[PS-1.72] Retrieving and encoding during gender-marked and unmarked European Portuguese reflexives

Luegi, P. 1, 3 , Leitão, M. 2 , Carvalho, M. 1 & Costa, A. 1, 3

1 Universidade de Lisboa
2 Universidade Federal da Paraíba - UFPB / CNPq
3 Centro de Linguística da Universidade de Lisboa

There is some debate about the effective use of binding principles during reflexive processing. Recently, it has been proposed that the effects that have been found may be due to retrieving rather than encoding interference.
We ran two self-paced reading experiments to contrast the effect of antecedents' gender-overlap/non-overlap on reflexive processing in European Portuguese (EP). EP has two possible reflexive forms: one gender-unmarked (SE-(1)) and one gender-marked reinforcement form (SE+verb+SI MESMO/A-(2)).

(1)A Ana_FEM/O Pedro_MASC disse que a Maria_FEM ontem se_Reflex_NO-GENDER vestiu?
(2)A Ana_FEM/O Pedro_MASC disse que a Maria_FEM ontem se_Refex_NO-GENDER vestiu a si mesma _REINFORCEMENT_FEM?
(Ana+Maria=gender-overlap; Pedro+Maria=no-gender-overlap)

32 items like (1) were used on Exp1 (Participants=32) and 24 like (2) on Exp2 (Participants=32) followed by: Did Mary dress herself?. Results show: less accuracy on questions after gender-overlap (Exp1&Exp2); slower reading time when answering to gender-overlap conditions (Exp1&Exp2); slower reading time when reading reflexives? region on gender-overlap conditions (Exp1). Overall results support retrieving interference, despite the unexpected effect found in the unmarked-reflexive region (only-Exp1). Two possible explanations for this effect are encoding interference or participants' adaptation to the experiment: since the experiment was long, participants fixated longer on the reflexive, on gender-overlap conditions, so as to properly answer the questions.