2020 LLL Conference Featured Talk: "The power of prediction in language processing and learning"

dc.contributor.authorGrüter, Theres
dc.date.accessioned2020-05-19T04:02:09Z
dc.date.available2020-05-19T04:02:09Z
dc.date.issued2020-05-18
dc.descriptionNOTE: We apologize but the first few minutes of this talk are missing from the recording.This featured talk was presented by Dr. Theres Grüter during the 24th annual College of Languages, Linguistics & Literature Graduate Student Conference on April 18, 2020. Dr. Grüter was the 2020 Junior/Mid-level Faculty LLL Excellence in Research award winner and was invited to give this talk.
dc.description.abstractTrying to anticipate, or predict, what will happen next is part of everyday human behavior, including language use. When a prediction turns out to be wrong, we are presented with an opportunity to learn and make better predictions next time. This constitutes learning through prediction error, a powerful mechanism of implicit (and explicit) learning. In this talk, I will begin by demonstrating that the timing of everyday communication between adult native speakers cannot be explained without appeal to prediction during real-time language comprehension and production. I will then discuss recent, and partially conflicting, evidence from research that has investigated the extent to which language users other than adult native speakers, such as second language (L2) users, engage in predictive processing. I will present findings from a visual-world eye-tracking study with native and L2 users of Mandarin Chinese, which suggest that both groups used information encoded by prenominal classifiers (measure words) predictively in real-time comprehension, but while native users drew primarily on formal grammatical cues, L2 users placed more weight on semantic information (Grüter, Lau, & Ling, 2020). In the second part of the talk, I will present findings from a structural priming experiment with Korean-speaking learners of English (Grüter, Zhu, & Jackson, in preparation). Structural priming refers to our tendency to use linguistic structures that we recently encountered in the input. For example, hearing an interlocutor say She gave her friend a cupcake (=double-object construction) will increase the chances that you will then also use a double-object (rather than a prepositional dative) construction to describe a different giving-event. The goal of our study was to test whether Korean learners of English, who are known to underuse double-object constructions in English, would increase their use of this construction when consistently seeing such sentences written by a (virtual) partner. We hypothesized that this increase would be greatest when participants are forced to first guess, or predict, what their partner would say, thus presenting them with opportunities to experience prediction error and adjust their expectations. Our findings support this hypothesis, suggesting that learning through prediction error, even at an explicit level, has the potential to contribute to L2 learning.
dc.description.sponsorshipCollege of Languages, Linguistics & Literature, University of Hawai‘i at Mānoa
dc.format.extent41 minutes and 4 seconds
dc.identifier.urihttp://hdl.handle.net/10125/68011
dc.language.isoen-US
dc.rightsCreative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
dc.rightsAttribution-NonCommercial-ShareAlike 3.0 United States*
dc.rights.urihttp://creativecommons.org/licenses/by-nc-sa/3.0/us/*
dc.subjectprediction
dc.subjectlanguage learning
dc.subjectlanguage processing
dc.subjecteye-tracking
dc.subjectstructural priming
dc.subject.lcshSecond language acquisition
dc.title2020 LLL Conference Featured Talk: "The power of prediction in language processing and learning"
dc.typeVideo
dc.type.dcmiMovingImage

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
2020_Featured_Talk_Dr_Theres_Grüter.mp4
Size:
1020.26 MB
Format:
Description:
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.73 KB
Format:
Item-specific license agreed upon to submission
Description: