Tracking multiple inputs
For many learners, language acquisition may entail acquiring more than a single language. Yet to date, much of the research on the fundamental mechanisms of language acquisition has been predicated, at least implicitly, on modeling monolingual acquisition. In this chapter, we explore statistical learning, the ability to track distributional properties of the input, through the lens of multilingual acquisition. This ability is thought to play a critical role in the early stages of language acquisition. We identify a set of theoretical challenges that need to be overcome in order to track multiple sets of statistics and develop multiple representations to accommodate each input language. We then review the limited number of empirical studies that have investigated how people keep track of statistics in multiple artificial inputs and explore the consequences of accruing statistics in multi-language input for infants raised in bilingual environments. We highlight the role that contextual cues may play in helping solve the problem of multiple inputs, pointing out that they may facilitate the forming of multiple representations. We conclude, based on the available data, that the consequences of bilingualism for statistical learning may be a greater propensity to posit multiple underlying causal models when the input is variable.