CLS Speaker Series - David Reitter (Penn State University) Syntactic Priming: Why it Exists, and How it Helps Dialogue
Mar 18, 2016
from 09:00 AM to 10:30 AM
|Where||127 Moore Building|
|Add event to calendar||
In this talk, I will discuss corpus-based, "big-data" methods to study a psycholinguistic process in naturalistic dialogue: syntactic priming. The data from corpora such as Penn TreeBank and Map Task motivate a cognitive model of priming in language production. This model, in ACT-R, explains syntactic choice as a declarative memory retrieval (Reitter, Keller, & Moore, 2011).
Syntactic priming (Bock, 1986) is of interest as it reveals syntactic processing, and also because it has been claimed to form the basis of interactive alignment (Pickering & Garrod, 2004). The theory posits that speakers mutually adapt to their linguistic choices, reaching a more efficient common language.
I will discuss some key questions surrounding interactive alignment: whether priming is a social signal rather than just a mechanistic effect, and whether divergence effects found by Healey, Purver, & Howes (2014) are truly an argument against Interactive Alignment. The ACT-R explains these effects, and new analyses of the large-scale Reddit dataset support these viewpoints empirically.