Atrito linguístico na produção vocálica de imigrantes brasileiros bilíngues (português-inglês) residentes em Londres.

Abstract

This thesis, grounded in the Complex Dynamic Systems Theory (CDST), aims to investigate First Language (L1) attrition, referring to changes in L1 resulting from the development of an Second Language (L2). More specifically, this work examines the production of the high back vowel /u/ and the mid-low vowel /ɔ/ in Brazilian Portuguese (BP-L1) by Brazilians residing in London, England. We operate under the assumption from CDST that bilinguals possess a single phonetic-phonological system and that the interaction and inclusion of new sounds or new acoustic cues will lead to alterations in the entire bilingual phonological system, including L1. Thus, it was expected that the high vowel /u/ in PB-L1 would undergo changes in contact with the English language, which includes two high back vowels in its phonetic-phonological inventory, the long vowel /uː/ and the short /ʊ/, i.e., a language that employs vowel duration (long vs. short) as a phonetic-phonological cue, as well as intensity, posteriority and height. Data (N=36) from Brazilian immigrants, English monolinguals, and BP monolinguals, extracted from Kupske's corpus (2016), were acoustically analyzed. Non-normalized and normalized F1 and F2 values, as well as absolute and relative duration values, were computed. Concerning the vowel development in English, the analysis reveals significant differences in both F1 (p<0.01) and F2 (p<0.001) between immigrants and English natives in the production of /uː/. The study also demonstrates significant differences in the duration of /uː/ (<0.001) between immigrants (167ms, SD=54) and English monolinguals (134 ms, SD=25). The same holds for the short vowel /ʊ/ (p<0.05), indicating that Brazilians do not produce English high vowels with values expected for monolinguals. In terms of differences related to the production of /uː/ and /ʊ/ by Brazilians, this study suggests that these vowels are produced differently in terms of F2 (p<0.01) and duration (p<0.001). This leads us to conclude that, although Brazilians exhibit different values from monolinguals for /uː/ and /ʊ/, they have already acquired the "duration" acoustic cue since they can differentiate between the long and short vowels in their production. Regarding PB-L1 sounds, the study reveals that the F1 and F2 values of /u/ in PB-L1 for immigrants are higher and exhibit greater variability when compared to values found for monolinguals. The analysis indicates significant differences in both F1 (p<0.001) and F2 (p<0.001) in the production of /u/ between Brazilian immigrants and BP monolinguals, indicating language attrition for this front vowel in PB-L1. Additionally, the analysis demonstrates statistical differences in the duration of the vowel /u/ between monolinguals and immigrants (p<0.01), with immigrants showing higher duration values (141ms, SD=26) compared to monolinguals (102ms, SD=18). For the mid-low vowel /ɔ/ in PB-L1, immigrants and monolinguals do not exhibit statistical differences in F2 values (p>0.05), but they do differ in F1 (p<0.05) and vowel duration, with immigrants having slightly longer durations (144ms, SD=31) compared to monolinguals (138ms, SD=30). Similar to /u/, immigrants do not produce the mid-low vowel like BP monolinguals. In this sense, attrition of PB-L1 was evident for all analyzed vowels, as expected, given that, according to CDST, environmental changes are among the multiple elements that can lead to changes in the phonological systems of L2.