Retour vers liste

Détail de la contribution

Auteur: Hannah WASHINGTON

Divergence in Romance Trends: Variable Object Clitic Placement in European Portuguese

Abstract/Résumé: While Brazilian Portuguese (BP) follows the Romance trend toward generalized proclisis for clitic objects, European Portuguese (EP) strongly favors enclisis as a general rule (Cunha & Cintra 2002, Perini 2002). However, following certain classes of words-namely conjunctions, negators, and certain adverbs-EP normatively requires proclitic placement. Variable enclitic placement can occur in the presence of proclitic triggers, although such placement is often ignored in the literature. This study investigates the factors that determine the selection of enclitic placement in two normatively proclitic contexts: following the adverb talvez ‘maybe’ and the complementizer que ‘that’. Using the Corpus do Português (Davies & Ferreira 2006-), 933 tokens of clitic pronouns following talvez and que were extracted and coded for fixed effects: trigger word (talvez vs. que), number of words between the adverb/conjunction and the pronoun, number of verbs in sequence, subject expression, verb form, tense, mood, person/number of the verb, pronoun, pronoun type, mode (oral vs. written), construction type, and year. The tokens were also coded for the random effects of speaker/author and verb, and the data was analyzed using a mixed-effects logistic regression model in R. Overall, 25% of the tokens reflect non-normative enclitic usage, and subject explicitness, construction type, trigger word, and mode all emerged as significant predictors of non-normative enclitic object placement. Much like what Schwenter & Torres Cacoullos (2010) show for Mexican Spanish, certain complex verbal constructions favor enclitic selection, with clear trends indicating a relationship between the extent of grammaticalization of compound verbal constructions and enclisis. Furthermore, post-verbal clitic placement in EP is affected by mode, such that oral data display significantly higher rates of enclitic placement than fiction, which itself has a higher rate than news and academic writing. Although the data show a relatively stable rate of enclitic usage over the past century, the results by mode and genre show clear differences based on the attention paid to speech, with oral language as the most divergent from the prescriptive norm. This may be indicative of a move toward categorical enclitic usage in EP, in contrast to the proclitic nature of BP and other Romance varieties. I conclude that the proclitic trends in Romance are not necessarily the inevitable result of linguistic change within the family. Similarly, grammatical trends in one standard variety of a pluricentric language do not necessarily result in similar outcomes in another standard variety.