Modeling Integration Effects in Language with Probabilistic Logic

Lead Research Organisation: University of Edinburgh
Department Name: Sch of Informatics

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description A syntax/semantics interface for a psycholinguistics model 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Primary Audience
Results and Impact The cognitive modeling community has expressed a great deal of

interest in simulating incremental syntactic processing. Syntax is, of

course, not the end of the language understanding process. However,

there has not been as much work on simulating word-by-word

incrementality in semantic and pragmatic processing. Before

developing a model of incremental semantic interpretation, though, it

is necessary to have a firm grasp on how syntactic structures are

converted to semantic ones. Indeed, existing computational models of

the syntax/semantics interface generally cannot produce a meaningful

semantic representation of an utterance until the syntax has been

completely processed -- even in cases when they assume the syntax is

analyzed incrementally.



In this talk, I will outline a sketch of an incremental

syntax/semantics interface, giving particular emphasis on how the

syntax/semantics interface lays the foundation for dealing with issues

such as argument structure ambiguity, quantifier scope ambiguity, and

syntactically-licensed presuppositions and implicatures. A critical

part of the model is how behavioural predictions are made from the

computational simulations. I assume that the syntax/semantics

interface is situated in a probabilistic model, using an incremental

probabilistic parser as the syntactic model, and, a "probabilistic

logic" for semantic interpretation. The semantic interpreter, instead

of attempting to prove the truth-conditional semantics of a utterance,

rather computes the probability of a (possibly ambiguous)

interpretation. This not only has an obvious role in disambiguation,

but also creates predictions on how "surprised" the sentence processor

is as new material is processed incrementally.
Year(s) Of Engagement Activity 2010
 
Description Inferring grounding through coherence and rationality 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Primary Audience
Results and Impact This paper analyses dialogues where understanding and agreement are problematic. We argue that pragmatic theories can account for such dialogues only by models that combine linguistic principles of discourse coherence and cognitive models of practical rationality. Presented at SemDial 2010: 14th workshop on the semantics and pragmatics of dialogue
Year(s) Of Engagement Activity