GGrantIndex
← Search

New Computational Tools for Language Processing Research

$74,300FY2006SBENSF

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

Language is one of several cultural/creative traits that distinguish human behavior from that of other animals. Not surprisingly then, understanding how we learn, comprehend, and produce language is an active research domain in numerous disciplines, including psychology, linguistics, computer science, philosophy, and education. However, language technology research in computer science has become increasingly applied, and has not maintained close ties to psychological investigations of language. Such disciplinary specialization has meant that the methodological tools developed within computer science have not been fully exploited by psychologists, nor have the computational tools been shaped by insights from current psychological research. This project will support interaction between the investigator (a psychologist) and computer scientists working on language, vision, and the interface between the two, during an 11-month fellowship. The goal is to develop/adapt new computational tools for the psychological study of language in two specific domains: (1) Extracting information from large corpora to characterize human knowledge about language. (2) Better understanding of how humans integrate information from spoken or textual language with visual information about the current situation or scene. The development of sophisticated computational tools, together with experimental measures of human language behavior, is critical for answering fundamental questions about language processes in normal adults: To what extent is linguistic knowledge isolated from knowledge about the world? To what extent is language comprehension isolated from other sensory processes? If linguistic knowledge and language processing are integrated with non-linguistic cognitive representations and mechanisms, how do we characterize the nature of the inter-relationships? These questions are central to a scientific understanding of the functional architecture of the human mind. As such, they have implications for investigations of developmental and adult language disorders, reading and foreign language pedagogy, and human computer interaction. These activities are supported by the Methodology, Measurement, and Statistics Program and the Linguistics Program under the Mid-Career Fellowship component of the MMS Program Solicitation.

View original record on NSF Award Search →