*******
Language, Cognition, and Computation Lecture Series *******
We use the interpretation of vague scalar predicates like
"small" as
an illustration of the ability of systematic semantic models of
dialogue context to derive useful, fine-grained utterance
interpretations from radically underspecified semantic forms.
Our
account involves two principles.
We model pragmatic reasoning as a
general process that infers consistent collaborative intentions to
explain agents' contributions to joint activity.
We interpret vague
predicates by recovering salient scales and relevant distinctions
along them from the dialogue context. Given this framework, we can
infer implicit standards of comparison for vague scalar predicates
through completely general pragmatics, yet closely constrain the
intended meaning to within a natural range.
Our account connects
closely with dynamic models from formal semantics, but we have
implemented it exactly in a natural language interface for
describing
spatial actions.
(The talk presents joint work with David DeVault, Rutgers.)
Bio
Matthew
Stone is an Assistant Professor in the Department of Computer
Science
and Center for Cognitive Science at Rutgers, the State
University
of New Jersey. He got his PhD in
Computer and Information
Science
at University of Pennsylvania, working under Mark Steedman in
the
Computational Linguistics Lab and the Institute for Research in
Cognitive
Science. His research centers
around computational
cognitive
models of natural language meaning and aims to characterize
how
meaning in conversation originates in and depends on
interlocutors'
collaborative real-world action. His
work therefore
bridges
computational logic, theories of agency and intention from
artificial
intelligence and philosophy, computational models of
embodied
action, and approaches to discourse context and semantic
representation
from formal and computational linguistics.
Recently,
he
has served on the Program Committee for IJCAI 2003, as the Tutorial
Forum
Chair of AAAI 2004, and program co-chair for TAG+ 2004, the
Seventh
International Workshop on Tree Adjoining Grammar and Related
Formalisms.