Sociolinguistics of phonological interference in Youba-English

The research seeks to demonstrate not only that variation exists in L2 speech but also that such variation is explainable in terms of describable linguistic and nonlinguistic variables. This contradicts the assumption, implicit in most existing contrastive analyses, that L2 speech of speakers with the same L1 background is homogenous and monolectal. In the present work, predictions of L1 interference based on a contrastive analysis of the phonological systems of English and Yoruba were verified from an analysis of actual English speech samples of Yoruba immigrants in England. Using a quantitative model analysis, frequencies of occurrence of target and nontarget language forms were then interpreted in relation to certain linguistic and extralinguistic problems. In the first, while the predictive-power rating of the contrastive analysis (on the Transformational Generative model) was very high the informants' L2 speech was not homogenous either at the intra- or inter- personal levels. Three types of phonological rules were identified with respect to rule variability. Finally, it was observed that the so called dynamic paradigm proved, in practice, an extension of the quantitative method. In the second, though none of the independent variables (sex, education, phonetic training and length of sojourn in England) reached significance in its effects on the informants' speech at one percent at least one was very close. Since the present work was limited only to the investigation of segmental phonology, it is suggested that similar investigations of the other linguistic levels are needed to fully evaluate language teaching programmes such as the 'Year-abroad' and 'Speech-training' schemes run for second- and foreign-language students in many educational systems world-wide.