A Statistical Modeling of the Correlation between Island Effects and Working-memory Capacity for L2 Learners

A Statistical Modeling of the Correlation between Island Effects and Working-memory Capacity for L2 Learners Euhee Kim Myung-Kwan Park Computer Science & Engineering Department of English School of IT Convergence Engineering College of Humanities Shinhan University Dongducheon City, Korea Dongguk University Seoul, Korea euhkim@shihan.ac.kr parkmk@dgu.edu Abstract The cause of island effects has evoked considerable debate within syntax and other fields of linguistics. The two competing approaches stand out: the grammatical analysis; and the working-memory (WM)-based processing analysis. In this paper we report three experiments designed to test one of the premises of the WM-based processing analysis: that the strength of island effects should vary as a function of individual differences in WM capacity. The results show that island effects present even for L2 learners are more likely attributed to grammatical constraints than to limited processing resources. 1 Introduction The role of memory in language learning has long received ample attention from researchers in first and second language acquisition (SLA) (Baddeley (1999), Ellis (2001), Juffs (2006)). At an intuitive level, it seems right to reason that individual differences among adult learners in their successful attainment of a second language (L2) are attributable to individual differences in memory capacity. In SLA, researchers have focused on short-term or working rather than long-term memory differences because they think short-term or working-memory (WM) plays a more instrumental role for individual differences in language development. The rationale for this belief is that WM is an on-line capacity for processing and analyzing new information (words, grammatical structures and so on). As a consequence, the bigger the on-line capacity an individual has for new information, the more information will settle into off-line, long-term memory. In this paper we concentrate on Korean leaners of English (KLEs) to examine the correlation between their individual WM capacity and their knowledge of island constraints on wh- dependencies in English. To this end we adopt the methodology that Sprouse, Wagers, and Phillips (SWP) (2012a, b) use for L1 speakers. 2 Hypothesis Testing The main focus of this paper is to examine the question of whether there is a correlation between KLEs' WM capacity and their knowledge of island constraints on wh-dependencies in English. In order to investigate this question, we need (i) a measure of WM capacity, and (ii) a measure of knowledge of wh-island constraints. The second measure is often termed a measure of „island effects‟, which refer to the relatively low acceptability ratings given to sentences with a wh- dependency between a wh-phrase and its gap position inside select syntactic environments (cf. Ross (1967), Rizzi (1990), and Chomsky (1995) among many others). Given the foremost interest in the role of such variables as GAP-POSITION (i.e. where a gap is) and STRUCTURE (i.e. PACLIC 29 422 29th Pacific Asia Conference on Language, Information and Computation pages 422 - 430 Shanghai, China, October 30 - November 1, 2015 Copyright 2015 by Euhee Kim and Myungkwan Park whether island structure is involved or not) in the instantiation of island effects, we want to bring …

Content truncated to 500 words