| 2 |
Numeral classifiers and the structure of DP
|
|
|
|
Abstract:
This dissertation investigates the structure of the Determiner Phrase from a crosslinguistic perspective, with a particular focus on English and Persian. Three main issues are addressed: the syntactic expression of number, the syntactic expression of specificity and the relationship between them. A numeral classifier is used in numeral+noun constructions, as in Mandarin yi/liang ge xuesheng 'one/two CL student', where the classifier is insensitive to the singular/plural distinction of the quantifying element. Other languages use number morphology, as in English table/tables, where number marking does make a singular/plural distinction. To account for the similarities and differences between classifiers and number morphology, this dissertation proposes that the heads of Number Phrase (NumP) and Classifier Phrase (CLP) house bundles of UG functional features. These features, arranged in a geometry based on Harley and Ritter's (2002) proposal for pronouns, assure syntactic composition of nP with classifiers and number morphology and of NumPs and CLPs with determiners. A classifier's feature, [individuation], only "individuates" the nP complement as a count noun; in contrast, plural morphology also has a [group] feature, which entails the presence if [individuation] but which also calls for a more-specific singular/plural distinction. This feature-based approach leads to the finding that despite descriptive claims that Persian has many "numeral classifiers", there is only one, ta; the others are in fact modifiers. On top of CLP and NumP, two Quantifier Phrases replace the traditional DP: Weak Quantifier Phrase (WQP) and Strong Quantifier Phrase (SQP). Following from earlier work, I argue that weak quantifying determiners denote a function from CLP/NumP predicates to a WQP generalized quantifier and that strong quantifying determiners denote a function from WQP generalized quantifiers to SQP generalized quantifiers. Crosslinguistic variation in the expression of number and specificity stems from slight differences in the feature bundles that appear in the functional heads and whether these bundles have overt form. This feature-based analysis permits the cooccurrence of a classifier and number marker, a cooccurrence ruled out in earlier theories but which occurs in a number of languages, including Persian. For determiners, particularly articles, the proposal is that the syntactic features in WQ and SQ are interpreted pragmatically with regard to the speaker's presuppositions about whether discourse participants know a referent. Necessarily, specificity and definiteness are analyzed into more primitive features. My working assumption is inspired by Cinque's (2002) hypothesis that functional syntax is the same across languages. But I diverge from the claim that exactly the same phrases are available in all languages. More precisely, I argue that what is universal are the functional features. These do tend to be associated with certain functional heads, but because the features can be bundled variously, some phrases may "fuse" under certain conditions (Bobaljik 1995, 2001) or may be always fused (Munn and Schmitt 2005, Schmitt and Munn 2002), with no evidence for the language learner that the two phrases are ever projected separately.
|
|
Keyword:
Linguistics
|
|
URL: http://pqdtopen.proquest.com/#viewpdf?dispub=3355664
|
|
BASE
|
|
Hide details
|
|
|
|
|
