Solving the Individuation and Counting Puzzle with λ-DRT and MGL If I Can Get a Book from the Library, It Saves Me from Needing to Buy It in the Bookshop

摘要

Individuation and counting present an open puzzle for lexical semantics. The key challenge posed by this puzzle is that polysemous words can be counted according to different facets, using different individuation criteria for each. Several solutions have been proposed and challenged in the past, and the complexity of the responses expected makes it an interesting and pertinent test for formal theories and automated systems. We present a simple solution for this puzzle using an integrated chain of analysis from syntax to semantics and discourse, with a partial implementation that uses publicly available tools and frameworks. This clarifies the status of logical, compositional formalisms for lexical semantics regarding their ability to handle quantification and individuation. We also discuss the ability of the same formalisms to handle the resolution of discourse-level anaphora and correctly parse utterances introducing multiple lexical facets that are later referred to in the discourse.