Comprehenders model the nature of noise in the environment
Name
nihms-1506288.pdf
Description
Accepted version
Size
975.88 KB
Format
Adobe PDF
Checksum (MD5)
60952a0d58de759671689764d6e2b3fd
Author(s)
Ryskin, Rachel A
Futrell, Richard
Kiran, Swathi
Gibson, Edward A.
Date Issued
September 2018
Journal
Cognition
Publisher
Elsevier BV
Citation
Ryskin, Rachel et al. "Comprehenders model the nature of noise in the environment." Cognition 181 (December 2018): 141-150 © 2018 Elsevier B.V.
Version
Author's final manuscript
Abstract
In everyday communication, speakers make errors and produce language in a noisy environment. Recent work suggests that comprehenders possess cognitive mechanisms for dealing with noise in the linguistic signal: a noisy-channel model. A key parameter of these models is the noise model: the comprehender's implicit model of how noise affects utterances before they are perceived. Here we examine this noise model in detail, asking whether comprehension behavior reflects a noise model that is adapted to context. We asked readers to correct sentences if they noticed errors, and manipulated context by including exposure sentences containing obvious deletions (A bystander was rescued by the fireman in the nick time.), insertions, exchanges, mixed errors, or no errors. On test sentences (The bat swung the player.), participants’ corrections differed depending on the exposure condition. The results demonstrate that participants model specific types of errors and make inferences about the intentions of the speaker accordingly. Keywords: Sentence comprehension; Noisy-channel; Rational inference; Adaptation; Error correction
Subjects
Linguistics and Language
Experimental and Cognitive Psychology
Cognitive Neuroscience
Developmental and Educational Psychology
Language and Linguistics
MIT Department
Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences
Terms of Use
Creative Commons Attribution-NonCommercial-NoDerivs License
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1016/j.cognition.2018.08.018
