Probabilistic Semantics and Pragmatics Uncertainty in Language and Thought
Noah D. Goodman, Daniel Lassiter · 2015
This chapter illustrates the use of probabilistic techniques in natural language pragmatics and semantics with a concrete formal model. This model shows that a probabilistic framework for natural language is possible and productive. The chapter provides background on probabilistic modeling and stochastic ?-calculus, and introduce a running example scenario: the game of tug-of-war. It describes a formal fragment of English suitable for the running scenario. Using this fragment the chapter illustrates the emergence of non-monotonic effects in interpretation and the interaction of ambiguity with background knowledge. The chapter then describes pragmatic interpretation of meaning as probabilistic reasoning about an informative speaker, who reasons about a literal listener. It discusses the role of semantic indices in this framework and shows that binding these indices at the pragmatic level allows us to deal with several issues in context-sensitivity of meaning, such as the interpretation of scalar adjectives.