| Linguistics
department
|
MellonFoundation Graduate ResearchWorkshopProgram Stanford
Semantics and Pragmatics Workshop
Ivan Sag
Abstract: In situ wh-interrogatives like (1)--(6) have received short shrift in the syntactic and semantic literature, either dismissed summarily as `metalinguistic' and not genuine questions (see Cooper 1983, Engdahl 1986, Haegeman 1991, Hornstein 1995), or else assumed to involve an intrinsically different logical form (Janda 1985). (1) A: Sandy loves
barack palinka.
Note that examples like (4)--(6) have no straightforward analysis as reprises of the prior utterance. Our basic claim is
that a reprise interrogative (including `echo' and `ref' uses) is metalinguistic
only in that the meaning that it gives rise to contains as a constituent
the illocutionary force of the (previous) utterance that it reprises. Beyond
that, we claim, there is nothing fundamentally different going on---syntactically
or semantically---from other uses of
We develop an account
of both reprising and non-reprising uses of in situ questions that incorporates
a uniform treatment of wh-expressions, one stated in terms of stored parameters
that are abstracted over in forming a question that serves as the content
of a superordinate phrase. Our analysis is construction-based, using
constraint inheritance to express the common properties of construction
families, e.g. wh-interrogatives, wh-relatives,
|