This study continues a series of reports on the work of the team which has carried out a sociolinguistic survey of multilingual communities. This study deals with an early sample of the results of the St. Lucian survey, and in particular with the extent to which they provide support for the theoretical model of linguistic choice and change, and rebut the linear, implicational-scale model of Bickerton (1975). The center of interest in this study has been the child in a multilingual community. The paper brings together data of three kinds: (1) from the social, demographic and economic history of St. Lucia; (2) from the linguistic behavior of 100 St. Lucian children; and (3) from social and economic observations on the family circumstances of those children. It was hypothesized that a child's linguistic behavior is symptomatic of his desire to identify with a particular group, in this case, speakers of Standard English, speakers of Bajanised English, or speakers of patois. At this time there seems to be no move towards a common acrolect in St. Lucia, but undoubtedly a local variety of English will emerge, and re-creolization will be seen to have played its part. (AM)