Menu Content/Inhalt
SFB 673 arrow Events
CRC Colloquium
show month-view show week-view   
Context Effects in Language Production: Models of Syntactic Priming in Dialogue Corpora
Montag, 30.06.2008 16:00 - 18:00
Context Effects in Language Production: Models of Syntactic Priming in Dialogue Corpora

David Reitter
University of Edinburgh


Successful communication depends on the alignment of the
interlocutors, that is, they need to be "on the same picture".
Pickering & Garrod (2004) have suggested that alignment is based on a
cascade of mechanistic repetition effects. In this talk I describe
experiments seeking confirmation of this hypothesis by looking at
repetition in dialogue corpora and whether repetition is correlated
with task success, as well as a cognitive model that shows how priming
emerges.

I develop two metrics for syntactic adaptation within a speaker and
between speakers in dialogue: one for short-term priming effects that
decay quickly, and one for long-term adaptation over the course of a
dialogue. Both methods estimate adaptation in large data-sets
consisting of transcribed human-human dialogue annotated with
syntactic information. Two such corpora in English are used:
Switchboard, a collection of spontaneous phone conversation, and HCRC
Map Task, a set of task-oriented dialogues in which participants
describe routes on maps to one another.

Map Task provides a measure of task success through the deviance of
the communicated routes on the maps. I will present results
describing whether there is a link between task success and
adaptation. Further evidence points to different cognitive bases for
short- and long-term effects. The data are consistent with the
hypothesis that semantic activity as the cause of short-term syntactic
priming.

A model of language production combines syntactic (Combinatory
Categorial Grammar, Steedman 2000) and general cognitive frameworks
(ACT-R, Anderson 2004). The model operationalizes syntactic decision-
making and provides a unified explanation of short- and long-term
priming effects. With the model, I argue that short- and long-term
adaptation results from a combination of general (base-level) learning
and cue-based memory retrieval (spreading activation).

http://www.david-reitter.com/
Contact
contact-person: Gerhard Jäger
homepage: wwwhomes.uni-bielefeld.de/gja