Grammar and information : a study of Turkish indefinites

Thesis Type: Doctorate

Institution Of The Thesis: Middle East Technical University, Turkey

Approval Date: 2010

Thesis Language: English

Student: Umut Özge



Turkish, along with many other languages, marks its direct objects in two distinct ways: overt accusative marking (Acc) versus no marking (∅). The research on the grammar and interpretation of Turkish indefinite descriptions has focused on the effects of this distinc- tion in case-marking on the interpretation of indefinite noun phrases. The overt accusative marker has been associated with discourse-linking (Nilsson 1985; Enç 1991; Zidani-Eroğlu 1997), specificity (von Heusinger 2002; von Heusinger and Kornfilt 2005), presuppositionality (Kennelly 1997; Kelepir 2001), individuation/particularization (Nilsson 1985; Taylan and Zimmer 1994; Bolgün 2005; Kılıçaslan 2006), and totality/delimitedness (Nilsson 1985; Nakipoğlu 2009). The common denominator of these proposals is that each draws a direct correlation between the accusative marker and a semantic or pragmatic category. The thesis argues that such direct associations are either too specific to receive full empirical support, or are too general to possess significant explanatory force. Instead, the thesis tries to come up with an account where the contribution of the accusative marker is minimized to create room for extra-grammatical resources to play explanatory roles. On the empirical side, we review and provide Turkish data that do not conform to the anal- yses of the accusative case as a specificity marker or as an existential presupposition trigger. The cases problematic for these proposals come from the interaction of accusative marked indefinites with various intensional operators. In an excursion to a closely related domain, we provide data and argumentation that challenge the tenability of a purely grammatically determined pattern of “neutral” intonation. Our discussion reveals that some information-theoretic concerns are effective in determining the “neutral” intonation of an utterance. We use these observations in clarifying the discourse-linking function attributed to the accusative marker. On the theoretical side, we search for a grammatical basis which not only gives rise to the interpretive effects attributed to the accusative marker, but also explains why certain types of noun phrases obligatorily receive the accusative marker in Turkish. On the basis of data from coordinating conjunctions, we argue that a “semantic incorporation” account that construes the difference between overt Acc versus ∅-marking as a type difference is not adequate in explaining the relevant Turkish facts. Instead, we propose to base the difference between Acc-marking versus ∅-marking on the distinction between properties and kinds. We take ∅-marked indefinites as existential quantifiers over instances of kinds that are licensed under string adjacency. We treat Acc-marked indefinites as referential objects based on contextually restricted properties, which are modeled as generalized Skolem terms (Steedman 2010). We show how the proposed distinction captures the empirical facts reviewed in the thesis.