The Use of Gender Information in Second Language Pronoun Resolution The Role of L1, Working Memory, and Distance

Abstract

Abstract

Aims and Objectives: This thesis investigates how the existence or absence of gender distinctions in pronouns in the first language (L1) affects the processing of English 3rd person singular pronouns, and to what extent working memory (WM) and the distance between the antecedent and the pronoun affect this process.

Methodology: In this study, L2 learners of English whose L1 has gender distinction in pronouns (e.g., Spanish) and participants whose L1 does not have such a distinction (e.g., Chinese) were recruited. They completed a self-paced reading task in which the pronoun and the antecedent were manipulated as congruent or incongruent regarding gender. The distance between the antecedent and the pronoun was also manipulated, with either short, long, or complex distance. The study also assessed participants on the WM with an N-back task.

Data and analysis: Data were collected using the Gorilla Experiment Builder (gorilla.sc) (Anwyl-Irvine, Massonnié, Flitton, Kirkham, & Evershed, 2020) and analyzed using R (version 4.4.0) with RStudio. Statistical tests were the Chi-squared test for English proficiency and WM, Welch’s t-test for statement accuracy, ANOVA for initial tests, and linear mixed-effects models for the interactions between variables. The reaction time (RT) serves as a dependent variable. Predictors include participants’ L1s (with vs. without gender distinction), distances between the antecedent and the pronoun, congruence between the antecedent and the pronoun, participants’ WM, and their English (L2) Proficiency.

Findings and conclusion: Our results did not show a significant mismatch effect between L1 groups with different backgrounds, which did not support the Competition Model (Bates & MacWhinney 1982,1989) nor the Feature Reassembly Hypothesis (FRH) (Lardiere, 2008; 2009a; 2009b). Also, our results did not support the idea that participants with better WM detected the incongruence faster or better. However, our observed results that the two L1 groups show statistically significant differences in long and short distance conditions but not in complex conditions support the Shallow Structure Hypothesis (SSH) (Clahsen & Felser, 2006).

Keywords: Cross-Linguistic Influence; Second Language Processing; Working Memory; Pronoun Gender Resolution; Shallow Structure Hypothesis; The Competition Model; The Feature Reassembly Hypothesis.