Back to list

Detail of contribution

Auteur: Mary Elizabeth BEATON

Titre:
The Role of Lexical Frequency and Morphosyntactic Processing: First Person Plural Expression in Rio de Janeiro Portuguese


Abstract/Résumé: The alternation between the first-person plural pronouns nós (P4) and a gente (P3) is a widespread phenomenon in Brazilian Portuguese. Previous studies (Beaton 2010, Lemos Monteiro 1991, Lopes 2003, Omena 1996a, Omena & Braga 1996, Naro et al 1999, Travis & Silveira 2009, Zilles 2005) consider many independent variables but do not investigate all of possibilities available in corpora research. This study looks at the roles of verb frequency and processing effects in the selection of the morphology for first person plural reference using the distance back to the previous P3/P4 reference as a means of understanding processing in casual speech. The data for the current study comes from 15 speakers from Rio de Janeiro recorded in 1980 and in 2000 for the Programa de Estudos sobre o Uso da Língua project. A total of 1,314 tokens were analyzed using a logistic regression mixed effects model with speaker and verb as random effects with an overall rate of 37% P4 and 63% P3. Independent linguistic variables included pronoun expression, construction type, tense, subject of the preceding clause, clause type, count back to previous P3/P4 reference up to 10 clauses and verb frequency. Building on recent work on the role of lexical frequency in variation (Erker & Guy 2012), I consider frequency as an orthogonal constraint by dividing my data by lexical frequency. A model using the five most frequent verbs does not select tense as a significant predictor. Other models (the model of the full data set, the ten most frequent, the 25 most frequent, etc.) indicate that tense has a significant effect on the variation. This suggests that more frequent verbs do not rely on tense as a determining factor for the variable. All previous studies suggest that tense is an important predictor and some studies equate this to the highly salient phonological differences between P3 and P4 forms in certain tenses (Omena 1996, Naro et al 1999). The current study adds a layer of complexity to the tense and phonological claims; when considered with the 10-clause processing counts, tense is not selected as significant. A more complete consideration of processing accounts for a great deal of the variation between nós P3 and P4 before accounted for by tense. In sum, this study emphasizes the importance of considering cognitive aspects of language in corpus research on morphosyntactic variation. In sum, this study emphasizes the importance of considering cognitive aspects of language in corpus research on morphosyntactic variation.