Syntax, information structure and social variation: The case of Spanish subject personal

This paper presents an analysis of internal and external factors that have an effect on the choice between null and overt subject pronouns in the Spanish spoken in Catalonia. A gender- and age-balanced sample of 54 Spanish speakers from Barcelona performed a gametask designed to elicit both declaratives and interrogatives. The transcriptions were syntactically tree-annotated. Subjects also responded to a sociocultural questionnaire based on the lifestyle operationalisation of Bourdieu (1979). The enveloppe of variation only includes clauses in which null as well as overt pronouns could occur which amounts to 8815 non-elliptical finite clauses. The dependent variable is the overt pronoun rate. I. Grammar-internal factors: 6 two-way ANOVAs were calculated with both the position of the overt pronoun (pre- vs. postverbal) and one of the following factors: root modality (declarative, polar, wh), verb class (transitive, unaccusatives, ditransitives...), polarity (affirmative, negative), object clitization (clitic vs. full DP), TopP (with or without topicalization/dislocation), person and number of the verb. The results show (i) factors that have a significant main effect that is independent of the subject pronoun position (person & number, polarity), and (ii) factors that only show an effect in interaction with the pronoun position (verb class, root modality)