La variation grammaticale en géolinguistique: les pronoms sujet en roman central (1997)

This thesis deals with subject pronoun variation in the geolinguistic continuum which extends throughout Northern Italy and much of Southern France. The subjects of finite verbs can be identified by pronouns (as in French [je] parle 'I speak') or by verbal morphology (as in Italian parl[i] 'I speak'). It has been postulated by Perlmutter 1971 and Chomsky 1981 that this distinction reflects the two possible settings of an innate parameter: human grammars must be either (-null subject), like French, or (+null subject), like Italian. The nonstandard linguistic varieties studied here, however, include grammars which do not fit neatly into either of these two categories. After general considerations concerning the importance of variation in linguistic theory, a methodology is developed to deal with data from traditional linguistic atlas surveys. Although almost completely ignored by 'theoretical' linguistics, the Atlas Linguistique de la France (Gillieron & Edmont 1902-1908) and the Sprach- und Sachatlas Italiens und der Sudschweiz (Jaberg & Jud 1928-1940) provide invaluable data for the investigation of morphosyntactic variation. Data are coded from a total of 438 points appearing on 101 maps from each of these atlases. These data are then synthesized in 15 different thematic maps, according to grammatical person and a variety of other linguistic factors. The quantitative analysis of this database also forms the foundation for some generalisations about the nature of subject pronoun systems. First, the evidence does not point to an abrupt transition, as the parametric hypothesis would predict, but rather towards a gradual transition zone. Secondly, the typology of subject pronoun systems is not completely unconstrained, since certain clear tendencies emerge among the intermediate varieties. Subject pronouns are least likely to occur in the first person singular or plural, and the second person plural, and most likely to occur in the third person singular or plural and (especially) the second person singular. Such morphological asymmetries cannot be readily represented in syntactic or featural terms, much less by parametric theory. Furthermore, many linguistic factors, including verb type and tense, clause type and position, other pronouns and negation, all contribute to the conditioning of subject pronoun usage. These facts strongly suggest that subject pronoun paradigms reflect complex morphosyntactic 'interface' phenomena (Chomsky 1995) rather than a straightforwardly syntactic parameter