L2 acquisition of [[??]], [[??]], and [[??]] in Spanish: impact of experience, linguistic environment, and learner variables.

ABSTRACT. This study examines the influence of L2 experience, learner variables, prosody, and orthography on the L2 acquisition of the approximants [[??]], [[??]], and [[??]] in Spanish by 36 English-speaking learners. The results suggest that accuracy in the production of [[??]], [[??]] and [[??]] was significantly related to the following factors: age of first exposure to Spanish, amount of formal instruction in Spanish, residence in a Spanish-speaking country, amount of out-of-class contact with Spanish, and concern with pronunciation.

Gender was not statistically related to accuracy. In addition, the results suggest that participants were significantly more accurate in producing [[??], [[??]], and [[??]] in unstressed syllables, compared to stressed syllables.

1. INTRODUCTION. One aspect of acquiring native-like pronunciation in a second language (L2) involves attaining control over allophonic processes in the L2. For learners of Spanish, this includes learning the patterns of allophonic variation between the voiced approximants [[??]], [[??]], and [[??]] (1) and the voiced stops [b], [[??]], and [g]. Previous research concerning the L2 acquisition of this contrast (2) in Spanish suggests that English-speaking late learners of Spanish in the lower and intermediate proficiency levels do not produce the approximants with target-like accuracy, but rather, frequently produce voiced stops in those environments in which Spanish native speakers produce approximants (Diaz-Campos 2004, 2006; Elliott 1997; Zampini 1994, 1998).

That the voiced approximants are difficult to learn is attested to by studies in which L2 learners have been found to make little or no improvement over time in the accurate production of these segments. Only minor improvements were observed in L2 learners who received focused pronunciation training in the classroom (Elliott 1997) or who spent one semester studying abroad in Spain (Diaz-Campos 2004, 2006). Furthermore, exposure to Spanish in foreign language classes over the course of two semesters resulted in only minor improvements in learners' approximant production (Zampini 1994, 1997, 1998). However, what is missing from previous research on this topic is an analysis of L2 learning over longer periods of time. Existing studies have primarily focused on one- to two-semester periods of learning, which may not be enough time to detect improvements in this area of pronunciation. In order to fill this gap, the present study compares L2 learners with greater differences in amount of experience learning Spanish than in previous studies and analyzes the impact on pronunciation accuracy of variables such as the amount of formal L2 instruction, the amount of out-of-class L2 contact, and the length of residency in an L2-speaking country.

In addition, this study considers two factors that have received little attention in the literature on the L2 acquisition of Spanish pronunciation: how accurate production of segments may be influenced by both prosody and orthography. In the first case, research with native speakers of Spanish indicates that the degree of closure and frication in the production of the approximants [[??]], [[??]], and [[??]] may be affected by word- and phrasal-stress placement (Cole et al 1999). Previous work also indicates that prosodic context impacts L2 learners' production accuracy with these segments (Zampini 1997, 1998), but the specific issue of stress has not been investigated. Secondly, research (Bassetti 2006, Detey and Nespoulous 2008, Young-Scholten 2002) has pointed to the influence of orthography on L2 production and perception, a factor that may also be relevant for the accurate production of stops and approximants in Spanish.

The present study examines the production of [[??]], [[??]], and [[??]] in the speech of 36 English-speaking late learners of Spanish in order to investigate the relationship between the production accuracy of these segments and L2 experience, prosodic context, and orthographic input. This analysis begins with an overview of the relevant literature, continues with a description of the methodology, and finally presents the results, discussion and conclusions.

2. REVIEW OF THE LITERATURE. Producing the stops [b], [d], and [g] in phonological environments in which native speakers of Spanish produce the approximant allophones [[??]], [[??]], and [[??]] contributes to the foreign accent observed in English-speaking L2 learners of Spanish. In Spanish, the voiced stops and approximants are position-sensitive allophones. The approximants occur in all phonetic contexts except in absolute utterance-initial position and after nasals. In addition, the stop [d] always occurs after the lateral /l/. Although several authors (Amastae 1986, Carreira 1998, Lipski 1994, Pineros 2002) point out that considerable dialectal variation exists in Spanish with regard to the distribution of stops and approximants, the one context that most authors generally agree does not exhibit dialect variation is the intervocalic position (i.e. V_V), in which approximants are universally produced (3) (Carreira 1998, Zampini 1998).