Retrieval dynamics and interference in anaphoric resolution: Skilled and unskilled readers

Haskins Laboratories, Inc., New Haven CT

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): According to the 2003 report on Adult Literacy and Life Skills (Statistics Canada & OECD, 2005) an estimated 51% of US adults aged 16-25 can read only at a below-basic level, which is insufficient for attaining advanced educational and occupational goals. Because, poor reading ability has tremendous negative personal, social, and economic consequences for individuals, it is an issue of real public health concern. The origins of reading difficulty are many and varied, but recently memory retrieval interference during incremental sentence processing has been identified as a source of comprehension difficulty (Van Dyke & Johns 2012). My project will investigate how retrieval interference during the construction of pronoun-antecedent dependencies contributes to poor comprehension. Pronominal dependencies are pervasive in natural language texts (between 25%-35% of all reference in language uses pronouns - Hardie 2007), and their successful resolution is crucial in constructing coherent and correct interpretations. Pronominal dependencies are subject to a number of structural constraints which restrict the positions of their antecedents. Though relational constraints play an indispensable role in theoretical descriptions of linguistic behavio, there has been little work to link the theoretical understanding of the distribution of pronouns with mechanisms of retrieval during incremental sentence comprehension. The current proposal focuses on investigating how effective individual readers are at resisting interference from structurally inappropriate but feature-matching NPs during antecedent retrieval. My Specific Aim 1 is to identify the mechanism that comprehenders employ to retrieve antecedents from memory. To date, no research has directly addressed this question, and the current proposal seeks to fill in this gap. The experiments will contrast two candidate mechanisms: a direct-access retrieval mechanism and a serial- search mechanism (e.g. Sternberg 1966) using the Speed-Accuracy Tradeoff paradigm. In my dissertation I proposed that skilled comprehenders are capable of employing strategies to resist retrieval interference. Specific Aim 2 will systematically investigate structural constraints of differing complexity to identify conditions under which retrieval interference occurs. Finally, Specific Aim 3 of the proposal investigates the extent to which poor and inhibitory control abilities are significant predictors of retrieval interference across all of the experiments proposed. In order to have broad impact, the studies will target a community-based sample comprised of both non-college-bound individuals (age 16-24) and college students of the same age. This broad population constitutes a more representative sample of reading ability in the population at large. I anticipate that the project will result in a new understanding of a significant source of individual variability in reading comprehension, which will perhaps inform clinical remediation methods.