This proposal uses new and original data from Picard, a little-known regional language of France, in an effort to develop a "socially realistic linguistics" (Wilson & Henry 1998). Picard is a Gallo-Romance language spoken in northern France and southern Belgium. Like French, it is directly descended from Latin. Throughout the Middle Ages, Picard and the other Gallo-Romance languages thrived in their respective regions of France. However, as a result of the concentration of political and economic power in Paris and the spread of education available solely in French, speakers have been abandoning their traditional regional languages as their everyday language. Currently, in Picardie, French is viewed as the language of social advancement, and most parents no longer speak Picard to their children. As a result, few speakers of Picard remain, and this may very well be our last chance to study Picard as a living language.
For this study, Vimeu, a region located in the western part of the Somme department of France, was selected as a focus of investigation because of the vitality which still characterizes Picard there. Over the past four years, a corpus of oral and written data was gathered. This corpus consists of sociolinguistic interviews, informal conversations, readings, museum guided tours, and cultural events, all in Picard or mostly in Picard, as well as a large quantity of books and magazines. It is probably the largest corpus of Picard available anywhere in the world. However, in order to make it truly accessible and useful for linguists and other researchers interested in Picard and to provide data for detailed analyses of the grammatical structure of this language, the corpus must be transcribed. Transcription of the corpus thus constitutes the first objective of this research proposal. The second objective consists in providing in-depth analyses of many grammatical constructions of Picard. These analyses will be carried out using an approach which combines a sociolinguistic methodology and the analytical tools provided by generative grammar. This approach is expected to shed new light on many longstanding issues within Romance and general linguistics (e.g., pronominal clitics, subject doubling), contribute crucial new data from language change in progress which will allow us to better understand well-known grammaticalization processes (e.g., how demonstrative determiners compete with definite articles and eventually replace them), and address new questions concerning constructions little known or unknown in the literature (e.g., why does Vimeu Picard have two subject neuter pronouns?). The third and fourth objectives pursued in this research proposal concern variation, both crosslinguistic and language-internal. Comparison of Vimeu Picard features with those of other varieties of Picard as well as other Gallo-Romance languages and varieties of French is expected to contribute significantly to the search for a better understanding of Universal Grammar, the limits it imposes on crosslinguistic variation, and how such variation functions. Finally, the issue of language-internal variation will be considered. I will combine recent developments in two different areas of Optimality theory, namely the study of phonological variation and that of morphosyntax, and will propose a model of grammar which generates variable morphosyntactic patterns and predicts relative frequencies of use for linguistic variants without having to assign probabilistic weights to linguistic constraints. This project will thus contribute to bridge the gap that currently exists between sociolinguistics and formal approaches to the study of language and "advance our understanding of both the underlying system of language available to all human beings, and the way in which this system becomes worked out in the social contexts of everyday life" (Wilson & Henry 1998:2-3).