The Perception and Production of Prominence in Spanish by Heritage Speakers and L2 Learners

摘要

Heritage speakers (i.e., descendants of immigrants that speak an ethnic minority language in a society where a different language is spoken as the majority language) are linguistically a unique population, because, while their early and immersed exposure to the heritage language puts them in an advantageous position with regard to their linguistic knowledge of the language, compared to those who learned it as adults, lack of intergenerational transmission is evident in this population, as demonstrated in the shift to the majority language of the society. Recently, the attempt to examine heritage speakers' linguistic knowledge from an acquisitional point of view has increased significantly, which has led to the creation of "heritage language acquisition" as an independent field of study. However, compared to other linguistic subfields, such as syntax and morphology, phonology has been an understudied area in heritage language research. Although phonology is the field in which heritage speakers have the most noticeable advantage over second language (L2) learners, their speech is nonetheless often perceived as accented by monolingual native speakers, which makes their speech unique in its own right. Thus, research on heritage language phonology is needed to gain a comprehensive understanding of heritage speakers' linguistic system.;This study examines U.S. Spanish heritage speakers' perception and production of two types of prosodic prominence in Spanish: word-level prominence (i.e., lexical stress), marking paradigmatic contrast (e.g., Canto. 'I sing.' vs. Canto. 'He/She/You (formal) sang.'), and sentence-level prominence (i.e., nuclear stress), marking information status and focus (e.g., ¿Que hizo Mariana? - Mariana canto. 'What did Mariana do? - Mariana sang.' vs. ¿Quien canto? - Canto Mariana. 'Who sang? - Mariana sang.'). Although lexical stress and nuclear stress exist in both Spanish and English, due to cross-linguistic differences between the two languages, English L2 learners are found to have great difficulties acquiring them in Spanish. The overarching goal of this study is to examine whether transfer from English (i.e., the majority language) is also observed in heritage speakers.;Spanish monolingual native speakers, Spanish heritage speakers, and English L2 learners of Spanish participated in four experimental studies: two forced-choice identification tasks for the perception of lexical stress and nuclear stress, a reading aloud task for the production of lexical stress, and a simulated interactive elicitation task for the production of nuclear stress. Results showed that, while the heritage speakers performed similarly to the monolingual speakers in the perception of lexical stress and nuclear stress, they showed a deviant pattern in their production, such as early alignment of f0 peak and elongation of unstressed final vowels in the production of paroxytones (lexical stress), and early alignment of f0 peak in the production of focused constituents (nuclear stress). Heritage speakers' discrepancy between perception and production is likely to be due to asymmetry in their use of Spanish. That is, heritage speakers speak Spanish much less frequently than they hear it. The L2 learners, on the other hand, showed divergent patterns from the monolingual speakers in both the perception and the perception. This suggests that, thanks to early exposure to the heritage language, heritage speakers have an advantage over L2 learners in their perception of heritage language speech sounds, but such an advantage is not guaranteed in their production. As nuclear stress is higher in the stress hierarchy and acquired later in life than lexical stress, the present study predicted that heritage speakers' use of nuclear stress would be affected to transfer from English to a larger degree than lexical stress. However, the nuclear stress results bore unexpected results that are not necessarily phonological in nature (e.g., bias toward focus on subject, use of cleft constructions to mark focus), making it difficult to make direct comparisons between the two linguistic features. Possible explanations to the unexpected results and suggestions for future research are presented.