Labov (1972) defines the speech community by participation in a shared set of norms regarding evaluation and usage, and by “the uniformity of abstract patterns of variation”. Guy (1980) further notices that uniformity in variable rules doesn’t entail equal usage rates of a variant by its members; rather, it’s defined by equal constraints regarding the set of correlated variables and the relative factor ranking within each variable. In Brazilian Portuguese, definite articles (DefArt) may be variably used before possessive pronouns (minha mãe vs. a minha mãe ‘(the) my mother’), and differ not only in usage rates between Northeastern and Southeastern regions (CALLOU; SILVA, 1997), but also in their variable rules. This paper thus analyzes the speech of 8 migrants from the Northeastern state of Paraíba living in the Southeastern city of São Paulo (PBSP), in contrast with 12 non-migrant speakers from their original community (PB) and 12 from the host community (SP), to verify whether their usage rates of DefArt increased and whether they employ it in the same contexts as native Paulistanos. From the three samples comparably balanced for speakers’ sex/gender, age group, and educational level, over 2,200 tokens were coded for social and internal predictors (i-Pronoun person; ii-Pronoun gender; iii-Pronoun number; iv-NP syntactic function; v-Presence/type of preposition; vi-NP semantics; vii-NP specificity). Overall, rates for DefArt are 42%, 51%, and 54% for PB, PBSP, and SP respectively, signaling increased usage for migrants. Results of conditional inference trees and multivariate mixed-effects models in R, including Speaker and NP nucleus as random effects, show that whereas non-migrant PB and SP share the same constraints for Pronoun person, semantics and specificity, they differ regarding the other four internal variables. PBSP align with SP in three out of these, and show an intermediary pattern in the fourth: for Syntactic function, compared to the reference level adverbial adjuncts, indirect objects favor the presence of DefArt for PB, but it’s not significantly different for SP and PBSP; predicates aren’t significantly different for PB, but disfavor DefArt for SP and PBSP; Pronoun gender is significantly correlated with DefArt for SP and PBSP, but not for PB; and Pronoun number correlates with DefArt for PB, but not for SP and PBSP. For Presence/type of preposition in SP, both contracted (pelo ‘for(the)’) and uncontracted prepositions (sobre-a ‘about(the)’) favor DefArt compared to absence of preposition, while in PB it’s only favored by contracted prepositions; here PBSP behave similarly to PB, but show an intermediary trend for uncontracted prepositions, signaling change. Social predictors only show correlation for Sex/gender in SP, with males disfavoring DefArt, revealing that while migrants have acquired SP’s internal variable rules, they don’t necessarily mirror SP’s social stratification. Individually, however, migrants also show striking assimilation of SP’s rules. Hence, this paper shows that migrants in dialect contact situation may not only raise their usage rates of the host community’s prototypical variant, but also acquire more abstract patterns of variation in the form of variable rules. References Callou, D.; Silva G. (1997) O uso do artigo definido em contextos específicos. In: Hora, D. (org.). Diversidade linguística no Brasil. João Pessoa: Idéia. 1997, p.11-26. Guy, G. R. (1980) Variation in the group and in the individual: the case of final stop deletion. In: Labov, W. Locating language in time and space. New York: Academic Press, p.1-36. Labov, W. (1972) Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.