Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Working memory and the maintenance of multiple inferences

Smith, Allison K. (2016) Working memory and the maintenance of multiple inferences. Undergraduate Thesis, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Primary Text

Download (238kB)

Abstract

The influence of working memory on facets of comprehension such as inferencing has been widely investigated. Working memory capacity for language (WMCL), a common measure of the central executive component of working memory, has been linked to inferencing and particularly to the maintenance or narrowing of inferences when multiple interpretations are generated. Conflicting conclusions are evident, however. Some work suggests that poor WMCL leads to overly narrowing inferences, attributing this performance to a lack of cognitive fuel to maintain simultaneous interpretations. Other work concludes that poor WMCL yields a difficulty in narrowing inferences to reach a single conclusion. And some data from the realm of syntax suggest that maintaining multiple options is a strength, available only to individuals with good WMCL. To further explore this issue, this experiment investigated inference narrowing in neurotypical adults with High vs. Low WMCL. In a thinking-outloud task, participants responded with whatever came to mind after hearing each sentence of 8-to-9 sentence narratives. Each response was coded as to whether it reflected a single inference, multiple inferences, or maintained inference.
This study also expanded the conceptual scope of prior investigations by assessing inference maintenance in relation to a newer component of the working memory system, the episodic buffer. The episodic buffer participates in combining and encoding the processes and outputs of other components of working memory and long-term memory. Its functioning was measured in terms of relative verbatim recall of a separate set of coherent versus scrambled stories.
One main result of this study was that the WMCL groups did not differ in the maintenance or narrowing of multiple inferences. Responses with multiple inferences declined for the Low WMCL group as the narratives progressed, but this result was not significant. Another major result was that the episodic buffer measure significantly predicted inference narrowing. Secondary analyses assessed potential differences between WMCL groups in inference revision and total inferences, neither of which were significant. Discussion centers around the limitations of relying solely on the central executive concept to predict inferencing outcomes, in light of its emphasis on the flexible allocation of cognitive resources, and suggests that a measure of episodic buffer functioning may be a better predictor, at least in some cases.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Smith, Allison K.aks77@pitt.eduAKS77
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Thesis AdvisorTompkins, Connie Atompkins@pitt.eduTOMPKINS
Committee MemberLundblom, Erinlunblom@pitt.eduLUNDBLOM
Committee MemberShaiman, Susanshaiman@pitt.eduSHAIMAN
Committee MemberHarris Wright, Heatherwrighth@ecu.edu
Date: 25 April 2016
Date Type: Publication
Defense Date: 30 March 2016
Approval Date: 25 April 2016
Submission Date: 11 April 2016
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 58
Institution: University of Pittsburgh
Schools and Programs: University Honors College
School of Health and Rehabilitation Sciences > Communication Science and Disorders
Degree: BPhil - Bachelor of Philosophy
Thesis Type: Undergraduate Thesis
Refereed: Yes
Uncontrolled Keywords: working memory, episodic buffer, inference
Date Deposited: 25 Apr 2016 16:44
Last Modified: 15 Nov 2016 14:32
URI: http://d-scholarship.pitt.edu/id/eprint/27605

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item