Douglas C. Walker

The new stability of unstable –e in French

Much recent work (e.g. Hansen 1994) is demonstrating that French schwa is becoming stable in many contexts where it formerly deleted freely. In this paper, I expand on Hansens study of the new stability of schwa, extending consideration to the lexicalized conditioning of the stability as indicated in four contemporary pronouncing dictionaries of Standard French. This increasing stability also provides evidence for a reanalysis of schwa as underlying /œ/; the implications of this analysis for the distinction between lexical and post-lexical rules are briefly explored.

Surface Structure and the Centrality of Syntax.

This paper examines the status of surface structure in transformational grammar, and the way that surface structure mediates the contacts between the phonological and semantic components of the grammar. Surface structure refers not to a single but to at least four distinct notions that do not necessarily define a homogeneous level of representation: output of the syntactic component, input to the phonological component, phonetic structure, and the level at which surface structure constraints are stated. Based on a survey of the literature, the conclusions include the necessity of direct links between deeper syntax and phonology, the influence of phonology on various syntactic operations, the need for phonetic information in certain semantic interpretation rules, and the lack of homogeneity among surface structure constraints. Finally, there is a recurrent influence of prosodic and morphological phenomena which motivate the revisions needed in the general organization of a grammar because they limit the types of interaction between the various grammatical domains. *

Lexical stratification in French phonology

Lexical stratification, indicated by diacritic features such as [foreign], [Romance], [learned], etc., is a widespread characteristic of contemporary phonological descriptions. This paper outlines some of the evidence distinguishing the learned from the non-learned lexical stratum in Modern French. When the learned vocabulary is incorporated into a standard phonological description of the language, the resulting analysis is highly abstract and not well correlated with data from language acquisition. Several aspects of the analysis of French phonology presented in Schane, French Phonology and Morphology, are discussed from this perspective, and an alternative dealing with stratal differences in a more concrete and lexically oriented manner is suggested.

Historical Romance linguistics: past, present and potential

Examination of the current status or future viability of disciplines (or perhaps better, subdisciplines) is a recurrent feature of the academic landscape, linked, at least partially, to the vagaries of academic (and other) fashion. One recent occurrence has focused on Historical Romance Linguistics, a collaborative and ultimately optimistic scrutiny in this journal ably coordinated by Steven Dworkin that drew a number of insightful responses from a baker’s dozen of scholars in both Europe and North America (2003, 31.2: 7-125). These surveys were then themselves thoughtfully reviewed Martin Maiden (2004), and it is no doubt presumptuous to think that dramatic new insights will emerge here. Nonetheless, “fools rush in...” Before that venture, however, one caveat is in order: my remarks will bear primarily if not exclusively on the North American situation and more specifically on Canada and the United States, the context I know best. This is obviously not to say that the discipline is not international in scope, nor that other, particularly European, scholars find themselves divorced from the North American milieu (or vice versa). But in the issues I intend to address, I believe that there are sufficient differences between the Old and New Worlds to justify a more focused examination. One need

Old French and the unordered rule hypothesis

Linguistic theory has been increasingly concerned with general principles that would permit the elimination of extrinsic rule ordering. Three such principles that have been proposed for phonology are proper inclusion precedence (PIP), obligatory- optional precedence (OOP), and the syllable structure principle (SSP), which requires rules that alter syllable structure to apply before ‘segmental’ rules. Old French presents alternations between forms such as jorn-jorz-jors ‘day’, phonetically [jorn-jorts-jors], where rules of denasalization and cluster simplification are required to describe the free variation between the latter two forms: ⧸jorn+s⧸↘↙[jorts][jors](by denasalization)(by cluster simplification) Much independent evidence indicates that [ts] is a single segment in Old French; hence, a solution with ordered rules of denasalization preceding cluster simplification (optional) is not possible, since the output of denasalization does not meet the structural description of simplification. Because of the nature of the alternation, however, one of the rules must be optional, but no matter which one is so specified, a violation of OOP results. Further, if cluster simplification is chosen as the optional rule, a violation of PIP also results. Finally, if one wished to interpret [ts] as a sequence rather than as a single segment on phonological grounds, a violation of SSP would be produced. In the latter case, each account of the alternation would contradict two general principles of rule application, and consequently, there would be no way of choosing between alternate solutions. This dilemma leads to two additional proposals: (1) the necessity of ranking hierarchically the general principles of rule ordering, so that in cases of conflict we can opt for a solution that violates a less highly ranked constraint, and (2) a reaffirmation of the traditional accounts of denasalization, with a new, theoretically based argument for a monosegmental interpretation of [ts]. Thus, the Old French data serve to falsify OOP, and to indicate the need for further study of the relationships among universal principles of rule application, if the unordered rule hypothesis is to be maintained.