The relationship between speech perception and word learning at the initial state of second language acquisition

  • Samra Alispahic

Western Sydney University thesis: Master's thesis

Abstract

The goal of most adult second language (L2) learners is to confidently and efficiently communicate in their target L2. However, this task is not easy. In order to produce new L2 words a learner first needs to perceive the sounds that comprise these words. Considered effortless in one's native language (L1), distinguishing novel L2 phonemes can be quite difficult for adult learners, and difficulties in L2 speech perception are often attributed to the negative transfer effects of the L1. Research suggests that the size of the L2 vowel inventory relative to the L1 inventory may affect the discrimination and acquisition of L2 vowels. Specifically, if the L1 has a smaller L1 vowel inventory than the L2 this may obstruct L2 vowel perception, while if the L1 has a larger vowel inventory it often facilitates vowel perception. However, the Second Language Linguistic Perception (L2LP) model specifies that it is the L1-L2 acoustic relationships that predict L2 vowel perception, regardless of L1 vowel inventory size. The model further posits that there is continuity between L2 perception and L2 word learning (Escudero, 2005, 2006, 2009; van Leussen & Escudero, 2015). In this view, sounds that are difficult to perceive would yield comparable difficulty in learning L2 words containing the same sounds. The present thesis investigates Australian English (AusE), Peruvian Spanish (PS) and Spanish adult listeners' non-native vowel perception and word learning of Dutch and the interrelation between these two abilities. In order to establish the initial state of learning the present thesis examines naà ve listeners (i.e., AusE, PS and Spanish monolinguals) with no prior knowledge of Dutch. This thesis comprises two studies, as well as an Introduction and General Discussion. Study 1 investigates the effects of vowel inventory size versus acoustic properties on non-native vowel perception. It compares XAB discrimination and categorization of five Dutch vowel contrasts between monolinguals whose L1 contains more (AusE) or fewer (PS) vowels than Dutch. It also examines whether cross-language discriminant analyses predict listeners' categorization patterns and whether these in turn predict their discrimination ability. Study 2 investigates the interrelation between listeners' initial perceptual and word learning abilities by comparing two sets of previously published data. AusE and PS listeners' XAB discrimination of five Dutch vowel contrasts, which was presented in Study 1, is compared to AusE and Spanish non-native word learning of minimal pairs containing the same vowel contrasts, which was previously reported in Escudero (2015). This comparison between perception and word learning data sets is done quantitatively, using the same statistical models, and also qualitatively. Results show that the size of participants' native vowel inventories did not affect performance in either task. Rather it is the L1-L2 acoustic relationships that predicted listeners' non-native categorisation and discrimination patterns and these in turn predicted their word learning difficulties. Specifically, minimal pair words containing perceptually difficult vowel contrasts were more difficult to discriminate, while word pairs containing perceptually easy contrasts were easier. Results presented in both studies confirm the L2LP proposal that there is continuity between perception and recognition in L2 development.
Date of Award2017
Original languageEnglish

Keywords

  • second language acquisition
  • speech perception
  • Dutch language
  • study and teaching
  • foreign speakers

Cite this

'