This talk outlines the differing results of two experiments investigating L3 acquisition of Germanic gram- matical gender concord by sequential Spanish/English bilingual learners, and emphasizes the key differences between true L3 initial state transfer and later developmental differences.
In Experiment 1 (Brown, 2020a), two sets of beginner L3 German learners (L1 English/L2 Spanish and L1 Spanish/L2 English) were compared in their ability to identify L3 grammatical gender agreement errors in a set of German sentences. Results showed that L3 learners who reported having Spanish, another gendered language, as an L2 significantly outperformed participants with L1 Spanish. On the surface, this finding might seem to suggest that these L2 Spanish participants are able to transfer their Spanish gender knowledge at the onset of L3 acquisition in a way that the L1 Spanish learners cannot, and could potentially fall in line with models of the L3 initial state such as the L2 Status Factor Model (Bardel & Falk, 2007). However, once it is considered that these participants had all had at least one semester of German instruction prior participating, it becomes unclear whether this L2 Spanish gender advantage is the result of initial transfer of some later developmental difference.
In order to test this, Experiment 2 (Brown, 2020b) considered true initial state L3 learners by exposing sequential English/Spanish and Spanish/English bilinguals to an artificial Germanic language and testing their ability to acquire the grammatical gender system. This study found no difference between groups, a result that contrasts with the findings in Experiment 1. This suggests that the Experiment 1 findings are not a reflection of initial state transfer, but rather demonstrate some form of later L3 developmental difference. These findings emphasize the importance of a distinction between L3 initial state transfer and early development.