The Effects of Pronunciation Instruction on L2 Production and L2 Perception in Spanish: A Comparative Analysis

Heather M Offerman, Purdue University

Abstract

Having historically received less attention than other linguistic structures (Derwing & Munro, 2005), second language (L2) pronunciation instruction represents an emergent area of research in the field (Thomson & Derwing, 2015). While several methods have been shown to be effective for improving L2 segmental production, including explicit instruction (Aliaga-García & Mora, 2009; Lord, 2010; Saito & Lyster, 2012) and inductive visual feedback instruction (Offerman & Olson, 2016; Olson, 2014b; Olson, 2019), there is a notable lack of empirically-based research comparing approaches (Derwing & Munro, 2015; Lee et al., 2015). Moreover, research has largely ignored the effects of instruction on L2 perception, due in part to the tacit assumption that L2 perception precedes L2 production (Levy & Law, 2010). Responding to these gaps, this study provides a large-scale comparative analysis of three types of pronunciation instruction (explicit instruction [EI], visual feedback [VF], and a combination instruction [CI] approach) on L2 segmental production and perception in Spanish. Production-oriented analyses focus on the change in voice onset time (VOT), and perceptual analyses focus on an L2 discrimination task (AXB task) and a nativenessjudgment task (Liker-scale rating task). Differences in VOT for the stops /p,t,k/ in wordinitial position exist for English (long-lag VOT) and Spanish (short-lag VOT) (Lisker & Abramson, 1964), causing notable accentedness for English-speaking L2 learners of Spanish (Lord, 2005), thus serving as the basis for L2 learner performance. Considering results for the production portion of the study, all three experimental groups were found to outperform the control (CO) group, along with each experimental group significantly improving individually from pretest to posttest. For the perception portion, participants did not display difficultly in discriminating between long-lag and short-lag productions at the pretest, and as such, showed no improvement following instruction. In contrast, results from the nativeness judgment task showed that participants were not able to categorize sounds as native-like (Spanish) or non-native-like (English), and significant improvement following training was found only for the CI group. Additionally, previous L2 perception theories largely focus on category formation and discrimination of sounds, such as the SLM (Flege, 1987), the PAM (Best, 1994), and the PAM-L2 (Best & Tyler, 2007), while this study considers perception as it applies to both discrimination and the social categorization of sounds in the L1 and L2. For the production and perception portions, the CI group largely outperformed all groups. It is proposed that the combination of two different modalities, auditory and visual (Baran-Łucarz, 2012), provides learners more resources for noticing (Schmidt, 1990) differences between their L1 and L2 for production and perception purposes. Moreover, the CI treatment is proposed to be most beneficial for teaching L2 pronunciation, although more research is to be done with comparative analyses for different segments in the L2, as it has been previously noted that not all pronunciation treatments are equally beneficial to L2 learners for different segmental features (Ruellot, 2011). This study adds to theoretical understanding of L2 phonetic acquisition, in both production and perception, while empirically testing pedagogical approaches in a classroom setting.

Degree

Ph.D.

Advisors

Olson, Purdue University.

Subject Area

Linguistics|Pedagogy|Language|Education|Foreign language education|Instructional Design

COinS