A proposed explanation of the specific/nonspecific TÙ constraint ranking

in Spanish

Richard Cameron

Sun. 9-10:40 A

The ranking of internal grammatical constraints on linguistic variation presents a considerable explanatory challenge. Such a challenge may be increased two-fold in cases where two regional dialects display different constraint rankings. A case in point is the different ranking of Specific and Nonspecific reference for the category of second person TÙ in Latin American and Iberian dialects of Spanish. As in other null subject languages, finite verbs in Spanish permit the variable expression of pronominal and null subjects. Second person TÙ subjects may be further analyzed per specificity of intended reference as Specific or Nonspecific. Specific TÙ occurs strictly in contexts of direct address or in narrative recreations, via reported speech, of direct address between two speakers. Nonspecific TÙ is used when a speaker reports a personal experience and generalizes such that it becomes applicable to anyone given a similar set of circumstances. In studies of San Juan, Buenos Aires, and Santiago, Nonspecific TÙ favors personal pronominal expression relative to Specific TÙ. In studies of Madrid and, more recently, Sevilla, the reverse pattern emerges with Specific TÙ favoring personal pronominal expression relative to Nonspecific TÙ. Although Cameron (1993) initially identified these dialect differences, no explanation was provided. In this paper, I propose two explanations for the ranking of these referential effects, one for the Latin American dialects, utilizing arguments from Accessibility Theory, and one for the Iberian dialects which relies on generative treatments of referential specificity and arbitrary reference (i.e., pro(arb)). However, neither explanation alone accounts for the different rankings. Consequently, I next propose that the Latin American dialects have reanalyzed the quantitative nature of Nonspecific TÙ by analogy to that of either Nonspecific UNO or Nonspecific USTED, both of which show relatively high frequencies of personal pronominal expression in all dialects. In short, we find paradigm leveling with respect to the frequency of pronominal expression for Nonspecific reference in singular subjects. This may have been facilitated by the fuzzy set quality of Nonspecific reference in a fashion similar to the ambivalence of the "pause" constraint on (t/d) deletion in English and its different ranking in Philadelphia and New York (Guy 1980). Moreover, because analogical change is not phonetically motivated, such analysis further permits us to explain why, contra the Functional Compensation Hypothesis, some variable /s/ dialects show a relatively high rate of pronominal expression whereas others, such as Sevilla, show a relatively low rate. Finally, review of the frequencies of second person TÙ pronominal expression permits a very precise prediction of which dialects, as yet not analyzed, will show which ranking of Specific and Nonspecific TÙ. Briefly, I submit that for any dialect of Spanish, if the overall rate of pronominal expression for second person TÙ is between 0% to 35%, then Specific TÙ relative to Nonspecific TÙ will favor pronominal expression. If the rate of pronominal expression is higher than 35%, the reverse pattern will obtain. Hence, in the Mexican-American Spanish spoken in Los Angeles, Specific TÙ relative to Nonspecific TÙ should favor pronominal expression because the average reported frequency for second person TÙ is 20% across the three groups studied by Silva-Corvalan (1994: 163). In contrast, the Spanish spoken in Malabo of Equatorial Guinea (Lipsky 1985), an /s/ retaining dialect, should show the reverse constraint ranking.