Components

• Framework:
  • An encoding of Combinatory Categorial Grammar (CCG)-based grammar fragments, as well as a CKY-style parser with string-based (instead of span-based) memoization.
  • A framework for typed λ-calculus based on δ-rules, which transform complex semantic representations into simpler (but equivalent) typed probabilistic models.
  • Code for transforming typed λ-terms into models in the Stan programming language.
• Theory:
  • An encoding of theories of formal pragmatics, including common grounds and Questions Under Discussion (QUDs).
• Analysis:
  • Analyses of particular natural language phenomena, including factivity and vagueness.