Regine Brandtner - Universität Stuttgart
Transcription
Regine Brandtner - Universität Stuttgart
ConSOLE XVIII Barcelona December 19, 2009 Events and their results – pragmatic constraints on copredication Regine Brandtner, Stuttgart University regine.brandtner@ling.uni-stuttgart.de 0. Introduction to copredication Copredication: Modifiers and predicates ranging over different semantic domains (events, states, objects etc.) can be simultaneously predicated of the same nominal: Die [langwierige]EVENT Übersetzung [verkaufte sich millionenfach]RESULT (1) OBJECT. ‘The tedious translation sold million-fold.’ Main questions: How can we compose sentences like this? Which reading does the nominal get in these contexts? Under which conditions can we create this structure? (2) ?Die Absperrung 1 [aus Holz]RO [dauerte zwei Stunden]EV. ‘The obstruction made of wood took two hours.’ (3) ?Die [regelmäßige]EV Lüftung der Kinderzimmer ist wichtig, aber [kaputt]MEANS. ‘The regular ventilation of the nursery is important, but damaged.’ Structure of the talk: Deverbal –ung nominals in German and their readings in context Copredication and its constraints A new viewpoint on copredication: predicate transfer and noteworthiness This research was done within the project “The formation and interpretation of derived nominals”(B1) headed by Prof. Klaus von Heusinger and Prof. Artemis Alexiadou, which is part of the collaborative research centre “Incremental specification in context” (SFB 732) funded by the German Research Foundation (DFG). Moreover, I would like to thank Daniel Hole for his comments on this talk. I am responsible for any shortcomings. 1 This nominal belongs to a special case, which is MEANS and RESULT OBJECT at the same time. However, for me the resultative feature is the crucial point here. 1 1. Deverbal nominals in German: German has various means to derive (deverbal) nominals, e.g. the suffixes –(er)ei, -er and –ung: (4) Dieser Vortrag beschäftigt sich mit der Nominalisierung von Verben, besonders mit der Endung –ung, welche den Betrachter oft in einen Zustand der Verwirrung und der Grübelei versetzt. Der Forscher findet diese Nominalisierungen in nahezu jeder Beschreibung oder Abhandlung in der Bücherei. ‘This talk deals with the nominalization of verbs, especially with the ending -ung, which often puts the observer into a state of puzzlement and of pondering. The researcher finds these nominalizations in nearly every description or paper/treatise in the library.’ (‘side table’, lit.: ‘addition’) (lit.: ‘demonstration, illustration’) However, -ung nominals are the only ones who can appear in copredication structures: (5) ?Der Flieger 2 [hat einen Flugschein]AGENT und [ist überfüllt]INSTRUMENT. ‘The aviator has a pilot’s licence and (the plane) is crowded.’ (6) ?Die Brauerei [liegt am Stadtrand]LOCATION und [ist sehr schwierig]EVENT. ‘The brewery is on the outskirts of the city and is very challenging.’ This behaviour might tell us something about the distribution of and relation between their readings, which are summarized for –ung nominals in table 1: 2 Literally: flyer, which can refer to the plane or to the pilot. 2 Table 1.readings of –ung nominals Entführung ‘kidnapping’ EVENT RESULT STATE Absperrung ‘obstruction’ ABSTRACT RESULT Übersetzung ‘translation’ RESULT OBJECT Bemalung ‘painting’ MEANS Lüftung ‘air conditioning’ AGENT Bedienung ‘waitress’ INSTITUTION/ Verwaltung ‘administration’ COLLECTIVE LOCATION Unterführung ‘underpass’ One nominal form mostly displays several of these readings They belong to different domains and have different features, shown in table 2 These features correlate with the selectional restrictions of certain modifiers and predicates (“indicators” shown in table 2 for events (EV) and result objects (RO)) 3 Table 2. Features and indicators of nominalization readings: F E A T U R E S Event Result state AbstractResult ResultObj. Means Agent/Coll. +duration +duration -duration -duration 4 -duration -duration +dynamic -dynamic -dynamic -dynamic -dynamic -dynamic +abstract +abstract +abstract -abstract -abstract -abstract +resultative +resultative +resultative +/-resultative -resultative -human -human -human -human +human -human I N D I C A T O R S Event readings Result object readings DP modifiers Dates: am 7.Juli ‘July 7th’ am Montag ‘on Monday’ Size, shape, weight etc.: lang, rot, schwer ‘long, red, heavy’ Process indicating adjectives: Iteration vorsichtig ‘cautious’ Physical change: zerrissen ‘disrupted’ wiederholt ‘repeated’ Internal structure: (200 Seiten / Teile) umfassend ‘consisting of 200 pages / parts’ 3 Cf. also Ehrich&Rapp 2000. Objects only have indirect durations, since we can say The book took 2 hours (to read, to write, to produce…), hence it is still an event that we are referring to. 4 3 Predicates Time frame predicates: Duration: beginnen/ aufhören/ weitergehen ‘begin’/ ’stop’/ ’continue’ dauert 6 Monate ‘takes 6 months’ Physical change: überreichen / zerstören ‘present’/ ‘destroy’ Posture verbs: (auf dem Tisch) liegen / stehen ‘lie/ stand (on the table)’ 2. Copredication and its constraints In many contexts, we find different indicators, but only one token of the nominal, to which all of them apply (Messung here): (7) Nur wenn man die genaue Bezeichnung des Videosystems kennt, kann man abschließend sagen, ob die [vorliegende]RO Messung [regelgerecht durchgeführt]EV wurde und somit [verwertbar]RO wäre. 5 ‘You can only tell whether the measurement at hand was conducted regularly and is hence usable, if you know the precise name of the video system.’ Is the nominal: underspecified? / able to shift its reference? / only disambiguated locally? Existing theories on nominalizations in context (e.g. Bierwisch 1989, Pustejovsky 1995): Focus on lexical structure Readings treated on a par no prediction of exceptions for copredication 2.1 Constraints concerning the readings involved: -ung nominals can in principle appear in copredication structures (cf. (1)), however we also find unacceptable cases here: 5 (8) ?Die [regelmäßige]EV Lüftung der Kinderzimmer ist wichtig, [kaputt]MEANS. ‘The regular ventilation of the nursery is important, but damaged.’ (9) ?Die Verwaltung ist [mühselig]EV und seit Montag [im Urlaub]COLLECTIVE. ‘The administration is troublesome and on vacation since Monday.’ aber http://www.frag-einen-anwalt.de/Polizeivideo-bei-Geschw.%C3%BCberswchreitung_f26038.html, 7.02.2008 4 (10) ?Die [sorgfältige]EV Reinigung der Kleider [ist ein Familienbetrieb]INSTITUTION. ‘The careful dry cleaning of the clothes is a family business.’ In these cases, the acceptability seems to depend on the readings involved, since other combinations are acceptable: (11) Die [gestern erfolgte]EV Sperrung der Foren [wird bald wieder aufgehoben]RESULT STATE. ‘The blocking of the bulletin boards carried out yesterday will be lifted soon.’ EV & RESULT STATE (12) Die Übersetzung [hat ein Jahr gedauert]EV und kann nun endlich [gedruckt werden]ABSTRACT RESULT. ‘The translation has taken one year and can finally be printed now.’ EV & ABSTRACT RESULT (13) Der Redakteur hat die [fehlerhafte]ABSTRACT RESULT Bekanntmachung [in den Mülleimer geworfen]RO. ‘The editor has thrown the faulty announcement in the dustbin.’ ABSTRACT RESULT & RESULT OBJECT Assumption 1 based on table 2: only if they share at least one feature, can different readings appear in a copredication. What about result objects then? They do not share any features with the event, but they result from the event and this relation can compensate this: (14) Die Zahnfüllung [dauerte drei Stunden]EV und [ist aus Gold]RO/MEANS. ‘The filling took three hours and consists of gold.’ However, even if we have two combinable readings we do not always get acceptable examples. 2.2 Constraints concerning the structure of the sentence: Acceptability does not only depend on the readings involved, but also on the placement of the indicators within the sentence structure: (15) Die [nur auf eine Stelle genaue] ABSTRACT RESULT Messung [liegt im Müll]RO. ‘The measurement only accurate to one decimal place lies in the trash.’ (16) Die Messung, die nur [auf eine Stelle genau]ABSTRACT RESULT ist [liegt im Müll]RO. ‘The measurement that is only accurate to one decimal place lies in the trash.’ (17) ?Die Messung [ist nur auf eine Stelle genau]ABSTRACT RESULT und [liegt im Müll]RO. ‘The measurement is only accurate to one decimal place and lies in the trash.’ Assumption 2: subordination of one indicator simplifies copredication 5 In copredications with an event and a result, a coordination structure is even worse, since the result cannot simultaneously exist with its event 6 : (18) ??Die Messung ist [mühsam]EV und [auf eine Stelle genau]RO. ‘The measurement is troublesome and accurate to one decimal place.’ However, the coordination of competing indicators is possible, if they are temporally disjoint (cf. also Jezek&Melloni 2009 for Italian): (19) Die Übersetzung [hat ein Jahr gedauert]EVENT und kann nun endlich [gedruckt werden]ABSTRACT RESULT. ‘The translation has taken one year and can finally be printed now.’ However, if these conditions concerning structure and readings are met, we still have no explanation for the oddness of the following examples. 2.3 Pragmatic constraints: The following examples involve event and result object indicators and are not coordinated, but still they are not acceptable: (20) ?Die Absperrung [aus Holz]RO [dauerte zwei Stunden]EV. ‘The obstruction made of wood took two hours.’ (21) ?Die [abblätternde]RO Bemalung [war mit roter Farbe durchgeführt worden]EV. ‘The scaly painting has been done with red paint.’ (22) ?Die [umständliche]EV Bestellung [wiegt 2 kg]RO. ‘The cumbersome order weighs 2 kg.’ To account for these constraints, I will first introduce a complementary approach to copredication by adopting Nunberg’s notion of predicate transfer (1995, 2004): Predicate transfer: instead of shifting the nominal, we transfer the second indicator to a predicate also applying to the first reading (here an event), as in: (23) Die [langwierige]EV Übersetzung [verkaufte sich millionenfach]RO. ‘The tedious translation sold million-fold.’ Die [langwierige]EV Übersetzung {hat ein Resultat, das sich [millionenfach verkaufte]RO}EV ‘The tedious translation has a result that sold million-fold.’ 6 This would only be acceptable, if it was the original text on the table, since the translated text depends on the completion of the event. This is not the case with simple nouns; cf. the example by Asher & Pustejovsky 2005: Lunch was delicious, but took forever. 6 The conditions for creating new (enriched) predicates are that there has to be a salient relation between their bearers and that: Assumption 3: the enriched version has to be noteworthy in the utterance situation for the identification or classification of the bearer: (24) Why not try some beer fed, hand massaged (?three legged) wagyu beef? 7 It is not the meat that is massaged, but the animal: we can create the predicate [hand massaged] applying to meat, since it acquires a noteworthy property in this context (being more tender etc.) Application to derived nominals: The salient relation between events and their results is the resultative aspect. In addition, e.g. a result predicate used for an event must be noteworthy in the context, i.e. for the kind of event: For examples (20) – (22) this means, it is not noteworthy: for the duration of an event of what kind of material its result is made for the shortcoming of an event, which colour its result has for the severity of an event how much its result weighs Hence, in examples (20) - (22) the predicate transfer could be licensed by a different indicator, which is noteworthy in this context: (25) Die [aus über 100 Teilen bestehende]RO Absperrung [dauerte 2 Stunden]EV. ‘The obstruction consisting of more than 100 parts took two hours.’ (26) Die [abblätternde]RO Bemalung [war mit alten Ölfarben durchgeführt worden]EV. ‘The scaly painting has been done with old oil paint.’ (27) Die [umständliche]EV Bestellung [wurde endlich verschickt]RO. ‘The cumbersome order was finally sent. 7 Acceptable part from: http://www.flyertalk.com/forum/diningbuzz/717174-whale-meat.html, Decembre 2, 2009. 7 Besides, there are event indicators like sorgfältig ‘careful’, aufwendig ‘elaborate’ and präzise ‘precise’ which generally have very strong implications for their results. If you do a measurement in a precise way, the values will be precise, too etc.: (28) Die [präzise]EV Messung der Studenten wurde im Aufsatz des Professors [veröffentlicht]ABSTRACT RESULT. ‘The precise measurement of the students was published in the professor’s paper.’ (29) Die [sorgfältige]EV Auflistung der Daten ist [zeitaufwendig]EV und [umfasst 5 Seiten]RO. ‘The careful listing of the data is time consuming and comprises 5 pages.’ (30) Die [aufwendige]EV Verpackung [war nach 5 Minuten zerrissen / landete schließlich im Müll]RO. ‘The elaborate packaging was damaged after 5 minutes / ended up in the trash.’ Assumption 4: Apart from the context, there are also indicators, for which this productive process of transfer between events and results is conventionalized support copredication between events and results Noteworthiness with nominalizations can also depend on our expectations about the correlations between events and results, e.g. that an event that takes a long time will result in something special: (31) Die [täuschend echte]RO Fälschung [dauerte lange]EV. ‘The deceptively real-looking imitation took a long time.’ (32) ?Die [schlecht gemachte]RO Fälschung [dauerte lange]EV. ‘The badly done imitation took a long time.’ (33) Die [schlecht gemachte]RO Fälschung [dauerte trotzdem lange]EV. ‘The badly done imitation still took a long time.’ Conclusion: Copredication cannot only tell us something about the distribution of nominalization readings, but needs its own analysis and clear conditions We cannot explain copredication and its constraints on semantic grounds concerning the nominal alone Nunberg’s notion of predicate transfer is extendable to nominalizations and their special distribution of readings, while noteworthiness constrains its application 8 References: Asher, N.& Pustejovsky, J., 2004. Word meaning and commonsense metaphysics. Semantics Archive: http:\\www.semanticsarchive.net. Bierwisch, M. 1989. “Event Nominalization: Proposals and Problems“. In: W. Motsch (ed.). Wortstruktur und Satzstruktur. Berlin, 1- 73 Brandtner, R. 2008. Meaning Transfer and the Compositional Semantics of Nominalizations. In: SinSpeC. Working Papers of the SFB 732 "Incremental Specification in Context", ed. Florian Schäfer. Online Publikationsverbund der Universität Stuttgart (OPUS) Cruse, D. A. 2000. Meaning in Language: An Introduction to Semantics and Pragmatics. New York: Oxford University Press Ehrich, V. & Rapp, I. 2000. Sortale Bedeutung und Argumentstruktur: ung- Nominalisierungen im Deutschen. Zeitschrift für Sprachwissenschaft 19: 245303. Jezek, E. & Melloni, C., to appear. Complex types in the (morphologically) complex lexicon. In: Proceedings of GL2009, 5th International Conference on Generative Approaches to the Lexicon, Sept. 15-17, 2009, Pisa, Italy. Nunberg, G. 1995. Transfers of Meaning. Journal of Semantics 12.2, 109132. Nunberg, G. 2004. The Pragmatics of Deferred Interpretation. In: The Handbook of Pragmatics, eds. Laurence Horn & Gregory Ward, 344364. Oxford: Blackwell Pustejovsky, J. 1995. The Generative Lexicon. Cambridge, MA: MIT Press. Appendix Nunberg’s predicate transfer (1995, 2004) In case of a sortal mismatch, we do not always have to shift the NP reference: (1) i. I am parked out back. ii. I am {the owner of a car that is [parked out back]} predicate transfer on the VP meaning instead of shifting the pronoun to car Test for transfer position: (2) i. Ij am {the owner of a car that is parked out back}j and have been waitingj for 15 minutes. ii. #I am {the owner of a cari that is parked out back} and may not starti. we cannot coordinate a predicate referring to the car, but only one referring to the driver (claim: the NP is still referring to him while the predicate changes its meaning) 8 Copredication with simple nouns (Nunberg 2004): (3) i. ii. Roth is [Jewish]person and [widely read]books. Roth is Jewish and {his books are [widely read]books}person both predicates refer to the person Roth, since he acquires a noteworthy property through his books being widely read (he might be famous, rich etc.) 8 In contrast, in the example This is parked out back (pointing to the keys) he would assume transfer on the demonstrative, since here we can go on with …and may not start etc. 9 10 11 12