From universals to learning

Abstract

Artificial grammar or language learning paradigms (henceforth ALL) have long been a core tool for understanding language acquisition. Findings from laboratory learning of miniature language systems have shed light on the cognitive mechanisms responsible for natural language learning, starting with studies on implicit learning in the 1960s (e.g.,

; ), and studies on statistical/implicit learning starting in the 1990s (e.g., ; ; ). While artificial grammar learning paradigms of the past often featured meaning-free sequences of symbols, presented to participants via passive exposure, now many studies include much richer linguistic content, as well as communicative and social context. This opens the door to new connections with second language acquisition (SLA) research. In this chapter, I survey recent findings from the experimental linguistics literature, where a major new area of research has emerged, using artificial language learning experiments to test predicted links between cognition and typological universals, or generalisations. Of course, universals have long had a place in SLA (e.g., ; ; ). However, the empirical basis for purported universals has begun to move forward in leaps and bounds. In light of this, I would like to argue here that the time is right for renewed interaction with the SLA community, in the service of generating and testing key predictions for theories coming from both sides.