant to help ‘make the familiar strange,’ in Stuart Hall’s words [22], helping researchers in human-computer interaction to identify the ideological commitments of the technology world. Analyzing a technology’s charisma helps us recognize ideologies that may otherwise be as invisible as water is to the proverbial fish. While concrete design suggestions are not the goal of this paper, this analysis may also help designers – who often hope to ‘do good’ through technological intervention in ways similar to those analyzed here – identify their own ideological commitments. Still, it is also important to recognize that charisma plays an important role in smoothing away uncertainties, contradictions, and adversities. As such, the purpose of this paper is not to ‘prove’ charisma ‘wrong.’ What matters is whether a technology’s charisma is ideologically resonant – whether it taps into deep-seated cultural values and identities, as OLPC does with childhood, school, play, and technology. This paper’s intention is likewise not to advocate for an eradication of ideologies; just as it is impossible to escape the bounds of our own subjective points of view, so too is it impossible to operate entirely outside of the frameworks of ideologies. However, a large body of Marxist theory (e.g. see [22]) notes that becoming cognizant of the ideological frameworks in which we operate allow us to evaluate whether they are really serving the purposes we hope or assume they are. Only by way of this cognizance can we shift them if they are not. To OLPC’s contributors, we saw that the charisma of the XO laptop affirmed their belief in the power of computers in childhood, imposed coherence and direction on their work, and gave them reasons to push back against doubters, even in the face of what might otherwise feel like overwhelming odds or ample counterevidence [2]. On the other hand, we saw that charisma could also have a blinding effect. It prevented those on the project from recognizing or appreciating ideological diversity, much less constructively confronting problems of socio-economic disparity, racial and gender bias, or other issues of social justice beyond most of their personal experiences [47]. OLPC’s XO laptop was charismatic to them because it mirrored their existing ideologies and promoted a social order with them at the top. ACKNOWLEDGMENTS I am indebted to many friends and mentors for helping me develop the ideas presented in this paper. In particular, Fred Turner and Eden Medina pushed me to articulate the importance of charismatic technologies, and Paul Dourish, Lilly Irani, Daniela Rosner, Ingrid Erickson, and Marisa Brandt helped me flesh out the salient features of charisma. I As a result, their narratives not only glorified childhood, they specified the kinds of learning that children are naturally inclined to do. In glossing over the many complexities of 118 also thank all of the participants involved in my research on One Laptop per Child. 17. Paul Dourish and Scott D. Mainwaring. 2012. Ubicomp’s colonial impulse. Proc. UbiComp ’12, ACM Press. REFERENCES 18. James Ferguson. 2006. Global Shadows: Africa in the Neoliberal World Order. Duke Univ. Press. 1. Madeleine Akrich. 1992. The De-Scription of Technical Objects. In Shaping Technology, Wiebe E Bijker and John Law (eds.). MIT Press. 19. Batya Friedman, Peter H. Kahn, and Alan Borning. 2006. Value-Sensitive Design and Information Systems. In Human-Computer Interaction in Management Information Systems: Foundations, P. Zhang and D. Galletta (eds.). M.E. Sharpe, Inc. 2. Morgan G. Ames, Daniela K. Rosner, and Ingrid Erickson. 2015. Worship, Faith, and Evangelism: Religion as an Ideological Lens for Engineering Worlds. Proc. CSCW 2015, ACM Press. 20. Gerald Futschek and Chronis Kynigos (chairs). 2014. Constructionism and Creativity conference. 3. Morgan G. Ames and Daniela K. Rosner. 2014. From drills to laptops: designing modern childhood imaginaries. Information, Communication & Society 17, 3, 357–370. 21. William Gibson. 1984. Neuromancer. Ace Books. 22. Stuart Hall. 1982. The Rediscovery of “Ideology”: The Return of the Repressed in Media Studies. In Culture, Society, and the Media, Michael Gurevitch, Tony Bennett, James Curran, Janet Woollacott (eds.). Methuen, 56–90. 4. Morgan G. Ames, Mark Warschauer, and Shelia R. Cotten. 2016. The “Costly Lesson” of One Laptop per Child Birmingham. In When School Policies Backfire, and What We Can Learn, Michael A. Gottfried and Gilberto Q. Conchas (eds.). Harvard Education Press. 23. Kristen Haring. 2007. Ham Radio’s Technical Culture. MIT Press. 5. Morgan G. Ames. Learning Consumption: Media, Literacy, and the Legacy of One Laptop per Child. The Information Society, in press. 24. Carlene Hempel. 1999. The miseducation of young nerds. http://jimhoeg.com/jimmy/proflife/miseducation.htm 25. Benjamin Mako Hill. 2002. The Geek Shall Inherit the Earth: My Story of Unlearning. http://mako.cc/writing/ unlearningstory/StoryOfUnlearing.html 6. Morgan G. Ames. 2014. Translating Magic: The Charisma of OLPC’s XO Laptop in Paraguay. In Beyond Imported Magic: Essays in Science, Technology, and Society in Latin America, Eden Medina, Ivan da Costa Marques, Christina Holmes (eds.). MIT Press, 369–407. 26. Diane M Hoffman. 2003. Childhood Ideology in the United States: A comparative cultural view. International Review of Education 49, 1-2, 191–211. 7. Walter Bender. 2007. OLPC: Revolutionizing How the World’s Children Engage in Learning. 27. Lilly Irani, Janet Vertesi, Paul Dourish, Kavita Philip, and Rebecca E. Grinter. 2010. Postcolonial Computing: A Lens on Design and Development. Proc. CHI 2010, ACM Press, 1311–1320. 8. Paulo Blikstein. 2013. Seymour Papert Tribute at IDC 2013 Panel Avant-propos: Thinking about Learning, and Learning about Thinking. 28. Kevin Kelly. 1999. Nerd theology. Technology in Society 21, 4, 387–392. 9. Chris Blizzard at OLPC Analyst Meeting. 2007. 10. Michael Burawoy. 1998. The Extended Case Method. Sociological Theory 16, 1, 4–33. 29. Scott Kirsch. 1995. The incredible shrinking world? Technology and the production of space. Environment and Planning D: Society and Space 13, 5, 529–555. 11. Heidi A. Campbell and Antonio C. La Pastina. 2010. How the iPhone Became Divine: New media, religion and the intertextual circulation of meaning. New Media & Society 12, 7, 1191–1207. 30. Rob Kling. 1996. Computerization and Controversy: Value conflicts and social choices. Morgan Kaufmann. 31. Pui-Yan Lam. 2001. May the Force of the Operating System be with You: Macintosh Devotion as Implicit Religion. Sociology of Religion 62, 2, 243–262. 12. Howard P. Chudacoff. 2007. Children at play: An American history. NYU Press. 13. Larry Cuban. 1986. Introduction. In Teachers and Machines: The Classroom Use of Technology Since 1920. Teachers College Press. 32. Annette Lareau. 2003. Unequal Childhoods: Class, Race, and Family Life. U. California Press. 14. Larry Cuban. 2001. Oversold and underused: Computers in the classroom. Harvard Univ. Press. 33. Bruno Latour. 2005. Reassembling the Social: An introduction to actor-network theory. Oxford Univ. Press. 15. Sidney I. Dobrin. 1997. Constructing Knowledges: The Politics of Theory-Building and Pedagogy in Composition. SUNY Press. 34. John Law. 1992. Notes on the Theory of the Actor Network: Ordering, Strategy and Heterogeneity. Systems Practice 5, 4, 379–393. 16. Susan Douglas. 2004. Listening In: Radio and the American Imagination. U. Minnesota Press. 35. Steven Levy. 1984. Hackers: Heroes of the Computer Revolution. Anchor Press Doubleday. 119 36. Donald McIntosh. 1970. Weber and Freud. American Sociological Review 35, 5, 901–911. 37. Scott Forsyth Mickey. 2013. Constructing The Prophet: Steve Jobs and the Messianic Myth of Apple. 55. Tim O Shea and Timothy Koschmann. 1997. Review, The Children’s Machine: Rethinking School in the Age of the Computer (book). Journal of Learning Sciences 6, 4, 401–415. 38. Steven Mintz. 2004. Huck’s Raft: A History of American Childhood. Harvard Univ. Press. 56. William A. Stahl. 1999. God and the Chip: Religion and the Culture of Technology. Wilfrid Laurier Univ. Press. 39. Vincent Mosco. 2005. The Digital Sublime: Myth, Power, and Cyberspace. MIT Press. 57. Maria Stavrinaki. 2010. The African Chair or the Charismatic Object. Grey Room 41, 88–110. 40. Albert M. Muniz Jr. and Hope Jensen Schau. 2005. Religiosity in the Abandoned Apple Newton. Journal of Consumer Research 31, 737–747. 58. Douglas Thomas. 2002. Hacker Culture. U. Minnesota Press, Minneapolis, MN. 59. Anna Tsing. 2000. The Global Situation. Cultural Anthropology 15, 3, 327–360. 41. Nicholas Negroponte. 1996. Being Digital. Vintage Books. 60. Fred Turner. 2006. How Digital Technology Found Utopian Ideology: Lessons from the first Hackers conference. In Critical Cyberculture Studies: Current terrains, future directions, David Silver, Adrienne Massanari, Steven Jones (eds.). NYU Press. 42. Nicholas Negroponte. 1998. One-Room Rural Schools. WIRED Magazine. 43. Nicholas Negroponte. 2006. No Lap Un-Topped: The Bottom Up Revolution that could Redefine Global IT Culture. 61. Fred Turner. 2006. From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism. U. Chicago Press. 44. David Nye. 1996. American Technological Sublime. MIT Press. 62. David Tyack and Larry Cuban. 1995. Tinkering toward Utopia: A Century of Public School Reform. Harvard Univ. Press. 45. Amy F. Ogata. 2013. Designing the Creative Child: Playthings and Places in Midcentury America. U. Minnesota Press. 63. Wayan Vota. 2007. Is OLPC the Only Hope to Eliminate Poverty and Create World Peace? OLPC News. http://olpcnews.com/people/negroponte/olpc_poverty_wo rld_peace.html 46. Ruth Oldenziel. 2008. Boys and Their Toys: The Fisher Body Craftsman’s Guild, 1930-1968, and the Making of a Male Technical Domain. Technology and Culture 38, 1, 60–96. 64. Wayan Vota. 2009. Negroponte Throws XO on Stage Floor “Do this with a Dell” - Everyone a Twitter. OLPC News Forums. http://olpcnews.com/forum/?topic=4597.0 47. Nelly Oudshoorn, Els Rommes, and Marcelle Stienstra. 2004. Configuring the User as Everybody: Gender and Design Cultures in Information and Communication Technologies. Science, Technology & Human Values 29, 1, 30–63. 65. Wayan Vota. 2010. Mesh Networking Just Doesn’t Work on XO-1.5 Laptop. OLPC News. http://olpcnews.com/har dware/wireless/mesh_networking_just_doesnt_wo.html 48. Seymour Papert. 1980. Mindstorms: Children, Computers, and Powerful Ideas. Basic Books. 66. Wayan Vota. 2011. XO Helicopter Deployments? Nicholas Negroponte Must be Crazy! OLPC News. http://olpcnews.com/people/negroponte/xo_helicopter_de ployments_nich.html 49. Seymour Papert. 1993. The Children’s Machine: Rethinking school in the age of the computer. Basic Books. 67. Mark Warschauer and Morgan G. Ames. 2010. Can One Laptop per Child Save the World’s Poor? Journal of International Affairs 64, 1, 33–51. 50. Seymour Papert. 2006. Digital Development: How the $100 Laptop Could Change Education. USINFO Webchat, via OLPC Talks. http://olpctalks.com/ seymour_papert/seymour_papert_usinfo.html 68. Max Weber. 1947. Charismatic Authority. In The Theory of Social and Economic Organization, Talcott Parsons (ed.). The Free Press. 51. Roy Pea. 1987. The Aims of Software Criticism: Reply to Professor Papert. Educational Researcher 16, 5, 4–8. 69. Langdon Winner. 1997. Technology today: Utopia or dystopia? Social Research 64, 3, 989–1017. 52. Brett Robinson. 2013. Apple and the religious roots of technological devotion. LA Times. 70. Geraldine Youcha. 1995. Minding the Children: Child care in America from colonial times to the present. Da Capo Press. 53. Daniela K. Rosner and Morgan G. Ames. 2014. Designing for Repair? Infrastructures and Materialities of Breakdown. Proc. CSCW 2014, ACM Press. 71. Joseph L. Zornado. 2001. Inventing the Child: Culture, Ideology, and the Story of Childhood. Garland Publishing. 54. Howard P. Segal. 2005. Technological Utopianism in American Culture. Syracuse Univ. Press. 120 An Anxious Alliance Kaiton Williams Information Science Cornell University kaiton@cs.cornell.edu ABSTRACT This essay presents a multi-year autoethnographic perspective on the use of personal fitness and self-tracking technologies to lose weight. In doing so, it examines the rich and contradictory relationships with ourselves and our world that are generated around these systems, and argues that the efforts to gain control and understanding of one’s self through them need not be read as a capitulation to rationalizing forces, or the embrace of utopian ideals, but as an ongoing negotiation of the boundaries and meanings of self within an anxious alliance of knowledge, bodies, devices, and data. I discuss how my widening inquiry into these tools and practices took me from a solitary practice and into a community of fellow travellers, and from the pursuit of a single body goal into a continually renewing project of personal possibility. Author Keywords self-tracking; autoethnography; experience; quantified-self ACM Classification Keywords H.5.m. Information Interfaces and Presentation (e.g. HCI): Miscellaneous INTRODUCTION At the beginning of 2012 I decided that it was time to lose some weight, and I turned to a set of digital health & fitness tools to help me on my way. I managed to lose the weight, but my journey with these systems continued and I substantially changed my patterns and everyday experiences in ways that ranged from the pedestrian to the sublime. Although this project was not initially intended as research but as part of an effort to address a real personal need, I believe that the journey has provided me with a new professional perspective on systems like these, and has taken me from a solitary practice into a community of fellow travellers, and from the pursuit of a specific body goal to a continually renewing project of personal possibility. I embraced a personal, existential, crisis as an opportunity to try solutions that I had been critical of professionally or avoided personally. I considered this as an opportunity to develop an experiential understanding that was tied to a real personal need. Surely, I told myself at the outset of this project, Copyright 2015 is held by the author(s). University and ACM. Publication rights licensed to Aarhus this would be better than critical analysis lobbed in from the outside. The route would prove uncertain. I wanted to be in better control of my self and I wanted to understand how it felt to be healthy and fit. Yet, as I began to make progress towards my goals, I wondered about the changes in mind and body that were accompanying my technologically guided, deliberate movement. Was the quest for more information about myself and my immediate world helping me find either? And how might any acquired tracking tools, and the ability to navigate them, fit into my understanding of what it meant to be a modern connected person? Furthermore, where did the approach I was using—experiments and experiences within one’s self— fit within the toolkit of methods in use within HCI? My retelling of that experience is presented here as a personal essay, building on Bardzell’s articulation of the value of the essay form for a CHI audience [2], and inspired by Taylor’s essay on the politics of research into cultural others in HCI [30]. My hope is that placing my self directly in the frame of analysis will highlight a tangle of personal and professional interests that are I believe should be embraced as we interact with, design, and sustain a discourse on systems that are ever more integrated into our everyday lives. These frames in which we pursue our health and fitness are neither fixed nor external but are created by, and depend on, our perceptions and beliefs, social norms, and our collective sense of technological possibility. These unavoidably intimate concerns lend an approach centred on the self-as-site, self-as-subject a ready allure. The autoethnographic approach used here (and to various degrees previously in [14, 8, 18]), with its synthesis of postmodern ethnography and autobiography, is intended to call into question the objective observer position, the realist conventions of standard ethnography, and to disturb notions of the coherent, individual self and its binaries of self/society and objective/subjective [26]. In doing so, it has the potential to provide a conduit to what McCarthy and Wright refer to as the “rich interior world of felt life” necessary for producing empathetic understanding [36]. But while undoubtedly productive, the approach presented significant challenges—which manifest here—both in the shaping of a research agenda through the course of a personal transformation, and in the reporting of those results. My primary goal is to generate an empathetic engagement with my experience, and through it, a connection with what 5th Decennial Aarhus Conference on Critical Alternatives August 17 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21146 121 could be the concerns and life paths of others in similar situations. But I intend my account of my experience to be an illustration, not general evidence or directly comparable to that of others. This is a distinctly idiosyncratic account, but it is through the production of idiosyncratic accounts that I believe we can find the empathy—as a view of our lives as entangled with others—that makes for a rich understanding of the role of personal devices in our lives. This requires highlighting, not suppressing, the vulnerability, motivations, personality traits, contradictions and stumbles that made up both my everyday experience and my attempts at analysing it. This is not to say that matters of concern of a larger scale are not important. It is vital that we understand the hybrid and structural phenomena around health and big data that knit together nations, social classes, forms of knowledge and inquiry with Latourian [16] hitches. In fact if others were not so ably investigating these matters, I would have no room to write an essay about the everyday and unspectacular. This is then an attempt to perform a balancing act of personal revelation within the tent of scholarly discourse. And just how big that tent should be is also a matter that this essay takes into consideration. My contention is that, to faithfully convey that messy, polyvocal world that finds its berth in personal experience, the tent’s boundaries need to be reshaped. Just more than 1,377,584 measured calories ago I embraced a simple goal: to lose weight. This is one made, I’m certain, by millions of others every day. Yet, more than any highminded investigation into technologies, or into a community and its practices, it was this that drove my interest and kept me going. In reflecting on the essay’s denigrated place in the CHI/HCI canon, Bardzell suggests that the essay is the best form for doing just this work: for revealing “a process of thinking” that is still “shaped and crafted as a work of writing”[2]. He argues that the essay is positioned as only opinion within CHI, but he places it as part of a millennia-long tradition of trying to deliver exactly the kind of analytic understanding of experience that the field solicits. And that interpretative efficacy is delivered through the playing up of a sense-making subject apprehending “an object of study...[that] must therefore, be...in some sense confused, completely, cloudy, contradictory.” This is necessarily the state of affairs in situations like these, where the subject and subjectivity are the sources of data. I made the decision after the Christmas holidays, when my sister had come to visit. I prepared a feast and we dined sumptuously for several days. At the end of the visit, we went to the Pacific shore and took a few photos on instant film. It was then, as I watched those photos develop, that I realised how bad things had gotten. I had been slowly but steadily gaining weight even though I thought I was in control of my diet and getting enough exercise. I looked, in my own estimation, terrible. This is more than an secondary note about construction. The essay’s production is embedded in the transforming frame of the journey it describes and reflects that. The shifts in tense and tone that follow reflect concerns that transform, in stages, from a matter of physicality, to one of self-understanding, and into an ontological project of expanding possibility pursued in alliance with others. At the same I remain faithful to an exploration of the pedestrian, and what Georges Perec referred to as the “infraordinary” or “endotic:” the small, quotidian, experiences that accumulate and become our lives. Perec charged his readers to join him and together “found our own anthropology, one that that will speak about us, will look in ourselves for what for so long we’ve been pillaging from others” [24]. He contrasted this exploration of small matters of concern against the big stories, the front page headlines and the spectacular. His exploration was to be one of the seemingly obvious and habitual: It matters little to me that these questions [of the everyday] should be fragmentary, barely indicative of a method, at most of a project. It matters a lot to me that they should seem trivial and futile: that’s exactly what makes them just as essential, if not more so, as all the other questions by which we’ve tried in vain to lay hold on our truth. BEGINNING Granted, as physical issues go, this was a minor calamity. Still, I decided then that I needed to not only get back in shape, but to find a place of steady balance. To be completely honest, I cannot say “back.” I was only ever in shape for brief moments. I had only passed, momentarily, through shape. I wanted to remain there; wherever there was. So as much as I loathed New Year’s resolutions, I decided to get going a few days later. The immediate problem was that I didn’t know how exactly to go about it. I knew the accepted wisdom was that I should eat less and exercise more but I didn’t really know what I should eat less of, or what exercises I should perform. I had done some research on health and fitness research previously and knew that a range of consumer tools existed. I knew about BMI, and even though I knew that number had contingencies, I knew that mine was too high. I knew what calories were but I didn’t know how many I should be consuming. I more or less still ate according to ratios and formulae mimicked from my childhood. But life was different in the here and now. I was surrounded by a stunning variety of foods, eating styles, and advertising schemes. A steady compass was hard to come by. I knew enough to avoid an all-cake diet, but where would I find salvation? Low-fat? Low-carb? I didn’t know which to follow, or how to keep in line. And what could I reasonably achieve? I had some vague goals from a variety of sources but I wasn’t sure what good targets and limits should be. And I definitely had little formal idea of how to manage my consumption to meet them. 122 There was an overwhelming amount of information to consume, much of it contradictory. In contrast, many of the applications I surveyed offered simple solutions. Download, launch, follow the prompts, and hit the suggested numbers. This was immensely attractive, even with my professional skepticism. This remains something that I consider with a fair amount of irony. I was among a group of researchers who had been critical of this persuasive and reductive logic that powered many of the popular diet control and tracking systems [25]. But now I found myself in need of them. tion. One was a simple social network designed to promote encouragement between users and to route encouragement to and from Facebook & Twitter, and the other was a sponsored motivation system of reminders, with badges (Figure 1) awarded to users at particular milestones, for example: the 10lb club; the regular user (every day for 2 weeks); the hardcore (26 weeks); the exercise buff, hound, and king; and, curiously, the Tea Time badge (“tea powered motivation brought to you by SPLENDA ESSENTIALS No Calorie Sweetener Products”) for those who recorded enough cups of tea. This was the time and place of the first of many conflicts. As a researcher seeking to modify my body how could I be both a participant in systems like these—systems backed by what I saw as rationalist agendas threatening self-knowledge, intuition and reliance—and still champion a resistance against them? How could I inhabit and speak from what [12] refers to as the normalised position of a diet participant, and one of a resister? Would I be taking these systems and their ideologies down from the inside? Maybe, I told myself, I’d get to that after I got my 6-pack, washboard abs. Then it would be down with the tyranny of rational reporting systems and selfsurveillance. But first, I needed a truce with that inside me that I thought was working against my best interest. And it appeared that the best way for me to do that was to swallow my pride and call in a mediator. And so, like any well-meaning, wellconnected person of my age, means, and technological background would, I chose the Internet and my smartphone to help me do so. And, as it turned out, there were many apps for that. LOSING IT It was 22 months into my project, and I had exceeded my early goals only to find myself frustrated with stalled progress in newer, ever-more fantastic schemes of self-transformation. In the pursuit of pinpoint diagnosis, I lay still on a hard counter in a dark room breathing into a small device of bleeps and bloops as it tried to measure my basal metabolical rate. 1950. Should I be relieved at this number? How did I get here? It began with a scale. I told it my age and my height, stepped on, and in reply it told me my weight and body composition: 5’ 8, 174lbs and 24% fat. This was, for me, an uncomfortable conversation. I felt a stranger to this self. A week later, I downloaded and started using Lose It!, a weight-loss and calorie tracking application for the smartphone and web. At the time, it had several features broadly representative of dominant calorie tracking and weight loss systems: the ability to track overall calories along with nutrients, budgeted according to reported weight and estimated metabolism; a large database of food items and the ability to scan packaged, bar-coded foods into that catalog; the ability to further extend that local database by creating new ingredients or saving food combinations as recipes. It also had two other features that I avoided, primarily around motiva- 123 Figure 1. A sampling of the author’s badges I never used these features; they were antithetical to why I used the tool in the first place. I had no desire to share my progress with others, and motivation through competition or badges held little appeal. I learned to carve out a space between calls to upgrade for premium features and the emails announcing my new badges. This work of finding spaces in a single app expanded into the formation of alliances, and then peace-keeping mandate within an ad hoc assembly. As the weeks and months went by, I auditioned, added, and managed many other mobile applications, physical tools, and fragments of knowledge. To name a few: calorie count, a nutritional database that I would consult to convert food into calorie and macronutrient values; Endomondo, my exercise tracking app that would report calories burned and detailed performance numbers from my bicycle rides; the Moves app and a FitBit device helped me keep track of places visited and steps taken; Gym Hero, PUSH, and Strong helped me track weightlifting workouts. Finally, a small scale, combined with measuring cups and spoons, provided predictability in the kitchen. With this shifting team I set out to change my life, navigating a landscape of other tools and systems, online discussions, and professional advice. I consulted with trainers, physicians and nutritionists. I read internet forums of every kind. I scheduled physicals and full body density scans in cuttingedge machines (See figure 2). Digital and paper journals helped me keep track of and reflect on my struggles and progress. Treating this an ethnographic encounter with my self, I began by taking frequent field-notes, recording and reflecting on goals, my progress towards them, and my developing relationship with these systems. I kept these notes in parallel with my expanding data recording, but those bars, curves, and data points were my reflections as well, my successes and trepidation encoded in their patterns and rhythms. The just-so upward slope of my Figure 2. The results of a Dual-energy X-ray absorptiometry (DXA) scan body-weight readings weeks after a holiday remains to me as evocative as a madeleine. As I recorded I worried about how to explain and recount an experience steeped in numbers yet both fleeting and embodied. Should I try to separate my selves as recorder and recorded? In [20, 36], the authors channel Bakhtin and Dewey to hold that this sort of aesthetic experience might be the paradigm for all experience, reflecting on Dewey’s definition of experience as a double-barreled “unanalyzed totality” that “recognises in its primary integrity no division between act and material, subject and object”[20]. [33] describes this as a circularity, “where knowledge comes to be inscribed by being with the ‘designing-being’ of a tool, thus in turn modifying (designing) the being of the tool-user”. I relaxed. I drew inspiration from Neustaedter and Sengers’ investigation of autobiographical design as a method that can provide detailed and experiential understanding of a design space [23]. My collection and curation of this menagerie of devices, applications, algorithms and APIs, and my fashioning and querying of reports and visualisations revealed itself as a form of design. I didn’t write the software in my conglomerate but I do understand my role here as a designer, and this as a form of research through design [10, 37]. I am designing myself, in collaboration with the systems I’ve cobbled together. In this we exist as anxious allies: body; device; data; cloud; others; algorithms; entrepreneurs. 124 This design is critical, reflective, ontological, and tinkering— what [33, 34] position as fundamental to being human wherein “we design...we deliberate, plan and scheme in ways which prefigure our actions and makings [and] in turn...are designed by our designing and by that which we have designed” [33]. This ontologically oriented understanding of design is “a philosophical discourse about the self—about what we can do and what we can be” necessarily “both reflective and political, looking backwards to the tradition that has formed us but also forwards to as-yet uncreated transformations of our lives together” [34]. This design is building as nurturing, not just constructing. And as we pervade the things we build, they in turn pervade us in a ways more fundamental than any single designer could consciously intend [33]. LIVING AMONG THE NUMBERS Rearrange the icons on my home screen and my thumb is lost. It quivers, bent searchingly and confused. We had a system, it seemed to say. (fieldnote excerpt) I enjoyed the everyday accounting. Entering in Lose It! the food I ate helped me reflect both on my food choices and on the events that surrounded them. The manual repetition of entries underwrote this: launch app, my thumb to the top corner of screen; “Add Food”, type in food; choose the amount, arrange ingredients and meals in orders temporal and strategic. 40 taps and scrolls to enter a meal of broccoli, roasted chicken, and an arugula salad dressed simply. The result: 461 calories. An immediate question might be, could this be faster? And maybe in the beginning, I would have wanted that. I can even imagine someday an app will allow me to take a picture of my food and return structured data in a flash. But even the seemingly tedious has its attractions. Out for a meal, far away from measuring cups and scales, I gradually found play in reverse-engineering a dish back into composite ingredients and macronutrients. The practice of recording is soothing and has its charms. I could reflect on the day’s flow, wordlessly. I could watch the numbers accumulate and balance, as charts filled in and lines extended; a retelling of my day in a new narrative form. I found myself launching the app several times a day, often only to scroll laps through the day’s huddled data. Untended data pervades me as anxiously as an untended garden might, resulting in a “thinking as being with” [33]. The daily weeding, pruning and nurturing of data not only pays dividends over the long term but is rewarding and meditative in its own right. It is action with rewards at once past, present, and future. I take a quiet pleasure from lying in bed at the end of the day and looking up and confirming locations I’ve visited, all tracked in Moves. I rarely look back at the data analytically; rarely seek to recall where I was on a Monday night, 8 weeks before. But I’m happy to know that data’s there, to feel its weight accumulating over time: the results of making paths in new cities, and retracing old ones in a familiar geography. They are all constellations in my private data mythology. An idiosyncratic self-knowledge Despite my success and its attendant numeric pleasures, I’ve yet to enjoy contemplating my self as a precarious balance of inputs and outputs: what I eat versus how much I exercise. I would have preferred to accomplish my self-transformation within broader measures, and I still long for that: to comprehend my body in longer and less granular scales—seasons instead of hours, calories banished altogether. A friend, at once svelte and blissfully ignorant, remarked in a conversation about my calorie goal that day: “I have no idea of calories / What amounts to 10 much less 1500.” Even with my fitness goal surpassed, my body seemingly in the balance I sought, I could only reply: “That’s actually beautiful.” Because it’s as hard to imagine going back as it is to imagine going forward. I can’t un-know the weight of things. As Dewey notes, “every experience both takes up something from those which have gone before and modifies in some way the quality of those which come after” [7, p27]. I had come to realise that an attention to ever smaller matters of concern could yield effective results, but at the same time I couldn’t help but imagine myself bound to a future where I could scarcely ignore them. At least, not if I wanted to remain in my present condition. To know was now, in many ways, to be bound to manage and optimise. I was coming to view the insights my tools had begun to provide me, and my developing dependence and relationship with them, as unsteady. But aren’t we all wrapped up in similarly conflicted relationships—what Bourdieu referred to as “the Benthamite 125 calculation of pleasures and pains, benefits and costs” [5]? Our modern life is as much defined by asceticism and denial as by gratification. To reach our goals we undertake broad changes in our lives: no meat; no dairy; no to the second drink; lock our WiFi router in a safe when we need to focus [28]. Why should this relationship be any different? Why did I continue with a program that contradicted my ideals and left me in an unsteady position? For one, it worked, and worked quickly. The efficacy of the approach was difficult to deny. After the first 4 weeks I had lost 5.6 lbs and I felt that I was arriving at a workable understanding of how what I ate affected my body. I felt healthy, and I was performing small experiments and getting valuable feedback: tuning the right amount of fibre, getting enough protein, monitoring my sugar intake. Still, I would tell myself that there were a number of other reasons that could have explained the changes to my body: I had a renewed awareness of my self and what I was eating; I was feeling less stressed now that I felt more in control; I was starting to use exact measures rather than estimating. The numbers and their precision continued to have a mesmerising effect over me. Intellectually, I had serious concerns about the value of calories as the sole or even dominant basis for weight management—that all calories were equal—and though I tried to focus more on the ratios of nutrients or on the quality of my food, Lose It!’s calorie budget readout, with its scary red zone indication, remained a metaphysical hurdle difficult to clear. Even as my attachment to these tools has waned, I still view that threshold warily, finding soothing reward in almost always being several calories under it—safe in the green zone (see figure 3)—even though I both theoretically and practically know it to be imprecise and perhaps even inaccurate. This was a surprising revelation. I knew intellectually that the numbers displayed in my app and used behind the scenes in its calculations were based on rough estimates or theories. And I knew how futile the notion of managing one’s self to precise calorie budgets must be. I knew that, by describing our bodies as precise systems that can go out of sync based on small discrepancies, the health industry and app creators benefit by this positioning of their tools and systems as indispensable and necessary guides in our lives. Yet, months after that internal conversation, I continued to ask myself, “can our bodies really optimise down to a few calories?” while measuring out exactly 7 evenly-sized almonds for a quick snack. I had begun to read non-weight-related features of my life through what [13] refers to as the “prism of weight loss.” My days had been reshaped through those units of measure. My tools became my oracles, and I consulted them before planning meals or an evening out. Through them I acquired a fluency in the forms and verse of macronutrients. Over the months I steadily made my life more calculable by streamlining my diet to in turn streamline how I input data into my tools. I prioritised certain foods and recipes, and avoided others to work best within the capabilities of the food database. I weighed my food or measured it by the fluid Figure 3. Calorie Thresholds: Day marked green are those with calories at least 1% below threshold; those marked red are more than 1% over the threshold; those in cream are within +/- 1% of the app’s suggested goal ounce. I groaned at foods that weren’t in the database or occasionally at recipes that were too complicated to assess calories per serving once complete. But still I found freedom in this calculation and control and room in its reduction. Regardless of whether the I/O, calories in vs. calories out, model was correct over the long term, I had found reassurance in its immediate results. Before I began this study I didn’t know how many calories I was eating overall or how many calories were even in particular foods. I wasn’t even aware of the mix of nutrients that I was eating. The apps provided me with expanded capabilities so that I was no longer eating “blindly.” Pushed to spend more time reading nutritional labels and scanning food databases to enter into my tools, I realised how little I knew about what I had been eating. I developed felt a strong sense of fidelity to my accumulated system; an ordained from Logos desire to keep the record true (read: often, precise). At the same time, I knew that I occasionally kept a false dialogue: over-estimating food items, or under-entering exercise. Why would I perform these small acts of rebellion even though they might ostensibly be against my self-interest? Why did I trust the system when it confirmed my “bad behaviour” but conspire to undermine its reports of when I was “good?” Even as I continued performing them, I wondered if these minor insurrections were how I came to terms with what seemed like an increasing rationality in my life. Together, we—my system and I—had co-constructed a digital model of my self that I fully bought into and managed: managing myself, it seemed later, by proxy. Within this, I had found a practical way to consume much of the advice and research I would come across. With my model in hand, I felt a relationship between my data and my body, though I struggled to decide how much of that connection was real or imagined. When I didn’t eat “enough” protein I felt weaker, and when I had too much sugar I felt fatter. These were delayed reactions—a re-reading of my body from the model. Had I sacrificed one balance for another, or balanced my model instead of my self? Did that final distinction even matter? Stepping Back viscera of being human was replaced with a gnawing worry about the feasibility about going it alone, even with my developing intuition. My reasons for continuing with this project were complicated: personal and professional, subject, steward, and researcher. Projects must end but what about the relationships I had developed within my alliance? What would I do without my calorie tracker? How would I wean myself away even now that I had met some of goals that I had developed? I still think a lot about the transformation. I am now “in shape.” I feel strong and healthy. This is a place that I had scarce considered before, but these numbers seem to hold as much fear as pleasure. Could I maintain this state without “outside” help? Relying on my existing alliance was no sure path to reassurance either: Back around 153, and it felt like it took too much doing. I’m not sure if I would have naturally made it back there or what the cause was. Was it a couple weeks eating [Lose It!’s] recommended 2000 calories? The flour of Denmark? The week in Seattle? There’s a lot that happens in between the numbers. [Two months later...] Confused with my weight. Moved too many variables and I don’t know how to unwind it all (fieldnote excerpt) And in the present, if I do cast these systems aside, would doing so really lead to any better engagement with my self? Granted, I am far better at mechanical feats of estimation and proportion, handily verifying my estimates when I enter them. And when I don’t track, I survive, mostly without worry, because I’ve internalised and made habit of my food behaviours. The notion of a pure originary sealed-off self felt false. As my project shifts from daily tracking and self-evaluations, and I continue to pursue personal goals, I realise that I can never truly divorce my personal and professional interests. I feel forever embroiled in an all-encompassing exploration of personal possibility and continuing improvement that pays no heed to disciplinary borders. CONTINUING IMPROVEMENT Yesterday, after the gym, the scale showed 152.0 lbs and 15% body fat. My weight has been hovering from 150 to 152 but with the fat level, it’s safe to say this is the trimmest I’ve ever been. It’s frightening to now reflect on what I carried around on my frame before. The task is now to keep keeping this weight off and to cement my progress with a good workout regimen. It’s going to be a long fight but hopefully a happy one. (fieldnote excerpt) After a year, I had lost 24lbs, but my will to improve continued and expanded beyond weight loss. On one morning, after some reading on the Internet, news of Anterior Pelvic Tilt1 arrived in my life. I wrote later: As my study’s initially projected run was coming to an end, my earlier fear of gazing too deeply at the messiness and 126 This is the weekend that I became aware, and seemingly a victim, of anterior pelvic tilt. The exercises to correct this have begun. Like many conditions, I doubt if I have it, though it would explain a great deal. And, as I’ve 1 A posture defect Figure 4. Tracked days: Both shades of green represent days that were tracked. Slices marked orange were days that I did not track. Days marked with a light green shade had their food entries edited retrospectively learned about other conditions, learning that you have something becomes an exhausting ride towards fixing it. (fieldnote excerpt) The universe of self-improvement was unfolding before me. Being healthy had become more than just avoiding obesity, sickness or early death. It had become what Rose refers to as the “optimization of one’s corporeality;” a model of fitness that is “simultaneously corporeal and psychological” [27]. As I came to revel, with some conflict, in my new-found selfunderstanding and weight loss, I also began to wonder what I couldn’t achieve given the right companion tools. I began weightlifting, tracking workouts and shifting my diet into performance mode; I experimented with tracking and optimising my sleep. Did coffee improve my focus? Wonder, track, and see. Track to peruse; then track to improve. This new model of fitness and health embraces what Rose refers to as “overall well-being,” inclusive of happiness, beauty, sexuality and more. What he concludes is that, spurred on by the privatisation of the health care industry and state efforts at health promotion, our personhood is increasingly being defined not just by ourselves, but by a complex of others; a complex that includes not just our immediate community but the application-creators, entrepreneurs and new media companies that are building on our desire to improve our selves. These collective decisions helped reshape how I perceived and contested my possibilities and limits. My algorithms and applications fit within the frame of this developing relationship with my self that was equally epistemological, attentive, and despotic [9]. By allowing insight into bodily processes, systems like mine increase our capacities and abilities: to critically assess “correct” functionality, to gauge inputs and outputs, and to help us regulate our processes or our models of them. At the same time though, as I felt, they encouraged our participation in networks of power that constrain us and might decrease our will or ability to exercise our capacities in the future [12]. That these networks of power are not just external (between us and the world) but also internal (between us and our selves) is what made a first-person evaluation so compelling yet simultaneously complex. In my case, losing weight by controlling diet and exercise required an ability to re-construct my body as obedient and submissive through attention to small details and newly revealed units of measure. This is a “relentless surveillance” [3] wherein I partitioned my body into new dimensions. Yet, through this control and restriction, I became aware of exactly what I consumed and its effects, and came to realise the positive embodied and ontological effects that can accompany changing patterns in my life [13]. And the accomplishment of those goals, even if those goals might have been subliminally directed by, and entangled with, a complex of exter- 127 nal sources—corporate or social—was still an enabling act of self-transformation that honed an ability to enhance capacities and to develop new skills. As my study progressed through its early phases, I had taken seriously Thoreau’s admonition to avoid becoming “the tools of our tools” or putting too much, and too often, between ourselves and a direct connection to the world through our intuition [31]. At the same time, I had to admit that as my self-tracking practice developed, it helped me increase my awareness and intuitive sense of my self and my world. That I had the advantage of a first person perspective is important to understand, but that I was party to an experience that transformed, over several months, the way in which I acted on, thought of, and was affected by the world through my body is equally important. Our bodies express our “double status as object and subject—as some-thing in the world and as a sensibility that experiences, feels, and acts in the world” [29]; yet the argument from that somaesthetic view is that the goal of objectivity and a nobler (in the enlightenment sense) version of humanity produces a one-sided focus on intellectual goals that marginalises studies of the body. But embodied considerations inform how we choose to form our existence, assess our quality of life, and make intellectual decisions. This is not just an anti-Cartesian position. Connecting with and attending to our bodies—even through these systems—might be a way to recuperate the kinds of aesthetic experience that can help us connect to others. My experience has shown me that such a path is far from straightforward. Criticisms of these technologies show how they, or at least the imaginaries that give birth to them, promote and cement a division of mind and body, and work to establish a self-legible body that is always at risk yet best addressed through objective “information” [32]. At the same time, even as a prosthetic, that information system expanded my ability to make sense of the world and in doing so increased my sense of freedom (see [17] for a related example). Even the idea of the prosthetic felt limited, and the disjunctions between body and mind, tool and human, felt hard to maintain. Data, device, and person at times felt instead transduced and interwoven, and enabling new possibilities, capabilities, and connections to emerge [19]. Transduced and transformed, not only can I not cleave the entanglements between my self as subject and researcher, I am no longer the same person who began this study. COMMUNING: THE IMAGINED BOUNDARY OF OUR SELVES Taking Thoreau’s notion of deliberate living seriously meant attending not only to the practical realities of my body and its immediate environment, but also to the social relations around me. For Thoreau, starting with the self was not a retreat but a strategy for re-engagement with the world. He wrote that “not till we are completely lost...not till we have lost the world, do we begin to find ourselves, and realise where we are and the infinite extent of our relations” [31]. My project changed my relationship to the world at large. In small ways, my conversations with those around me often centred on changes to my physical proportions, or changes to my diet. To others, my dietary choices and my measuring and data entry habits were, at best, curious, and, at their worst, alienating. Still, my admission of these quirks and personal challenges provided a common ground to discuss other struggles. They may not have been self-trackers or body-hackers but they all had goals they wanted to achieve. As I presented ongoing work and told friends and family about my approach, responses were varied. Two responses stood out. I was either pitied: seen as someone crippled through an embrace of technology, bound by a system of measures and restraints. Or I was envied, my control of my diet and exercise taken as signs of a far reaching mastery of my larger life (no such power exists). For those with whom I found similar ground, we were connected not because we used the same suite of tools, but because of a shared commitment to a subjunctive understanding of our systems. This was a view of the tools at our disposal as a range of approaches that provided craggy but passable conduits to an old self, a new self, or even a better grasp of a current self. Focusing on my personal, experiential knowledge provided a starting point for those conversations. Thoreau’s pursuit of self-knowledge had become a vehicle for self confrontation and a vantage point from which to confront his world with fresh perspective. His deliberate stripping down of his life and activities was done in order to recreate a more empathetic one, and he built to political action through this alternating retreat from and engagement with society. His concern was the development of a capacity to improve ourselves and our world. Our role as citizens to engage a wider culture can begin from movements both big, and in my case, small. I believe a cycling “retreat” into first-person studies is a similar and valid methodology for engagement with particular communities, one that fits within the frame and support demonstrated by third-person and structural approaches such as [11, 25, 6]. My evolving relationship with my self and system has required me to be more critical of how I imagine my role. Is it as a critic and “defender”? Is it as a willing (and perhaps pitied) general participant? Is it as someone attempting a fused role somewhere between? As critics, if we decide to defend users against technology and social formations that we see as bad, where do we position ourselves? Are we part of the laity? Or are we standing apart, too savvy to be caught in the same traps? Does our writing, strategies, and commitments reflect that we too are in liminal positions? In [30], Taylor calls for us to view technology and its design as a “means of participating in unfolding ways of knowing, being and doing” rather than standing back here reporting on others out there. Might deeply personal accounts help us focus on what he refers to as the mutuality of our “unfolding enactments of ordering, classifying, producing and ultimately designing technology”? When I began my project, I knew nothing of the Quantified Self (QS), a community of users and makers of self-tracking tools2 perhaps best known through its stated commitment to “self-knowledge through numbers.” And yet, 2 years later, with only slightly more of a clue, I found myself giving a talk and running a workshop at an annual QS conference. In the years since I began this study, popular conversation and critical discourse on the QS movement in particular, and on the wider practices of self-tracking in general, have increased, while its related devices and approaches have entered the mainstream. It’s now commonplace to meet people equipped with some form of automated personal tracking clipped to waist or wrist. An outsider view of this community might likely mirror the view that others sometimes take of me, and that I had initially taken of it: as narcissistic or self-indulgent; as trapped in a corporate designer’s ideology; or as a member of a cabal determined to increase rationalisation in our lives and to make data of our bodies. It was perhaps too easy for me as a critic to see my charge as rescuing the naive from the great evil of such an imagined neo-liberalist scheme of big data, surveillance, and individualising institutions. Of course, there are reasons to decry and worry about such an encroaching future of personal panopticons, and individuals alienated from supporting institutions that are divesting themselves from social responsibility for issues of health. But instead in QS I found positive possibilities among selves and other—where other might be a device, a database, or a community of people. And it helped me see the boundaries that I initially considered between myself and my devices, and between myself and others, as imaginary and limiting. I have now begun to rethink tracking practice as something more akin to a cyborg communion: with devices, others, and self. Does it stretch the bounds of the scholarly tent too far to think of us as bridged by data, or as connected by a joint recognition and quest for the sublime? I don’t mean to paint the pursuit of personal informatics as a religious one, although I’m sure it has its zealots. And I don’t maintain that all or even many members of the Quantified Self community would agree with my characterisation. In fact, my presence and that of other critics who wished to inject more reflection on QS’s effect on the wider public provoked some understandably negative reactions. After all, why should individuals in a community have to worry about the consequences that might spiral indirectly from their individual practice? I’ve wondered the same. Friends and family have now followed my lead, tracking activity and intake. They’ve seen great results but I worry that I’ve caught them in a subjunctive net that I can give little advice on how to fully escape. What I found QS to be was markedly different from the majority if the popular media and critical academic discourse about the movement. Their relationship with data was pro2 128 http://quantifiedself.com/about/ foundly different from what I had come to expect, yet profoundly similar to how I felt. Our questions were similar: what data is important? Who gets to say so? What does it mean for me? In a breakout session I led—provocatively titled Can Data Make Us More Human?—I asked what values and considerations were important to the participants. The responses were instructive: ease of use, transformation, openness, freedom, honesty and directness, reflexivity, presence, constancy, comfort, anchor, exchange, to find better questions, to create my own narrative, how do I want to live? objectivity, to make order, tangibility, discipline, selfcontrol, autonomy, diversity, transparency, accountability, value-free, judgement-free, guilt-free, living better in the moment, diversity, fragility, empathy world” rather than a reliance on rigid distinctions between the mechanics of capture and the formation of knowledge. The alternative—“the possibility of pure transport”—is an illusion: we cannot escape the methods we use to frame and capture our data, nor can we as self-trackers “ever be quite the same on arrival at a place as when [we] set out.” My earlier understanding of the community and my own developing practice as a collection of numbers missed this rich ontological designing, and saw transport in attempts at wayfaring. CONCLUSION I had a single, simple goal when I began this project: to lose some weight. To achieve this goal, I brought a range of systems within the limits of my self, and we made more goals in collaboration; mutually designed as we were designing. At various points in this journey I felt lost: out of touch with my numbers, my narrative, and even my collective. But I found that losing myself was essential to continuing onward. Nafus and Sherman [22] argue that QS practices are actually “soft resistance” to institutional models of producing and living with data: “a readiness to evolve what constitutes meaning as it unfolds.” With this form of resistance, they find that QS participants “work to dismantle the categories that make traditional aggregations appear authoritative”. Using systems like Lose It!, that were built with seemingly helpful badges, gamification, and databases flush with particular kinds of food, helped me understand firsthand those struggles of being enrolled in, while simultaneously wrestling with, embedded ideologies of how best to monitor and manage one’s health and person. The community members I met were in a similar struggle, also attempting to work around systems largely designed to meet purposes with which they did not completely identify. This was not a site where self-authority was being ceded nor one where a joint search for an objective, external locus of meaning making was being launched. It was a place where data and devices were being used to reflect on what may or may not be happening inside our bodies, and to insist on the “idiosyncrasy of individual bodies and psyches” [22]. This difference echoes Ingold’s [15] distinction between wayfaring and transport. For Ingold, transport is distinguished from wayfaring not through the use of devices (mechanical means) but by what he refers to as the “dissolution of the intimate bond that in wayfaring couples locomotion and perception.” The transported traveller is a passenger, one who “does not himself move but is rather moved from place to place,” while the wayfarer goes along and threads a way through the world rather than routing across the surface from point to point. This leads to a fundamental difference in the integration of knowledge, where wayfarers “grow into knowledge of the world about them” while in transport, knowledge is upwardly integrated. The QS discussions and presentations that I’ve been privy to are less about reading results and more about following a trail where participants, following Ingold, know as they go: sharing an always-bound trifecta of what they did, how they did it, and what they learned. This a way of knowing which reflects what he refers to as “a path of movement through the 129 The use of my self as subject was an important vehicle for getting close to the difficulty of mapping and communicating such an experience: a model of knowledge acquisition equally epistemological and existential. Knowing logically what to expect during this project did little to prepare me for the daily mixture of sadness, elation, and fear that I experienced. I knew that numbers, and the intimate devices that deliver them, have a powerful force on us, but to experience that somehow-embodied tug was important. And that I did this with personal and professional consequences at stake is a productive complication that I am still unraveling. The last years have been more than an academic research opportunity. They also represented an important effort and transformational experience linked to my own ideas of self-worth. My particular choices helped me reveal and begin to come to terms with the nuances of an internally conflicted position. Long-term usage coupled with an existing personal need modulated my position on health and fitness technologies, confirming some of my criticisms but substantially reframing my prior estimations. While I remain cautious of the approaches of many of these systems, I learned that my conflicted position need not be hypocritical, and reflects the joys and capabilities that can be found in reduction, and the freedom to be found in control. I had been wary of the attractions of these technologies, and found that confirmed in my connection to them; I was not in a comfortable space of ironic distance. Many of the theoretical issues that concerned me prior to my project, such as the reductionism inherent in a calorie-based point of view and the limitations of the modelled “truths” that these technologies are predicated on, remained with me throughout, but were experienced anew. My intent has not been to present a thorough record of this journey but to demonstrate a struggle to make sense of it. What I tried to provide was a form of extensive, genuine usage that responds to what [23] refers to as “detailed, nuanced, and experiential understanding of a design space.” But this is a design space in which I exist as subject, object, and verb—a perspective that, at once privileged and problematic, elastically collapses subject and researcher. Through the essay I tried to make two main contributions (topical, methodological), but without distinct separations of section which I feared would too easily mirror these binaries against which the essay also labours. Given the underlying approach (autoethnography, research through an ontological design), only discussing the method without showing it in action would likely not make a compelling argument, nor would deploying it without also arguing significantly for its validity. As such, this essay leaves much unresolved. In this though, it is faithful to my underlying experience, and responsive to the calls to action in [20] and [30]. 2. Bardzell, J. HCI and the Essay: Taking on “Layers and Layers” of Meaning. In CHI 2010 Workshop on Critical Dialogue (Sept. 2010). I hope that writing about this project in a first person, autobiographical voice helps to fulfil a commitment to support a “fuller range and richness of lived experiences” [4], wherein we draw on aesthetic principles and techniques [1, 21, 35] to achieve that goal. As a method for both understanding others and the self, I believe this voice marries well with a desire to deeply describe and engage with those moral, political, emotional and ineffable experiences in our work. That said, the value for me has not been in whether I could reveal different things than a more traditional third-person approach might, but whether I might be allowed a look at the same things in a different way [3, 26]. 5. Bourdieu, P. Distinction: a social critique of the judgement of taste. Harvard University Press, Cambridge, Mass., 1984. I don’t assume though that my experiences will be the same as others. I hope to read other accounts that will be markedly different. This has been an unabashedly person-centred attempt to present a perspective on these systems and communities that is not common within HCI, and in doing so to hopefully expand its tent. I tried to present as much of an inner view possible but that view is inescapably complex, obscured, and contradictory. Instead of attempting to transform an idiosyncratically personal account into a directly generalizable one, I opted to draw out and demonstrate a connection between a personal and embodied foundation, and a broader politically and socially engaged position—one that should not require us to gloss over or reduce our individual distinctions. 8. Efimova, L. Weblog as a personal thinking space. In Proceedings of the 20th ACM Conference on Hypertext and Hypermedia, HT ’09, ACM (New York, NY, USA, 2009), 289–298. It may be that to live with these technologies is to be necessarily—not just until the version next—in alliances marked by these capabilities and productive contradictions, and by the frequent negotiating and renegotiating of perspectives and boundaries. In this, they are perhaps but another set of relationships among many others in our lives; a collection of others that have always been expanding the limits of our selves. ACKNOWLEDGMENTS This work was funded through the Intel Science & Technology Centre for Social Computing, and NSF Grant IIS1217685. I’m particularly grateful for the thoughtful feedback and support of Phoebe Sengers over the years of the study, and to the helpful comments, criticisms, and suggestions of the anonymous reviewers. REFERENCES 3. Bartky, S. L. Toward a Phenomenology of Feminist Consciousness. Social Theory and Practice 3, 4 (1975), 425–439. 4. Boehner, K., Sengers, P., and Warner, S. Interfaces with the ineffable: Meeting aesthetic experience on its own terms. ACM Trans. Comput.-Hum. Interact. 15, 3 (Dec. 2008), 12:1–12:29. 6. Brynjarsdottir, H., Håkansson, M., Pierce, J., Baumer, E., DiSalvo, C., and Sengers, P. Sustainably unpersuaded: How persuasion narrows our vision of sustainability. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’12, ACM (New York, NY, USA, 2012), 947–956. 7. Dewey, J. Experience and education. The Macmillan company, New York, 1938. 9. Foucault, M., Martin, L. H., Gutman, H., and Hutton, P. H. Technologies of the self : a seminar with Michel Foucault. University of Massachusetts Press, Amherst, 1988. 10. Gaver, W. What should we expect from research through design? In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’12, ACM (New York, NY, USA, 2012), 937–946. 11. Grimes, A., Tan, D., and Morris, D. Toward technologies that support family reflections on health. In Proceedings of the ACM 2009 International Conference on Supporting Group Work, GROUP ’09, ACM (New York, NY, USA, 2009), 311–320. 12. Heyes, C. J. Foucault goes to weight watchers. Hypatia 21, 2 (2006), 126–149. 13. Heyes, C. J. Self transformations: Foucault, ethics, and normalized bodies. Oxford University Press, Oxford, 2007. 14. Höök, K. Transferring qualities from horseback riding to design. In Proceedings of the 6th Nordic Conference on Human-Computer Interaction: Extending Boundaries, NordiCHI ’10, ACM (New York, NY, USA, 2010), 226–235. 15. Ingold, T. Lines: a brief history. Routledge, 2007, ch. Up, across and along, 72–103. 1. Bardzell, J. Interaction criticism and aesthetics. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’09, ACM (New York, NY, USA, 2009), 2357–2366. 16. Latour, B. We have never been modern. Harvard University Press, Cambridge, MA, 1993. 130 17. Leshed, G., Velden, T., Rieger, O., Kot, B., and Sengers, P. In-car gps navigation: Engagement with and disengagement from the environment. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’08, ACM (New York, NY, USA, 2008), 1675–1684. 31. Thoreau, H. D. Walden, Or, Life in the Woods. Penn State Electronic Classics Series. The Pennsylvania State University, 2006. 18. Ljungblad, S. Passive photography from a creative perspective: ”if i would just shoot the same thing for seven days, it’s like... what’s the point?”. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’09, ACM (New York, NY, USA, 2009), 829–838. 19. Mackenzie, A. Transductions: bodies and machines at speed. Technologies. Continuum, London, 2006. 20. McCarthy, J., and Wright, P. Technology as experience. MIT Press, Cambridge, Mass., 2004. 32. Viseu, A., and Suchman, L. Wearable Augmentations: Imaginaries of the Informed Body. In Technologized Images, Technologized Bodies, J. Edwards, P. Harvey, and P. Wade, Eds. Berghahn, 2010. 33. Willis, A.-M. Ontological designing: Laying the ground. In Design Philosophy Papers Collection Three. Ravensbourne (Qld, Aust): Team D/E/S Publications, 2007, 80–98. 34. Winograd, T., and Flores, F. Understanding computers and cognition: a new foundation for design. Ablex Pub. Corp., Norwood, N.J., 1986. 35. Wright, P., and McCarthy, J. The value of the novel in designing for experience. Future Interaction Design. Springer (2005), 9–30. 21. McCarthy, J., and Wright, P. Putting ‘felt-life’ at the centre of human–computer interaction (hci). Cognition, Technology & Work 7, 4 (2005), 262–271. 36. Wright, P., and McCarthy, J. Empathy and experience in hci. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’08, ACM (New York, NY, USA, 2008), 637–646. 22. Nafus, D., and Sherman, J. Big data, big questions — this one does not go up to 11: The quantified self movement as an alternative big data practice. International Journal of Communication 8, 0 (2014). 37. Zimmerman, J., Forlizzi, J., and Evenson, S. Research through design as a method for interaction design research in hci. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’07, ACM (New York, NY, USA, 2007), 493–502. 23. Neustaedter, C., and Sengers, P. Autobiographical design in hci research: Designing and learning through use-it-yourself. In Proceedings of the Designing Interactive Systems Conference, DIS ’12, ACM (New York, NY, USA, 2012), 514–523. 24. Perec, G. Species of spaces and other pieces. Penguin Books, London, England; New York, N.Y., USA, 1997. 25. Purpura, S., Schwanda, V., Williams, K., Stubler, W., and Sengers, P. Fit4life: The design of a persuasive technology promoting healthy behavior and ideal weight. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’11, ACM (New York, NY, USA, 2011), 423–432. 26. Reed-Danahay, D. Auto/ethnography: rewriting the self and the social. Berg, Oxford, 1997. 27. Rose, N. The politics of life itself. Theory, Culture & Society 18, 6 (2001), 1–30. 28. Schüll, N. D. The folly of technological solutionism: An interview with evgeny morozov. http://www.publicbooks.org/interviews/the-folly-oftechnological-solutionism-an-interview-with-evgenymorozov, September 2013. 29. Shusterman, R. Thinking through the body, educating for the humanities: A plea for somaesthetics. The journal of aesthetic education 40, 1 (2006), 1–21. 30. Taylor, A. S. Out there. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’11, ACM (New York, NY, USA, 2011), 685–694. 131 ! 132 The User Reconfigured: On Subjectivities of Information Jeffrey Bardzell Informatics and Computing Indiana University Bloomington, IN 47408 jbardzel@indiana.edu Shaowen Bardzell Informatics and Computing Indiana University Bloomington, IN 47408 selu@indiana.edu theorizing subjectivity in the humanities. Using critical and empirical methods, we explore sexual subjectivity as designed, and reflect back on the benefits and limits of subjectivities of information as a formulation of computing use. ABSTRACT Foundational to HCI is the notion of “the user.” Whether a cognitive processor, social actor, consumer, or even a nonuser, the user in HCI has always been as much a technical construct as actual people using systems. We explore an emerging formulation of the user—the subjectivity of information—by laying out what it means and why researchers are being drawn to it. We then use it to guide a case study of a relatively marginal use of computing—digitally mediated sexuality—to holistically explore design in relation to embodiment, tactual experience, sociability, power, ideology, selfhood, and activism. We argue that subjectivities of information clarifies the relationships between design choices and embodied experiences, ways that designers design users and not just products, and ways to cultivate and transform, rather than merely support, human agency. USER AND SUBJECT Although at first blush “the user” might seem to be very obvious in its meaning—it’s the person who is using a system!—in fact HCI researchers and practitioners operate with evolving understandings that have been informed for years by technical understandings, grounded in epistemological, ethical, and methodological commitments. Even the following brief survey makes this clear. In their 1983 classic, Card, Moran, and Newell [8] characterize the user thus: a scientific psychology should help us in arranging this interface so it is easy, efficient, error-free—even enjoyable…. The key notion is that the user and the computer engage in a communicative dialogue whose purpose is the accomplishment of some task…. The human mind is also an information processing system. (pp.1, 4, 24) Author Keywords User; subjectivity; sexuality; theory; criticism ACM Classification Keywords H5.m. Information systems: Miscellaneous. In 1985 Gould and Lewis [16] proposed a three-part methodology for systems design: an early focus on users, empirical measurements, and iterative design. And in 1988, Norman [26] advised, “make sure that (1) the user can figure out what to do, and (2) the user can tell what is going on” (p.188). Common to these is the user understood as an individual completing a task with a system, where the human is abstracted as a cognitive processor well defined needs; tasks are understood as behavioral sequences that can be explicitly defined; and interaction is understood as a dialogue between this cognitive processer and the task needing to be done—and all of the above were available to and measurable by empirical methods. INTRODUCTION A decennial conference beckons us to reflect with a wider perspective than our everyday research often affords. We reflect in this paper on the ways that HCI has theorized the user. We begin in the early 1980s, when notions of “usability” and “user-centered design” were first consolidating a notion of the user—one that would quickly be criticized and reworked, beginning an ongoing activity that continues today. We will turn our attention to an emerging formulation in HCI of the user that is informed by recent philosophy: the user as a subjectivity of information. After characterizing what this formulation means, we then seek to test it. We do so by investigating a marginal use of computing, digitally mediated sexuality. We choose this use because it brings onto the same stage a diverse range of concerns that HCI is increasingly seeking to take on: design, embodiment, tactual experience, sociability, power, ideology, selfhood, and activism. Sexuality has also centrally figured in But it was not long before these formulations were critiqued and expanded. In 1986 and 1987, respectively, Winograd and Flores [46] and Suchman [40] challenged the view of the human constructed by these earlier notions of the user, leveraging philosophy to rethink what it means to act in a situation. Bannon & Bødker [1] wrote that in classic HCI, the user had been “hacked into” a “disembodied ratiocinator,” referring to cognitive modeling disengaged from social and material praxis. Kuutti [22] and Cockton [10] would later trace three HCI “users” emerging in the research discourse from the 1980s to the 2000s: the user as source of error, as social actor, and as consumer, broadly corresponding to changes in computing paradigms over the decades, Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21298 133 from one-user-one-terminal systems, towards systems supporting workplace cooperation, and finally towards experience-oriented systems. More recent work has added the “non-user” to the mix as well [31]. requiring another: teacher-student, mother-daughter, policecitizen, designer-user. Subjectivities are more diverse and situated: two students can occupy the same subject position and yet perform distinctly different behaviors and attitudes. A subject position thus constrains possible behaviors, attitudes, etc., but it also provides generative “rules” for creative, stylized, and individual performances within them. Any individual occupies many different subject positions: a schoolchild is also a daughter, a citizen, a goalie, etc. Subject positions can come into conflict, thus shaping (and helping explain) how an individual negotiates them. More radically, Cooper & Bowers [11] analyzed “the user” as a discursive structure, rather than an actual person: The representation of the user [in HCI discourses] can be analytically broken down into three constituent and mutually dependent moments [as follows…] In the first, the user is created as a new discursive object that is to be HCI’s concern […] In the second, HCI tends to construct cognitive representations of the user […] In the third, HCI represents the user in the political sense of ‘representation’: as Landauer puts it, people within HCI should be the ‘advocates of the interests of users’. [p.50] We can apply this concept to computer use. Common IT subject positions include power users, novices, early adopters, and non-users; social actors; gamers, trolls, and n00bs; standard users and admins; expert amateurs and everyday designers; quantified selves; makers and hackers; social media friends; and many more. It is clear that all of these positions comprise both technical and social roles. The concept of subjectivities gets at how these roles are diversely enacted, embodied, or performed. For example, we understand how step- and calorie-counting apps construct quantified self subject positions, but how those apps are subjectively experienced (e.g., as a source of empowerment, anxiety, and/or curiosity) and how they are enacted or stylized (e.g., integrated into one’s lifestyle, displayed and/or hidden from the view of others) is a different type of question. Subject positions are analytically available to research, while subjectivities are empirically available. The present work continues this thinking, analyzing the user in relation to abstract structural positions or roles that are inhabited by actual people using systems in real situations. Situated Actors and/or Structured Roles The key conceptual move is to distinguish the user qua an individual person—an embodied actor in a given situation—from the user qua structural roles, with anticipated needs and capabilities. We make this move because the former—the user qua embodied individual—maps (perhaps too intuitively) onto everyday notions of persons as already given: that is, that the individual is already there, and our job as designers is to support that individual in achieving a need. What the latter—the user qua structural role—gives us is not only a type of individual whom we can support, but also an analytic means to think about designing the role itself, i.e., designing not just the system but also the user. (We believe that designers already design both systems and users, but some HCI theory, fusing with ordinary language notions of personhood, obfuscates that we also design the user—consider, again, how most will parse Norman’s dictum: “make sure that the user can figure out what to do….”) We understand subjectivity of information to refer to any combination of (structural) subject positions and (felt and performed) subjectivities relevant to the context of information technology use. We do not mean to reify a subjectivity of information as somehow distinct or independent from other subjectivities. We simply understand the concept in its usual theoretical sense, applied in the domain of IT design and use. That is, we seek to understand “users” insofar as they are structurally subjected to certain IT systems and associated social practices, and how they become subjects of (i.e., agents of) those systems and social practices. These are analytically distinct from “users” as commonsense persons because a given person can inhabit multiple subject positions and can experience/perform them not only differently from other individuals, but even differently from themselves at different times or in different situations. Our argument is that to resist the analytical conflation of situated actors and discursive constructs into “the user,” that we should analyze “the user” as a “subjectivity of information,” that is, structural roles that are designable and designed, which users-as-situated-actors take up and perform. In what follows, we will argue that such a view has practical advantages for HCI researchers and designers, as well as important methodological implications. We argue that the notion of subjectivities of information has three benefits. It can give us a tighter analytic coupling between specific design decisions and particular social experiences, because the design decisions and social experiences can both be analyzed using an analytic vocabulary common to both, i.e., the structures that comprise that subject position, understood as an amalgam of computational rules and social structures, including rights and responsibilities, laws, mores, etc. Second, it makes clear that we can (and do) design subjects as well as interfaces, products, and services (while recalling that the parallel concept of subjec- Subjectivities of Information The concept of subjectivity is derived from the philosophy of Lacan, Althusser, Foucault, Delueze & Guattari, Butler, and others. Two related concepts are at the core of this theory: the notion of subject positions, which are social roles that people are thrust into, and subjectivity, which is the felt experience and creative agency of individuals within that situation. Subject positions are typically relational, that is, 134 tivities means that individuals have plenty of agency beyond what we can prescribe as subject positions). And finally, such a view supports design practices aimed at cultivating and transforming, rather than merely supporting or extending, human agency; one use of subjectivity in critical theory has been to serve activist purposes, to imagine and pursue the transformation of society. the way she did. But as a feminist, Radway must acknowledge and contend with the value that actual women find in these texts. Radway’s dilemma is HCI’s as well: if researchers cede the authority of identifying the sociotechnical significances of systems to actual users, we risk overlooking to the very subtle roles of ideology and false pleasure in systems, an approach that is ultimately regressive. If HCI researchers instead become more active critics, we risk speaking for users, substituting our own values for theirs. Thus we believe something akin to the methodology that Radway sought to undertake, which stages a dialogue between equally weighted critical and empirical studies of the same phenomenon, might be an appropriate strategy for investigating the meanings of interactive products—and the subject positions and subjectivities inscribed within them. Methodologies for Information Subjectivities The user as subjectivity of information has methodological implications, because “the subject” is shaped by sociotechnical structures that include individual algorithms and system structures (e.g., permissions, task sequences), and also social structures, including shared practices, intersubjective understandings, policies, and broader sociological structures including gender, racial, and class ideologies. As such, the subject is not a simple given, and cannot be rendered available to understanding merely with data, but rather it must be interpreted, a critical activity that makes sense of multiple sources, including systems, empirical studies of users, and analysis of data—all of which can be supported by relevant theories. One methodological challenge, then, becomes how we weigh differing theoretical, data, and critique-based resources in interpreting the subject. A CRITICAL-EMPIRICAL STUDY OF SOCIOTECHNICAL SEXUAL SUBJECTS The shift to envisioning users as sociotechnical subjects is anything but new: it is the outcome of a generation of research and theory throughout HCI, explored in CSCW, participatory design, in traditional user research, and so on (see e.g., [12]; the term “subjectivities of information” itself was coined by Paul Dourish). Because we see analysis of sociotechnical subjects as involving (at least) three planes of meaning (i.e., the values and subject positions inscribed in designs; the skilled practices of actual users; and the critique of domain experts), we undertook a research initiative that engages all three. Adapting from feminist film theory [43], we can distinguish three sources of the subject: cultural objects (i.e., interactive technologies for our purposes) that express expectations about their ideal user who will get the most out of them; actual users, who have their own experiences, purposes, and perspectives about their use of an object; and researchers/critics, who offer a trained expert perspective on objects. Most research only accesses one or two of these positions. For example, if a work of interaction criticism [2,5] asserts that a system manipulates users, two of the three positions are active—the critic and the inscribed user—but not the second (actual users). When HCI researchers study the actual aesthetic judgments that users make (e.g., [41]), they are activating the second but not the first and third positions. The research focuses on a paradigmatic change that began about a decade ago in sex toy design. Prior to it, sex toy design was linked to the porn industry, marketed to men, manufactured cheaply, and called “novelty products” to avoid regulation under medical standards. Since then, the industry has been transformed by designers with backgrounds in high-end industrial design, including Apple, Ericsson, Frog Design, and so on; the toys are manufactured according to the standards of modern consumer electronics (e.g., MP3 players) and in most cases using medical grade, body-safe materials; they are marketed to women as luxury accessories; their price point is generally higher and in some cases much higher; and they align themselves with feminist sex positive activists who seek to transform the meanings and consequences of sexual body practices (see [3]). Broader cultural meanings of sex toys also changed in this time, with episodes of television shows, such as Sex & the City, presenting them in a positive light. Once we acknowledge these three positions, another research challenge comes into focus: they often do not agree. The “actual” meaning of the design, then, must be some negotiation of these different positions. A methodological strategy proposed by feminist researchers to study subjectivity has been to deploy a critical-empirical approach. For example, in a landmark study of romance novels, feminist literary critic Janice Radway [29] read the romance novel genre as patriarchal (because it offers narratives that reassure and satisfy women while preserving an unequal status quo), and then she ethnographically studied working class women readers. Not surprisingly, their literary understandings did not agree. The dilemma for Radway, then, is this: as a scholar trained in feminist theory, Radway is in a position to see “false pleasures” that everyday readers may not be able or want to see; one cannot dismiss Radway’s reading simply because the everyday readers did not see things Our overall research program on sex and HCI has included an interview study with sex toy designers and feminist sex activists, critical analysis of sex toys, and empirical studies of digitally mediated sexual life in virtual worlds, on social media, and with digital sex toys. Some of this research has been reported elsewhere [3], so we focus here on a criticalempirical study of sex toys and their sexual subjectivities. 135 In conducting this research, we are able to focus on the interrelations of critical and empirical epistemologies for design and user research, and we use this as an opportunity to better understood subjectivity holistically in its designed, technological, embodied, political, and social dimensions. trix of interpersonal humanity and responsibility…. Erotic experience is experiencing the other as a person in and through experiencing oneself as a person. The epiphany of the other is the foundation of ethics [30, pp.76, 85] Sexual subjectivity. Part of the process of socio-sexual maturity is acquiring skills of sexual perception. Sexual perception is skilled, because it must be learned, and it is mediated by our learning of related conceptual vocabularies [39]. For example, a discriminating palette for tasting wine requires not just tasting many different wines but also developing a conceptual vocabulary to name the distinctions that one learns to perceive (e.g., notes of oak, chalkiness, cherry; the mouthfeel, the finish). For Shusterman, sensory pleasure, aesthetic appreciation, and the formation of the self are all contingent on one’s skill with one’s body. We develop our body skills through training, understood as “actually engaging in programs of disciplined, reflective, corporeal practice aimed at somatic self-improvement” [37], including diets, exercise programs, and erotic arts. From Cosmo to The Joy of Sex, there is plenty of evidence that millions cultivate their sexual abilities. Theory Background: Sex and the Subject Studies of sexuality in HCI have increased since 2005, albeit sometimes in the face of resistance, which, along with dominant social norms, often wants to bracket sexuality aside irrelevant to computing, usability, and software design, except for very narrowly construed sexual technologies. But sexuality researchers in health sciences [18], sociology [6,33,38], philosophy [4,30,36], and the fine arts [45] offer a different point of view, one in which sexuality is not compartmentalized but rather is a basic aspect of our social life, body knowledge, and body habits. Attention to sexuality allows us to recuperate “a range of often discounted or forgotten social actors, movements, and landscapes” [15]. Our research program leverages two related ideas from this literature and explore their applicability to HCI: sexual sociability and sexual subjectivity. Sexuality research helps us understand the interrelations among a complex set of tactual and social skills, body-body interactions, and body habits. HCI, CSCW, and ubiquitous computing research is already taking up such themes. To them we add that the embodied subject of interaction is gendered and sexual all the time, not just in the bedroom. It is only through such a formulation that we can begin to hear the voices of sexual subjects: “Those who have been marginalized, oppressed and pathologized begin to speak, begin to make their own demands, begin to describe their own lives, not in a spirit of scientific objectivity, but with passion, indignation, and exhalation” [18, p.187]. Sexual sociability. Sexuality is not merely a private, but it also has an outward-turning dimension. Yet, physical pleasures are commonly assumed to be private and even antisocial. Philosopher Richard Shusterman challenges this view: “Most pleasure,” he writes, “does not have the character of a specific, narrowly localized body feeling (unlike a toothache or stubbed toe). The pleasure of playing tennis cannot be identified in one’s running feet, beating heart, or swing of the racket hand” [36, p.41]. Shusterman adds that “Feeling bien dans sa peau can make us more comfortably open in dealing with others” [36, p. 41]. Shusterman tightly couples physical pleasure and social competence as inseparable aspects of human body practices, and he concludes that we should cultivate, rather than repress, our capacity to experience socio-physical pleasure. This contrasts with traditional morality, in which sexual pleasures are seen as selfish and antisocial, exemplified by the myth, frequently critiqued by sex educators, that vibrators are fine for when one is alone, but not when one is in a relationship [34]. On the sociability of sexuality, psychotherapist Horrocks writes, Critical-Empirical Methodology For this study, we deployed both a critical analysis of sex toys and an empirical investigation of how different people react to them, partly in hopes of better understanding how these two planes of meaning related to one another, and specifically to understand how sexual subjectivities are projected by, and how they contribute to an understanding of, these toys. I construct my sexual identity not simply to obtain pleasure or love, but also to communicate who I am…. But I do not exist simply as an individual: the reason I am able to use sexuality in this complex manner is because I take part in a socialized sexual system” [18, p.191] Critical analysis. We offer an expert analysis of five of the toys used in the empirical portion of the study, three recognized as in the “designerly” paradigm and two from the earlier paradigm. From that analysis we produce an interpretation that helps us understand the ways that this new generation of sex toys inscribed within its toys a new (to sex toys, if not to discourse) sexual subjectivity. We use critique in the same ways it is used in the humanities, i.e., as an interpretative activity that seeks to accomplish some combination of the following: to provide a value judgment supported with reasons and evidence and to help a particular audience or public appreciate (dis)value in works of the cultural domain [9]; to reveal values and assumptions inscribed This social view of sex is mainstream in the scholarly community. It stresses that characteristics deemed “natural” about sex are in fact historically contingent and hegemonic [33]. The scholarly view has deep ethical implications: [In] erotic life, we experience our bodies as the locus of the communication between us, and the site of our answerability to each other’s perspective…. Sexuality is not ‘animal,’ and it is not ‘amoral.’ It is … the originary ma- 136 in works and subject them to interrogation [19]; to account for a work’s most general organization and the ways that it “presents its subject matter as a focus for thought and attention” [13]. We refined this general activity to generate a set of critical prompts (summarized below), by leveraging the sexuality and subjectivity theories introduced above. Empirical analysis. This portion of the study was labbased, in which subjects interacted with 15 different sex toys chosen, in consultation with an AASECT-certified sex educator, to represent the range of toys available, when the toys were purchased. Subjects participated in two activities: a self-guided talk-aloud exploration of the 15 toys over the course of roughly 20-30 minutes, followed up by a semistructured interview lasting 15-30 minutes. 17 study sessions were conducted with a total of 25 subjects in 9 individual sessions and 8 done in pairs, at their preference. The subject population consists of 9 male and 16 female, 22 to 33 years of age. 11 of the subjects identify themselves as heterosexual, 10 as homosexual, and 3 as queer/other, and 22 out of the 25 total subjects have prior experience with sex toys. Sessions were video-recorded with subjects’ permission; afterwards, the audio from the video sessions was transcribed, and videos were used to annotate the transcriptions to clarify what was going on (e.g., which toy a participant was talking about when not obvious from the audio). Figure 1. The five toys that were critiqued granted or constrained; how do toys assert a relationship to the human body; what demands on perception does the toy make; what subject positions are implied? What sociocultural contingencies are expressed as norms in these toys? What would likely happen if a friend or parent discovered it; in what ways does the toy reflect, reify, or resist attitudes toward sex and issues of sexual social justice; in what ways are social relationships inscribed? The five toys we analyzed (Fig. 1) are from left to right, the OhMiBod Freestyle, a wireless vibrator that converts MP3 audio into vibration patterns; the Jimmyjane Form 6, a jelly bean-shaped vibrator; the VixSkin Johnny, a non-vibrating dildo that simulates in detail a penis; the Maximus Power Stoker Viper, a vibrating penis sleeve; and Nomi Tang’s Better Than Chocolate, a clitoral vibrator with a touchsensitive UI. Following industry experts, we divide these toys into two groups. Freestyle, Form 6, and Better Than Chocolate all represent the “designerly” turn in sex toys, which brought about the dramatic change in the industry. Johnny and Maximus represent the earlier era. They range from USD $90-$185, so all might be seen as on the highend of the mainstream commercial spectrum. We acknowledge that lab-based interactions with 15 sex toys do not provide the same insight that in-the-wild-based interactions would. Because we obviously could not observe and interview participants during sex with each of the toys, we pursued this as a practical alternative. That said, given the expansive view of sexuality that we have just summarized—one that comprises much more than actual sexual acts and sees it instead as a basis for sociality and perception—we also believed that participants would not “turn off” their sexualities in the lab. That many of them brought their sexual partners, used language of sexual attraction/repulsion during the study, and linked their perceptions of these toys with prior experiences—all provide evidence that this is true. Moreover, the Kinsey Institute—one of the most renowned sexuality research labs in the world— conducts most of its studies in labs, for the same reasons. So while we do not provide data concerning reports of sexual acts with these toys, we do have reason to believe that our subjects engaged these toys as sexual subjects. Group 1: “Designerly” Sex Toys OhMiBod Freestyle. The Freestyle’s distinctive feature is its wireless connectivity to an iPod, so that the audio signals from the iPod can be converted into vibrations: the idea is that the user masturbates with it to music. Its physical form is abstract and inorganic. Its shape is geometrically cylindrical, and it is symmetrical along its axes. At 21cm, it is considerably longer than the average human penis length, which is 12-13cm. The silver band that visually separates the handle end (i.e., the end that one holds, which also features the UI and the electrical jack) has no anatomic analogue, and it is clearly branded and contains UI icons of its own. The visual language of the Freestyle is that of a consumer electronic at least as much as it is of the penis. THE CRITICAL ANALYSIS Our critical goal in this project is to offer an interpretation about how designs propose relationships between technology and the body-as-subject in sexually specific ways— organized around a sexual subjectivity. To interpret the designs on these terms, we began our analyses of each toy with three orienting questions: How might one characterize this as an object? What are the product features, semantics, functions; what does this design claim as pleasure and what does it ask users to give up; in what ways is it precise? How does the product design subjects? What agencies are The penis-like form, its size, and the iPod connectivity all suggest that OhMiBod is targeted at a solo female—solo because of the personal music player, female because of its pink color and form suitable for clitoral or vaginal massage. It literalizes cultural links between pop music and sexuality. The success of the device largely depends on the music one selects; beyond personal preferences, we noticed that music 137 with a well defined beat produces a more structured vibration massage than music without. But the design also asks its users to give up something: social intimacy with a partner, because it invites its user to retreat into a private world of personal music and individual genital pleasure. By associating itself with iPods and consumer electronics, it also asserts itself as a middle class product, hip and youthful. In categorizing itself as another iPod-like device, it denies stigmas about sex and specifically masturbation. part of the body, or in any other orientation than the single intended one; of all the designs, it is arguably the most striking in its visual form—it looks like a cross between a medical device and a sculpture. Its implied subject is a solo female. Better Than Chocolate constrains the user to a single (albeit very clever) use—clitoral stimulation. With its clever name and formal beauty, it would make a nice gift, again proposing that masturbation can be sophisticated and tasteful as well as immediately pleasurable. In this way, like the Freestyle, it also rejects stigmas about masturbation. Jimmyjane Form 6. This design is like the Freestyle in its color and consumer electronic aesthetic. But Form 6 is ambiguously shaped, not clearly resembling a body part nor obviously intended to couple with any particular body part. Its curvy asymmetrical shape connotes without denoting the organic. Form 6 is all about ambiguity, with its two motors, nearly invisible controls, lack of any obvious grip or handle, and diverse affordances. The larger end is large, heavy, and powerful enough to serve as a non-erotic massager, e.g., for the back. It also has rounded ridges running along its length, which do not in any literal sense represent penile ridges, but they add an organic look-feel and suggest that they might feel good when the product is inserted and rotated; the ridges also make it easier to hold, which is important for a device that is likely to have lubricant on it. Taken as a group, several themes emerge. All three designs are tasteful and sophisticated. Though “pleasure objects” (as designer Lelo puts it), they are not explicit. They come in feminine colors, look like cosmetic products, and merely allude to their function. Their names are playful and girlie or pretentious—but not kinky or raunchy. All are made with body safe materials—and heavily marketed as such. They not only resemble consumer electronics, but they are made using their manufacturing processes [3]. They are expensive and they look expensive. Their subject is a classy and sophisticated woman who attends to her sexual needs like other body needs, say, skincare or feminine hygiene. Sex, then, is not special or hidden; it is just another aspect of the body needing care, and sexual taboos don’t even seem to exist, or ever have existed, in this world. At USD $185, the Form 6 is by far the most expensive of the six toys we critiqued. It is full featured and heavy, connoting product quality. Its price, weight, formal ambiguity, and pretentious name all connote luxury and taste: this toy suggests that sex is valuable enough to invest in, both financially and imaginatively. It demands of the user sufficient imagination to try it out in different places and different ways. But it also asks its user to give something up: its pampered, sensual, refined sexuality is intimate like a bubble bath, and it thereby resists the raunchy and gritty connotations of sexuality—bestial passion, getting dirty, etc. Group 2: Traditional Sex Toys We now turn to two more traditional designs, which propose different different sexual subjects. VixSkin Johnny. The Johnny is spectacularly literal: it has an anatomically modeled glans, veins, circumcision scars, a wrinkly scrotum, and human skin color. Whereas the three preceding toys have a thin soft skin covering an extremely firm device, the Johnny is rigid but gives realistically to pressure. It is also an unusually large penis: 21cm tall and 4cm in diameter, made of medical grade materials. At USD $126, the Johnny is not cheap, and yet it sells briskly. Nomi Tang’s Better Than Chocolate. Better Than Chocolate is placed over the vulva to mediate between the hand and the clitoris. Its bottom half conforms to the pubic mound, while its top half is fit for the hand. Its dramatic curves connote the body. One of the key features of the Better Than Chocolate is that rather than controlling vibration intensity or pattern through buttons, it uses a 4cm touch slider: stroking the slider changes vibration. When the device is in place, the slider is located directly above the clitoris (on the top side of the device). Thus, controlling the device is used with similar gestures and in a similar place to stroking the clitoris itself. Sex expert Cory Silverberg (personal communication) describes this design technique as “kinaesthetic onomatopoeia,” because its user interface mimics the physical action it augments. In this way, Better Than Chocolate introduces another reinterpretation of the organic, but through its interactive vocabulary and not just its form. Johnny’s user is someone who straightforwardly wants a big penis to play with. In its formal and visual fidelity, it seems pornographic. As a dismembered body part, it can be used hand-held or with a strap-on. In its literalness, it asks users to give up any sense of mystery; it has no pretentions to be other than what it is. Its sheer size suggests that some people may view it as a sexual challenge: what is it like to be penetrated by something so big? It implies a view of sexuality that is based around anatomy, athleticism, and animal intensity. It neither denies nor seeks to undermine the stigma of sex: it embraces it as a source of heat. Maximus Power Stroker Viper. The Maximus is a “beaded power masturbator” featuring a black, white, and gray color scheme, a symmetrical design, visibly cheap materials and construction (indeed, it caught on fire during a lab session). The opening in which the penis is to be inserted is white plastic with 5mm-long nubs in ordered rows and columns. The penis sleeve itself is surrounded by cables with 8mm Better Than Chocolate’s ergonomics are clever, but also inflexible; that is, the device won’t fit well with any other 138 steel bearings on them, which massage the penis. Its mechanisms are all visible through its exterior shell. The product is inorganic in its design language: its shape, color, and materials do not reference the body. Its color scheme, material connotations (e.g., steel bearings), perfect rows and columns all suggest a mechanical aesthetic. But these toys can also be read as a form of denial: for they deny that their users are subject to sexual taboo, and by extension they also deny both the political imperative to resist repressive sexual ideologies (or suggest that by buying these one already is), and also, perversely, the sexual pleasures of transgressing them. Moreover, they frame their denial in a classist way, where high-class sexuality (reflected in the pricing, naming, and taste of the toys) is barely recognizable as sex, and by implication not subject to social powers that govern sex, while (by implication) low-class sexuality is raunchy sex and thereby subject to social censure. Such a reading suggests that the new paradigm sex toys can be read as regressive, reinforcing both a social class system and the social control of sexual life—just not for those who have the income to buy a Form 6. We will not try to resolve the progressive vs. regressive readings of these toys here, but we hope we have demonstrated that designers design subjects as well as objects, and that analysis helps us get at some of the sociopolitical consequences of that design. This toy suggests a masculine subject, from its masculine name to its color scheme and look-under-the-hood aesthetic. In calling itself a “masturbator” (as opposed to, say, “pleasure object”) it is direct about what it is for. Thus, in its product form, semantics, and marketing, this toy asks its user to give up tenderness, allusiveness, and femininity. Indeed, like many machines, it favors functionality over aesthetics, and this functionality is well understood, well defined, and very clearly solved: put the penis in the hole, turn it on, masturbate, remove and clean. Sexuality seems to be an engineering problem with an engineered solution. It is not our intention to argue that the theory of sexuality inscribed in one category of designs is intrinsically superior to that of another. But it certainly is clear that these two toys differ from the first grouping of toys. They are much more explicitly sexual, understood in the sense of genital attraction and stimulation and the pleasures immediately associated with them. Both of these products represent and indulge “base” or “raunchy” notions of sexuality. THE EMPIRICAL ANALYSIS Our empirical study was designed to help us understand sexual subjectivities: our research question was not, which toys did users like?, but rather, how did users express themselves as sexual subjects through their interactions with these toys? Linking back to the theory summarized earlier, we focus here on participants’ perceptiveness, social meaning-making, and boundary-drawing. Discussion of Critiques Obviously, sexual subject positions are not created solely by the sex toy industry. Sociocultural norms have a lot to say about how we understand sexuality. Sex toys reify these norms in concrete and specific ways, which relate to but are also distinct from sexual norms presented in Sunday sermons, porn movies, advice columns, therapy sessions, and so on. In the west, there are many taboos surrounding our understanding of sexuality: the notion that sex is shameful and should be kept private is pervasive. Much of sexual practice, especially visible in consensual BDSM, involves exploring the sexual power of taboos—that which is forbidden becomes especially hot. Johnny and Maximus both position themselves within such a sexual subjectivity. Perceptiveness Though study participants obviously made use of their perceptual abilities to notice the objective features of toys— colors, shape, texture—such characterizations only infrequently appear in their words. Instead, the focus of their perceptions was on the fit of toys in their lifeworlds, and objective features of the toys, rather than being seen in and of themselves, were raised in a more subjective sense. A simple example of this is how subjects articulated size. The pocket vibrator like it could be hung, it could carry that around in a purse, you could almost carry that around in those little itty bitty like go to a bar purses. [F, 24, heterosexual, Y] What becomes clear in contrast is that the new paradigm toys do not. They associate themselves with non-sexual forms of body care (and increasingly they are showing up in body care aisles of department stores, where one would never find Johnny), which they define in allusive and vaguely positive ways: “pleasure,” “feeling sexy,” “lifestyle.” As critics, we are concerned with alienation, how products reify ideologies that alienate us from goods that are rightfully our own. As sex positive feminists, we view traditional sexual morality as having alienated us from sexual pleasures that are rightfully ours. But the politics here are complex. We welcome the new paradigm toys for their celebration of sexual pleasure, and they can be (and frequently are) read as exemplars of sex positive feminism. So on the side of my bed I have a little box… And it’s a really nice box and I’m never like oh like I have to hide this. And oh so when I’m shopping for one I look for something that’s going to fit in there. [F, 24, heterosexual, Y] [Note: the annotation after each quote refers respectively to gender (M/F/GQ [gender queer]), age, orientation (subject’s own self-description), and whether or not they have used sex toys in their personal life (Y/N).] In both quotes, the objective quality is size, but this quality is understood in relation to other artifacts in the toy’s anticipated ecology, here the containers that participants imagine themselves putting the toys in. These evaluations 139 tacitly imply intentions (e.g., portability and making available) and social predispositions (e.g., the need for discretion). spending the money to get in the range where it is dishwashable—a culturally specific sexual hygeine practice if ever there was one. Of course, like any other design, a sex toy is meant to be used, and participants’ commentary on toys showed some specific ways that people anticipate use: These three sets of examples (toy size and its place in everyday life, anticipated uses, and concerns about hygiene) make clear that toys are never interpreted in terms of “just sex.” This helps explain why the designerly toys—which situate themselves in relation to personal care and pleasure—are more desirable; they speak to a more holistic sociosexual subject. The one negative comment (that VixSkin is too literal) indicated that the toy is not legible beyond sex acts. We also see here that participants’ imaginative exploration of the toys is temporally structured (to be carried and stored; to be used; to be cleaned up). Such characterizations are compatible with people’s understandings of intimacy [28] and experience itself [24]. The toys’ meanings say as much about the participants as it does about the toys themselves. Yet our participants did not use the language of critical theory; although in our critiques we discussed ideology, false pleasures, semantics, inscribed users, etc., such issues seldom explicitly came up with the users. Yet user comments that express or respond to sexual norms do bridge the two discourses, as we see in the following sections. I don’t like vibrating things because the motors make them sound like they’re always screaming. I don’t want to have something inside of me that’s yelling unless it’s yelling in pleasure. [M, 22, homosexual, Y] You know, if you’re masturbating or having some sort of sex, do you really want to change the batteries like midstream? It just makes everything unsexy. [F, 30, lesbian, Y] There seems to be too much of an imitation of the penis, and the penis…doesn’t do that much. So mimicking the penis is not very imaginative because it doesn’t really get you what females really want or need. [M, 26, straight, Y] A range of sexual norms are articulated in these quotes: that vibrators should not be too loud, that sexual activity is optimally experienced as a flow, and that sex toys should be imaginatively designed. Yet none of these norms is “natural” or “animal”; all of them are situated in everyday social life. Sociologist William Simon writes, “Desire is the scripting of potential futures … drawing upon the scripting of the past desire as experienced in the contingent present. Desire, in the fullness of its implicit ambiguity, can be described as the continuing production of the self” [38, p.139]. These comments do not merely reference such scripts, as if the scripts were already whole and stored somewhere: rather, they enact and embody those scripts. So, for example, the participant is not saying that a noisy vibrator would probably turn her off were it to be introduced into a hypothetical sexual situation; instead, it is already turning her off in the moment she is saying it—even in the lab. Nor does this scripting end with the sex act: Social Meaning-Making The theory of self as an outward-facing sexual subject outlined earlier this paper emphasizes the role of intersubjective experience as an input to and outcome of our sexualities. We saw plenty of evidence that intersubjective experience is central to how participants think about toys: My partner and I for a really long time … didn’t use sex toys because it wasn’t necessary…. It wasn’t until we had been together for a while that we were like, hey, let’s throw something else into the mix. [F, 30, lesbian, Y] It’s just kind of one more thing for me to offer my partner or myself… Yeah it’s just kind of enhancing what I’m able to give her [F, 22, lesbian, Y] I’d be worried about getting water and moisture in here…. that’s like my number one thing is can I clean it…. I look for nice, solid, something that says it’s waterproof that you can submerge the entire object in water because then I know the entire surface is able to be cleaned. [F, 28, straight, Y] If I’m looking for something for me and my partner, then we pick something that we’re both going to like, and her tastes are a little bit different than mine, so we…compromise. [F, 30, lesbian, Y] Sex toys here are meaningful inasmuch as they enhance or mediate their sexual partnership with another. Each speaker has a collaborative, not individual, relation to the toy. This sort of concrete interaction defines the self as a subject [27]. As [32, p.204] writes, “embodied subjects develop direction and purpose on the basis of the practical engagements they have with their surroundings and through the intentionality they develop as a result of the situatedness of embodied existence.” The subject’s practical engagements are not limited to those between partners; our bodies are also socially positioned. The presentation and interpretation of our bod- When I [have] money I will get the ones that are $100 or so and you just toss it in the…dishwasher…. [Y]ou wash them like that, and you can use them without having to go through the steps of putting a condom on and cleaning it out. I think mine is starting to stain now, even though I use a condom…so I think yeah, that’s a good criteria. [M, 22, homosexual, Y] These comments focus on hygiene, which is a practical consideration for sex toys. But hygiene is a culturally specific activity [35,44]. Thus, hygiene is not an afterthought, but primary (“my number one thing”); hygiene justifies 140 ies is restrained by external forces [7,32]. One external force is body norms, as the following quotes exemplify: mantics, provide more room for expressive agency (e.g., most people would presumably rather be discovered with a Form 6 than a VixSkin Johnny). The problem with things like this is the belts are not made…for people who are fat. [GQ, 21, Lesbian, Y] Boundary-Drawing The sex toy assumes a normative weight, and when the participant falls outside of it, she feels “fat.” Quotes like these help bridge to the critiques, because they so clearly point to ideological issues (here, norms of slenderness as beauty). We also wanted to explore the language of acceptance or rejection of sex toys to understand what they reveal about the self and its boundaries as a sexual subject. We begin with several rejections of sex toys: It’s not adjustable in any way, so for a man to get full pleasure out of that, he has to be a particular size not any bigger not any smaller. [F, 28, Straight, Y] it’s lime green and purple so like I’m already putting something unnatural inside my body. [F, 24, heterosexual; Y] Here again hegemonic norms are foregrounded. Penis size is meaningful, frequently linked to notions of masculinity, self-esteem, and racism (e.g., the perceived threat or special appeal of the stereotypically well-endowed black man). Our bodies are not merely biological but are “inscribed, marked, engraved by social pressures external to them” [17], and our subjects fluently understood such inscriptions. I don’t know if I like the size of it, and it’s really hard also so I don’t know if I would enjoy putting it in any holes because it might end up hurting me. [M, 22, homosexual, Y] She hates this material. She says it like absorbs bacteria […], and over time it will discolor. [F, 28, straight, Y] In each of these quotes, the relationship between the toy and the body is extremely intimate, as if the toy is a candidate to become a part of the body. Theorist Julia Kristeva proposed the concept of the abject to refer to objects in the world that challenge the borders between our bodies and the outside world: body excretions, blood, sweat, corpses, vomit, etc. [21]. Our repulsion to the abject is protective mechanism, a way of reinforcing the boundaries between self and outside when those borders are threatened. As Body theorist Blackman writes, the abject “demarcates a number of key boundaries, such as the inside and outside, the natural and the cultural, and the mind and the body…. [The body] is engendered or brought into being through two key concepts: individuation and separation.” [6] Shame, implied in both of the size quotes above, is an intersubjective emotion, defined as the reaction when one sees oneself through the eyes of another in a negative way [23]. In addition to being ashamed of body properties, people can also be ashamed of its needs or pleasures: I think that a lot of my friends and I would much more go for something smaller; something more compact something that you could hide. [F, 24, heterosexual, Y] I recently found out my friends would find these in their parents’ rooms, their mothers’ rooms. The divorced mothers, and I think everyone has that dirty little secret. [M, 22, homosexual, Y] the cat actually found this in my sister’s room. Yeah, she had gone away to college and like the cat was playing with it for a good month or so and my mom came home and was like, ‘What is this?’… If I knew what I know now back then, I would have not thought it was funny but disgusting. [M, 22, homosexual, Y] It is interesting to see that dislike was often specifically expressed as a threat, and rejection in terms of self-protection; sex toy likes and dislikes are often far stronger than mere preference statements, e.g., for flavors of ice cream. Our application of the abject also suggests specifically where participants perceive their own borders between what could be accepted into the self versus those that could not. In the above quotes, the unnatural, the physically large, and the discolored are all offered as boundaries. Yet as Kristeva’s theory suggests, boundaries are movable. The same physically large toy that is intimidating to one is an object of curiosity or even a sexual challenge to another. The following quotes all explore boundary reactions, in which some aspect of the toy is threatening, but rather than rejecting it, the participant seeks ways to accommodate it: These quotes suggest three ways vibrators and masturbation are social: by involving a partner, inscribing social norms (and provoking personal reactions), and causing embarrassment/amusement when discovered. Social structures— from body ideals to racism—are part of the toy, and so the toys not only bear the capacity for erotic arousal and intimate creativity, but also pride/shame, disgust, indignation, and hilarity. Moreover, none of these perceptions or reactions are mutually incompatible; the same person can be subjected to and become subjects of any or all of them. An analytic understanding of such a complex response benefits from a conceptual vocabulary that can account for the copresence of diverse selves (qua subjects) as well as the complex enactments and performances that they occasion in different social situations. In such a view, it becomes clear why the more designerly toys, with their more nuanced se- But it seems like you would have to really lubricate well just because it like, it kind of like pulls. The texture of it kind of pulls so you’d have to definitely makes sure that you use lots of lube so there’s no like pulling or tearing. [F, 22, lesbian, Y] 141 this one seems like it would be a little too much because I think like too much vibration can kind of just be numbing [M, 20, homosexual, Y] SUBJECTIVITIES OF INFORMATION Subjectivity theory, developed in the context of postmodernism, might come across as radical, proposing a “fragmented self” that lacks unity or coherence, in stark contrast to our intuitive sense of ourselves as unified wholes. On the other hand, the idea that our sense of self is shaped by our participation in social structures (parent-child, teacherstudent, partner-partner), and that we have some agency in how we embody and perform such roles, seems intuitive. We argue for subjectivity theory in HCI not in a metaphysical sense, as if to assert that “subjectivities of information” somehow corresponds better to reality as the “hidden truth” about how we “really are.” Rather, we argue that this formulation has pragmatic benefits for research and design. So if it’s any bigger than that then I probably wouldn’t want it at least not for penetration purposes. [F, 23, heteroflexible, Y] Size, texture, and vibration are all qualities that in some measure are acceptable and in other measures are rejected. Along the borders, then, toys perceived as challenging (as opposed to outright threatening) can be managed, e.g., with more lubricant or by changing its use. A possible outcome of management efforts is an increasingly inclusive and more specific drawing of the boundaries, which appear to reflect a certain amount of self-awareness: The strength of this theory is its ability to account for the mutability of selfhood, as it negotiates the boundaries of internal experience and intention and our (designed) environments and social reality. If we view the self as unified and even fixed, it limits design to “supporting” and “augmenting” existing capabilities (and much of HCI literature uses these terms). If we can understand the mechanisms by which selves change—as embodied social beings situated in sociotechnical environments and practices—then we can look beyond supporting and augmenting towards the hope and the ideal of social change. From environmental sustainability to social media, and from participatory design to action research and feminist HCI, the field has been aspiring to contribute to social change. In doing so, it has already begun to explore many humanistic concepts and methodologies. For this reason, we believe that many in the field are already embracing, if tacitly, some version of subjectivities of information, from Turkle’s Life on the Screen to research on the “quantified self.” Our hope here is to offer an argument, a set of concepts, an example, as a provocation that helps frame and consolidate these developments. I would rather have a lot, a larger range of vibrations rather than a larger range of pulsing. [F, 30, lesbian, Y] For vaginal stimulation it’s not really a big deal; but for clitoral stimulation I have to have a big range…just so I can have a good variation [F, 23, heteroflexible, Y] These are not “natural” embodied reactions but rather learned and practiced ones. Each speaker has a developed sensibility regarding the nature of vibration and its tactual effects on her body. That the toy challenges boundaries— not only physically as body penetration, but somatically as an extended body part—helps explain why sex toys are greeted with such extreme reactions—desire, repulsion— and giving rise to the sober skill of managing the toy’s power in one’s own bodily ecosystem. Discussion of Empirical Findings The critical perspective was strong at analyzing links between design choices to sociocultural issues, e.g., how toys embody and replicate class norms and ideologies, how toys participate in activist challenges to traditional morality, and how sexual experiences can be offered up as/in products. The empirical findings also provided linkages between specific design choices and experiences, but in more concrete terms. They revealed how design choices exposed (or pushed) the boundaries of participant sexuality: what was desirable, what was possible under what conditions, and what must be rejected outright both from a physical perspective (e.g., too big, wrong color) and from a social perspective (e.g., too shameful to acknowledge, reifies regressive body ideals). The findings also reveal how sexuality relates to other parts of life, e.g., storage and hygiene. The emphatic reactions emphasized how much these toys are literally incorporated, taken into the body, and how powerful, for better and for worse, that move is. The body is changed, both to oneself and to the social world. This change is a powerful experience to the self; the discovery that one’s partner (or mother!) has been subjected to such change is socially significant; as is the realization that one’s body is ineligible for such change (e.g., one is “too fat” or “too small”). Near the beginning of this essay, we argued that this theory provides a tight analytic coupling between specific design decisions and particular socio-subjective experiences; a vocabulary to reason about the ways that we can design subjects as well as interfaces, products, and services (without denying human agency enacted in subjectivities); and a means to get at design aimed at cultivating and transforming, rather than merely supporting or extending, human agency. We conclude by returning to these three claims. Linking Design Choices with User Experiences Aristotle in his Poetics exemplifies an analysis that links formal features of literary genres (e.g., tragedy and epic) to subjective experiences and social significances, e.g., how the form of a tragic reversal in a tragedy causes feelings of pity and fear in the audience, which then purge them of their own emotions, reestablishing their rationality. Our critical analysis likewise seeks to understand how the product semantics of certain contemporary sex toys cohere with certain particular kinds of experiences, themselves situated in certain social meanings. More specifically, we followed 142 experts in distinguishing between two categories of sex toys—earlier adult novelty products vs. more recent “designerly” lifestyle accessories—to interpret how the designs project subject positions as part of their meaning. ing becomes a reality, we are constructing new environments and human organisms will adapt. Part of this means the acquisition of new skills of perception and a precise conceptual vocabulary to grasp and to express them: no one is born with a capacity to explicate a sex toy feature’s physical, sexual, romantic, social, and ideological consequences (e.g., reification of body weight norms and penis size). Part of it also means researchers and designers improving their own conceptual grasp: for example, the way that our subjects perceived and interpreted the significance of toys as (prospectively) part of their bodies would seem to raise potentially generative concepts about wearable technologies. In our empirical study, we gained a sense of individuals perceive, interpret, and tell stories about sex toys. A common statement involved a participant pointing to something specific about the design (a shape, color, texture) and then relating it to experience (how it would feel, its implications for hygiene, how one might hide it). Such statements can add up to a poetics in their own right. Thus, even though we view subject positions as inscribed in objects and available to researchers analytically through interpretation, and we view subjectivities as part of human experiences and available to research via empirical studies, they need not be incommensurate. To the extents that these two modes of analysis overlap increases confidence in the findings. Our study of sex toys, which comprises expert interview studies, critical analysis of toys, and empirical studies of sexual experiences, has helped us grasp not only the emerging paradigm of designerly sex toys compared to what came before it in design historical terms, but more importantly to understand its tremendous range of significance: from intensely personal experiences to forms of public activism, literal orgasms to public shame, product design history to third wave feminism, and law enforcement to pop music. Each of these is an opening to forms of agency: to try out a new sexual experience, to become active in sexual politics, to jump into sex toy design, to become more sex educated and/or an educator. It is not our position that design alone causes such transformations, but it is our position that design can contribute meaningfully to such transformations, and subjectivity theory, and a critical-empirical research methodology, can help us understand them better. Because designers in HCI are hoping to contribute such transformations, we believe a reconceptualization of the user as a subjectivity of information will be salutary. Designing Subjects as Well as Interfaces Both our critical and empirical studies pointed to two different sexual subjectivities. Earlier sex toys propose a raunchy and transgressive sexual subject, who finds pleasure in genital attraction/sensation in precisely the ways that are taboo in traditional morality. The toys construct this subjectivity through their blunt visual and functional appeal to sex acts, including literal replication of anatomy, pornographic packaging, overt problem-solving functionalism, as well as more subtle features, such as vibrator controls that face away from a woman (i.e., so her partner can control them). As sex positive feminism gained momentum over the decades, however, an alternative sexual subject was offered: a body-conscious woman, who takes care of her physical needs with taste, discrimination, and consumer choice. The designerly sex toys speak to this subject, in their forms, materials, and marketing. Such features include more abstract shapes that connote but do not denote sexual organs, feminine color palettes instead of skin tones (especially bright pinks and purples), formal allusions to cosmetics (e.g., the silver band on the Freestyle), emphasis on health (e.g., body safe materials), and euphemistic marketing. ACKNOWLEDGMENTS We thank Cory Silverberg, Gopinaath Kannabiran, and Katie Quehl for their contributions to this study. We thank our peers in the Intel Science and Technology Center for Social Computing for their contributions to this research, and to Intel for supporting the research. We also thank our reviewers for their tough questions. It is important to stress that any individual can choose between these subjectivities, embracing one, the other, both, or neither—and that this choice can be made again and again, changing over time. One can be in the mood for raunchy and transgressive sex one day, and be in the mood for something more sensual and synesthetic then next. It is also possible to start predominantly in one place and then to change one’s tastes, or form new desires, over time. Designs—and the images of sexual life inscribed within them—can help effect such changes. REFERENCES 1. Bannon, L., and Bødker, S. (1991). Beyond the interface: Encountering artifacts in use. In Designing Interaction: Psychology at the Human-Computer Interface. Carooll, J. (eds.). Cambridge UP, 227-253. 2. Bardzell, J. (2011) Interaction criticism: an introduction to the practice. Interacting with computers 23(6): 604621. 3. Bardzell, J., and Bardzell, S. (2011). “Pleasure is your birthright”: Digitally enabled designer sex toys as a case of third-wave HCI. Proc. of CHI'11. Cultivating and Transforming Human Agency Dewey frequently reminds us in his writings that humans are organisms, that we adapt to our environments. As ubiquitous computing, Internet of Things, and wearable comput- 4. Bell, S. (2010) Fast Feminism. Autonomedia. 5. Bertselsen, O. & Pold, S. (2004). Criticism as an ap- 143 proach to interface aesthetics. Proc. of NordiCHI’04, ACM Press, 23-32. 26. Norman, D., 2002. The Design of Everyday Things. Basic Books, New York. 6. Blackman, L. (2008). The Body. Berg. 27. Nussbaum, M. (1999). Sex and social justice. Oxford UP. 7. Butler, J. (1993). Bodies that matter: On the discursive limits of sex. Routledge. 28. Pace, T., Bardzell, S., and Bardzell, J. (2010). The rogue in the lovely black dress: Intimacy in World of Warcraft. Proc. of CHI’10. ACM. 8. Card, S., Moran, T., Newell, A., 1983. The Psychology of Human–Computer Interaction. Lawrence Erlbaum Associates Publishers. 29. Radway, J. (1984). Reading the romance: Women, patriarchy, and popular literature. The University of North Carolina Press. 9. Carroll, N. (2009). On criticism. Routledge. 10. Cockton, G. (2008). Revisiting usability's three key principles. CHI2008 EA. 2473-2484. 30. Russon, J. (2009). Bearing witness to epiphany: Persons, things, and the nature of erotic life. SUNY Press. 11. Cooper, G., and Bowers, J. (1995). Representing the user: Notes on the disciplinary rhetoric of humancomputer interaction. In Thomas, P. (ed.). The social and interactional dimensions of human-computer interfaces. Cambridge UP. 31. Satchell, C., and Dourish, P. (2009). Beyond the user: Use and non-use in HCI. Proc of OzCHI'09. 9-16. 32. Schilling, C. (1993). The body and social theory. Sage. 33. Seidman, S. (2003). The social construction of sexuality. W.W. Norton & Company. 12. Dourish et al. (2012). Research Themes: Intel Science and Technology Center for Social Computing. http://socialcomputing.uci.edu/content/research-themes 34. Semans, A. (2004). The many joys of sex toys: The ultimate how-to handbook for couples and singles. Broadway Books. 13. Eldridge, R. (2003). An introduction to the philosophy of art. Cambridge UP. 35. Shove, E. (2003). Comfort, cleanliness, and convenience: The social organization of normality. Berg. 14. Goodman, E. and Vertesi, J. (2012). Designing for X? Distribution choices and ethical design. CHI’12 EA. ACM Press. 36. Shusterman, R. (2008). Body consciousness: A philosophy of mindfulness and somaethestics. Cambridge UP. 15. Gordo López, Á., and Cleminson, R. (2004). Technosexual Landscapes: changing relations between technology and sexuality. Free Association Books. 37. Shusterman, R. (2011): Somaesthetics. In: Soegaard, M. and Dam, R.F. (eds.). Encyclopedia of HumanComputer Interaction. Aarhus, Denmark: The Interaction Design Foundation. 16. Gould, J. & Lewis, C. (1985) Designing for usability: Key principles and what designers think. Communications of the ACM 28(3), ACM Press, 300–11. 38. Simon, W. (1996). Postmodern sexualities. Routledge. 39. Sonneveld, M., and Schifferstein, H. (2008). The tactual experience of objects. In Schifferstein, H., and Hekkert, P. (eds.). Product experience. Elsevier. 17. Grosz, E. (1994). Volatile bodies: Toward a corporeal feminism. Indiana University Press. 18. Horrocks, R. (1997). An introduction to the study of sexuality. Macmillan Press, Ltd. 40. Suchman, L., 1987. Plans and Situated Actions: The Problem of Human–Machine Communication. Cambridge University Press, Cambridge, UK. 19. How, A. (2003). Critical Theory. Palgrave Macmillan 20. Irani, L., Vertesi, J., Dourish, P., Philip, K., and Grinter, B. (2010). Postcolonial computing: A lens on design and development. Proc. of CHI’10. ACM. 41. Sutcliffe, A. (2010) Designing for user engagement: Aesthetic and attractive user interfaces, Synthesis Lectures on Human-Centered Informatics. Morgan & Claypool Publishers. 21. Kristeva, J. (1980). Powers of horror: An essay on abjection. Columbia UP. 42. Taylor, A. (2011). Out there. Proc. of CHI'11. ACM. 22. Kuutti, K. (2001). Hunting for the lost user: From sources of errors to active actors – and beyond. Cultural Usability Seminar, Media Lab, University of Art and Design Helsinki, 24.4. 43. Thornham, S. (1997). Passionate detachments: An introduction to feminist film theory. Arnold. 44. Vigarello, G. (1998). Concepts of cleanliness: Changing attitudes in France since the Middle Ages. Cambridge UP 23. Laine, T. (2007). Shame and desire: Emotion, intersubjectivity, cinema. Peter Lang. 45. Watson, G. (2008). Art & Sex. I.B. Tauris. 24. McCarthy, J., and Wright, P. (2004). Technology as experience. MIT Press. 46. Winograd, T., Flores, F., 1986. Understanding Computers and Cognition: A New Foundation for Design. Ablex Corporation, Norwood, NJ. 25. Nielsen, J. (1993). Usability Engineering. Morgan Kaufmann. 144 Creating Friction: Infrastructuring Civic Engagement in Everyday Life Matthias Korn & Amy Voida School of Informatics and Computing Indiana University, IUPUI {korn, amyvoida}@iupui.edu Civic engagement encompasses the myriad forms of both individual and collective action that are geared toward identifying and addressing issues of public concern. Human-Computer Interaction (HCI) researchers employ diverse methods and design strategies to study, support, and provoke civic engagement. Some researchers work within mainstream politics to improve the efficiency with which citizens can engage with the state through e-government services [6, 73]; to improve access to voting [22, 69]; to seek input and feedback from citizens on public planning issues [24, 42]; or to foster dialogue, debate, and deliberation among citizens and with the state [5, 41, 65]. Researchers also work to foster civic engagement outside of the political mainstream, supporting the work of activists, protestors, and grassroots movements [2, 18, 35, 37, 51]. The sites in which political life takes place are not only sites of government, but also cities, neighborhoods, and communities. ABSTRACT This paper introduces the theoretical lens of the everyday to intersect and extend the emerging bodies of research on contestational design and infrastructures of civic engagement. Our analysis of social theories of everyday life suggests a design space that distinguishes ‘privileged moments’ of civic engagement from a more holistic understanding of the everyday as ‘product-residue.’ We analyze various efforts that researchers have undertaken to design infrastructures of civic engagement along two axes: the everyday-ness of the engagement fostered (from ‘privileged moments’ to ‘product-residue’) and the underlying paradigm of political participation (from consensus to contestation). Our analysis reveals the dearth and promise of infrastructures that create friction— provoking contestation through use that is embedded in the everyday life of citizens. Ultimately, this paper is a call to action for designers to create friction. Individual technologies and systems, and the civic engagement they support, are enabled by and based upon a myriad of different, ‘layered’, interwoven, and complex socio-technical infrastructures [37, 47, 52, 68]. Yet, infrastructures of civic engagement are a particularly challenging site for HCI as there are competing forces at play. On one hand, infrastructures of civic engagement are fundamentally about engaging people; and even more so, they may be designed to engage people to enact change. On the other hand, infrastructures are typically invisible; they remain in the background and are taken for granted by their various users [68]. Even further, Mainwaring et al. warn that infrastructures, which are so conveniently at-hand, can breed complacency and stasis [52]. Infrastructures of civic engagement, then, must counter not only the challenges of provoking civic engagement through everyday life; they must also overcome challenges of complacency and stasis. These are the dual challenges that we take up in this research. Author Keywords Civic engagement; everyday life; infrastructuring; friction. ACM Classification Keywords K.4.2. Computers and Society: Social Issues; K.4.3. Computers and Society: Organizational Impacts. INTRODUCTION The object of our study is everyday life, with the idea, or rather the project (the programme), of transforming it. [..] We have also had to ask ourselves whether the everyday [..] has not been absorbed by technology, historicity or the modalities of history, or finally, by politics and the vicissitudes of political life. ([49]: 296, vol. 2) The distinction between everyday life and political life is a highly problematic one for social theorists such as Lefebvre. And it is a distinction that is increasingly being challenged by researchers designing to support or provoke civic engagement in everyday life. We advocate for the construct of friction as a design strategy to address the dual challenges of civic engagement. Following Tsing [70] and Hassenzahl et al. [30], we maintain that friction produces movement, action, and effect. Friction is not exclusively a source of conflict between arrangements of power; it also keeps those arrangements in motion [70]. In the infrastructuring of civic engagement, we believe that frictional design can help to expose diverging values embedded in infrastructure or Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21198 145 values that have been left aside during its design. We also contend that frictional design can help to provoke people not only to take up more active roles in their communities but to question conventional norms and values about what it means to be a citizen, as well Lefebvre contends that contact with the state has become a superficial and apolitical one in modern society: Not only does the citizen become a mere inhabitant, but the inhabitant is reduced to a user, restricted to demanding the efficient operation of public services. [..] Individuals no longer perceive themselves politically; their relation with the state becomes looser; they feel social only passively when they take a bus or tube, when they switch on their TV, and so on— and this degrades the social. The rights of the citizen are diluted in political programmes and opinion polls, whereas the demands of users have an immediate, concrete, practical appearance, being directly a matter for organization and its techniques. In the everyday, relations with the state and the political are thus obscured, when in fact they are objectively intensified, since politicians use daily life as a base and a tool. The debasement of civic life occurs in the everyday [..]. ([49]: 753–754, vol. 3) In the remainder of the paper, we introduce the theoretical lens of the everyday to extend the emerging bodies of research on contestational design and infrastructures of civic engagement. Our analysis of social theories of everyday life suggests a design space that distinguishes ‘privileged moments’ from the more holistic ‘productresidue’ of everyday life. We analyze various efforts that researchers have undertaken to design infrastructures of civic engagement along two axes: the everyday-ness of the engagement fostered and the underlying paradigm of political participation. Our analysis reveals the dearth and promise of infrastructures that create friction—provoking contestation and debate through use that is embedded in the everyday life of citizens. Ultimately, this paper is a call to action for designers to create friction. For Lefebvre, one of the most fundamental concerns about civic engagement in modern society is that it has been confined to “privileged moments” ([49]: 114, vol. 1)— special occasions or punctuated feedback cycles on public servants and service provision. Civic engagement has been degraded in the product and residue of everyday life. Lefebvre argues that “use must be connected up with citizenship” ([49]: 754, vol. 3)—that the everyday demonstration of concern for public services are as essential a facet of being a citizen as, e.g., voting or debate. THEORETICAL FOUNDATIONS Our research synthesizes strands of scholarship about everyday life, infrastructuring, and contestational design. We describe each strand of scholarship below, with a focus on their relationships to civic engagement. Everyday Life and Civic Engagement As technology has been woven into “the fabric of everyday life” ([77]: 94), it surrounds us in ever smaller and more invisible ways [20, 25] and reaches into multiple spheres of our lives [10]. But what is everyday life? Two social theorists—Lefebvre and de Certeau—have been most central in emphasizing everyday life as a legitimate site of study—a site that follows an underlying logic that is both relevant and discoverable. The significant tensions in everyday life that Lefebvre characterizes as playing out between personal and social rhythms are reprised somewhat, albeit with different framing, in the work of de Certeau. De Certeau [13] theorizes about the everyday as interplay between the social forces of institutional rituals and routines, “strategies,” and the “tactics” opportunistically employed by ordinary people, who subvert these strategies as they go about their everyday lives. The performative and embodied nature of everyday life is also an emphasis taken up by more recent work in cultural studies [25]. According to Lefebvre—operating in post-war democratic France—the everyday is the space in which all life occurs ([49]: 686ff., vol. 3).1 It is not merely the stream of activities in which people engage over the course of their days (thinking, dwelling, dressing, cooking, etc.). Rather, everyday life more holistically understood is also the space between which all those highly specialized and fragmented activities take place. It is the residue. And further, everyday life must also be understood as the product of these activities, the conjunction and rhythms of activities that render meaning across fragmented activities. Where strategies accumulate over time through the exertion of power (e.g., standardization), tactics live in and capitalize on what is in the moment ([13]: 37). The institutional rituals, routines, and their underlying representations, which influence and are influenced by the performance of the everyday, also, then, begin to form infrastructures for everyday life. Even amidst the push-andpull interdependence of the personal and the political, the tactics and the strategies, people still manage to carve boundaries amidst and around the everyday. And these boundaries, which are “normally taken for granted and, as such, usually manage to escape our attention” ([80]: 2), are important factors for consideration in the infrastructuring work that needs to be done if civic engagement is to be released from its confinement to privileged moments. For Lefebvre, everyday life is dialectically defined by contradictions between the body’s personal rhythms and the rhythms of society or at the “intersection between the sector man [sic] controls and the sector he does not control” ([49]: 43, vol. 1). Everyday life includes the political; however, 1 The three volumes of Lefebvre’s Critique of Everyday Life [49] were published in 1947, 1962, and 1981, respectively. 146 Infrastructuring Civic Engagement disempowerment [52], which is particularly problematic for the infrastructuring of civic engagement. Infrastructures are the predominantly invisible and takenfor-granted substrates, both technical and social, that enable and support local practices [48, 68]. Infrastructures of civic engagement are those socio-technical substrates that support civic activities. Contestational Design and Civic Engagement Critical Design Roots There is a strong interest within the field of HCI to critique and provoke, to question and reflect in and through design —either on established social and cultural norms in general [3, 30, 66], or within socio-political domains more particularly [1, 17, 37]. Dunne and Raby [21] discern between two modes of design—affirmative design and critical design: Infrastructures are relational. They are embedded in social structures, learned as part of membership in social structures, and shaped by the conventions of social structures [68]. The relationship between infrastructures and practices of use is dialectic, characterized by a dynamic interplay between standardization and the local practices they support at each site of use [68]. Infrastructure must, therefore, be understood as processual, evolving over time. What is infrastructure in one moment may become a more direct object of attention and work in the next, particularly when breakdowns occur—either when the technology ceases to work or when local practices change and deviate from standards implemented in infrastructure [58, 68]. The former reinforces how things are now, it conforms to cultural, social, technical, and economic expectation. [..] The latter rejects how things are now as being the only possibility, it provides a critique of the prevailing situation through designs that embody alternative social, cultural, technical, or economic values. ([21]: 58) Because infrastructures reflect the standardization of practices, the social work they do is also political: “a number of significant political, ethical and social choices have without doubt been folded into its development” ([67]: 233). The further one is removed from the institutions of standardization, the more drastically one experiences the values embedded into infrastructure—a concept Bowker and Star term ‘torque’ [9]. More powerful actors are not as likely to experience torque as their values more often align with those embodied in the infrastructure. Infrastructures of civic engagement that are designed and maintained by those in power, then, tend to reflect the values and biases held by those in power. Critical design aims to question the status quo, revealing hidden values and agendas by designing provocative artifacts that adopt alternative values not found in mainstream design. Critical design aims to “enrich and expand our experience of everyday life” ([21]: 46) by “generat[ing] dilemmas or confusions among users in such a way that users are encouraged to expand their interpretative horizons or rethink cultural norms” ([4]: 289). By so doing, critical design directs attention to issues of public concern. In the context of civic engagement, critical design introduces dilemmas and confusions into the experience of civic issues with the aim to surface multiple, alternative values and spur debate. By problematizing values that are ‘hard-wired’ into infrastructures of civic engagement, critical design can counter the stasis, complacency, and inertia that can result from the at-hand experience of infrastructure. Infrastructures are also relational in a second sense. As they increasingly extend into everyday spaces, infrastructures are shaped by the spaces of everyday life and shape our encounters with those spaces; they are increasingly experienced spatially [19]. Infrastructures of civic engagement, then, must also contend with these spatiorelational experiences insofar as they engage with everyday space and the physical environment (e.g., [40]). Critical Design for the Everyday Dilemmas and confusions can also be designed into artifacts of everyday use—furniture or government ID cards, for example [15, 21]. Hassenzahl et al. advocate for designing everyday artifacts following an ‘aesthetics of friction’ as opposed to an aesthetics of convenience and efficiency [29, 30, 44]. Based on the psychology of motivation, they recommend designing artifacts that will lead to momentary transformation, subsequent reflection and meaning making, and, taken together, longer-term behavior change [30]. In contrast to many strategies of persuasive design, the emphasis here centers on building from small, mundane, everyday moments of transformation: The processual and evolving character of infrastructure has led participatory design researchers to examine the tentative, flexible, and open activities of ‘infrastructuring’ with the goal of empowering sustainable change [47, 58, 67]. Researchers exploring civic engagement have increasingly focused on similar issues of empowerment and sustainability, moving beyond individual applications to engage more systematically with the infrastructures that support civic engagement (e.g., [47, 74]). This research has emphasized challenges associated with the fragmentation and interoperability of infrastructures currently supporting civic engagement [74]. However, the research community’s critical engagement with infrastructure has also found that the full-service, ready-to-hand nature of many infrastructures may also invite complacency or The primary objective is, thus, not necessarily maximizing change (e.g., reducing energy consumption) per se, but supporting people with realizing the goals they find worthwhile to pursue, but hard to implement.” ([44]: 2) 147 These momentary transformations are provoked by a class of technology that they term ‘pleasurable troublemakers’ [30]. These troublemakers create small but pleasurable obstacles to targeted moments of everyday life, i.e., they create friction. Rather than designing to help or to make everyday life easier, Hassenzahl et al.’s designs engineer pause and provoke reflection, using friction to emphasize the active role individuals should have in constructing meaning of their experiences [30, 44]. ‘product-residue.’ These two perspectives form the first axis of our design space. Privileged Moments Everyday life includes political life [..]. It enters into permanent contact with the State and the State apparatus thanks to administration and bureaucracy. But on the other hand political life detaches itself from everyday life by concentrating itself in privileged moments (elections, for example), and by fostering specialized activities. ([49]: 114, vol. 1) There is genealogical resonance between theories of the everyday and theories of critical design. Lefebvre’s Critique of Everyday Life [49] was a foundational influence on the Situationists, whose art tactics have been taken up by designers and critical theorists in HCI [46]. Lefebvre’s critique of the dilution of citizens’ everyday agency in politics and civic engagement, which we take up here, parallels his criticism of the passive role of individuals imposed on them by a consumer society. Whereas the Situationists created spectacles to raise awareness and intervene in the consumerist status quo, we advocate for creating friction to raise awareness and intervene in the depoliticized status quo. According to Lefebvre, the depoliticization of everyday life confines civic engagement to privileged moments. Foremost, privileged moments are privileged through their status as special activities that occur only infrequently. Yet, privileged moments are also privileged through invitation by institutions of power. When citizens are mere users, “the state is of interest almost exclusively to professionals, specialists in ‘political science’” ([49]: 754, vol. 3). That is, Lefebvre argues, the state extends the privilege to participate to citizens only when needed. Such privileged moments may be activities during which structures of power designed by experts are merely refined with input and feedback provided by users. From Critical Design to Contestational Design Contestational design examines how design can provoke and engage ‘the political’—the conflictual values constitutive of human relations [17, 31]. It aims to challenge beliefs, values, and assumptions by revealing power relationships, reconfiguring what and who is left out, and articulating collective values and counter-values [17]. Contestational design takes a more explicit confrontational approach than critical design does, reflecting an activist stance: “it is an openly partisan activity that advances a particular set of interests, often at the expense of another” ([31]: 11). Based on the political theory of Chantal Mouffe [56], contestational design sees dissensus and confrontation as inherent yet productive aspects of democracy, drawing attention to the plurality of viewpoints that fundamentally can never be fully resolved. Rather than working to resolve differences, contestational design embraces pluralism and seeks ways to engage critically with contentious issues of public concern. Privileged moments can further be understood reflexively by returning to de Certeau’s ‘strategies’ ([13]: 35ff.). For de Certeau, those with power—i.e., with proprietorship over a space and able to delineate a space of their own distinct from others—are the ones who employ ‘strategies’ (e.g., laws and regulations, urban design). Strategies of the state, then, define the frame within which citizens can go about their lives. Citizens do not themselves have power over these structures; in order for citizens to influence the frame itself, i.e., the structures of power, the state has to extend opportunities of influence to citizens and it does so only as deemed appropriate—i.e., in privileged moments. Consequently, civic engagement and the political are often relegated to the periphery of everyday life and to specialists who define strategies and identify and construct privileged moments. Product-Residue An understanding of the everyday as product-residue, in contrast, emphasizes the ways in which political life is everyday life ([49]: 114, vol. 1). For Lefebvre, an everyday political life is much more than privileged moments; it is both the product of meaning constructed across specialized and fragmented civic activities (including privileged moments) and the residue of civic life lived between these activities. DIMENSIONS FOR THE INFRASTRUCTURING OF CIVIC ENGAGEMENT Synthesizing the literature about contestational design and the everyday, we introduce two cross-cutting dimensions that, together, provide a framework for understanding the infrastructuring of civic engagement. Our analysis foregrounds a design space of untapped potential for HCI researchers to provoke civic engagement through contestational design in everyday life. An understanding of Lefebvre’s product-residue can be expanded through theoretical resonances with de Certeau’s ‘tactics’. Tactics embody the product-residue. De Certeau characterizes tactics as the practices by which people appropriate the structures they are confronted with, i.e., structures framed through ‘strategies’ ([13]: 35ff.). By Everydayness Theories of the everyday foreground two perspectives on how politics and civic engagement can be experienced—as confined to ‘privileged moments’ or as experienced through 148 Contestation and Critique putting these structures to use, people invariably produce understandings, interpretations, and opportunities—spaces become neighborhoods, from entertainment people derive values, and grammar and vocabulary are put to use in language [13]. People, then, are not mere consumers of structures of power (of politics). Within and through the ordinary and everyday, people are producers of their own civic lives, engaging with these political structures and thereby actively appropriating them. A contrasting perspective on civic participation understands democracy as a condition of forever-ongoing contestation and ‘dissensus’ [17, 56]. Political theorist Chantal Mouffe has called this ‘agonistic pluralism’ [56]: a fundamental multiplicity of voices inherent in social relations that are forever contentious and will never be resolved through mere rationality. Agonistic pluralism is a perspective that seeks to transform antagonism into agonism, moving from conflict among enemies to constructive controversies among ‘adversaries’ forever in disagreement. Agonistic pluralism emphasizes the non-rational and more affective aspects of political relations. It sees contestation and dissensus as integral, productive, and meaningful aspects of democratic society. As DiSalvo [17] points out: De Certeau’s tactics embody the product and the residue because they concern conscious and unconscious practices, the activities and their results, the given structures and their derived meanings. In being tactical, people actively engage with, adapt, and appropriate the physical, social, cultural, and political structures in the ordinary and everyday. This appropriation and enactment is fundamentally political— moving people form passive consumers to active producers of issues of public concern. From an agonistic perspective, democracy is a situation in which the facts, beliefs, and practices of a society are forever examined and challenged. For democracy to flourish, spaces of confrontation must exist, and contestation must occur. ([17]: 5) Paradigms of Political Participation Scholars in contestational design have pointed toward alternative understandings of civic engagement and the role that technology and design might play in countering the political status quo. Drawing from the political theory of Mouffe [56], they argue for an understanding of civic engagement based on two contrasting theories of and approaches to democracy and participation—what might best be described as a consensual and a contestational view [2, 17, 31]. Designers have drawn from Mouffe’s theoretical position in multiple ways: directly in the form of contestational or adversarial design [17, 31] and indirectly in designing for activist technologies and technologies for protest [2, 33, 35, 37]. In this view, reflection and critical thinking are at the core of civic processes and activities, and provocation and contestation are seen as means to attain these values. FOUR APPROACHES ENGAGEMENT Consensus and Convenience TO DESIGNING FOR CIVIC The two cross-cutting dimensions described above provide the theoretical framework for understanding the infrastructuring of civic engagement. Through our analysis of socio-technical research in the domain of civic engagement, we characterize each of the four quadrants in that design space. These four approaches—deliberation, situated participation, disruption, and friction—have been taken up, albeit unevenly, by designers or researchers to foster civic engagement (Figure 1). The consensus and convenience paradigm of political participation emphasizes rationality and consensus as the basis for democratic decision-making and action (e.g., [26, 60, 61]; see [55]). It subscribes to the idea that rational compromise and consensus can be arrived through the deliberation of diverging arguments and viewpoints. Efforts to foster civic engagement following this paradigm typically focus on involving citizens in an efficient and inclusive manner. Deliberation As DiSalvo [17] argues, a typical trope of e-democracy initiatives within this paradigm is to improve mechanisms of governance generally and to increase participation of the citizenry through convenience and accessibility. In order to better support the administrative operations of government, initiatives within this trope often translate traditional democratic activities into online tools for participation (e.g., e-deliberation, e-voting, etc.; see [2]). The main concerns of such initiatives center around issues of efficiency, accountability, and equitable access to information and means of ordered expression and action such as petitions, balloting, or voting. They seek to retrofit or replace existing civic activities in order to realize established political ideals and maintain the status quo (see [2]). E-government and e-democracy research has embraced opportunities offered by emerging ICTs and broader-based internet access to translate offline activities of civic engagement into computer-mediated online counterparts [63]. Novel platforms that support the deliberation of civic issues such as urban planning or public policy are common within these bodies of research [7, 23]. Here, the focus of design is often on fostering discourse among stakeholders with differing viewpoints in order to arrive at some form of actionable consensus. Other prototypical systems include platforms for collective decision making—often a more formal conclusion to deliberative processes either affirming, rejecting, or choosing among various alternatives—in the form of e-voting [22, 62]. 149 into infrastructures supporting ongoing dialogue (temporal embedding) about a variety of civic issues, from neighborhood living to public service provision. This continuous involvement stands in contrast to civic participation only in ‘privileged moments’. Everydayness Product-residue Privileged moments Paradigms of Political Participation Consensus/convenience Contestation/critique Deliberation Disruption Situated Participation Friction With the rise in popularity of social media, other research in this vein integrates civic engagement and particularly advocacy into people’s online and offline social networks (social embedding) [16, 28, 53, 65]. Such research seeks to build, extend, or connect communities of interest around shared causes. Its focus ranges from generating awareness of civic issues [28], to enabling debate and discussion [53, 65], to motivating and rallying for action [16, 53]. A third strand of embedded approaches to fostering civic engagement is to spatially align engagement opportunities with people’s whereabouts in the city (spatial embedding) —particularly in domains where spatiality plays a central role such as urban planning [40, 64]. Spatially situated engagement seeks to make relevant and meaningful to people the issues and topics of discussion that are in their close proximity as they move about their day [40]. Spatially situated technologies for civic engagement range from being stationary (installed at places of interest; e.g., [36, 64, 69, 75]), to mobile (typically location-based and often with rich media capturing capabilities; e.g., [5, 27, 41, 50]), to ubiquitous (more deeply embedded into the fabric of the city in the form of sensors, smaller pervasive displays, ubiquitous input/output modalities, etc.; e.g., [43, 72]). Figure 1. Approaches to designing for civic engagement. This strand of research often reflects an understanding of civic engagement that is extended as an explicit invitation for participation from the state to the citizen, invitations that are proffered only periodically, either at major electoral junctures or for significant public projects. We see less research supporting discourse about citizen-originated issues in this quadrant; those discourses are more commonly initiated, instead, through everyday, productresidue infrastructures of engagement. Research supporting deliberation has been foundational in moving civic engagement online and much of this research has shown promise in finding ways to increase the participation in and quality of civic discourse, enabling more broad-based, actionable forms of consensus. But this strand of research still relegates citizenship to the periphery of everyday life taking place only in privileged moments. While researchers design for situated participation are concerned with unobtrusively integrating civic engagement into people’s daily lives, they, at the same time, are nudging and reminding people of participation opportunities that may be relevant to their interests through strategies such as personalization or notifications [5, 40]. Some of the research focused on embedding civic engagement into everyday life manifests a view of civic engagement as a form of one-way issue or bug reporting on public services and infrastructures [24, 42]. Citizens are encouraged to report the small, mundane nuisances that get in the way of leading a ‘productive’ everyday life. This research often mines local knowledge, relying on citizens— as ‘users’ of a city—to maintain, repair, and improve the efficient operation of public services and infrastructures. A related, but larger-scale, strand of research leverages large crowds of people in a more automated fashion (e.g., as ‘sensors’) to contribute data about issues of civic relevance such as problems with physical infrastructure or shared environmental concerns [54, 57]. In contrast to these two strands, other approaches seek to foster active dialogue, community, and more permanent relationships among citizens and with the state [5, 41, 50]. Situated Participation The strand of research fostering civic engagement through situated participation has capitalized on opportunities to embed civic engagement in everyday life through novel networked, mobile, and ubiquitous technologies. These systems resonate with Weiser’s vision of ubiquitous computing by emphasizing technology that dissolves into the everyday [77], interleaving civic engagement between and across temporal, social, and spatial contexts of activity. Through this embedding, research supporting situated participation has aimed at aggregating local knowledge and understanding local needs of citizens that can be valuable resources in planning processes but are often difficult to obtain. Increasingly, research in e-democracy and e-government has shifted from facilitating state-initiated invitations to deliberation to leveraging online platforms for anytime, ongoing civic interactions [6, 11, 38, 73]. This research has transformed previously discrete engagement mechanisms In sum, research in this quadrant focuses on rationale dialogue and harmonious relationships—both among communities and with the state. And it does so while 150 embodying the product-residue. This research emphasizes the in-between aspects of everyday life. Designs target people’s commuting, going about, and everyday curiosity, seeking to only minimally disrupt by providing options for quick feedback and the capturing of small memory cues for later. Research here also targets the holistic product of the fragmented activities of everyday life by fostering productive dialogue, community building, and sustained relationships among citizens and with the state. Privileged moments of civic disobedience are also reflected in Participatory Design research [15, 18, 51]. Inspired by Agre’s [1] ‘critical technical practice’ and Ratto’s [59] ‘critical making’—which emphasize linking socio-political critique with technical projects through hands-on construction and situated reflection—this research invites and encourages citizens to co-construct alternative values during participatory design workshops. It concerns itself with exploring alternative values related to environmental sustainability, local neighborhood communities, or privacy in governmental identification schemes [15, 18]. Disruption Research fostering disruption provides mechanisms for citizens to reveal, address, reflect on, and/or call into question the status quo of values, assumptions, and beliefs held by a community. It does so by focusing on ‘privileged moments’ of dissensus, protest, and civic disobedience. According to Castells [12], moments of collective civic disobedience arise when people overcome anxiety through moments of outrage and anger, when others experience similar moments of injustice that they also share, and when enthusiasm emerges and hope rises through identification and togetherness ([12]: 13–15). Whereas this contestational strand of research in civic HCI began by supporting individual protest activities and moments of civic disobedience, it has increasingly acknowledged recurring practices, the re-appropriation of technologies, and the need for infrastructural support for continued and ongoing activism. Friction Research employing friction as a design strategy embodies both an engagement with the product-residue of the everyday and with a philosophy that politics is fundamentally contestational. Here, civic engagement is emancipated from privileged moments—those fragmented or, even, recurring fragmented activities—and interleaved into the product and residue of everyday life. This transition shifts the unit of analysis from the activity to the smallerscale gaps and spaces in between activities and the largerscale implications of those activities. The contestational design approach provokes citizens to reflect and question the status quo with respect to the conditions of the productresidue of civic life. We find a dearth of research that embodies this approach to designing for civic engagement, and so we engage with the only two examples we can find in more detail. Researchers have studied the use of technologies during demonstrations, occupations of public squares, and protest actions at sites of interest to the local community [2, 35, 71]. Research in this vein frequently focuses on communication and coordination practices. Such practices concern, e.g., the sharing, mobilization, and dissemination of causes and protest actions on social media [2, 71], or the coordination work required during decentralized forms of protest, e.g., via FM radio, mobile phones, and text messaging [33, 35]. Activist moments of civic engagement typically entail responding opportunistically to dynamic political, legal, and technical environments [33]. Hence, activist technologies often support immediate, short-lived campaigns and events that result in a number of individual protest actions [2, 33]. However, research has also begun to focus beyond the individual system to larger infrastructural goals. Hirsch [32] has found that while activists appropriate technologies for individual projects in a quick-and-dirty fashion, they often aim to serve multiple communities or protest actions on the basis of a single development effort. In addition, activists adopt and implement these technologies in ways that create infrastructural redundancies and resilience [33]. Asad and Le Dantec [2] have similarly argued for open, flexible, and underdetermined design of platforms for activists that support a range of practices over time. Clement et al. [15] create friction through the adversarial redesign of identity infrastructures that receive everyday use. Their speculative overlays for government-issued ID cards allow citizens to temporarily black out information unnecessarily exposed by default in numerous situations (e.g., buying alcohol requires sharing your photo and date of birth but not your name or address). The overlays present small obstacles in a larger infrastructure that citizens are confronted with on an everyday basis, provoking them to question the means of government identification and reflect on privacy more generally. We find another example of friction in research that has explored alternative and oppositional media [32, 34, 51]. Activists have long sought to facilitate the exchange of alternative voices—particularly, but not exclusively so within state-controlled (i.e., monitored and/or censored) media landscapes. Research in alternative media explores ways to bypass government control and allow the dissemination of potentially dissenting information via alternative news channels. By building on top of other In addition to supporting individual or repeated activities of protest and activism, research has also studied situations of ongoing crisis, military occupation, and open conflict [71, 78, 79]. Wulf et al. [78], e.g., found that social media came to be used as a way to organize regular demonstrations, to communicate and interact with local and global networks of supporters, and to offer information to the broader public. 151 stable and pervasive infrastructures (e.g., the web [34] or mobile phone networks [32]), at times in parasitic relationships, activists subvert these infrastructures and put them to new and alternative use. Here, alternative media use escapes privileged moments by undercutting the power relationship that allows the state to define what constitutes privilege. residue between activities and in doing so counter the stasis and complacency caused by typical infrastructures. 3. Designs for friction are naïve and inferior. They are not intelligent. Designs for friction do not take agency away from the citizenry. They do not make use of ‘intelligent’ algorithms to anticipate and represent individuals in civic participation. Rather than staking a claim to values and ideals identified a priori by designers, they provoke individuals to articulate and stake a claim to their own values and ideals. Within infrastructures of civic engagement, designing for friction is not primarily about exerting or extending more power per se, but about making room for and creating opportunities for citizens to take power and ownership over issues of public concern. Both examples of design for friction foreground the significant role of infrastructuring when the everyday is emphasized. Yet, as infrastructures are slow to change and susceptible to inertia, particular design strategies have been employed to work through and around existing infrastructures, applying contestation, provocation, and critique to question the status quo and counter inertia. Given the dearth of examples of friction applied in designs for civic engagement and the promise suggested by these initial examples, we turn next to expand on ‘friction’ as a strategy for bringing everyday provocation to the infrastructuring of civic engagement. CREATING FRICTION: INFRASTRUTURING ENGAGEMENT IN EVERYDAY LIFE 4. Designs for friction are not absolute. They acknowledge that although change is desirable, it still lies in the hands of individuals with agency. Designs for friction, because they acknowledge the agency of citizens, cannot impose change. Citizens ultimately retain agency over if and when change happens and if or when reflection about civic life happens. Designs for friction provide opportunities, not mandates. There is always an alternative to acknowledging, responding to, or using infrastructures of civic engagement. Frictional infrastructures do not stop citizens from carrying on as intended; they do not break down completely in the face of inaction or disinterest. Rather, they serve to make citizens pause, to be disruptive without bringing things to a halt. Friction happens in a flicker of a moment in the product-residue of everyday life. CIVIC Hassenzahl et al. have identified four principles for designing persuasive artifacts within an ‘aesthetics of friction’, which we take as a starting point for exploring how to create friction through the infrastructuring of civic engagement [29, 30, 44]: 1. Designs for friction take a position or a stance. They are not neutral. No technologies are value-neutral. And, certainly, any designers who intend to foster civic engagement have taken something of an activist stance in their research, even if implicitly. But designs for friction take an explicit stance toward their users. In the case of infrastructuring civic engagement, designs for friction take the stance that users should be citizens and that the most foundational and requisite work of infrastructuring civic engagement has to come from building the social infrastructure of citizenry—provoking individuals to identify themselves not as users but as citizens in the most active sense possible. Design Strategies Previous research that we see as embodying infrastructural friction, both within and outside the domain of civic engagement, has employed a variety of different strategies for doing so. Synthesizing this existing research, we identify and characterize an initial suite of design strategies for infrastructuring civic engagement by creating friction. • Infrastructuring through intervention When citizens do not maintain control over existing infrastructures, strategies of subversive intervention have been used to create what we would identify as friction. For example, research has sought to intervene in established, large-scale infrastructures by manipulating the representations of infrastructure that citizens hold in their hands (e.g., overlays for government ID cards [15]). This case of everyday appropriation of ‘imposed’ infrastructure is reminiscent of de Certeau’s ‘tactics’, and emphasizes infrastructure’s relational and potentially evolving character at the interlay between standardization and local practices. Research has also taken to “graft a new infrastructure onto an existing one” ([37]: 616) in order to reveal previously invisible relationships between various actors within the existing infrastructure, 2. Designs for friction want to cause trouble. They do not want to help you; rather, they place little obstacles in your way. Designs for friction do not merely blend activities of civic engagement seamlessly into everyday life, enabling participation ‘anytime, anywhere’. They do not make civic engagement more efficient or convenient for citizens. Rather, in the vein of contestational design, creating friction in infrastructures of civic engagement means making citizens pause and reflect—reflect on alternative civic values, reflect on the status quo of civic doing and acting, reflect on the viewpoints of others, and reflect on one’s agency as a citizen… not user. The little obstacles of friction carve out space for reflection in the 152 empowering a central but previously disenfranchised stakeholder population in their everyday (work) lives. Strategies of intervention, then, shift the arrangements of power and, in doing so, nudge the infrastructures themselves toward change. infrastructures of civic engagement in the hands of those in power, and in their best interest, in order to counter issues of stasis and complacency and to reach toward the ideal of the ‘more active citizen’. • Infrastructuring by creating alternatives Other research has created what we would identify as friction by building alternative infrastructures in parallel to existing ones to facilitate a pluralism of voices (e.g., alternative media channels [32, 51] or parallel communication infrastructures [33]). Such alternative infrastructures typically ‘piggy-back’ on existing, stable, and pervasive infrastructures [33], introducing additional layers and ways around established forms of power embedded in infrastructure. This design strategy, then, suggests a mechanism for undercutting and questioning power relationships that define what constitutes privilege and, thus, the structures underlying privileged moments that constrain civic engagement. We see three open questions as being productive areas for future research. FUTURE WORK Facilitating a shift from user to citizen is a cultural process through which peripheral participants are scaffolded or provoked toward more expert participation [45]. While Lave and Wenger caution against understanding situated learning as a movement from the periphery to the center, we do imagine that the movement from user to citizen, from periphery to expert, will interact in some significant ways with where one stands with respect to the institutions of influence over infrastructure. Bowker and Star [9] warn that individuals farther removed from these centers of power will more strongly experience torque from the values embedded into the infrastructure. Studies of infrastructures of civic engagement, then, will need to pay particular attention to understanding how legitimate peripheral participation can be supported while avoiding the disenfranchisement that comes with being too far at the periphery. Frictional infrastructures are likely to be a good first step as the little obstacles they place in the everyday may more readily suggest possibilities of opting out than other infrastructures (see also [52]), but this is a question that necessitates more empirical work. • Infrastructuring by making gaps visible To foreground and leave open (rather than to close and remove) gaps or seams within and between infrastructures is another strategy that we identify from previous research as embodying friction (e.g., [52]). ‘Seamful design’ does so by “selectively and carefully revealing differences and limitations of systems” ([14]: 251) and the underlying infrastructural mechanisms between them. If appropriated for friction, moments of infrastructural breakdown can become moments of awareness, reflection, and questioning about the activities that infrastructures enable and the values inscribed in them, moving beyond the mere feedback cycles of ‘users’ of public services. For infrastructures of civic engagement, then, seams and gaps provide both the space in which to design for friction as well as the substance of everyday residue on which citizens may be provoked to reflect. In this paper, we advocate for frictional design contesting the status quo of citizen–state relationships. Yet, Lefebvre —and the Situationists he inspired—as well as de Certeau also deeply engage with questions of ‘passive’ consumers as political actors [13, 49]. Research has already sought to inject and intervene in the prevailing metaphor of markets [37, 39]. Infrastructuring corporate arrangements of power, then, is a direction for which future research may fruitfully mobilize frictional design strategies. • Infrastructuring by using trace data for critique A final strategy amenable for creating friction is one of employing trace data of infrastructural use in order to critique those infrastructures, or even to reveal them in the first place. For example, researchers have visualized the traces individuals leave behind when surfing the web in order to uncover and ultimately critique large-scale, commercial data mining practices [39]. Working with trace data hints at such a strategy’s power to emphasize the relationship between the residue of activities and their product. Working with trace data related to civic engagement could foreground issues of data ownership and opportunities for challenging power dynamics [76]. Lastly, we assume quite pragmatically that the research community will need to explore a balance between the benefits and annoyances of friction. Friction is likely not to scale, for example, to every situation of civic engagement in which we expect individuals to find themselves. An empirical question for future research, then, is for whom, in what contexts, and over what periods of time is friction useful—and for whom, in what contexts, and after what periods of time does it stop being useful—for addressing the dual challenges of infrastructuring civic engagement. There are also related questions about how one’s identity and participation as a citizen evolves, as we assume that designs for friction will need to be nimble enough to engage citizens through the ebb and flow of engagement. From an infrastructuring perspective, this is ultimately a question about sustainability. The four design strategies described above primarily act as subversive forces, working to enact change from outside the centers of power over these infrastructures. However, frictional design does not necessarily have to be applied exclusively from the outside. Rather, we posit friction to be a mechanism that is also applicable to the design of 153 CONCLUSION indicators for a large-scale urban simulation. In Proc. ECSCW 2005. Springer, 449–468. 8. Bowker, G. (1994). Information mythology: The world of/as information. In L. Bud-Frierman (Ed.), Information Acumen: The Understanding and Use of Knowledge in Modern Business (pp. 231–247). London: Routledge. 9. Bowker, G.C. & Star, S.L. (1999). Sorting Things Out. Cambridge: MIT Press. 10. Bødker, S. (2006). When second wave HCI meets third wave challenges. In Proc. NordiCHI 2006. ACM Press, 1–8. 11. Carroll, J.M. & Rosson, M.B. (1996). Developing the Blacksburg electronic village. Commun. ACM 39(12), 69–74. 12. Castells, M. (2012). Networks of Outrage and Hope: Social Movements in the Internet Age. Cambridge: Polity Press. 13. de Certeau, M. (1984). The Practice of Everyday Life. Berkeley: University of California Press. 14. Chalmers, M. & Galani, A. (2004). Seamful interweaving: heterogeneity in the theory and design of interactive systems. In Proc. DIS 2004. ACM Press, 243–252. 15. Clement, A., McPhail, B., Smith, K.L., & Ferenbok, J. (2012). Probing, mocking and prototyping: participatory approaches to identity infrastructuring. In Proc. PDC 2012, ACM Press, 21–30. 16. Crivellaro, C., Comber, R., Bowers, J., Wright, P.C., & Olivier, P. (2014). A pool of dreams: Facebook, politics and the emergence of a social movement. In Proc. CHI 2014. ACM Press, 3573–3582. 17. DiSalvo, C. (2012). Adversarial Design. Cambridge: MIT Press. 18. DiSalvo, C., Nourbakhsh, I., Holstius, D., Akin, A., & Louw, M. (2008). The Neighborhood Networks project: a case study of critical engagement and creative expression through participatory design. In Proc. PDC 2008. ACM Press, 41–50. 19. Dourish, P. & Bell G. (2007). The infrastructure of experience and the experience of infrastructure: meaning and structure in everyday encounters with space. Environment and Planning B: Planning and Design 34(3), 414–430. 20. Dourish, P. & Bell, G. (2011). Divining a Digital Future: Mess and Mythology in Ubiquitous Computing. Cambridge: MIT Press. 21. Dunne, A. & Raby, F. (2001). Design Noir: The Secret Life of Electronic Objects. Basel: Birkhäuser. 22. Everett, S.P., Greene, K.K., Byrne, M.D., Wallach, D.S., Derr, K., Sandler, D., & Torous, T. (2008). Electronic voting machines versus traditional methods: improved Bowker [8] has argued for research to undertake ‘infrastructural inversion’—subverting the traditional figure/ground relationship that perpetuates the analytic invisibility of infrastructures. In this research, we reposition infrastructural inversion as an approach to design rather than as ethnographic practice. The use of friction when infrastructuring civic engagement is a designerly enactment of infrastructural inversion. Friction is positioned to foreground infrastructures through everyday obstacles that counter the potential of stasis and complacency. In this paper, we have introduced theories of the everyday to the emerging bodies of research on contestational design and infrastructures of civic engagement. Our research contributes a design space distilling and describing four distinct approaches to designing for civic engagement, including deliberation, disruption, situated participation, and friction. We argue that there is untapped potential for designing for friction—for leveraging critique and contestation as means of re-unifying politics and the everyday. ACKNOWLEDGMENTS We thank Susann Wagenknecht for feedback on early drafts of the paper. This material is based upon work supported by the National Science Foundation under Grant Number 1360035. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. REFERENCES 1. Agre, P. (1997). Toward a critical technical practice: Lessons learned in trying to reform AI. In G. Bowker, S.L. Star, W. Turner, & L. Gasser (Eds.), Social Science, Technical Systems, and Cooperative Work: Beyond the Great Divide, Hillsdale: Erlbaum, 131–158. 2. Asad, M. & Le Dantec, C. (2015). Illegitimate civic participation: supporting community activists on the ground. In Proc. CSCW 2015. ACM Press, 1694–1703. 3. Bardzell, J. & Bardzell, S. (2013). What is “critical” about critical design? In Proc. CHI 2013. ACM Press, 3297–3306. 4. Bardzell, S., Bardzell, J., Forlizzi, J., Zimmerman, J., & Antanitis, J. (2012). Critical design and critical theory: the challenge of designing for provocation. In Proc. DIS 2012. ACM Press, 288–297. 5. Bohøj, M., Borchorst, N.G., Bødker, S., Korn, M., & Zander, P.-O. (2011). Public deliberation in municipal planning: supporting action and reflection with mobile technology. In Proc. C&T 2011. ACM Press, 88–97. 6. Borchorst, N.G., Bødker, S., & Zander, P.-O. (2009). The boundaries of participatory citizenship. In Proc. ECSCW 2009. Springer, 1–20. 7. Borning, A., Friedman, B., Davis, J., & Lin, P. (2005). Informing public deliberation: value sensitive design of 154 preference, similar performance. In Proc. CHI 2008. ACM Press, 883–892. 23. Farina, C., Epstein, D., Heidt, J., & Newhart, M. (2014). Designing an online civic engagement platform: Balancing ‘more’ vs. ‘better’ participation in complex public policymaking. International Journal of E-Politics 5(1), 16–40. 24. Foth, M., Schroeter, R., & Anastasiu, I. (2011). Fixing the city one photo at a time: mobile logging of maintenance requests. In Proc. OzCHI 2011. ACM Press, 126–129. 25. Galloway, A. (2004). Intimations of everyday life: Ubiquitous computing and the city. Cultural Studies 18(2/3), 384–408. 26. Habermas, J. (1996). Between Facts and Norms. Contributions to a Discourse Theory of Law and Democracy. Cambridge: MIT Press. 27. Halttunen, V., Juustila, A., & Nuojua, J. (2010). Technologies to support communication between citizens and designers in participatory urban planning process. In S. Wallin, L. Horelli, & J. Saad-Sulonen (Eds.), Digital Tools in Participatory Planning (pp. 79–91). Centre for Urban and Regional Studies, Aalto University. 28. Han, K., Shih, P.C., & Carroll, J.M. (2014). Local News Chatter: augmenting community news by aggregating hyperlocal microblog content in a tag cloud. International Journal of Human-Computer Interaction 30(12), 1003–1014. 29. Hassenzahl, M. (2011). Towards an Aesthetic of Friction. TEDxHogeschoolUtrecht. http://youtu.be/ ehWdLEXSoh8 (Retrieved: Dec. 10, 2014). 30. Hassenzahl, M. & Laschke, M. (2015). Pleasurable troublemakers. In S.P. Walz & S. Deterding (Eds.), The Gameful World: Approaches, Issues, Applications (ch. 6, pp. 167–195). Cambridge: MIT Press. 31. Hirsch, T. (2008). Contestational Design: Innovation for Political Activism. Unpublished PhD Dissertation, Massachusetts Institute of Technology. 32. Hirsch, T. (2009). Communities real and imagined: designing a communication system for Zimbabwean activists. In Proc. C&T 2009. ACM Press, 71–76. 33. Hirsch, T. (2009). Learning from activists: lessons for designers. interactions 16(3), 31–33. 34. Hirsch, T. (2011). More than friends: social and mobile media for activist organizations. In M. Foth, L. Forlano, C. Satchell, & M. Gibbs (Eds.), From Social Butterfly to Engaged Citizen (pp. 135–150). Cambridge: MIT Press. 35. Hirsch, T. & Henry, J. (2005). TXTmob: Text messaging for protest swarms. In Ext. Abstracts CHI 2005. ACM Press, 1455–1458. 36. Hosio, S., Kostakos, V., Kukka, H., Jurmu, M., Riekki, J., & Ojala, T. (2012). From school food to skate parks in a few clicks: using public displays to bootstrap civic engagement of the young. In Proc. Pervasive 2012. Springer, 425–442. 37. Irani, L.C. & Silberman, M.S. (2013). Turkopticon: Interrupting worker invisibility in Amazon Mechanical Turk. In Proc. CHI 2013. ACM Press, 611–620. 38. Kavanaugh, A., Carroll, J.M., Rosson, M.B., Zin, T.T., & Reese, D.D. (2005). Community networks: where offline communities meet online. Journal of ComputerMediated Communication 10(4). 39. Khovanskaya, V., Baumer, E.P.S., Cosley, D., Voida, S., & Gay, G. (2013). “Everybody knows what you’re doing”: a critical design approach to personal informatics. In Proc. CHI 2013. ACM Press, 3403–3412. 40. Korn, M. (2013). Situating Engagement: Ubiquitous Infrastructures for In-Situ Civic Engagement. Unpublished PhD Dissertation, Aarhus University. 41. Korn, M. & Back, J. (2012). Talking it further: from feelings and memories to civic discussions in and about places. In Proc. NordiCHI 2012. ACM Press, 189–198. 42. Korsgaard, H. & Brynskov, M. (2014). City bug report: urban prototyping as participatory process and practice. In Proc. MAB 2014. ACM Press, 21–29. 43. Kuznetsov, S. & Paulos, E. (2010). Participatory sensing in public spaces: activating urban surfaces with sensor probes. In Proc. DIS 2010. ACM Press, 21–30. 44. Laschke, M., Hassenzahl, M., & Diefenbach, S. (2011). Things with attitude: Transformational products. In Proc. Create11: The Interaction Design Symposium, London, UK, 23rd June 2011, http://www.createconference.org/storage/create11papersposters/Things%2 0with%20attitude.pdf (Retrieved: Dec. 10, 2014). 45. Lave, J. & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press. 46. Leahu, L., Thom-Santelli, J., Pederson, C., & Sengers, P. (2008). Taming the Situationist beast. In Proc. DIS 2008. ACM Press, 203–211. 47. Le Dantec, C.A. & DiSalvo, C. (2013). Infrastructuring and the formation of publics in participatory design. Social Studies of Science 43(2), 241–264. 48. Lee, C.P., Dourish, P. & Mark, G. (2006). The human infrastructure of cyberinfrastructure. In Proc. CSCW 2006. ACM Press, 483–492. 49. Lefebvre, H. (2014). Critique of Everyday Life. London: Verso. 50. Lehner, U., Baldauf, M., Eranti, V., Reitberger, W., & Fröhlich, P. (2014). Civic engagement meets pervasive gaming: towards long-term mobile participation. In Ext. Abstracts CHI 2014. ACM Press, 1483–1488. 51. Lievrouw, L.A. (2006). Oppositional and activist new media: remediation, reconfiguration, participation. In Proc. PDC 2006. ACM Press, 115–124. 155 52. Mainwaring, S.D., Cheng, M.F., & Anderson, K. (2004). Infrastructures and their discontents: Implications for Ubicomp. In Proc. Ubicomp 2004. Springer, 418–432. 53. Mascaro, C.M. & Goggins, S.P. (2011). Brewing up citizen engagement: the Coffee Party on Facebook. In Proc. C&T 2011. ACM Press, 11–20. 54. Monroy-Hernández, A., Farnham, S., Kiciman, E., Counts, S., & De Choudhury, M. (2013). Smart societies: from citizens as sensors to collective action. interactions 20(4), 16–19. 55. Mouffe, C. (2000). Deliberative Democracy or Agonistic Pluralism. Political Science Series 72. Vienna: Institute for Advanced Studies. 56. Mouffe, C. (2013). Agonistics: Thinking the World Politically. London: Verso. 57. Paulos, E., Honicky, RJ, & Hooker, B. (2009). Citizen science: Enabling participatory urbanism. In M. Foth (Ed.), Handbook of Research on Urban Informatics (ch. XXVIII, pp. 414–436). Hershey, PA: IGI Global. 58. Pipek, V. & Wulf, V. (2009). Infrastructuring: Toward an integrated perspective on the design and use of information technology. Journal of the Association of Information Systems 10(5), 447–473. 59. Ratto, M. (2011). Critical making: Conceptual and material studies in technology and social life. The Information Society 27(4), 252–260. 60. Rawls, J. (1971). A Theory of Justice. Cambridge: Harvard University Press. 61. Rawls, J. (1993). Political Liberalism. New York: Columbia University Press. 62. Robertson, S.P. (2005). Voter-centered design: Toward a voter decision support system. ACM Trans. Comput.Hum. Interact. 12(2), 263–292. 63. Robertson, S.P. & Vatrapu, R.K. (2010). Digital government. Ann. Rev. Info. Sci. Tech. 44(1), 317–364. 64. Schroeter, R., Foth, M., & Satchell, C. (2012). People, content, location: sweet spotting urban screens for situated engagement. In Proc. DIS 2012. ACM Press, 146–155. 65. Semaan, B.C., Robertson, S.P., Douglas, S., & Maruyama, M. (2014). Social media supporting political deliberation across multiple public spheres: towards depolarization. In Proc. CSCW 2014. ACM Press, 1409–1421. 66. Sengers, P., Boehner, K., David, S., & Kaye, J. (2005). Reflective design. In Proc. Aarhus 2005: Critical Computing. ACM Press, 49–58. 67. Star, S.L. & Bowker, G.C. (2006). How to infrastructure. In L.A. Lievrouw & S. Livingstone (Eds.), Handbook of New Media: Student Edition (pp. 230–245). Thousand Oaks: Sage. 68. Star, S.L. & Ruhleder, K. (1996). Steps toward an ecology of infrastructure: Design and access for large information. Information Systems Research 7(1), 111– 134. 69. Taylor, N., Marshall, J., Blum-Ross, A., Mills, J., Rogers, J., Egglestone, P., Frohlich, D.M., Wright, P., & Olivier, P. (2012). Viewpoint: empowering communities with situated voting devices. In Proc. CHI 2012. ACM Press, 1361–1370. 70. Tsing, A.L. (2005). Friction: An ethnography of global connection. Princeton: Princeton University Press. 71. Tufekci, Z. & Wilson, C. (2012). Social media and the decision to participate in political protest: Observations from Tahrir Square. Journal of Communication 62(2), 363–379. 72. Vlachokyriakos, V., Comber, R., Ladha, K., Taylor, N., Dunphy, P., McCorry, P., & Olivier, P. (2014). PosterVote: expanding the action repertoire for local political activism. In Proc. DIS 2014. ACM Press, 795– 804. 73. Voida, A., Dombrowski, L., Hayes, G.R., & Mazmanian, M. (2014). Shared values/conflicting logics: working around e-government systems. In Proc. CHI 2014. ACM Press, 3583–3592. 74. Voida, A., Yao, Z., & Korn, M. (2015). (Infra)structures of volunteering. In Proc. CSCW 2015. ACM Press, 1704–1716. 75. Wagner, I., Basile, M., Ehrenstrasser, L., Maquil, V., Terrin, J.-J., & Wagner, M. (2009). Supporting community engagement in the city: urban planning in the MR-tent. In Proc. C&T 2009. ACM Press, 185–194. 76. Weise, S., Hardy, J., Agarwal, P., Coulton, P., Friday, A., & Chiasson, M. (2012). Democratizing ubiquitous computing: a right for locality. In Proc. UbiComp 2012. ACM Press, 521–530. 77. Weiser, M. (1991). The computer for the 21st century. Scientific American 265, 94–104. 78. Wulf, V., Aal, K., Kteish, I.A., Atam, M., Schubert, K., Rohde, M., Yerousis, G.P., & Randall, D. (2013). Fighting against the wall: social media use by political activists in a Palestinian village. In Proc. CHI 2013. ACM Press, 1979–1988. 79. Wulf, V., Misaki, K., Atam, M., Randall, D., & Rohde, M. (2013). 'On the ground' in Sidi Bouzid: investigating social media use during the Tunisian revolution. In Proc. CSCW 2013. ACM Press, 1409–1418. 80. Zerubavel, E. (1991). The Fine Line: Making Distinctions in Everyday Life. Chicago: University of Chicago Press. 156 Not The Internet, but This Internet: How Othernets Illuminate Our Feudal Internet Paul Dourish Department of Informatics University of California, Irvine Irvine, CA 92697-3440, USA jpd@ics.uci.edu built – the Internet, this Internet, our Internet, the one to which I’m connected right now? ABSTRACT What is the Internet like, and how do we know? Less tendentiously, how can we make general statements about the Internet without reference to alternatives that help us to understand what the space of network design possibilities might be? This paper presents a series of cases of network alternatives which provide a vantage point from which to reflect upon the ways that the Internet does or does not uphold both its own design goals and our collective imaginings of what it does and how. The goal is to provide a framework for understanding how technologies embody promises, and how these both come to evolve. I ask these questions in the context of a burgeoning recent interest in examining digital technologies as materially, socially, historically and geographically specific [e.g. 13, 15, 36, 37]. There is no denying the central role that “the digital,” broadly construed, plays as part of contemporary everyday life. Wireless connectivity, broadband communications, and computational devices may be concentrated in the urban centers of economically privileged nations, but even in the most “remote” corners of the globe, much of everyday life is structured, organized, and governed by databases and algorithms, and “the digital” still operates even in the central fact of its occasional absence, the gap that renders particular spots “off grid.” Where digital technologies were once objects of primarily engineering attention, their pervasive presence has meant that other disciplines – anthropology, political science, communication, sociology, history, cultural studies, and more – have had to grapple with the question, “what are the cultural consequences of ‘the digital’?” The problem, however, has been to get a grip up on what ‘the digital’ itself is. Rather too often, ‘the digital’ is taken simply as a metaphor for regimentation and control, or it is used to name some amorphous and unexamined constellation of representational and computational practices. The somewhat casual assertion that “the Internet” has some particular properties runs this danger. It’s difficult to know how to read or assess these assertions without, for example, understanding something of the scope of what is being claimed through some kind of differential analysis of “the not-Internet.” We need to take an approach to “the Internet” that begins with an understanding of its technical reality, although not one that reduces it to simply that. In other words, we want to maintain a focus on the social and cultural reality of technologies such as Internet, but in a way that takes seriously its material specificities. Author Keywords Network protocols; network topology; naming routing; media infrastructures. ACM Classification Keywords H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous. INTRODUCTION What does it mean to say, as many have, that “the Internet treats censorship as damage and routes around it” [14]? Or what does it mean to argue, as is also common, that the Internet is a democratizing force as a result of its decentralized architecture [8]? Or that it’s a platform for grass-roots community-building rather than mass media messaging [35]? Statements like these have been critiqued for their treatment of topics such as democracy, censorship, broadcasting, or community – all complicated terms that are perhaps invoked and conjured more than they are critically examined [19]. However, the term that I want to focus on in these statements is “the Internet”. When we attribute characteristics to the Internet, specifically, what do we mean? Do we just mean “digital networks”? Do we mean digital networks that implement the Internet protocols? Or do we mean the one very specific network that we have Two Transitions To make this argument more concrete, let me begin by describing two cases from my own working history, two digital transitions that illustrate the complexity of talking about the Internet as a force for decentralization. Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark The first of these transitions occurred in the early 1990s when I worked for Xerox. Xerox had been a pioneer in DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21200 157 digital networking. Early research on Ethernet and distributed systems constituted an important precursor to the development of the Internet Protocol (IP) suite, and the research systems such as PUP [4] and Grapevine [3] had subsequently given rise to a protocol suite called XNS (Xerox Network Services), which was the basis of a line of digital office systems products that Xerox sold in the marketplace. Xerox’s own corporate network spanned the globe, linking thousands of workstations and servers together using the XNS protocols. In the 1990s, as Unix workstations began to dominate the professional workstation market, and as the arrival of distributed information services such as Gopher, WAIS, and WWW spurred the accelerated growth of the Internet, many inside Xerox became interested in using TCP/IP on internal networks too. Particular at research and development centers, various groups began to run TCP/IP networks locally and then increasingly looked for ways to connect them together using the same leased lines that carried XNS traffic between sites. What began as renegade or illicit actions slowly became organizationally known and tolerated, and then organizationally supported as TCP/IP became a recognized aspect of internal corporate networking within Xerox. operated. Since XNS was the dominant technology for organizational communication, it wasn’t entirely possible for people using TCP/IP to “route around” the corporate network, but it started to provide some independence. The second transition was also a transition to IP, but in a very different context. This transition was going on while I briefly worked at Apple in the late 1990s. As at Xerox, the rise of the Internet in general was reflected in the increasing use of TCP/IP in a network that had originally been put together through a different network protocol – in this case, AppleTalk. AppleTalk was a proprietary network suite that Apple developed to connect Macintosh computers; it had evolved over time to operate over the Ethernet networks commonly deployed in corporate settings, although it had originally been developed for linking computers together in relatively small networks. One important feature of the Appletalk networking protocols is their so-called “plugand-play” approach, which allows a network to be deployed with minimal manual configuration. For example, Appletalk does not require that network addresses be preassigned or that a server be available for network resource discovery; these features are managed directly by the networked computers themselves. Accordingly, setting up Appletalk networks requires little or no administrative intervention. TCP/IP networks, on the other hand, do require some services to be set up – DHCP servers to allocate addresses, name servers to resolve network addresses, and so on. (In fact, the contemporary networking technologies known as Bonjour or Zeroconf are mechanisms designed to re-introduce Appletalk’s plug-andplay functionality into TCP/IP networking.) So, where the transition from XNS to TCP/IP was a decentralizing transition at Xerox, one that increased people’s independence from corporate network management, the transition from Appletalk to TCP/IP at Apple moved in the other direction, creating more reliance on network infrastructure and hence on network administration. XNS was a protocol suite designed for corporate environments. Although technically decentralized, it depended on an administratively centralized or managed model. The effective use of XNS was tied together by a distributed database service known as the Clearinghouse, which was responsible for device naming, address resolution, user authentication, access control, and related functions. Users, servers, workstations, printers, email lists, organizational units, and other network-relevant objects were all registered in the Clearinghouse, which was implemented as a distributed network of database servers linked via a so-called “epidemic” algorithm by which they would keep their database records up to date. Access control mechanisms distinguished administrators, who could update Clearinghouse databases, from regular users, who could look up names but couldn’t introduce new ones. The Clearinghouse service was central enough to the operation of XNS services that this administrative access was needed for all sorts of operations, from adding new users to installing new workstations. These examples illustrate two important concerns that animate this paper. The first is that statements about “the Internet” and its political and institutional character suffer for a lack of contrast classes (“critical alternatives,” if you will). The Internet may well be decentralized – but compared to what, exactly? More decentralized internetworks could be imagined and have existed. We might be able to make some more fruitful observations if the “Internet” that we are trying to characterize weren’t such a singular phenomenon. In this paper I will briefly sketch some cases that might serve as points of comparison that provide for more specific statements about contemporary phenomena by showing how they might be, and have been, otherwise. The first concern, then, is to provide a framework within which the characterizations can take on more meaning. The second concern is that the singular nature of the Internet makes it hard for us to distinguish the conceptual object of our attention, the object that we are characterizing. Given that the Internet – the By contrast, the TCP/IP network, and the Unix workstations that it generally linked, was much less centrally administered. For the Unix machines, users could be defined locally for each computer, and similarly, workstations could maintain their own machine addressing and network routing information. Even when systems were interconnected, much less coordination was required to get machines connected together effectiveness in the IP network than in the XNS network. As a result, the rise of the IP network provided a route by which people could to some extent become more independent of the corporate IT management structure through which the XNS network was 158 specific Internet to which we can buy or obtain access today – is generally the only Internet we have known, it’s hard to be able to pin down just what object it is that we are characterizing when we talk about, say, decentralization. Do we mean that a world-spanning network-of-networks is inherently decentralized? Or is decentralization a characteristic of the specific protocols and software that we might use to operate that network (the Internet protocols)? Or is it rather a characteristic of the specific network that we have build, which doesn’t just use those protocols, but implements them in a specific network made of particular connections, an amalgam of undersea fiber-optic cables, domestic WiFi connections, commercial service providers, and so on? Is it our Internet that’s decentralized, while we could still imagine a centralized one being built? Or is our Internet actually less decentralized than it might be, failing to achieve its own promise (if that’s something we want)? read and post messages. In the US, where local phone calls were generally free, BBSs flourish in the late 1970s and early 1980s, as the home computer market grew. With very simple software, they allowed people to communicate by dialing in to the BBS at different times and posting messages that others would read later. While one certainly could make a long-distance call to connect to a BBS, most BBS use was local to take advantage of toll-free calling, so that most BBS activity was highly regionalized. Fido was the name of BBS software first written by Tom Jennings in San Francisco in 1984 and then adopted by others elsewhere in the US. Before long, Fido was updated with code that would allow different Fido BBS systems to call each other to exchange messages; Fidonet is the name both of this software and of the network of BBS systems that exchanged messages through this mechanism. Fidonet’s growth was explosive; from its start in 1985, Fidonet had around 500 nodes by the end of 1985, almost 2,000 by 1987, 12,000 by 1991 and over 35,000 by 1995. Each of these nodes was a BBS that server ten to hundreds of users, who could exchange email messages, files, and discussions on group lists. In order to provide the resources to think about these questions fruitfully, I will approach the topic from two perspectives. The first is to briefly catalog some alternatives to “the” Internet. Some of these are entirely alternative networks; some are small components of the broader Internet that do not always operate the same way as the whole. The second is to take in turn some key aspects of network function – routing, naming, and so on – and examine their contemporary specificities, with particular focus on the relationship between specific commercial and technical arrangements and the openness or range of possibilities encapsulated by the network design. Taken together, these allow us to develop an understanding of the landscape of potential network arrangements within which our current arrangements take their place, and perhaps more accurately then target or assess statements about what “the Internet” is or does. Fidonet’s design was (with one critical exception) radically decentralized. Based as it was on a dial-up model rather than an infrastructure of fixed connections, it employed a model of direct, peer-to-peer communication. The Fidonet software was originally designed with a flat list of up to 250 nodes; system of regions, zones, and networks was introduced within a year of the original software when it became clear that the system would very soon grow beyond that capacity. This structure, which mapped the topology of the network geographically, provided a message routing structure which reduced costs (by maximizing local calling and by pooling messages for long-distance transfer) but with a minimum of fixed structure; direct communication between nodes was always central to the Fidonet model. The structure essentially exhibited a two-level architecture; one of conventional structures (that is, the conventional pattern of daily or weekly connections between sites) and an immediate structure, made up of those nodes communicating with each other right now. (The Internet – being made up in its current form primarily of fixed infrastructure and broadband connectivity – largely conflates these two.) A CATALOG OF OTHERNETS Our starting point are what I have been calling “othernets” – non-Internet internets, if you will. Some of these are networks of a similar style but which happen to use different protocols; some of them are radically different arrangements. Some of them are global networks, and some more localized; some are specific networks and some are ways of thinking about or approaching network design. The collection listed here is far from comprehensive.1 What they do for us here, though, is to flesh out an internetworking space of possibilities, one that helps to place “the Internet” in some context. Within a year, Fidonet was in in danger of hitting an initial limit of 250 nodes, and so the network protocols and software were redesigned around the concepts of regions and nets. Until this point, Fidonet had used a single, “flat” list of nodes, which was directly maintained by Jennings. The introduction of regions and nets allowed for a more decentralized structure. This was simultaneously an addressing structure and a routing structure, linked together – the structure of the network was also the structure that determined how messages would make their way from one place to another. Fidonet A Bulletin Board System (BBS) hosts messages and discussions, generally on a quite simple technology platform such as a conventional home PC with a modem that allows people to call into it over regular phone lines to 1 Most problematically, local and community “DIY” networks usefully illuminate the issues in question but had to be omitted for space [e.g. 1, 16, 21, 30]. 159 Fidonet was built around a file transfer mechanism that would allow files to move from one node to another. Other facilities could be built on top of this mechanism, such as electronic mail. One of the major features of Fidonet from the user perspective was the discussion groups known as “echoes”. Echoes allowed users to post messages that would be distributed to all users on the system interested in a topic. A “moderation” mechanism allowed echo managers to discourage off-topic posts, but this was a post-hoc mechanism (rather than an approach which required explicit approval before a message was sent out). As in systems such as The Well [39], the topically-oriented discussion forums provided by echoes were the primary site of interaction and community building across Fidonet (although unlike The Well, Fido echoes were networked rather than centralized.) connection; our example bang path only works if “mcvax” is one of the computers that “seismo” regularly connects to directly. One couldn’t, for instance, route a message along the path “seismo!ukc!itspna!jpd” because seismo only dials up to certain other computers, and ukc isn’t one of them. Two aspects of this are worth noting. The first concerns the dynamics of network structure in the presence of this routebased addressing mechanism. Route-based addressing via bang paths means not just that you need to understand how your computer is connected to the network; everybody needs to understand how your computer is connected to the network, if they want to reach you, and they need to understand how all the intermediate computers are connected to the network too. This arrangement does not allow, then, for frequent reconfiguration. Should seismo stop talking to mcvax, then people using that connection as part of their routing process will find their routes break. Usenet Usenet is a somewhat informal term for a worldwide network of computers linked by a series of mechanisms built on top of a facility provided by versions of the Unix operating system [18]. The facility was known as uucp, which stands for “Unix-to-Unix copy.” In Unix’s command-line environment, “uucp” was a command that users could use to copy files between computers. It was designed by analogy with the standard “cp” command for copying files; just as a user might use “cp” to copy a file from a source to a destination filename, they might also use “uucp” to copy a file from a source to a destination, where either the source or destination file location was in fact on a different computer. The second noteworthy aspect, a partial consequence of the first, is that within the open, pairwise connection structure afforded by Usenet, a backbone hierarchy of major sites soon arose, at first through conventional practice and later through explicit design. These were major sites that engaged in significant data transfer with each other or with other groups, including ihnp4 at AT&T’s Indian Hill site, seismo at the Center for Seismic Studies in northern Virginia and mcvax at the Mathematics Centrum in Amsterdam, which effectively became the primary transAtlantic gateways, and national sites such as ukc at the University of Kent at Canterbury, which effectively became the gateway to the United Kingdom. Significantly, some of these also served as “well-known nodes” for routing purposes; one might quote one’s email address as “…!seismo!mcvax!itspna!jpd”, with the “…” essentially standing for the directive “use whatever path you use to regularly use to reach here.” The design of the network, then, is unstructured but the practice requires a commitment to some form of well-understood centralization. Uucp was developed as a basic user file-copy facility, but the core idea of file-based interconnections between minicomputers was general enough that many other facilities could be built on top of it. For instance, uucp could be used to support email messaging between sites, exchanging individual messages as files that the remote system would recognize as drafts to be delivered by email locally. The same mechanisms that named remote files, then, might also be used to name remote users. Uucp mail, with its explicit use of bang paths, was not the primary or most visible aspect of Usenet, however. A distributed messaging service which came to be known as Usenet news was first deployed in 1980, using the same underlying uucp mechanism to share files between sites. Unlike email messages directed to specific users, however, articles in Usenet news were open posts organized into topically-organized “newsgroups.” Via uucp, these would propagate between sites. Usenet connected many of the same academic and industrial research sites that came to be incorporated into ARPAnet or its successors, and so over time, internet protocols became a more effective way for messages to be exchanged, at which point, Usenet newsgroups were distributed over TCP/IP rather than uucp. Since it had been designed initially simply to provide a user interface to a dial-up mechanism for file exchange, Usenet provided no global naming mechanism to identify sites, files, objects, or users. Rather, its naming mechanism was a sequence of identifiers (separated by exclamation points or “bangs”) that explained how a message should be routed. So, for instance, the path “seismo!mcvax!ukc!itspna!jpd” directs a computer to deliver the file first to a computer called “seismo”, at which point the path will be rewritten to “mcvax!ukc!itspna!jpd”; subsequently, it will be delivered to a computer called “mcvax”, then to one called “ukc,” and so on. “jpd” is the user, whose account is on the computer “itspna”. To send a message correctly, then, required that one knew not only the destination, but the route that the message should take – the series of peer-to-peer connections that must be made. Each path describes a direct DECnet Digital Equipment Corporation (DEC, or just “Digital”) was a leading designer and vendor of minicomputers through the 1970s and 1980s. Indeed, many of the 160 computers connected to the early ARPAnet/Internet were DEC machines, especially systems in the DEC-10, DEC-20, PDP, and VAX ranges, which were widely used in the academic research community. At the same time, DEC had its own networking system, delivered as part of its RSX and VMS operating systems. This network system, known as DECnet or the Digital Network Architecture (DNA), was initially introduced in the mid-1970s as simple point-topoint connection between two PDP-11 minicomputers. Subsequently, the network architecture evolved to incorporate new technologies and new capabilities. The design and development of DECnet was, then, largely contemporaneous with the development of the Internet protocols. The last fully proprietary versions were DECnet Phases IV and Phase IV+, in the early 1980’s [10]; DECnet Phases V and V+ maintained compatibility with the proprietary DECnet protocols but moved more in the direction of support for the ISO-defined OSI protocol stack. computer on a network needs to have a different address, the size of the address is a limit upon the size of the network. With 16-bit addresses, DECnet network implementations were limited to 64449 hosts. These three of these features of the DECnet design speak to a particular context of use. They highlight the expectation that DECnet deployments would be uniform, well-regulated and actively managed. This makes perfect sense in the context of DEC’s sales in corporate settings, where network implementation can be phased, planned, and centrally directed. Effective use of the shared file facilities, for instance, require a coordinated approach to the layout and conventions of filesystems across machines, while the network management infrastructure suggests that this was a key consideration in the settings for which DECnet was designed. An odd little implementation quirk in DECnet Phase IV similarly supports this. To make routing and discover easier, the network software running on a DECnet computer operating over an Ethernet network would actually reset the hardware network address of the computer to an address that conformed to the DECnet host address. This would cause considerable difficulty when DECnet was running in the same environment as other protocols, but in a highly managed environment where uniformity could be guaranteed, it was less troublesome. Given that DECnet was designed at roughly the same time as the Internet protocol suite, and given that it connected many of the same computer system types as the Internet protocols, it is again a useful point of comparison. DECnet was based on much the same “layered” protocol model that was the contemporary state of the art, and its basic architecture – a point to point connection layer, a routing layer, a layer for reliable sequenced delivery, and so on – is similar to that of systems like PUP, XNS, and TCP/IP. However, some key differences reveal the distinct character to the contexts in which DECnet was expected to operate. In sum, although DECnet was based on the same decentralized, peer-to-peer approach to network connectivity that characterizes the Internet protocol suite, its specific configuration of that approach is one that was in practice designed for the highly managed, highly controlled setting of corporate IT. One of these is that DECnet incorporated a sophisticated management interface, and indeed, that facilities for network management were designed into the protocol stack from an early stage. That is, DECnet was entirely imagined to be deployed in a managed environment. TCP/IP has to be managed too, of course, but the management of TCP/IP networks is not a network function in itself. (The Internet protocol suite includes a protocol, SNMP, for network management uses, but network-wide, management is not a key consideration.) CSNET The conventional historical account of the emergence of today’s Internet traces its beginnings to the ARPANET. ARPANET was both a research project and a facility; that is, the ARPANET project developed the networking technologies that are the underpinnings of the contemporary Internet and it also operated a facility – a network – based on those underpinnings. So this was not simply a research project funded by ARPA; it was also a facility that supported ARPA research. ARPA is the research arm of the US Department of Defense (DOD), and so in order to qualify as a site to be connected to ARPANET, one had to be a DOD site or a DOD contractor. At the time, this included many prominent research universities and computer science departments, but by no means all, and not even most. A second suggestive distinction lies within the sets of services standardized within DECnet. These included services, like network terminal access, similar to TCP/IP, but also some that the Internet protocol suite did not natively attempt to support, such as seamless remote filesystem access, in which disks attached to one computer would appear to be virtually available to users of other, connected computers. Remote file access of this sort (which was also a feature that had been part of the Xerox network system) goes beyond simply file transfer by providing users with the illusion of seamless access to both local and remote files. (Email, on the other hand, was not one of the standardized protocols, although networked email services were available through operating system applications.) Recognizing the value that the DOD-contracting universities were deriving from their participation in the ARPANET effort, wanting to expand beyond the smallerscale network arrangements upon which they were already depending [6], and concerned that ‘the ARPANET experiment had produced a split between the “haves” of the ARPANET and the “have-nots” of the rest of the computer science community’ [9], a consortium of US computer A third – although relatively trivial – distinction was that DECnet addresses were just 16 bits long. Since each 161 science departments partnered with the National Science Foundation and other groups to put together a proposal for the Computer Science Research Network, or CSNET. CSNET integrated multiple different network technologies, with a store-and-forward email system over regular phone lines as the base level of participation, but leased line and X.25 networking available for higher performance connections. other important, user-facing senses, distinct networks, suggesting intriguingly that there may be more to being “an Internet” than running TCP/IP on interconnected networks. FACILITIES Briefly examining some of these “othernets” lets us place our contemporary experience of the Internet – in its multiple capacities as a configuration of technologies, a constellation of services, and an object of cultural attention – in context. Some of that context is a “design space” – that is a space of possibilities that are available as outcomes of specific design decisions. Some lies in “historical circumstance” – that is, ways that different configurations arose reflecting their own historical trajectories. In order to reflect on these in a little more detail, we can take a different cut at the same question – of the nature of the Internet within a space of alternatives – by approaching the it in terms of the kinds of facilities that works can offer. While the development of CSNET produced many important technical contributions, CSNET’s most significant legacies might be historical and institutional, in that, first, CSNET represented a significant expansion in the spread of TCP/IP in the academic research community, and second, it was designed in collaboration with and in order to support the mission of the National Science Foundation (rather than, say, the military backers of ARPA and ARPANET). The CSNET effort laid the foundation for a later NFS-funded network called NSFNet, and it was the NSFNet backbone that was later opened up to commercial traffic, and then in 1995 replaced entirely by private service providers. The importance of this point is the explicit link at the time between institutional participation and connectivity, and the very idea that a network is conceived around an understanding of quite who and what will be permitted to connect. Naming Consider even this simple question: does a computer “know” its own name? In the naming arrangement of traditional TCP/IP, the answer is no. A computer knows its address but not necessarily the name by which it might be addressed by others, and in fact can operate in the TCP/IP environment quite effectively without one, as signaled by the fact that both TCP nor IP use addresses, not names, internally. So, facilities are built into the Domain Name Service (DNS) protocols [26] that allow a computer, on booting, to ask a network server, “What’s my name?” Naming is entirely delegated to the name service; that is, the network authority that answers the query “what is the address at which I can reach www.cnn.com?” is also the authority that tells a computer that it is www.cnn.com in the first place. The more interesting technical point for us here though concerns the relationship between ARPANET and CSNET. In a minor sense, CSNET was “in competition with” ARPANET; it was, after all, designed as a network for those who were institutionally denied access to ARPANET. However, in all the ways that matter it was entirely collaborative with ARPANET; key players in the ARPANET project, such as TCP/IP co-designer Vint Cerf at Stanford, participated in the CSNET effort, and the networks were bridged together. The basis for that bridging was CSNET’s adoption of the TCP/IP protocols that had been designed within the ARPANET effort. Through this bridging, ARPANET became a subnet of CSNET. However, we normally think of arrangement of subnetting and internetworking as providing seamless interconnection, but the interconnection between CSNET and ARPANET was not quite so seamless, since they adopted different protocols for delivering services at the higher levels. For instance, the MMDF network messaging facility developed as part of CSNET [7] was needed to be able to bridge between the “phonenet” and TCP/IP components of CSNET, and that meant that messages destined for CSNET recipients would need to be routed explicitly to CSNET rather than simply dispatched using the usual protocols used on purely TCP/IP networks (such as SMTP, the standard Internet email transfer protocol). In other words – both ARPANET and CSNET implemented the Internet protocols (TCP/IP) but not all the other protocols of what we is sometimes called the “Internet protocol suite”; accordingly, even though they were connected together through a common TCP/IP infrastructure, they remained in some By contrast, other networking protocols – like AppleTalk, for example – delegate to individual computers the right to assign themselves names. This can be implemented in different ways – by having each computer register a name with a naming facility, for example, or by entirely distributing the name lookup process rather than electing particular computers to implement a name or directory service – and these different mechanisms have their own differences in terms of both technological capacities and organizational expectations. The abilities for a computer to assign itself a name and to advertise that name to others are facilities that the designers of the IETF “Zeroconf” protocol suite felt it important to add, in order to support the server-free “plug-and-play networking” approach of AppleTalk [38]. The issue of naming raises a series of questions that highlight the relationship between specific technical facilities and the social organization of technological practice. Who gets to name a computer? Who is responsible for the names that network nodes use or call themselves? How is naming authority distributed? How visible are names? How centralized is control, and what temporalities 162 shape it? Even such a simple issue as naming presents a microcosm of my larger argument that questions of online experience need to be exampled in their technical and historical specificities. Routing Elsewhere [12], I discuss the case of Internet routing and explore the history of the development of routing protocols as being entwined with the history of the expansion of the Internet infrastructure. That paper contains a fuller argument, with a particular focus on the materiality of routing, but I want to summarise from it one relevant element here, which is the role of so-called “autonomous systems” in today’s Internet routing. Figure 1. The Arpanet in 1972, exhibiting a net-like structure. If there is one cornerstone of the argument that the Internet is a “decentralized” phenomenon, it is the case of network routing. Rather than requiring a central authority to understand the structure of the network and determine how traffic should flow, the design of the IP protocol, which gets Internet packets from one host to another, relies on each computer along a path deciding independently which way a piece of information should travel to take it closer to its destination, and creates the conditions under which coherent behavior results from this distributed process. Routing protocols are the protocols by which the information is distributed upon which these local decisions are to be made. the practical realities of the technology as it had developed; but in the evolutionary cycle of network protocol development, this meant baking those arrangements right into the protocol. Structure The emergence of autonomous systems as a consideration for routing information has an analogy in the emergence of structure within the interconnection arrangements of the network itself. The abstract model of interconnection that the Internet embodies is one of decentralized and amorphous structure. Indeed, the very term “network” was originally coined to convey not idea of computers connect together, but the topology of their connection as a loose and variable net-like mesh, in contrast to the ring, star, or hierarchical arrangements by which other connection fabrics had been designed. The basic idea of Internet-style networking is that computers can be connected together in whatever arrangement is convenient, and data units (“packets”) will nonetheless find their way from one point to another as long as some path (or route) exists from point A to point B. In a fixed topology, like a ring, star, or hierarchy, the path from one node to another is fixed and pre-defined. The key insight of the packet-switching design of the Internet is that there need be no prior definition of the route from one place to another; packets could find their own way, adaptively responding to the dynamics of network topology and traffic patterns. Amorphous topology is arguably one of the key characteristics of internet-style technologies. Early internets relied on fairly simply protocols for distributing routing information. As the Internet has grown from a small-scale research project to an international utility, the complexities of routing, and of the distribution of routing information, have also grown, and different protocols have arisen to solve emerging problems. The major routing information protocol operative in our current Internet is BGP, the Border Gateway Protocol. The particular relevance of BGP here is that, like its predecessor protocol (the Exterior Gateway Protocol or EGP), BGP is designed not to distribute the information necessary to route between networks but rather to route between so-called “autonomous systems.” The essential consideration here is that corporations, network infrastructure providers, universities, and so on, have networks that are themselves both quite complicated and autonomously managed. This in turn leads to the idea that the unit of network routing should be one of these systems. This is in many ways a recognition of a fundamental feature of our Internet, which is that it is not simply a network of networks, but a network of institutions and semi-corporate entities, each of which wish to maintain control over the organization of and access to their own networks. The protocols by which routing information are passed around are protocols that reflect this particular arrangement, encoding and producing “autonomy” as much as they encode “connection.” As our Internet grew, the “network” was no longer an effective unit at which routing information could be encoded, and “autonomous system” arose as an alternative that reflected Early network diagrams from the days of ARPAnet, the predecessor to the Internet, do indeed display this net-work character (figure 1). Over time, though, more hierarchical arrangements have arisen. These hierarchical patterns arise not least from two sources – geography and economics. The geographical considerations produce a distinction between metropolitan and long-haul networks, where metropolitan networks support relatively dense connectivity within small regions (for instance, the homes and businesses connected to the network in a town or neighborhood), while long-haul networks provide high-speed and high-capacity connections between urban regions [24]. The different demands upon 163 these different kinds of connection are best met with different technologies. This is where the economic considerations also come into play. Ever since the decommissioning of the NSFNet backbone in 1995 and the emergence of the commercial Internet, Internet service provision has been largely in the hands of commercial operators. To function as a competitive marketplace, network provision must involve multiple competing providers, each of whom in turn specialize in different aspects of service provision. One result of this specialization is a hierarchical segmentation of network service provision. Netflix paying ISP Comcast for access to its facilities and networks have largely ignored the fact that the problem to which Netflix was responding was a breakdown in relations between Comcast and Cogent, one of the transit carriers that Netflix pays to transmit its traffic to ISPs [32]. A dispute arose between Comcast and Cogent concerning whose responsibility it was to address bandwidth limitations when Cogent began to send more traffic onto Comcast’s network than their peering agreement allowed. In the face of this ongoing dispute, Netflix arranged to locate its servers with direct access to Comcast’s subscriber network, eliminating their dependence upon Comcast’s transit network. While this raised the specter in the popular press and in technical circles of an end-run around “net neutrality” arguments, the source of the problem in a dispute between carriers – and in particular at the boundary between an ISP and a transit carrier – is in some ways more interesting. Broadly, we can distinguish four different commercial entities – and hence four different forms of service provision and four different technological arrangements – in contemporary Internet service. The four are content providers, content distribution networks (CDNs), internet service providers (ISPs), and transit carriers. While the inspiration for the design of the internet protocols was to allow the creation of a network of networks, the emergent structure is one in which not all networks are equal. It’s not surprise that networks of course come in different sizes and different speeds, but the structure of contemporary networking relies upon a particular structural arrangement of networks and different levels of network providers, which in turn is embodied in a series of commercial and institutional arrangements (such as “settlement-free peering” [23]. As in other cases what we see at work here is an evolutionary convergence in which the network, as an entity that continues to grow and adapt (both through new technologies becoming available and new protocols being designed) does so in ways that incorporate and subsequently come to reinforce historical patterns of institutional arrangements. Content providers are corporations like Netflix, Amazon, Apple, or others whose business comprises (in whole or in part) delivering digital material to subscribers and consumers. Internet Service Providers, by and large, sell internet services to end-users, either individual subscribers or businesses. We think of our ISPs as delivering the content to us, but in truth they are generally responsible only for the last steps in the network, in our local cities or neighborhoods. Transit carriers are responsible for getting data from content providers to the ISPs. While the names of ISPs like Verizon and Comcast are well known, the names of transit carriers – such as Level 3, Cogent, or XO – are much less familiar, despite the central role that they play in bringing information to any of our devices. Content providers will typically contract directly with transit carriers for their content delivery needs. Transit carriers and ISPs connect their networks through a set of technical and economic arrangements collectively known as “peering” (as in the connection between two “peer” networks). End-to-End Gillespie [17] has discussed the importance of the so-called “end-to-end principle” [33] in not just the design of the Internet platform but also in the debates about the appropriate function of a network. The end-to-end principle essentially states that all knowledge about the specific needs of particular applications should be concentrated at the “end-points” of a network connection, that is, at the hosts that are communicating with each other; no other network components, such as routers, should need to know any of the details of application function or data structure in order to operate effectively. This implies too that correctly routing a packet should not require routers to inspect the packet contents or transform the packet in any way. So, for instance, if data is to be transmitted in an encrypted form, the intermediate nodes do not need to understand the nature of the encryption in order to transmit the packets. There is a trade-off in this of course; on the one hand, the network can be simpler, because every packet will be treated identically, but on the other, distinctions that we might want to make, such as between real-time and non-real-time traffic are unavailable to the network. Peering arrangements are almost always bilateral, i.e. an agreement between two parties. They are generally bidirectional and often cost-free, although they may have traffic or bandwidth limits beyond which charges are levied. The term “peer” is a hold-over from the military and non-profit days preceding the development of the commercial Internet, and speaks directly to the notion of “inter-networking” (ie, connecting together different networks and transmitting traffic between them). The “peers” that are now linked by peering arrangements, though, are not peer academic or research networks but commercial entities engaging in economic relations. These arrangements are largely hidden from the view of end users, but become broadly visible when disputes arise around the adequacy of peering relationships, corporate responsiveness to changing conditions upon them, or the responsibilities of carriage associated with them. For example, recent (early 2014) debates around Internet streaming movie provider 164 The end-to-end principle was originally formulated as a simple design principle, but it had, as Gillespie notes, many consequences that are organizational and political as well as technical. For instance, the separation between application semantics and packet transit also essentially creates the opportunity for a separation between application service providers (e.g. Netflix or Facebook) and infrastructure providers (e.g. ISPs like Time Warner or Verizon) by ensuring that applications and infrastructure can evolve independently. Arguably, these aspects of the Internet’s design have been crucial to its successful evolution. However, the end-to-end principle is under assault as a practical rubric for design. Two examples of obstacles to the end-to-end principle are the rise of middleboxes and assaults on net neutrality. deploying technologies like so-called “deep packet inspection” (mechanisms that examine not just the headers of a packet but its contents in order to decide how it should be treated) that violate the end-to-end principle. As discussions such as those of Saltzer et al. [33] make clear, the end-to-end principle was a foundational and significant principle in the design of the Internet, representing not just a specific technical consideration but also a commitment to an evolutionary path for network service provision. There is no question then that the Internet was designed around this principle. However, as the rise of middleboxes and challenges to network neutrality make clear, that doesn’t imply that the end-to-end principle is a design feature of our (contemporary) Internet. More generally, this case illustrates that technical design principles are themselves subject to revision, reinterpretation and revocation as the technology is implemented and evolves (c.f. [28]). Middleboxes is a generic term for devices that intervene in network connections. Unlike routers, which, designed according to the end-to-end principle, never examine packet internals or transform transmitted data, middleboxes may do both of these. Perhaps the most common middleboxes are gateways that perform Network Address Translation or NAT. Almost every residential network gateway or WiFi base station is actually a NAT device. NAT is a mechanism that partially resolves the problem of scarce network addresses. Most people buy a residential internet service that provides them with just a single network address, even though they have multiple devices using the service (e.g. a laptop and a smartphone). Network Address Translation allows both of these devices to use a single network address simultaneously. To the outside world, the laptop and the smartphone appear like a single device that is making multiple simultaneous connections; the NAT device keeps track of which connection are associated with which device and delivers responses appropriately. NAT violates the endto-end principle, however, with various consequences. For exactly, because the individual devices are not technically end-points on the public internet, one cannot make a connection to them directly. Further, network flow control algorithms may be confused by the presence of multiple devices with essentially the same network address. Further, NAT gateways achieve their effects by rewriting packet addresses before the packets have been delivered to their end-point. THE FEUDAL INTERNET The emergent structure of our Internet – the market niches of transit carrier and ISP, the practical solution of CDNs, the fragmentation produced by middleboxes, the relationship between mobile carriers, telecommunications firms, and media content producers, and so on – draws attention to a simple but important truth of internetworking: the Internet comprises a lot of wires, and every one of them is owned by someone. To the extent that those owners are capitalist enterprises competing in a marketplace, then the obvious corollary is that, since all the wires can do is carry traffic from one point to another, the carriage of traffic must become profit-generating. The mechanisms by which traffic carriage over network connections becomes profitable is basically either through a volume-based or per-byte mechanism – a toll for traffic – or through a contractual arrangement that places the facilities of one entity’s network at the temporary disposal of another – a rental arrangement. This system of rents and tolls provides the basic mechanism by which different autonomous systems, each of which provisions its own services, manages its own infrastructures, and then engages in a series of agreements of mutual aid. If this system seems familiar, it is not so much that it encapsulates contemporary market capitalism but more that it is essentially feudal in its configuration. Marx argued for feudalism and capitalism as distinct historical periods with their links to material means of production – “The handmill gives you society with the feudal lord; the steam-mill society with the industrial capitalist” [25]. Recent writers, though, have used the term “neofeudal” to describe the situation in late capitalism in which aspects of public life increasingly become private, gated domains – everything from toll lanes on the freeway and executive lounges at the airport, on the small end, to gated communities and tradeable rights to pollute the environment issued to large corporations at the other (e.g. [34]). The essential A second assault on the end-to-end principle is the emergence of threats to network neutrality. Net neutrality is a term to express the idea that network traffic should be treated identically no matter what it contains, and no matter where it originates or terminates. However, many organizational entities have reasons to want to violate this principle. For instance, a network provider which is owned by a media company might want to prioritize the delivery of its own media traffic to its customers, giving it a higher priority than other traffic to ensure a smoother service, or an ISP might want to limit the bandwidth available to hightraffic users in order to encourage more equitable access to the network. These kinds of situations often involve 165 consideration here is the erasure of public infrastructure and the erection of a system of tariffs, tolls, and rents that govern the way we navigate a world made up of privatized but encompassing domains, within which market relations do not dominate. ways – in the provision of specific linguistic content [36], in the regional caching of digital content, in the question of international distribution rights for digital content (e.g. which movies can be viewed online in which countries), in assertions of national sovereignty over information about citizens (e.g. Vladimir Putin’s public musings that data about Russian citizens should be stored only on servers in Russia [22]), in the different regimes that govern information access (e.g. the 2014 EU directive known popularly as the “right to be forgotten”), and in local debates about Internet censorship (from China’s “Great Firewall” and Singapore’s self-censorship regime to discussions of nationwide internet filters in Australia and the UK). The very fact that a thriving business opportunity exists for commercial Virtual Private Network (VPN) services that allow users to get online “as if” they were located in a different country signals the persistent significance of nation-states and national boundaries in the experience of our Internet. Similarly, significant debate has surrounded the role of national interests in Internet governance [e.g. 27], and the International Telecommunications Union or ITU – a United Nations organization whose “members” are not technical experts or corporations but nation-states – remains a significant body in the development of network technologies and policy. Just as feudalism reinforced and made all aspects of everyday life subject to the boundaries of the manor, the shire, and the nation, so too does our Internet – not necessarily any Internet, but certainly our Internet – retain a significant commitment to the relevance of similar geographical, national, and institutional boundaries. Beyond the general use of the term “neofeudal” to refer to the privatization of public goods, let me take the metaphor of the feudal Internet more seriously for a moment to point to a couple of significant considerations. The first is that operating mechanism of feudalism is not the market transaction but rather long-standing commitments of fealty, vassalage, and protection. These are not the instantaneous mutual engagements of market capitalism but temporally extended (indeed, indefinite) arrangements with little or nothing by way of choice or options. Indeed, the constraints upon feudal relations are geographical as much as anything else: infrastructural, if you will. One can see, arguably, some similarities to the way that geographical and infrastructural constraints lead to a pattern of relations between internet providers that also relies upon long-term, “residence”-based, partnerships. The ties that bind individuals to their service providers in semi-monopolistic conditions of the US broadband market, or perhaps even more pointedly, the links that connect large-scale data service providers such as Netflix with transit carriers like Level 3 are not simply conveniently structured as long-term arrangements, but rather can only operate that way because of the infrastructure commitments involved (such as the physical siting of data stores and server farms.) Similarly, the need for physical interconnection between different networks makes high-provider-density interconnection nodes like One Wilshire in downtown Los Angeles (see, e.g., [12, 40]) into “obligatory passage points” in Callon’s language [5] – that is, to get interconnection between networks, you need to be where all the other networks are, and they all need to be there too. For all that we typically talk about the digital domain as fast-paced and everchanging, these kinds of arrangements – not simply commitments to infrastructure but commitments to the institutions relationships that infrastructure conditions – are not ones that can change quickly, easily, or cheaply. These relations are more feudal that mercantile. The third point to be emphasized here is the way that these are simultaneously social and technical arrangements – not social arrangements that give rise to technological designs, nor technological designs that provoke social responses. Middleboxes, deep packet inspection, and the autonomous systems of BGP should be regarded as, and analyzed as, both at once. This requires, then, a view that both takes specific technological configurations (rather than principles, imaginaries, and metaphors) seriously as objects of analytic attention, and that identifies and examines the social, political, and economic contexts within which these come to operate. To the extent that a feudal regime of hierarchical relations based on long-term structures of mutual commitment can be invoked as an account of our Internet, it can be done only within with context of a sociotechnical analysis that is both historically specific and symmetric. The second interesting point that a feudal approach draws attention to is the persistence of pre-existing institutional structures – perhaps most obviously, the nation-state. Although John Perry Barlow’s [2] classic “Declaration of the Independence of Cyberspace” famously argues that the “governments of the industrial world… [have] no sovereignty” in the realm of the digital, and notwithstanding the IETF’s famous motto that “rejects kings [and] presidents” in favor of “rough consensus and running code” [20], the truth is that governments and presidents continue to manifest themselves quite significantly in not just the governance but the fabric of our Internet. National and regional concerns arise in a variety of CONCLUSIONS The larger project of which the current exploration forms a part is an attempt to take seriously the materialities of information and their consequences. It is critical to this project that we move beyond accounts simply of information infrastructure, but also recognize the relevant materialities of representation, and their consequences [11, 13]. As Powell [31] has argued in her study of open hardware projects, patterns of historical evolution torque 166 the design principles that often provide not only technical arrangements but also the social imaginaries that are mobilized in our discussions of technology. To bring this perspective to contemporary networking, then, means not simply that we need to think about the pragmatics of undersea cables [37], satellite downlinks [29], and server farms [40], but also the way that specific representations of action and encodings of data are designed to be manipulated, transmitted, and moved in particular sorts of ways, with consequences for the resultant experience of the network as offering particular opportunities for individual and collective action. REFERENCES 1. Antoniadis, P., Le Grand, B., Satsiou, A., Tassiulas, L., Aguiar, R., Barraca, J.P., and Sargento, S. 2008. Community building over Neighborhood Wireless Mesh Networks. IEEE Society & Technology, 27 (1): 48-56. 2. Barlow, J. 1996. A Declaration of the Independence of Cyberspace. Davos, Switzerland. Downloaded from https://projects.eff.org/~barlow/Declaration-Final.html on June 4, 2014. 3. Birrell, A., Levin, R., Needham, R., and Schroeder, M. 1982. Grapevine: An Exercise in Distributed Computing. Communications of the ACM, 25(4), 260274. In other words, the primary concern is to see networked arrangements as historically particular crystallizations of not just technical but also institutional, economic, and political potentialities. To do this, particularly with respect to the technology that we rather glibly call “the Internet,” I have suggested that two moves are needed. 4. Boggs, D., Shoch, J., Taft, E., and Metcalfe, R. 1980. IEEE Transactions on Communication, 28(4), 612-624. 5. Callon, M. 1986. Elements of a sociology of translation: Domestication of the Scallops and the Fishermen of St Brieuc Bay. In Law (ed.), Power, Action and Belief: A New Sociology of Knowledge? 196-233. London: Routledge. The first move is from the idea of “the Internet” to that of “an Internet” – that is, to re-encounter our contemporary network as not the only possible Internet that could have been built, but as one of a range of possible networks. When we consider a number of possible networks, we start to pay attention to the range of expectations, institutional arrangements, policies, technical configurations, and other dimensions that might characterize the space of potentials. The second move is from “an Internet” to “this Internet” – that is, to narrow down once more and so grapple with the historical, geographical, political and social specificities that constrain and condition the particular network with which we are engaged at any particular moment in time. This Internet is not the one that was conceived of by those like Paul Baran or Donald Davies, designed by those like Vint Cerf or Bob Taylor, or opened to commercial operation by the NSF – it has elements of each of those, but it is a historically specific construction which has encompassed, transformed, extended, and redirected any of those particular networks. This Internet is something that we can grapple with empirically. We are not in much of a position to make general statements about “the Internet”; but when we ask questions about “this Internet,” we may have a starting point for investigation. 6. Comer, D. 1983. The computer science research network CSNET: a history and status report. Communications of the ACM, 26(10), 747-753. 7. Crocker, D., Szurkowski, E., and Farber, D. 1979. An Internetwork Memo Distribution Facility – MMDF. Proc. ACM Conf. Computer Communication SIGCOMM, 18-25. 8. Dahlberg, L. 2001. Democracy via cyberspace: examining the rhetorics and practices of three prominent camps. New Media & Society, 3, 187–207. 9. Denning, P., Hearn, A., and Kern, W. 1983. History and overview of CSNET. Proc. Symp. Communications Architectures & Protocols (SIGCOMM '83), 138-145. 10. Digital Equipment Corporation. 1982. DECnet DIGITAL Network Architecture (Phase IV): General Description. Order number AA-N149A-TC. Maynard, MA: Digital Equipment Corporation. 11. Dourish, P. 2014. NoSQL: The Shifting Materialities of Databases. Computational Culture. 12. Dourish, P. 2015. Packets, Protocols, and Proximity: The Materiality of Internet Routing. In Parks, L. and Starosielski, N. (eds), Signal Traffic: Critical Studies of Media Infrastructures, 183-204. University of Illinois Press. ACKNOWLEDGMENTS Many key ideas here arose in conversation with Jon Crowcroft, Alison Powell, and Irina Shklovski. I am indebted to Morgan Ames, Panayotis Antoniadis, Sandra Braman, Marisa Cohn, Courtney Loder, Lilly Nguyen, Katie Pine, and Chris Wolf, each of whom offered valuable critique at key junctures. This work is part of a larger project on the materialities of information in which Melissa Mazmanian has been a key partner, and it has been supported in part by the Intel Science and Technology Center for Social Computing and by the National Science Foundation under awards 0917401, 0968616, and 1025761. 13. Dourish, P. and Mazmanian, M. 2013. Media as Material: Information Representations as Material Foundations for Organizational Practice. In Carlile, Nicolini, Langley, and Tsoukas (eds), How Matter Matters: Objects, Artifacts, and Materiality in Organization Studies, 92-118. Oxford, UK: Oxford University Press. 167 14. Elmer-Dewitt, P. 1993. First Nation in Cyberspace. Time, 49, Dec 6. 28. Oudshoorn, N. and Pinch, T. 2003. How Users Matter: The Co-construction of Users and Technology. Cambridge, MA: MIT Press. 15. Fernaeus, Y. and Sundström, P. 2012. The Material Move: How Materials Matter in Interaction Design Research. Proc. ACM Conf. Designing Interactive Systems, 486-495. 29. Parks, L. 2012. Satellites, Oil, and Footprints: Eutelsat, Kazsat, and Post-Communist Territories in Central Asia. In Parks and Schwoch (eds), Down to Earth: Satellite Technologies, Industries, and Cultures. New Brunswick: Rutgers University Press. 16. Gaved, M., & Mulholland, P. 2008. Pioneers, subcultures, and cooperatives: the grassroots augmentation of urban places. In Aurigi, A. and De Cindio, F. (eds.), Augmented urban spaces: articulating the physical and electronic city, 171-184. England, Ashgate. 30. Powell, A. 2011. Metaphors, Models and Communicative Spaces: Designing local wireless infrastructure. Canadian Journal of Communication. 31. Powell, A. 2014. The History and Future of Internet Openness: From ‘Wired’ to ‘Mobile’. In Swiss and Herman (eds), Materialities and Imaginaries of the Open Internet. Routledge. 17. Gillespie, T. 2006. Engineering a Principle: ‘End-toEnd’ in the Design of the Internet. Social Studies of Science, 36(3), 427-457. 18. Hauben, M. and Hauben, R. 1997. Netizens: On the History and Impact of Usenet and the Internet. Wiley/IEEE Computer Press. 32. Rayburn, D. 2014. Here’s How the Comcast & Netflix Deal is Structured, With Data and Numbers. Streaming Media Blog, February 27 2014. Downloaded from http://blog.streamingmedia.com/2014/02/heres-comcastnetflix-deal-structured-numbers.html 19. Hindman, M. 2008. The Myth of Digital Democracy. Princeton University Press. 33. Saltzer, J., Reed, D. and Clark, D. 1984. End-to-End Arguments in System Design. ACM Transactions on Computer Systems, 2(4), 277–88. 20. Hoffman, P. 2012. The Tao of IETF. Downloaded from http://www.ietf.org/tao.html on June 4, 2014. 21. Jungnickel, K. 2014. DIY WIFI: Re-imagining Connectivity. Palgrave Pivot. 34. Shearing, C. 2001. Punishment and the Changing Face of Governance. Punishment and Society, 3(2), 203-220. 22. Khrennikov, I. and Ustinova, A. 2014. Putin’s Next Invasion: The Russian Web. Downloaded from http://www.businessweek.com/articles/2014-0501/russia-moves-toward-china-style-internet-censorship on June 4, 2014. 35. Shirky, C. 2008. Here Comes Everyone: The Power of Organizing Without Organizations. Penguin. 36. Shklovski, I. and Struthers, D. 2010. Of States and Borders on the Internet: The Role of Domain Name Extensions in Expressions of Nationalism Online in Kazakhstan. Policy and Internet, 2(4), 107-129. 23. Laffont, J.-L., Marcus, S., Rey, P., and Tirole, J. 2001. Internet Peering. Annual Economic Review, 91(2), 287291. 37. Starosielski, N. 2015. The Undersea Network. Durham, NC: Duke University Press. 24. Malecki, E. 2002. The Economic Geography of the Internet’s Infrastructure. Economic Geography, 78(4), 399-424. 38. Steinberg, D. and Cheshire, S. 2005. Zero Configuration Networking: The Definitive Guide. Sebastopol, CA: O’Reilly Media. 25. Marx, K. 1847 (1910). The Poverty of Philosophy. Tr: Quelch, H. Chicago, IL: Charles H. Kerr & Co. 39. Turner, F. 2006. From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism. Chicago, IL: University of Chicago Press. 26. Mockapetris, P. 1987. Domain Names – Concepts and Facilities. RFC 1034, Network Working Group. 27. Mueller, M. 2010. Networks and States: The Global Politics of Internet Governance. MIT Press. 40. Varnelis, K. 2008. The Infrastructural City: Networked Ecologies in Los Angeles. Barcelona: Actar. 168 Interacting with an Inferred World: The Challenge of Machine Learning for Humane Computer Interaction Alan F. Blackwell University of Cambridge Computer Laboratory afb21@cam.ac.uk internal models) and the economics of network connectivity (the range of data from which models can be derived). One set of concerns about those changes of scale is now familiar as an implication of the phrase ‘big data’ [9], but I contend that the real problem is more insidious - that it results from changes in the technical underpinnings of artificial intelligence research, which are in turn changing designers’ conceptions of the human user. The danger is not the creation of systems that become maliciously intelligent, but of systems that are designed to be inhumane through neglect of the individual, social and political consequences of technical decisions. ABSTRACT Classic theories of user interaction have been framed in relation to symbolic models of planning and problem solving, responding in part to the cognitive theories associated with AI research. However, the behavior of modern machine-learning systems is determined by statistical models of the world rather than explicit symbolic descriptions. Users increasingly interact with the world and with others in ways that are mediated by such models. This paper explores the way in which this new generation of technology raises fresh challenges for the critical evaluation of interactive systems. It closes with some proposed measures for the design of inference-based systems that are more open to humane design and use. The structure of the paper is as follows. The first section discusses the nature of the intellectual changes that have accompanied developments in AI technology. For many years, researchers in Human-Computer Interaction (HCI) have been concerned about the consequences of AI, but I argue that it is time for some fundamental changes in the questions asked. The second section takes a specific research project as a focal case study. That project was not designed as an interactive system, but suggests future scenarios of interaction that seem particularly disturbing. I use the case study as a starting point from which several key concerns of humane interaction can be explored, introducing questions of human rights, law and political economics. The final section is more technically oriented, describing practical tests available to the critical technical practitioner, followed by a selection of research themes that might be used to explore more humane interaction modes. Author Keywords Machine learning; critical theory ACM Classification Keywords H.5.2. User interfaces; I.2.0 Artificial intelligence INTRODUCTION Anxiety about artificial intelligence (AI) has pervaded Western culture for 50 years - or perhaps 500. As I write, the BBC News headline reads Microsoft's Bill Gates insists AI is a threat, while the Cambridge Centre for Existential Risk, led by past Royal Society President and Astronomer Royal Sir Martin Rees, lists AI as the first in its list of the technologies that may represent ‘direct, extinction-level threats to our species’. The legend of the Golem and other examples from fiction demonstrate human unease about machines that act for themselves. Of course, this ability is shared by devices as humble as the thermostat, the sat-nav, or predictive text – all somewhat mysterious to their users, and all possessing some kind of ‘mind’ – that is, an internal model of the world that determines system behavior. BACKGROUND In recognition of the long view taken at this conference, I will set out the development of these concerns over a relatively extended time-frame. The phrase big data has only recently become popular as a marketing term, promoting the potential for statistical analysis to model the world by extracting patterns from data that is becoming more readily available. The Machine Learning techniques used to create such models have supplanted earlier AI technology booms such as symbol-manipulation methods in the Expert Systems of the 1980s – now remembered nostalgically as GOFAI, for ‘Good Old-Fashioned AI’. The goal of this paper is to analyze the ways in which our relationship with such devices and systems is changing, in response to the continued developments arising from Moore's Law (namely, the scale and complexity of their Copyright© 2015 is held by the author(s). Publication rights licensed to Aarhus University and ACM These changing trends are significant to Human-Computer Interaction (HCI) because GOFAI has long had a problematic relationship with HCI – as a kind of quarrelsome sibling. Both fields brought together 5th Decennial Aarhus Conference on Critical Alternatives August 17 – 21, 2015, Aarhus Denmark DOI: http://dx.doi.org/10.7146/aahcc.v1i1.21197 169 knowledge from Psychology and Computer Science, to an extent that in the early days of HCI, it was difficult to distinguish HCI from AI or Cognitive Science. The program of work at Xerox PARC that resulted in the seminal book The Psychology of Human-Computer Interaction [10] was initiated by a proposal from Allen Newell filed as Applied Information-Processing Psychology Project Memo Number 1[25], while Norman and Draper's UCSD book (User Centered System Design) [28] punningly emerged from the Cognitive Science department at UCSD (UC San Diego). Many early HCI researchers had taken degrees in AI or Cognitive Science (including the current author). Even today, popular accounts of HCI research, referencing both psychology and computing, often result in the naive assumption that we must be doing AI. Empirical Motivation Intersectional Community Technical Research Business Opportunity Critical Question 1980s-90s 2000s-10s Experimental Psychology Functional Imaging Cognitive Science Neuroscience Artificial Intelligence Machine Learning Expert Systems Big Data Situated Cognition Humane Interaction Table 1. Structural analogy of the new technical context for HCI, with the focus of this paper as a new critical question. However, the years of the 1980s Expert Systems boom were also associated with a critical reaction. The ‘strong AI’ position anticipated computers that could pass the Turing Test and simulate people. This possibility was challenged not only by philosophers of mind, but by researchers concerned with HCI such as Winograd and Flores [39], Gill [16], and Suchman [35], all of whom noted the ways in which symbolic problem-solving algorithms neglected issues that were central to human interaction, including social context, physical embodiment, and action in the world. Each of these researchers had distinct concerns, but these can be summarized in their reception by AI researchers as problems of situated cognition – the failure of formal computational models of planning and action to deal with the complexity of the real world, where actions seem more often improvised in response to events and situations [1, 29]. To summarize those changes, and also the central analogy leading to the concern of this paper, Table 1 outlines the relationship between the technoscience landscape of GOFAI that laid the context for HCI in the 1980s, and the corresponding scientific, technical and business trends to which we should be responding today. The critical challenges to symbol-processing GOFAI were that the symbols were not grounded, the cognition was not situated, and there was no interaction with social context. These are not the critical problems of the ML landscape. In contrast to those earlier critiques, machine learning systems operate purely on ‘grounded’ data, and their ‘cognition’ is based wholly on information collected from the real world. Furthermore, it appears that ML systems are interacting with their social context, for example through the use of big data collected from social networks, personal databases, online business and so on. However, this question of interacting with social data introduces the most important critical concern of this paper. Although the phrase ‘situated cognition’ has become tainted by debate and controversy, these debates between HCI and symbol-processing AI have been underlying concerns of major theoretical contributions in HCI, such as Dourish's discussions of context and embodiment [14, 15], and concepts of interaction offered as design resources in a critical technical practice [2, 20]. This concern can be framed in terms of the Turing Test. In that test, judges are asked to distinguish between a human and a computer. In the classic formulation, we wish to see if the computer will appear sufficiently human-like that the two cannot be distinguished. However, I am more concerned with the reverse scenario. What if the human and computer cannot be distinguished because the human has become too much like a computer? The New Critical Landscape The goal of this paper is to explore the ways in which this continued central tension in HCI is now changing in fundamental ways, because of the technical differences between the methods of GOFAI, and those that are now predominant in Machine Learning (ML). Where symbolprocessing approaches failed to take account of the rich information available in the world, ML algorithms have access to huge quantities of such information. This has resulted in enormous changes, by comparison to the technical and commercial context in which the field of HCI was formed. The original version of the Turing Test fails for all the reasons identified in the 1980s HCI critiques of AI. The new version ‘passes’ the Turing Test, but in a way that demands a new critique. My concern is that reducing humans to acting as data sources is fundamentally inhumane. A serious additional concern is that technical optimists appear to be blind to this problem – perhaps because of their excitement as they finally seem to be approaching the scientific ‘goal’ of the Turing Test. HCI researchers and other critical practitioners should be alert to these technical developments, and be ready to draw attention to their consequences. 170 This is particularly important because, whereas Expert Systems attracted much technical excitement in the 1980s, they were not widely deployed – at least, not to the extent that ordinary people would interact with them every day. In contrast, the statistical techniques of ML are now widely deployed in interactive commercial products. Everyday examples include the Pagerank algorithm of Google, Microsoft's Kinect motion-tracking interface, many different mobile predictive text entry systems, Amazon's recommendation engine and so on. If there is a critical problem, it is not simply an academic or philosophical concern about a speculative future of intelligent machines. On the contrary, I believe this may be a fundamental problem for contemporary society. model that purports to explain the user, yet cannot explain itself. This second concern has been elaborated in debate between Norvig and Chomsky [30]. Chomsky questions the value of models that contain no symbols for us to inspect, while Norvig observes that these models clearly work, and have supplanted symbolic models as a result. As with Breiman’s Two Cultures essay, this is partly an appeal to technological pragmatism over epistemological reservations – Breiman himself says “the goal is not interpretability, but accurate information” [7, p 210]. However, if this is the case, it is reasonable to ask whether the most pragmatic approach will necessarily be the most humane. These questions are located in a complex nexus of technical and psychological considerations. Before exploring them further, I illustrate that context with a specific case study of research within this nexus. In the following discussion, the case study also serves as a useful (and, by intention, slightly distanced from HCI) concrete example. To summarise this section, it has drawn comparisons between the technical trends that inspired the HCI critiques of the 1980s, and the technical trends of today. Much theory in contemporary HCI originally emerged from those 1980s critiques of symbol-processing AI. But whereas the core problem of symbol-processing AI was its lack of connection to context – the problem of situated cognition – the core problem of machine learning is the way in which it reduces the contextualised human to a machine-like source of interaction data. Rather than cognition that is not situated, our new concern should be interaction that is not humane. CASE STUDY: READING THE MIND My case study comes from the work of Jack Gallant’s research group, described as ‘reconstructing visual experiences from brain activity’ [27]. Generating wide public interest from the disturbing suggestion that his team had created a mind-reading machine, this ML system retrieved film clips from a database according to the similarity of EEG readings taken while people watched the films. Press reports and publications were accompanied by images such as that in Fig. 1, showing (on the left) a still from the film that had been shown to an experimental subject, together with (on the right) a ‘mind-reading’ result that was often interpreted as the rendering of an image captured from within the brain of the subject. AN OUTLINE OF THE PROBLEM The key question is whether ML systems, which create their own implicit internal model of the world through inference rather than an explicit symbolic model, carry new consequences for those interacting with them. There are some existing concerns regarding the epistemological status of such systems. One is summarised by Breiman [7] as resulting from ‘two cultures’ of statistical modeling. Breiman contrasts the traditional practice of a statistician building and testing a mathematical model of the world, to ML techniques in which the model is inferred directly from data. He observes that “nature produces data in a black box whose insides are complex, mysterious, and, at least, partly unknowable” [7, p.205]. As a result, rather than following Occam’s razor, he believes that “the models that best emulate nature in terms of predictive accuracy are also the most complex and inscrutable” [7, p 209]. Breiman’s analysis, if correct (there are objectors), suggests that predictive accuracy is more important than interpretability. Systems built around such models therefore predict the state of the world and the people in the world, including the actions of the user, without offering an explicit symbolic explanation. One consequence of collecting data without explaining why is to trigger the now-familiar concerns of surveillance and privacy associated with ‘big data’. More subtle is the consequence of interacting with the world through the mediation of a Figure 1. A visual image reconstructed from brain activity, as reported in [27] (image reproduced with permission of the authors, extracted from https://www.youtube.com/watch?v=nsjDnYxJ0bo, and reused under CCA license). The left side of the figure shows a film scene presented to an experimental subject. The right side is a reconstructed image informed by EEG measurements from the subject’s brain. 171 The visual rhetoric in these scientific illustrations is compelling. The captured mind-image on the right of Fig. 1 resembles an oil painting by Goya or by Turner, an artistic rendering of the deep unconscious. The image on the left is a scene from a Steve Martin film – reminding us of Martin’s classic riffs on disembodied brains (The Man with Two Brains, 1983) and mind-body exchange (All of Me, 1984). But on closer inspection, the resemblance between the right and left halves is rather slim. Why does the reconstructed image have dark hair rather than white? Why is it not wearing a white shirt? In fact, all we can say for sure is that there appears to be a human figure on the right of the frame. We are sure because we see its face – as plainly as we see a face in the moon! essential considerations in the critical assessment of ML models. I will not dwell further on the visual rhetoric of scientific discourse, despite the fact that the case raises fascinating questions in that area too. Of course, humans (in the Northern hemisphere) observe a face in the moon because human brains, as with the brains of all social mammals, have evolved to detect faces well. If one were to record EEG signals while humans observe a wide range of different images, we might be confident that images of faces would result in some of the most distinctive signals. We also know that each hemisphere of the brain responds to only one half of the visual field, so the simplest possible inference is to determine the side of the head on which a signal has been detected. The central part of this paper explores the resulting questions. The following sections draw out these questions in turn, starting with those that arise directly from the domain of film – art works and their readers – but then moving on to the economic and psychological networks in which artworks are embedded. From Case Study to Critical Questions This case study has illustrated typical techniques from ML and neuroscience that might (if they worked in the manner implied) provide a basis for new models of user interaction that would be able to predict the user’s needs. A ‘natural’ brain-computer interface of this kind has not yet been proposed – though it may come soon! However, the case study can also be used as a starting point for critical enquiry. QUESTION 1: AUTHORSHIP GOFAI systems maintained a clear distinction between data (symbolic representations of the world), and algorithms that processed this data. In the modern digital economy, business and legal frameworks still maintain a clear distinction between (data-like) content and (algorithm-like) services. However, the behavior of ML systems is derived from data, through the construction of the statistical model. So, this is a point at which one might ask more closely how this image of the unconscious was painted. Might it be possible to construct such an image, if the ML model were so crude as to encode only the presence of a face on one side or other of the visual field? It seems, from close reading of the research described, that the right hand image is a blurred average of the 100 film library scenes most closely fitting the observed EEG signal. That is, it is a blurred combination of 100 film scenes in which a face appears in the right hand side of the frame. The other visual features, given this starting point, are unexceptional. The location of the face within the right hand side follows a completely conventional film frame composition. The face is located by the rule of two thirds, and the gaze of the figure is directed inward to the center of the frame. In fact, this could be a scene from any film at all. The image in Fig. 1 appeared meaningful because it was derived from actual movie scenes. Similarly, ML-based interactive technologies such as the Microsoft Kinect game controller are able to recognize human actions because their models have been derived from a large database of human actions [33]. In one sense, these statistical models can be regarded as an index of the content that created them, allowing the system to look up an interpretation that was originally created by a human author. This close relationship between index and authorship has been a focus of critical inquiry in the past, for example in Borges’ Library of Babel [8], which contained every possible book in the universe that could be written in an alphabet of 25 characters. If there were a catalogue to this library, its index would have to contain the full text of the book itself, to distinguish each book from the volume that is identical apart from a single character. Borges’ Library was a thought experiment, but we do now have ML algorithms that index (nearly) every book in the world, meaning that their models incorporate a significant proportion of those books’ content. However, the algorithm collecting an index does not care whether the data is copyrighted – despite the fact that the copyrighted content is in some way mashed-up into the structure of the resulting model. The rhetorical skill of the researchers in constructing this striking image is to have chosen just the right number of film scenes to average. A composite of only one or two scenes would retain sufficiently clear visual features to make it obvious if we were not looking at the right film. On the other hand, an average of 10,000 scenes would be no more than a brown smear, losing the powerful hints of contour and shadow that remind us of dream paintings. Even the level of detail used in the averaging process seems carefully chosen – the unit pixels, the hints of vertical lines suggesting scenery – are the size and shape of brush strokes to further evoke painterly rendering. In this case study, we can admire the skill of the researchers in presenting such a compelling story. However, this example also provides an opportunity to discuss some To consider the implications, imagine a dynamic audio filter trained on a single song – perhaps John Lennon's 172 Imagine. If allowed to process sufficient random noise, this filter could select enough sounds to reproduce the song. If applied to any other song, it selects only those parts of the song that resemble Imagine. The model is not far from being a copy. But what if we trained the model on two Lennon songs, or on the whole Beatles repertoire, to the extent that it selects or simulates that repertoire? This is not dissimilar to the ‘reconstructed’ film scene described in the experiment above, although we are closer to technical feasibility with audio than with film (e.g. [12]). song recorded with it. If ML models are interpreted (and applied) as processes, this is a challenge for attribution. Although they resemble a purely algorithmic construct, they are also a kind of intertextual content derived from the data used to train them (just as a John Lennon ‘filter’ represents all possible John Lennon songs, even including those not written, but predicted from his body of work). This situation is exacerbated by the fact that, when we interact with computer systems, we often over-attribute intelligence to the system, failing to recognize the fact that apparent intelligence, in a system that is in some way faulty or ambiguous, may arise from our own judgments. Collins and Kusch describe this dynamic as ‘Repair, Attribution and all That’ (RAT) [11]. RAT explains the enthusiastic popular reception of the mind-reading demonstration in Figure 1. Although clearly a poor reproduction of the original scene, the human observer ‘repairs’ this, interpreting the computer output as an accurate portrayal of a dream, and attributing their own subsequent interpretation of that ambiguous image as resulting from the apparent intelligence of the system that produced it. The ethics of copyright are a common enough topic in critiques of digital media. However, rather than purely considering the professional arts (loudly defended by copyright holders in the media industries), we should also consider the ways in which every digital citizen is an ‘author’ of their own identity, because of the ways that persistent traces of our experiences and interactions with others are now extending beyond our own memories into digital representations. The human self is a narrative of experiences and relations with others, and ownership of this narrative is a critical condition of being human. I return to this issue later. In symbolic systems, where the system behavior has been specified and encoded directly by a human designer, the user can apply a semiotic reading in which the user interface acts as the ‘designer’s deputy’ [34]. In contrast, if the system behavior is encoded in a statistical model, derived by inference from the work of many authors, and presented to users in a context where any faults are repaired and attributed to the intelligence of the system itself, then this humane foundation of the semiotic system is undermined. QUESTION 2: ATTRIBUTION The logic of digital copyright would assert that the content of the original material captured in an ML model or index should still be traced to the authors – the scale of the appropriation is not the key legal point. However, indexing of the Internet is still heavily influenced by rather utopian perspectives derived from champions of the public domain in which it is often asserted that ‘information wants to be free’ (attributed to Stewart Brand). But if we acknowledge that ‘information’ represents the work and property of individuals, then ‘freedom’ might simply mean the freedom to appropriate that work by those wishing to encode it in the form of ML models, especially if there are no copyright holders leaping to defend their license revenue. Models derived from big data defy symbolic representation, because the scale and complexity of the data processing algorithms makes it very difficult (or even impossible) to ‘run them backward’ and recover the necessary links for human attribution. We can choose to treat such dynamics as a necessary sacrifice for the public good of the creative commons, supporting a massive global project of postmodern collaging. However, if we are accelerating the death of the author through technical means, then who will get paid? Attribution is already problematic in digital media, as a result of postmodern collaging practices – remixes, mashups and so on. Experts in forensic musicology report that court decisions are contingent on the availability of uncontestable symbolic representations [3]. Lyrics are easy to defend. Melodies likewise, so long as they can be written out as notes on a scale. However, reprocessed samples are more ambiguous, and distinctive digital production techniques almost impossible to verify without separate evidence of provenance. The law is a symbolic system, and it works well only with symbolic data. QUESTION 3: REWARD Attribution of authorship is considered an inalienable human right [37]. However, as noted by Scherzinger [32], the digital public domain pays lip service to attribution of authorship, while actually providing unfettered access to commercial interests. Global indexing and data-centric service models represent a new era of enclosures, echoing the enclosure of common grazing land by the British aristocracy. In particular, Scherzinger notes that there is a tendency for the information assets of the global South to be incorporated into the ‘public domain’ administered from the North, while the revenue in services derived from those assets continues to flow from the South to the North. The commercial logic applied in digital music licensing (in particular, within sample-based genres) is a logic of contamination – the inclusion of any data fragment results in a derived work, meaning that attribution is required and license fees payable. In contrast, the application of processes and algorithms (whether Autotune, or a fuzzbox) does not imply that the inventor of the fuzzbox owns the 173 I have already discussed the way in which the separation of data and algorithms in the systems of the GOFAI era has become far less distinct in the statistical models that underlie the behavior of ML devices such as Kinect. There is a commercial analog to this technical change, in the relationship between content and services. In the contemporary digital economy, we retain a notional separation between content and services. However, in practice, the corporations responsible for digital infrastructure ‘ecosystems’ find it useful to blur those boundaries. Apps for the iPad and iPhone often prevent the user from inspecting stored data other than through the filters of the application itself. The market models of interactive digital products (such as the AppStore) are gradually integrated with those of digitized content delivery (such as iTunes), and the company deploys proprietary services such as cloud storage, user account authentication and so on, on top of these. The ecosystem players – Apple, Google, Facebook and Microsoft – are all attempting to establish their control through a combination of storage, behaviour and authentication services that are starting to rely on indexed models of other people’s data. QUESTION 4: SELF-DETERMINATION The previous section has offered a relatively traditional Marxian analysis of what we might consider to be humane, in relation to economic exploitation of many people by a few. However, the original case study also draws attention to the ways in which ‘mind-reading’ technologies hold implications for the psychological identity of the user. This section considers implications of Machine Learning technologies for the self, as a psychological rather than purely legal and economic construct. Sense of Agency The first of these is the perception of agency – the sense of one’s self as an undivided and persisting entity, in control of one’s own actions. This is a key element of mental health, and is often disrupted in those suffering from delusional syndromes such as schizophrenia. In previous research, we have shown that diagnostic devices used to measure reduced sense of agency in psychiatric illnesses can also be used to measure the user’s sense of agency when interacting with ML systems [13]. The behavior of many ML-based systems is determined, not only by models of the external world, but by statistical user models that predict the user’s own actions. Those predictions may be based on data collected from other people (as in the case of Kinect), or one’s own past habits. However, in all of these cases, one frequent outcome is that the resulting system behavior becomes perversely more difficult for the user to predict. Rather than an explicit rule formulated and symbolically expressed by a designer, the behavior is encoded in the parameters of a statistical model – the kind of model that Breiman [7] describes as “complex, mysterious, and, at least, partly unknowable”. The result is frequently useful, but can also be surprising, confusing or irritating – as often noted of auto-correct [22]. This is a more serious problem than the commonplace observation that “if you are not paying for it, you’re not the customer; you're the product being sold” (e.g. [19]). While national and international legal frameworks focus on outdated models of content protection (through copyright and licensing) and service provision (through free trade and tariff agreements), the primary mechanism of control over users comes through statistical index models that are not currently inspected or regulated. The regulated revenue models of whole industries are being disrupted by digital alternatives that bypass retail distribution with proprietary indexing and access (e.g. Spotify, YouTube and others). The underlying models are transnational, with the corporations increasingly resembling the zaibatsu of William Gibson’s cyberspace, rendering national jurisdictions irrelevant in comparison with the internal structure of ML models. Whether or not the result is immediately useful, the work of Coyle et al [13] shows that these ML-based predictions reduce the user’s sense of agency in his or her own actions. Furthermore, some classes of user may be excluded from opportunities to control the system, because the prior data from which the trained model has been constructed does not take account of their own situation. One such example is those Kinect users whose body shapes are not consistent with the training data of the system, for example because they have absent limbs. This is a concern that is likely to become far more pressing after implementation of the proposed Transatlantic Trade and Investment Partnership, which would allow corporations to sue a country whose laws obstruct their interests. As a result, the significance of this analysis extends beyond the purely commercial, to the political foundations of our economic system [24]. When we replace content with services that obscure individual authorship through the construction of statistical models, Carlo Vercellone [38] observes that we are developing a new form of cognitive capitalism, in which access to the proprietary infrastructure encoded in models and indexes operates as a rent – extracting surplus value from the labor of the majority. The owner of the statistical model used to index and rank content, as in the case of Google’s PageRank, thus becomes a rentier, rather than a proprietor, of digital content [31]. These cases draw attention to the ways in which, when interacting with an inferred model, the user is effectively submitting to a comparison between their own actions and those of other people from which the model has been derived. In many such comparisons, the statistical effect will be a regression toward the mean – the distinctive characteristics of the individual will be adjusted or corrected toward the expectations encoded in the model. 174 A well-known early example of this challenge was expressed in the Wall Street Journal through advice on ‘If TiVo thinks you are gay, here's how to set it straight’ [40]. The choice of this specific theme drew attention to the implications of surveilling sexual preference, thereby conflating the concerns of privacy with those of control. In later reports, and in a classic meme formulation, the problem has been simplified purely as a matter of surveillance: ‘my TiVo thinks I’m gay’. However the TiVo developers at the time experimented with a ‘Teach TiVo’ function that could be used to modify the inferred model. Although briefly released in a beta version, the company eventually focused on refining the algorithms, rather than offering explicit control to users1. Construction of Identity This is the second of the ‘psychological’ problems that I suggest arise when interacting with an inferred world. The processes through which we construct our own individual identity depend on the perception that we are distinct individuals rather than interchangeable members of a homogeneous society. Processes of individuation, in which we separate ourselves from family and from others, are central to the achievement of maturity and personhood. These processes can be damaged through institutional and systemic constraints, leading to a widespread concern for self-determination as a fundamental human right. If the construction of one’s personal identity is achieved substantially through narratives in digital media – through Facebook profiles, photo streams, blogs, Twitter feeds and so on – then the behavior of these systems becomes a key component of self-determination. To some extent, digital media users are highly aware of the need to ‘curate their lives’ in the presentation of such content. However, they are less able to control the ML models that are inferred from their online behavior. At a trivial level, these models may record unintentional or private actions that the user would prefer to disown. At a more profound level, regressions to the mean result in personal identities that are trivialized (cute animals and saccharine sunsets), or have pandered to a lowest common denominator of mass-market segmentation (prurience and porn). Contracts The problem of how a user can express the behavior they want extends also to the legal relationship between users and service providers. As already noted in the earlier discussion of authorship, the structure of legal frameworks relies on symbolic representation rather than statistical patterns. Service providers now offer end-user license agreements that describe the procedures for collecting data rather than the implications of the model that will be trained from that data. Observation of user behavior, and datamining from content, have become completely routine, to the extent that it is hardly possible for users to opt out of this functionality if they want the product to work. At present, these universal license agreements do not help the user to understand what benefits they (or others) will obtain from the resulting inferred models [18]. They also provide no option for customizing or restricting either the model or the contract – the user may opt in or out, but not select any other trade-off between self-determination and convenience. This appears to be a joint failure of technical and legal systems, failing to recognize the interdependence of the two that arises from interacting with the world through an inferred model. QUESTION 5: DESIGNING FOR CONTROL The previous sections have referred to examples of interactive software, although the original case study was not itself proposed as an interactive system. This section considers two specific issues that arise when a user needs to operate products that have been built using ML techniques Control The anxieties regarding loss of control in inference-based interfaces are already well-established: small-scale behaviors such as Microsoft’s Clippy, Amazon recommendations, or predictive text errors become the object of popular ridicule, tinged only slightly with the anxiety of disempowerment. TOWARD HUMANE INTERACTION The ‘mind-reading’ case study that provided my initial example has the character of a parlor trick, albeit one that I have used to introduce some genuine ethical and legal issues. To reiterate the real concern for interactive systems: when an ML system sees the world we see, and also observes our responses, the expectation is that it will make predictions about us, and about what we want and need. This inferred model mediates our interaction with the world and with others. However, case studies in which the user’s own needs are modeled point to the issues that arise when more complex or larger-scale system behaviors are determined through ML techniques. Since the system behavior is derived from data, if the user wishes to control or modify that behavior, they must do so at one remove, by modifying the data rather than the model. As argued in [4], if the user wishes to change the behavior of the system more permanently, then ML-based techniques can inadvertently make the task more challenging. Rather than simply learning the conventions of a scripting or policy language in which to express the desired behavior, the user must second-guess an inference algorithm, trying to select the right combination of data to produce a model corresponding to their needs. I have drawn attention to numerous ways in which the shift from direct symbolic expression to inferred statistical models of the world has changed the nature of the relationship between interactive digital systems and the 1 175 Personal communication, Jennifer Rode, 20 Feb 2015. people that use them. These include questions about the source of the data in those statistical models, attributions of authorship and agency, the political consequences of shifts to non-symbolic expression, and the effects on the identity of the individual as constructing their own self and controlling their digital lives. I have suggested that these shifts result in systems that are less humane, because of the ways in which the relationship between the system behavior and human activities has become obscured through the scale and complexity of the modeling process. ML models have become less accountable, open to exploitation by commercial actors, closed to legal inspection, and resistant to the directly expressed desires of the user. Figure 2. Two images synthesized such that they will result in a classification judgment of ‘school bus’ (from [26], reproduced with permission of the authors). perception. A recent publication illustrated this phenomenon with images such as Fig. 2, and the title ‘Deep neural networks are easily fooled: High confidence predictions for unrecognizable images’ [26]. However, the goal of this paper has been to offer a technically-informed critical commentary, as a supplement to the many existing critiques that address more traditional dystopian anxieties of control and surveillance. Rather than simply sounding further warning alarms, or lamenting the loss of a golden age of symbolic transparency, a technically-informed critique should be able to draw attention to opportunities for technical adjustment, caution, correction and allowance. Although neither of the images in Fig. 2 is recognizable as a school bus, the image on the right offers a more visually salient account of those features in the training data set which have been encoded in the training process. They allow a human viewer to recall, for example, that school buses in the USA are painted in a characteristic color, and furthermore to speculate that this ML classifier might not be effective in other countries where those colors are not used. In contrast, the image on the left in Fig. 2 does not provide a basis for this kind of assessment (the original paper includes many similar noise-based images that represent different categories, while being indistinguishable to a human viewer). In the following sections, technical considerations are therefore presented as ways in which the structure of the inferred model might be opened up – more open to understanding by users and to critical assessment by commentators. Where symbolic systems offered direct representations of knowledge, ML systems must be inspected in terms of their statistical structure. Contrasts of this kind point toward the design opportunity for more humane ML systems that reveal the nature of the features from which judgments have been derived. It is often the case that such features do not satisfy the symbolic expectations underlying representational user interfaces. The semiotic structure of interaction with inferred worlds can only be well-designed if feature encodings are integrated into that structure. Features A classic student exercise in the GOFAI days was to ask how a machine vision system might recognize a chair. If the initial answer described four legs, with a seat and a back, the tutor would ask what about a stool, or one with three legs, or a bean bag2? Eventually the student might offer a functional description – something a person is sitting on. But then what about a person sitting on a table, or resting on a bicycle? The discussion might end with Wittgensteinian reflections on language, but the key insights are a) that judgments are made in relation to sets of features, and b) that accountability for a judgment is achieved by reference to those features. Labeling The models underlying many ML-based systems have been constructed from sets of training data, where each example in the data set is labeled according to a so-called ‘ground truth,’ often an expert judgment (this is characteristic of supervised learning algorithms – I discuss unsupervised learning below). The inferred model, however complex it might be, is essentially a summary of those expert judgments. However, it is expensive to label large data sets, so the availability of suitable labels often determines the models that can be obtained. For example, in the case of natural language processing systems, the ML model that is most often used to assign a part-of-speech (POS) to individual words (POS-tagging) is based on a training set from the Wall Street Journal. When applied to the speech of people who do not talk like the WSJ, the accuracy of this In the case of statistically inferred models, the features can often be far more surprising. One of the greatest technical changes in the transition from GOFAI to ML systems has been the discovery that many very small features are often a reliable basis for inferred classification models (e.g. [21]). However, the result is that it becomes difficult to account for decisions in a manner recognizable from human 2 This example is taken from a class taught by Peter Andreae at Victoria University of Wellington in 1986. 176 model will be reduced. But it is unlikely that expert judges would invest expensive time labeling (for example) the street patois of a Brazilian favela, even if there were a standard textbook description of its linguistic syntax. interaction, including predictive text and search engines, offer the user a list of choices that have been ranked according to relative confidence. However, these systems do not currently scale the ranked choices in proportion to the magnitude of the prediction. MacKay’s Dasher is an alternative example of a model-based predictive text entry system that directly exposes the confidence of the prediction in the user interface, by varying the size on screen of the different choices [36]. The phrase ‘ground truth’ implies a degree of objectivity that may or may not be justified for any particular labeled data set. If the interpretation of a case is ambiguous, then that training item must either be excluded from the data set (a common expedient), or labeled in a way that discounts the ambiguity – perhaps because the expert has not even noticed the problem. Furthermore, expert judges may approach a data set with different intentions from those who will interact with the resulting model. One might ask whether these experts are representatives of the same social class as the user, or whether their judgments are dependent on implicit context, perhaps uninformed by situations that they have never experienced. The challenge for incorporating confidence in an interactive system is to do so unobtrusively, allowing the user to take account of relevant cues without information overload. However, in order to establish this as a design opportunity, we first need to acknowledge that confidence does vary, and that probabilistic inferred models should not be presented as though they were deterministic. Errors Moreover, labeling large data sets is tedious and expensive, to a degree that those with broader expertise of judgment might be reluctant to spend their own valuable time in such activity. As a result, many researchers resort to the use of online labor markets such as Amazon’s Mechanical Turk (AMT), to commission ‘human intelligence tasks’. This strategy casts further doubt on the presumption of ground truth, through the economic relations in which it is embedded. For example, when the AMT service was introduced, it was formally available only to users in North America, with the result that statistical models labeled by AMT workers might incorporate an embedded form of cultural imperialism, perhaps of the kind illustrated in the black and yellow ‘school bus’ category of Fig. 2. Decisions made on the basis of an inferred model will include errors. Research results in ML conventionally report the degree of accuracy in the model (80%, 90%, 99% etc). However, the user’s experience of such models is often determined by the consequence of the errors, rather than the occasions on which the system acts as expected. 90% accuracy is considered a good result in much ML research, but using such models in an interactive system means that one in ten user actions will involve correcting a mistake. User experience researchers understand the need to focus on breakdowns, rather than routine operation, although in the past these have tended to result from indeterminacy in human behavior, rather than in the behavior of the system itself. It is important to recognize that departures from routine are more costly to manage than routine operation, because they require conscious attention from the user [5]. A system that mostly behaves as expected, with occasional departures, may be less useful than one that has no intelligence at all. Furthermore, it is possible that a 1% error rate will be even more dangerous than a 10% error rate, because the operator may become complacent or inattentive to the possibility of error. Confidence Many of the symbolic models created in the early days of GOFAI were deterministic – a particular output was known to have resulted from a given range of inputs, and those inputs were guaranteed always to produce the same output. In contrast, the behavior of inference-based systems is probabilistic, with a likelihood of error that varies according to the quality of the match between the inferred model and the observed world. This match will always be approximate, because the model is only a summary of the world, not an exact replica. (In fact, training a model until it does replicate the observed world is ‘over-fitting’ – it results in a fragile model that performs poorly at handling new data). Deep Learning The above discussion of features and labeling applies to the ML research techniques most popular in the early 2000's (and now widely applied in commercial practice), but it should be noted that recent algorithms in the broad category of ‘Deep Learning’ (including deep belief networks, convolutional neural networks and many others) raise somewhat different issues. Deep Learning techniques aim to be less dependent on explicit feature encoding, and also emphasise the importance of unsupervised learning, so that a labeled training set is not needed. However, each of these attributes leads to further questions for the critical technical practitioner. Despite the fact that inferred judgments can carry varying degrees of confidence, many interactive systems obscure this fact. In situations where the behavior of the system results from choosing the most likely of several possibilities, it might benefit the user to know that this behavior resulted from one 51% likelihood being compared to a 49% alternative, as distinct from another case where the model predicts one choice with 99% likelihood. Some of the most successful examples of inference-based 177 The first problem is that, just as it is not possible for a human to gain information about the world unmediated by perception, it is difficult for a Deep Learning algorithm to gain information about the world that is unmediated by features of one kind or another. These ‘perceptual’ features may result from signal conditioning, selective sampling, optical mechanics, survey design, stroke capture – because every process for capturing and recording data implicitly carries ‘features’ that have been embedded in the data acquisition architecture through technical means. If the features have not been explicitly encoded as a component of the ML system, then it is necessary for the critic to ask where they have been encoded. The questions already asked with regard to obfuscation of the model perhaps become more urgent, in that only the designers of the associated hardware may be able to provide an answer. a summary that is directly readable in the manner of a symbolic representation. The second challenge in assessing Deep Learning systems is that, if the judgments are not made by humans, they must be obtained from some other source. In one of the most impressive applications of convolutional neural networks, the staff of DeepMind Technologies [23] demonstrated a system that can learn to play early Atari video games without any explicit human intervention (other than drawing attention to the score – which is a crucial factor). DESIGN RESOURCES FOR INTERACTIVE ML It is possible for users to interpret and interact with such models in a way that places more emphasis on human concerns, but this requires designs that communicate essential characteristics of the model. Important aspects include the features that have been used to train the model, the source of the data in which those features were observed, the expert judgments that were applied when labeling the ground truth, the degree of confidence in any particular application of the model, the specific likelihood of errors in the resulting behavior, the infrastructure through which input data was acquired, and the semiotic status of the representational worlds in which an unsupervised model apparently acts. This paper has presented an historical argument for a new critical turn in HCI, that steps aside from the preoccupations of symbolic GOFAI and draws attention to the consequences of interacting with the inferred models derived from ‘big data’. It has investigated a number of specific problems that arise in such models, where these problems are also attuned to the ways in which distinctions between data and control, or content and services, are changing in the digital economy and regulation of the Internet. Examples of this kind are often discussed with the expectation that the next step after a video game will be action in the real world. Similar assumptions were often made in the GOFAI era, although that focus on ‘toy worlds’ was eventually abandoned, in recognition that operating ‘in the wild’ was overwhelmingly more challenging. This quite obviously applies to the case of Atari game worlds, and perhaps such toy applications do not seem a matter for serious concern. However, we do have reason to be concerned if similar algorithms are applied to some of the other representational ‘games’ played in contemporary society, such as the representational game worlds of corporate finance, audience ratings, or welfare benefits, and the ‘scores’ that are assigned to them in financial markets or monetarist government. This analysis suggests the need for design considerations that might help users to engage with inferred models in a way that is better informed, more effective, and supports their human rights to act as individuals. We need improved conceptual constructs that can be used to account for a new designed relationship between user intentions and inferred models. The following suggestions are drawn from work carried out in the author’s research group, in order to provide concrete illustrations of the kind of design research that might take account of these considerations. One such construct is the notion of agency – if the machine acts on the basis of a world model that is derived from observations of other users (or from the assumptions of an expert labeler), then this will be perceived by the user as a proportionate loss of personal agency through control of one's own actions. Fundamental human rights of identity, self-determination and attribution are implicated in this construct. If the inferred model obfuscates such relations, then they should be restored through another channel. So critical questions in the analysis of Deep Learning systems can be set alongside those of earlier ML techniques: 1) what is the ontological status of the model world in which the Deep Learning system acquires its competence; 2) what are the technical channels by which data is obtained; and 3) in what ways do each of these differ from the social and embodied perceptions of human observers? Each of these questions represents a deeply humane concern with respect to the representational status of inferred models and the degree to which we are obliged to interact with such models. A second construct is the interaction style previously described as programming by example – where future automated behaviors are specified by inference from observations of user action. Often promoted as an idealised personal servant, many such systems struggle to allow the basic courtesies of personal service, such as asking for confirmation of a command, or responding to a change of mind. Empowering users through such techniques will Summary The relationship between inferred models, and the data that they are derived from, is complex. The model is already a summarized version of the original data, although this is not 178 involve explicit representation of the inferred requirements and actions. Doing so requires a philosophical framework in which labour, identity and human rights are recognized as central concerns of the digital era – concerns that are directly challenged by recent developments in engineering thinking. In short, we need a discipline of humane computer interaction. A third construct is the recognition that, although approximate and errorful inferred models of the user’s intentions are problematic and worrisome, humans themselves also develop world models on the basis of incomplete and selective data. Kahneman's investigations of heuristics and biases in human decision-making [17] offer a mirror to the inferred world model of ML systems. There is a valuable opportunity to create user interfaces that acknowledge and support such human reasoning styles, rather than attempting to correct the user on the basis of unseen data or expert design abstractions. ACKNOWLEDGMENTS I am grateful for feedback and suggestions from Luke Church, Moira Faul, Leonardo Impett, Elisabeth LeedhamGreen, David MacKay, Becky Roach, Advait Sarkar, and anonymous reviewers of earlier versions. REFERENCES A fourth construct is to reconsider the role of the state, in an era when neither intellectual property nor legal policies need be explicitly formulated as symbolic representations. The new status of content that underlies inferred models throws new light on the role of public service broadcasters, who should be in a position to establish and protect genuine public value in the public domain [6]. 1. Agre, P. (1997) Computation and Human Experience, Cambridge University Press. These four illustrative examples are not proposed as the basis for a unified theoretical framework to be adopted by future design researchers. The intention is rather to provide a relatively pragmatic set of observations and suggestions, showing connections between the ideas in this paper and established topics within mainstream interaction design and digital media studies. Hopefully there are many other such opportunities, which may indeed come together to offer a basis for future design frameworks and methods. 3. Bennett, J. (2013) Forensic musicology: approaches and challenges. In The Timeline Never Lies: Audio Engineers Aiding Forensic Investigators in Cases of Suspected Music Piracy. Presented at the International Audio Engineering Society Convention. New York, USA, October 2013. 2. Agre, P. E. (1997). Towards a critical technical practice: lessons learned in trying to reform AI. In Bowker, G. Star, S. L & Turner, W. (Eds) Social Science, Technical Systems and Cooperative Work, Lawrence Erlbaum, Mahwah, NJ, pp. 131-157. 4. Blackwell, A.F. (2001). SWYN: A visual representation for regular expressions. In H. Lieberman (Ed.), Your wish is my command: Giving users the power to instruct their software. Morgan Kauffman , pp. 245-270. CONCLUSION 5. Blackwell, A.F. (2002). First steps in programming: A rationale for attention investment models. In Human Centric Computing Languages and Environments, 2002. Proceedings. IEEE 2002 Symposia on (pp. 2-10). IEEE. The central technical assumptions that underpinned the design of software applications for the first 50 years of the computer industry are now largely outdated. The intellectual agenda of data processing and communications, in which users either interact with each other or make choices between defined system functions, has not been succeeded by the autonomous human-like AI that was anticipated in the 1950s. Of course, HCI has always resisted such ambitions, drawing attention to the pragmatic human needs of social conversation and embodied usability. 6. Blackwell, A.F. and Postgate, M. (2006). Programming culture in the 2nd-generation attention economy. Presentation at CHI Workshop on Entertainment media at home - looking at the social aspects. 7. Breiman, L. (2001). Statistical modeling: the two cultures. Statistical Science 16(3), 199–231 In the new technical environment of the 21st century, users increasingly interact with statistical models of the world that have been inferred from a wide variety of data sources rather than explicit design judgments. This situation forces us to attend to the politics of information and action, as well as the attributes and limitations of the inference systems themselves. Just as the technical competence required of engineers is shifting from data and algorithms to information theory and stability analysis, so user experience designers must reconceive the relationship between content and services as constituting an ‘inferred world’ that stands in rich semiotic relation to individual and collective experience. 8. Borges, J.L. (1941/tr. 1962) The Library of Babel. Trans. by J.E. Irby in Labyrinths. Penguin, pp. 78-86 9. boyd, d. and Crawford, K. (2012). Critical questions for big data. Information, Communication & Society, 15(5), 662-679, 10. Card, S.K. Allen Newell, and Thomas P. Moran. 1983. The Psychology of Human-Computer Interaction. L. Erlbaum Assoc. Inc., Hillsdale, NJ, USA. 11. Collins, H. and Kusch, M. (1998). The Shape of Actions: What Humans and Machines Can Do. MIT Press. 12. Cope, D. (2003). Virtual Bach: Experiments in musical intelligence. Centaur Records. 179 13. Coyle, D., Moore, J., Kristensson, P.O., Fletcher, P. & Blackwell, A.F. (2012). I did that! Measuring users' experience of agency in their own actions. Proceedings of CHI 2012, pp. 2025-2034. predictions for unrecognizable images. arXiv:1412.1897 [cs.CV] http://arxiv.org/abs/1412.1897 27. Nishimoto, S., Vu, A. T., Naselaris, T., Benjamini, Y., Yu, B., & Gallant, J. L. (2011). Reconstructing visual experiences from brain activity evoked by natural movies. Current Biology, 21(19), 1641-1646. 14. Dourish, P. (2001). Where the Action Is: The Foundations of Embodied Interaction. MIT Press. 15. Dourish, P. (2004). What we talk about when we talk about context. Personal and Ubiquitous Computing 8(1), 19-30. 28. Norman, D. A., & Draper, S. W. (1986). User centered system design. Hillsdale, NJ. 29. Norman, D.A. Cognition in the head and in the world: an introduction to the special issue on situated action. Cognitive Science 17, 1-6 (1993). 16. Gill, K.S. (Ed.) (1986). Artificial Intelligence for Society. Wiley. 17. Gilovich, T., Griffin, D., & Kahneman, D. (Eds.). (2002). Heuristics and biases: The psychology of intuitive judgment. Cambridge University Press. 30. Norvig, P. (2011). On Chomsky and the two cultures of statistical learning. http://norvig.com/chomsky.html 31. Pasquinelli, M. (2009). Google’s PageRank algorithm: a diagram of cognitive capitalism and the rentier of the common intellect. In K. Becker & F. Stalder (eds), Deep Search. London: Transaction Publishers. 18. Gomer, R., schraefel, m.c. and Gerding, E. (2014). Consenting agents: semi-autonomous interactions for ubiquitous consent. In Proc. Int. Joint Conf. on Pervasive and Ubiquitous Computing: (UbiComp 14). 19. Kepes, B. (2013). Google users - you're the product, not the customer. Forbes Magazine, 4 December 2013. 32. Scherzinger, M. (2014). Musical property: Widening or withering? Journal of Popular Music Studies 26(1), 162-192. 20. Leahu, L., Sengers, P., and Mateas, M. (2008). Interactionist AI and the promise of ubicomp, or, how to put your box in the world without putting the world in your box. In Proc. 10th int. conf. on Ubiquitous computing (UbiComp '08), pp. 134-143. 33. Shotton, J., T. Sharp, A. Kipman, A. Fitzgibbon, M. Finocchio, A. Blake, M. Cook, and R. Moore. (2013). Real-time human pose recognition in parts from single depth images. Communications of the ACM 56(1), 116124. 21. Lowe, D. G. (1999). Object recognition from local scale-invariant features. In Proc 7th IEEE Int. Conf. on Computer Vision, pp. 1150-1157. 34. de Souza, C.S. (2005) The Semiotic Engineering of Human-Computer Interaction. The MIT Press. 35. Suchman, Lucy (1987). Plans and Situated Actions: The Problem of Human-machine Communication. Cambridge: Cambridge University Press. 22. Madison, J. (2011). Damn You, Autocorrect! Virgin Books. 23. Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., and Riedmiller, M. (2013). Playing Atari with deep reinforcement learning. http://arxiv.org/abs/1312.5602 36. Ward, D.J., Blackwell, A.F. & MacKay, D.J.C. (2000). Dasher - a data entry interface using continuous gestures and language models. In Proc. UIST 2000, pp. 129-137. 37. United Nations General Assembly. (1948). Universal Declaration Of Human Rights, Article 27. 24. Monbiot, G. (2013). Transatlantic trade and investment partnership as a global ban on left-wing politics. The Guardian, 4 Nov 2013. http://www.monbiot.com/2013/11/04/a-global-ban-onleft-wing-politics/ 38. Vercellone, C. (2008). The new articulation of wages, rent and profit in cognitive capitalism. Paper presented at The Art of Rent Feb 2008, Queen Mary University School of Business and Management, London. 25. Newell, A. (1974). Notes on a proposal for a psychological research unit. Xerox Palo Alto Research Center Applied Information-processing Psychology Project. AIP Memo 1 39. Winograd, T. and Flores, F. (1986). Understanding computers and cognition: A new foundation for design. Intellect Books. 40. Zaslow, J. (2002) If TiVo thinks you are gay, here's how to set it straight. Wall Street Journal online Nov. 26. http://www.wsj.com/articles/SB1038261936872356908 26. Nguyen, A., Yosinski, J. and Clune, J. (2014). Deep neural networks are easily fooled: High confidence 180