"What is word meaning, really? (And how can distributional models help us describe it?)"
In this talk, I will argue in favor of reconsidering models for word meaning, using as a basis results from cognitive science on human concept representation. More specifically, I argue for a more flexible representation of word meaning than the assignment of a single best-fitting dictionary sense to each occurrence: Either use dictionary senses, but view them as having fuzzy boundaries, and assume that an occurrence can activate multiple senses to different degrees. Or move away from dictionary senses completely, and only model similarities between individual word usages. I argue that distributional models provide a flexible framework for experimenting with alternative models of word meanings, and discuss example models.
Katrin Erk is an assistant professor in the Department of Linguistics at the University of Texas at Austin. She completed her dissertation on tree description languages and ellipsis at Saarland University in 2002, under the supervision of Gert Smolka. From 2002 to 2006, she held a researcher position in the Salsa project at Saarland University, working on manual and automatic meaning analysis of natural language text. Her current research focuses on computational models for word meaning and the automatic acquisition of lexical information from text corpora.