Statistical learning and innate knowledge in the development of second language proficiency: evidence from the acquisition of gender concord

Empirical studies show that adult second language (L2) learners are unconsciously sensitive to statistical properties of input, such as frequency of forms. Such properties have typically not ! gured in ?nativist? accounts of L2 pro! ciency. One potential challenge for nativist accounts is explaining pervasive optionality in the use by L2 speakers of morphological variants of the same form (e.g. the le vs la de! nite article contrast in French, or the rojo vs roja ?red? adjective contrast in Spanish). This chapter considers two accounts of such optionality. On the basis of evidence from a study of the L2 acquisition of Dutch noun-adjective gender concord by Blom, Poli?ensk� and Weerman (2008), it is argued that statistical properties of input do play a role.