[Japanese | English]
It is cross-linguistically common for a single particle to serve as a part of a wh-indeterminate and a disjunction marker (Jayaseelan 2001; Slade 2011; Szabolcsi 2015). Among such multi-functional particles, one of the most well-studied one is the Japanese particle “ka” (Kuroda 1965; Hagstrom 1998; Shimoyama 2006, i.a.). However, none of the current compositional semantic analysis of “ka” (Hagstrom 1998; Shimoyama 2006; cf. Slade 2011) can successfully capture the fact that its function is conditioned by its syntactic position, both in its wh-indeterminate use and its disjunction use, in a parallel fashion. Specifically, both in wh-indeterminates and disjunctions, when the “ka”-phrase is embedded within a CP, its semantic contribution is an existential quantifier (without the question force); on the other hand, when it forms a matrix CP, its semantic contribution is to form a question. In this talk, I will propose a unified semantics for “ka” in wh-indeterminates and disjunctions that can properly capture this parallel effect.
According to the analysis, which I formalize in terms of two-tier alternative semantics (e.g., Rooth 1985, Beck 2006), “ka” is analyzed as an operator that always projects a set of alternatives. The crucial claim is that this set has to be closed into an existential quantifier if (and only if) it cannot by itself enter the semantic composition with the rest of the sentence without a type-mismatch. This accounts for the fact that a “ka”-phrase in a sub-CP position is interpreted existentially while it is interpreted as a question when it forms a CP. The set projected by “ka” in the former case has to be type-shifted to an existential quantifier to avoid a type-mismatch, while the set projected by “ka” in the latter case will not be type-shifted since the set itself can be interpreted as a question.
Discontinuous nominal expressions, expressions in which two or more subparts are separated in phrase structure (e.g. an adjective and a noun) are attested in a number of morphologically rich languages (e.g. Warlpiri, Wambaya, Latin, Russian, Polish). In all of these languages, word order is largely determined by information structure, and information structure has also been claimed to play a role in the occurrence of discontinuous nominal expressions (e.g. Fanselow & Fery 2006, Schultze-Berndt & Simard 2012), leading to the subparts of the same semantic unit having different information structural status. In this talk I address the role of information structure in discontinuous nominal expressions and the way that this relates to syntactic structure. Moreover, I discuss how discontinuity can straightforwardly be accounted for by glue semantics, the linguistic theory of semantic composition commonly used in the Lexical-Functional Grammar framework (Dalrymple et al. 1993). Discontinuity forms a problem for semantic composition as two parts of the same semantic unit are not adjacent in phrase structure. In this discussion I compare cases of discontinuity in ‘non-configurational’ languages to cases of discontinuity in other languages such as English relative clause extraposition and quantifier float.
Semantics Research Group
Last modified: 2016-04-01 15:13:32 JST