Past research suggests that when learners speak in a second language (L2), their first language (L1) is co-activated and needs to be inhibited. The present study examined the mechanism of non-target language inhibition when Japanese learners of Spanish spoke in their L1 Japanese and L2 English as well as in their third language (L3) Spanish. It was predicted that the degree of non-target language inhibition would be modulated by relative language proficiency.