M.A. Thesis: Theoretical implications of schwa deletion in French

Abstract: The phonological status and behaviour of schwa ('e-caduc') constitues one of the most widely investigated topics in the synchronic study of French phonology. A considerable amount of research is directed towards the attempt to account for the phonetic level alternatin of schwa with zero in certain environments. Proposed explanations vary considerably depending on the theoretical approach taken in the treatment of the synchronc data. Little attention has been given, however, to the study of the associated diachronic phenomena which constitute a body of substantive evidence from which valuable insights can be drawn for an analysis of the synchronic state. This thesis presents a description of the historical development of schwa from the Early Old French period up to the present day. This long-range perspective allows the examination of the various historical processes involving schwa deletion in word-final, interconsonantal and hiatus contexts. Their effects on the organizaton of the grammar (e.g. lexical and phonotactic restructuring) are discussed as well as their evolution into Modern French. The diachronic survey serves as an informative background for a reanalysis of the structure and functioning of the synchronic schwa deletion and insertion rules formulated in Dell (1973). Modifications are made in the optional sycope rules found there in order to more adequately express their variable nature. A different scope of application is proposed for the obligatory word-internal syncope process. An alternative, more concrete analysis is also presented for the treatment of final schwa. In addition, modifications are made in Dell's formulation of the rules involving schwa in hiatus so as to reflect their function and role both as phonotactic constraints and as phonological rules. The investigation also raises a number of problems of both a diachronic and synchronic nature, problems whose analysis involves several fundamental issues in generative phonological theory. The data investigated provide examples of mechanisms assumed to account for historical change, such as the addition, loss, generalization, morphologization and inversion of rules. Evidence in support of certain notions on the way in which phonological change is implemented, e.g. lexical diffusion, variable rules, is also presented. The notion of linear ordering and principles governing rule orderings, such as feeding/bleeding relationships, rule transparency/opacity, receive support from the account. The problem of multiple applicaition of rules is encountered and solved by an alternative proposal to the Simultaneous Application Principe, viz., the Principle of Iterative Rule Application as formulated by Tranel (1974). Finally, the concept of natural phonological rules and the metatheoretical criteria advanced by Schane for determining whether or not a rule is natural receive corroboration from the analysis of schwa. In addition to the basic categories of such rules, the frequent resort to Schane's 1972 proposal for the inclusion of processes involving stress dynamics in the typology of natural rules points to the theoretical well-formedness of this hypothesis.