This course examines the hypothesis that surface sentences are directly built by the combinatory syntactic rules, working in tandem with the compositional semantic rules (cf., Montague, 1974). In other words, surface sentences are "directly" assigned a model-theoretic interpretation, without recourse to a level of LF. (Note that despite intense debate on this issue in the 70's and 80's, the view that LF is necessary has become almost axiomatic in much current work within linguistic semantics. This course will challenge the LF view.) We will look in detail at some empirical phenomena which have been thought to necessitate LF, and will develop alternative accounts. Part of the focus will be on my own research on variable-free semantics, which removes arguments for LF (and "reconstruction") which are based on binding phenomena. Additional material will focuson quantifier scopes Pied-Piping, and Ellipsis phenomena. The material draws heavily from work within Categorial Grammar and systems inspired by Combinatory Logic.
Session 1: Background; Motivation for direct compositionality; Mechanics of variable-free semantics and of direct compositonality.
Sessions 2 and 3: Some initial payoffs: functional readings in a variety of constructionss; binding and coordination; paycheck pronouns; weak crossover (and i-within-i effects) and some apparent exceptions to weak crossover.
Sessions 4 and 5: Further applications: Direct compositional analyses of ellipsis, Pied-Piping, and focus; the interactions of focus and paycheck pronouns; variable-free treatments of focus. We will end with a discussion of open problems and challenges to the direct compositional approach.
At least one course in formal semantics. Familiarity with some version of Categorial Grammar is helpful, but not absolutely necessary.
Dept. of Cognitive and Linguistic Sciences
Providence, RI USA 02912
phone: (401) - 863-3037
fax: (401) - 863-2255