Cantonese Speakers and the Acquisition of French Consonants

Abstract

This article reports the findings of research on the acquisition of French consonants by native speakers of Cantonese. An error analysis based on a careful phonetic transcription of production data resulted in a scale of difficulty of consonants in both initial and final positions. Several of the major patterns of difficulty are explained by the Markedness Differential Hypothesis. On the basis of errors found, and which are not perfectly predicted by this theory, it is shown that there is an interaction of language acquisition with markedness reversals.

Markedness Universals and the Acquisition of Voicing Contrasts by Korean Speakers of English

“…Also, Table 6 shows that the voiced allophonic variants already existing in Korean, Ibl, HI, Igl, and 1)1, did not cause relatively fewer errors than the other obstruents, Ivl, Izl', Izl, and Ibl, as might be expected, even though they do not exist in Korean, Izl and Ivl had relatively high p values. The fact that an explainable hierarchy of order according to phoneme was not found in this study is consistent with Cichocki et al (1993), who found that voiced final stops and fricatives were devoiced at similar rates by Cantonese speakers of French. These findings may suggest that the voicing feature has identical markedness relationships regardless of whether it concerns stops, fricatives, or affricates.…”

“…the word list there are some differences in stops, fricatives, and affricates, but they only varied between 33% and 41%. This is consistent with other L2 acquisition data in which no clear stop > fricative > affricate order of acquisition existed (Cichocki et al, 1993;Edge, 1991). For the text, however, the stops exhibited a significantly higher percentage of accuracy (83%) than that of the affricates (67%) or fricatives (60%).…”

“…In addition, Eckman (1984) found similar results in Farsi native speakers (NSs) learning English. To date, one of the most thorough L2 studies of voice contrasts as a function of word position (initial and final) is that by Cichocki, House, Kinloch, and Lister (1993), who studied Cantonese NSs' acquisition of French obstruents (stops and fricatives). Cantonese has only /p/, /t/, and /k/ in final position, whereas French has voiced and voiceless stops and fricatives in all positions.…”

Development and Recoverability of L2 Codas

“…For example, when syl-lables are being closed, voiced word-final obstruents typically become devoiced, and this simplification process normally remains in the system up to the age of 4 (Yavas, 1994, p. 274). Likewise, L2 learners do not go from epenthesis to entirely correct pronunciation, but many of the nonreduced or nonepenthesized forms that learners produce deviate from the target with respect to the exact quality of the coda, and obstruent devoicing has been shown to be one of the more prevalent processes in L2 acquisition (e.g., see Cichocki, House, Kinloch, & Lister, 1993;Eckman, 1981;Edge, 1991;Flege & Davidian, 1984;Major & Faudree, 1996;Weinberger, 1987;Yavas). Devoicing of the final consonant-which in fact preserves the underlying syllable template as well as most of the feature complex of the coda-results in less ambiguity than deletion but more than epenthesis.…”

Laryngeal contrasts in second language phonology

“…Greater accuracy in the production of final voiceless obstruents has been found not only for learners of English, but also for Cantonese-speaking learners of French (Cichocki et al 1993) and German-speaking learners of Swedish (Hammarberg 1990). While we cannot eliminate the native language as the (possibly sole) source of this asymmetry, it is important to note that final voicing is realized very differently in French and Swedish than in English.…”