Summary: | This dissertation is a sociolinguistic study examining the phonetic variation in the speech of four generations of Cajun French women, living in four Louisiana parishes: Avoyelles, Lafourche, St Landry and Vermillon. Based on the Gold 1975 corpus and the Dubois 1997 Cajun French corpus available at LSU, a sample of 29 speakers was chosen for the analysis of five traditional phonetic variables: O in front of nasals MM/NN and liquids R and L, in front of R, E in front of R and at the third person ending of the imparfait, J in lexical words, personal pronoun JE, and Z in liaison structures like nous-autres.
Following the quantitative methodology, a total of over 20 000 tokens were codified and analyzed with analytical softwares such as Statview 4.5 and Goldvarb to determine the direction of the variation observed. After the first frequency analysis, we found an apparent rise in the use of the dialectal features in the younger generations. Because Cajun French is a minority language in an endangered situation due to a lack of constitutional support and a dramatic decreasing number of speakers, it appeared rather impossible that the tendency observed could be a true language change. We decided to further our analysis by comparing the two interviews in French available in the Dubois 1997 corpus for each speaker. One interview was lead by a native Cajun French speaker and the other one by a student speaking academic French. This comparison aims to measure the degree of adaptation to a specific linguistic situation by Cajun French speakers.
The results of the comparative study shows that the rate of dialectal features used by each speaker significantly drops when they speak to the outsider interviewer. This proves that what we observe is not a language change but rather the fact that Cajun women in our sample have maintained their ability to adapt stylistically to the variety of French being spoken to them. This goes against the theory of linguistic shrinkage, stating that when a language is dying, the speakers lose their ability to detect and adjust to different ranges of styles.
|