Inchoative verbs and adverbial modification: Decompositional and scalar approaches Walter A. Pedersen Department of Linguistics McGill University, Montreal A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy (cid:2)c Walter A. Pedersen 2014 To my parents, Keith and Joanne. ii ABSTRACT This dissertation provides an in-depth comparison of the BECOME/small-clause and scalar approaches to the syntax and semantics of inchoative verbs. It is argued that the traditional BECOME/small-clause approach faces a number of insurmountable problems; these include, among others, the overgeneration of adverbial ambiguities, incorrect truth-conditions for sentences with adverbs in decomposition-internal attachment sites, and incompatibility with certain kinds of modifiers (like directional measure-phrases). A scalar approach, which does not require positing multiple clausal levels in the representation of what on the surface appearstobeamonoclausalsentence, isshowntoavoidmanyoftheproblemsinherentinthe decompositional approach. However, moving from a decompositional analysis of inchoative verbs to a scalar one requires new explanations for phenomena, like again-ambiguities, that were previously handled in decompositional terms. It is shown that, in many ways, scale- sensitive meanings can take the place of decompositional structures in the explanations of such phenomena, resulting in analyses that are preferable on both empirical and theoretical grounds. iii ´ ´ RESUME Cette dissertation pr´esente une comparaison d´etaill´ee de deux approches de la syntaxe et s´emantique des verbes inchoatifs: l’approche BECOME/proposition r´eduite et l’approche scalaire. J’affirme que l’approche traditionnelle de BECOME/proposition r´eduite fait face a` des probl`emes insurmontables; ceux-ci incluent la g´en´eration d’ambigu¨ıt´e adverbiale inexis- tante, des conditions de v´erit´e erron´ees pour les propositions avec un adverbe `a l’int´erieur de la structure d´ecompos´ee, et l’incompatibilit´e avec certains modificateurs, tels que les syntagmes de mesures directionnels. Je d´emontre qu’une approche scalaire, qui n’exige pas plusieurs niveaux propositionnels dans la repr´esentation des phrases apparement sim- ples, ´echappe `a beaucoup des probl`emes inh´erentes `a l’approche d´ecompositionelle. Cepen- dant, une approche scalaire des verbes inchoatifs n´ecessite une nouvelle analyse de certains ph´enom`enes, comme l’ambigu¨ıt´es avec again, qui recevaient une explication sous l’analyse d´ecompositionnelle. Je d´emontre que des interpr´etations sensibles aux ´echelles peuvent rem- placer des structures de d´ecomposition pour expliquer de tels ph´enom`enes, et que ces expli- cations sont pr´ef´erables pour des raisons empiriques et th´eoriques. iv ACKNOWLEDGEMENTS While this dissertation by necessity has a single author, it would not exist at all were it not for the guidance and support of many other people, who I wish to take this opportunity to acknowledge. First of all, I would like to thank my supervisor Brendan Gillon. For the past seven years he has challenged me to formulate my ideas as clearly as possible, to understand the formal tools I was applying to problems, and to think about my own work in a greater scientific, philosophical and historical context; I could not have asked for a more fulfilling education than the one I have received under his supervision. I would also like to thank my committee members Bernhard Schwarz and Ed Stabler. Bernhard Schwarz provided countless comments on many drafts of many papers, abstracts, and dissertation chapters; I have benefited greatly from his attention to detail and his gift for seeing how seemingly disparate ideas can be put together into a cohesive whole. I was also fortunate to spend two quarters at UCLA, where I learned much from Ed Stabler about computational and mathematical approaches to language, and in turn discovered new ways of thinking about the goals of theoretical linguistics. In addition to my dissertation committee, I would like to thank all of the professors at McGill, UCLA and elsewhere who have contributed in various ways to my education and thus to this dissertation. I am particularly indebted to Junko Shimoyama for introducing me to adverbial ambiguities in my first semester at McGill, and thus planting the seed which later grew into this dissertation; and to Alan Bale, who many times was able to point me to just the right resource at just the right time. Much of the enjoyment of graduate studies comes from experiencing the challenges it involves with others who are doing the same. I would thus like to thank all of the graduate students at McGill with whom I have shared offices, drinks, and many good times with. A special thanks to Jozina Vander Klok and David-Etienne Bouchard for help with French translation. I thank Naoko Tomioka for her constant willingness to discuss my work, for prodding me when I felt stuck, and just for bringing so much joy to my life. Finally, my parents have done so much to enable me to attain my goals, academic and otherwise, and have always been interested in what I was doing and where I was going. Nothing I have accomplished here or anywhere else would have been possible without their unwavering love and support. CONTENTS Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Re´sume´ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Overview of the dissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2 Predicate decomposition in linguistic theory . . . . . . . . . . . . . . . . . 6 2.1 Semantic markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Generative semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3 Montague Grammar and metalinguistic expansion . . . . . . . . . . . . . . . 15 2.4 The Y-model: Logical Form and Phonetic Form . . . . . . . . . . . . . . . . 20 2.4.1 Interpreting Logical Forms . . . . . . . . . . . . . . . . . . . . . . . . 24 2.4.2 The Y-model and decomposition . . . . . . . . . . . . . . . . . . . . 28 2.5 Syntactic predicate decomposition in GBT/ Minimalism . . . . . . . . . . . 30 3 Ambiguity and the Y-model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.1 Ambiguity and the ‘alternate truth value judgment test’ . . . . . . . . . . . 36 3.2 Sources of ambiguity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.3 Ambiguity vs. indeterminacy, vagueness and deixis . . . . . . . . . . . . . . . 48 3.3.1 Indeterminacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3.2 Vagueness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3.3.3 Deixis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4 Syntactic decomposition of inchoative verbs . . . . . . . . . . . . . . . . . . 52 4.1 Inchoative verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.2 The BECOME/small-clause analysis . . . . . . . . . . . . . . . . . . . . . . . . 63 4.2.1 Interval Semantics for BECOME . . . . . . . . . . . . . . . . . . . . . 66 4.2.2 BECOME/small-clause decompositions and adverbial modification . . . 70 4.3 Issues with the BECOME/small-clause analysis . . . . . . . . . . . . . . . . . . 75 4.3.1 Ambiguity overgeneration . . . . . . . . . . . . . . . . . . . . . . . . 75 4.3.2 Cross-linguistic and cross-speaker variation . . . . . . . . . . . . . . . 77 v 4.3.3 Incorrect readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.3.4 The gradable nature of change . . . . . . . . . . . . . . . . . . . . . . 81 4.3.5 PF issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.3.6 Multiple decompositions and a missed generalization . . . . . . . . . 86 4.3.7 A delineation approach to BECOME . . . . . . . . . . . . . . . . . . . . 87 4.3.8 Closing remarks about the BECOME/small-clause analysis . . . . . . . 89 4.4 Dowty’s (1979) Metalinguistic Expansion (ME) analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 5 Foundations of gradability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 5.1 Gradability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 5.1.1 Preorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 5.1.2 The move to abstract scales and degrees . . . . . . . . . . . . . . . . 99 5.2 Infinite difference systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 5.2.1 IDSs and gradability . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 5.2.2 Differential modifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 5.2.3 Antonymy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 5.2.4 Closed and open scales . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.3 IDS measurements as ‘degrees’ . . . . . . . . . . . . . . . . . . . . . . . . . . 118 5.A Sassoon’s (2010) measurement-theoretic analysis . . . . . . . . . . . . . . . . 120 6 A scalar analysis of inchoative verbs . . . . . . . . . . . . . . . . . . . . . . 124 6.1 Gradable Adjectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 6.1.1 Gradable and non-gradable adjectives . . . . . . . . . . . . . . . . . . 124 6.1.2 Relative and absolute adjectives . . . . . . . . . . . . . . . . . . . . . 125 6.1.3 Explaining the relative-absolute distinction . . . . . . . . . . . . . . . 132 6.1.4 A semantics for adjectival bare forms . . . . . . . . . . . . . . . . . . 135 6.2 A scalar analysis of verb meaning . . . . . . . . . . . . . . . . . . . . . . . . 139 6.2.1 Scales and inchoative verbs . . . . . . . . . . . . . . . . . . . . . . . 139 6.2.2 Verbs with fully-closed scales . . . . . . . . . . . . . . . . . . . . . . 145 6.2.3 Variable telicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 6.2.4 Degree-pairs vs. difference degrees . . . . . . . . . . . . . . . . . . . . 147 6.3 Comparison of the scalar and BECOME/small-clause approaches . . . . . . . . 152 6.3.1 Result-state inferences . . . . . . . . . . . . . . . . . . . . . . . . . . 152 6.3.2 Adverbial modification . . . . . . . . . . . . . . . . . . . . . . . . . . 155 6.3.3 Result-state/pre-state asymmetries . . . . . . . . . . . . . . . . . . . 163 vi 6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 7 A scalar analysis of again-ambiguities . . . . . . . . . . . . . . . . . . . . . 165 7.1 Again-ambiguities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 7.1.1 Repetitive and reversative readings . . . . . . . . . . . . . . . . . . . 167 7.1.2 The syntactic position of again . . . . . . . . . . . . . . . . . . . . . 169 7.2 Previous accounts of again-ambiguities . . . . . . . . . . . . . . . . . . . . . 171 7.2.1 BECOME and again . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 7.2.2 Counterdirectional again . . . . . . . . . . . . . . . . . . . . . . . . . 176 7.3 A scalar analysis of again-ambiguities . . . . . . . . . . . . . . . . . . . . . . 177 7.3.1 Two meanings for again . . . . . . . . . . . . . . . . . . . . . . . . . 178 7.3.2 Repetitive readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 7.3.3 Reversative readings . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 7.4 Discussion and comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 7.4.1 Structural properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 7.4.2 Inter-language and cross-linguistic variation . . . . . . . . . . . . . . 187 7.4.3 More complex sentences with again . . . . . . . . . . . . . . . . . . . 190 7.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 vii Chapter 1 Introduction Pedersen (2014) The question often arises in theoretical linguistics, when attempting to analyze some linguistic phenomenon, whether one should enrich the syntactic representation of a sentence in some way, or instead whether one should posit more enriched denotations for the expressions involved. A case in point is the ‘result-state’ inferences that inchoative verbs like open give rise to. For example, an utterance of (1a) leads one to naturally infer that (1b) is true. (1) a. The door has opened. b. The door is or was open. An early but highly influential approach to result-state inferences like the above was to posit that the syntactic representation of sentences like (1a) contain unpronounced mate- rial that make the inferential link between that sentence and another a matter of ‘syntactic containment’. This approach, referred to here as syntactic predicate decomposition, dates back to the time of the Generative Semantics movement (e.g. Morgan 1969; McCaw- ley 1971). The most influential decomposition of inchoative verbs is what can be called a ‘BECOME/small-clause’ decomposition; when given an analysis of this kind, a sentence like (1a) is assigned a syntactic analysis something like that in (2). (2) [ [ The door ] [ [ BECOME ] [ t open ] ] ] S NP 1 VP V SC 1 ADJ In this type of analysis, the verb-phrase of the sentence is ‘decomposed’ into a BECOME operator and a small-clause consisting of the adjective open and the subject of the sentence; a transformation of some kind raises the small-clause subject to the subject-position of the matrix clause (the original small-clause position of the subject is shown in (2) with a trace t ). Theimportantthingtonoteaboutthistypeofanalysisisthatresult-stateinferenceslike 1 the one described above result from a kind of ‘syntactic containment’; the stative clause that 1 Chapter 1. Introduction 2 characterizes the result-state of a sentence is assumed to be an actual syntactic constituent of that sentence. Evidence in favor of a BECOME/small-clause decompositional analysis of inchoative verbs has been argued to come from certain kinds of sentences involving adverbial modifica- tion, where the interpretation of a particular sentence can apparently be explained with a decompositional-internal position for the adverb involved (e.g. Morgan 1969; McCawley 1971,1973b;vonStechow1996). Oneadverbwhichisclaimedtoappearbothdecomposition- internally (i.e. below BECOME) and decomposition-externally (i.e. above BECOME) is again; these different attachment sites for again are used to explain the two readings of a sentence like (3), which is claimed to be ambiguous between a ‘repetitive’ reading and a ‘restitutive’ reading. (3) The door opened again. a. repetitive: The door opened, and the door opened before. b. restitutive: The door opened, and the door has been open before. (4) a. repetitive: [ [ The door ] [ [ [ BECOME ] [ t open ] ] again ] ] S NP 1 VP(cid:2) VP V SC 1 ADJ b. restitutive: [ [ The door ] [ [ BECOME ] [ [ t open ] again ] ] ] S NP 1 VP V SC 1 ADJ When again attaches below BECOME and modifies the small-clause, a restitutive reading re- sults; when again attaches above BECOME and modifies the whole verb-phrase, a repetitive reading results. The fact that a BECOME/small-clause decomposition makes available attach- ment sites for adverbs like again – attachment sites which actually appear to be utilized – is taken to be strong corroborating evidence for a decomposition approach to inchoative verbs (c.f. von Stechow 1996). Decompositional analyses like the BECOME/small-clause analysis were developed before the widespread adoption of formal semantics in linguistics research; at the time, ‘syntactic containment’ was viewed as the main method for generating inferences between sentences. With the advent of the application of formal semantics to natural languages (e.g. Mon- tague 1970a), it become possible to approach phenomena like ambiguity and inter-sentential inferences in ways that did not require positing substantial abstract syntactic structure. Re- garding inchoative verbs specifically, one type of semantic analysis that has emerged in more recent times is what can be called a scalar analysis (see e.g. Hay et al. 1999; Winter 2006; Kennedy and Levin 2008). In this type of analysis, inchoative verbs are assigned formal denotations that involve scales consisting of an ordered set of degrees, as well as a func-
Description: