Achievement

Inspired by recent experimental work on word learning (Trueswell et al., 2013), IGERT faculty Trueswell (Psychology), Gleitman (Linguistics & Psychology) and Yang (Computer Science and Linguistics) have embarked on a computer modeling enterprise that examines how learners (both children and adults) learn the meanings of words from the immediate situational context. When a child hears a novel word, it could refer to many possible things present in the environment (or even absent entities). These researchers have developed and tested experimentally a model in which learners need only maintain a single meaning for each word, seeking verification at the next occurrence of that word (propose-but-verify). This stands in contrast with popular word learning accounts in which vast numbers of possible word-meaning pairs are tracked simultaneously in a memory intensive fashion. This work benefited from numerous discussions that occurred at IGERT-funded graduate seminars.