Onnis, L. 1 & Thiessen, E. 2
1 University of Hawaii
2 Carnegie Mellon University
Adults' linguistic background influences their sequential statistical learning when exposed to an artificial language characterized by conflicting forward-going and backward-going transitional probabilities. English-speaking adults favor backward-going transitional probabilities to group words, consistent with the head-initial structure of English. Korean-speaking adults favor forward-going transitional probabilities, consistent with the head-final structure of Korean. The current experiments assessed when infants develop this directional bias. Seven-month-old infants showed no preference for forward-going or backward-going regularities. By 13 months, though, English-learning infants favor backward-going transitional probabilities over forward-going transitional probabilities, consistent with English-speaking adults. This result indicates that sequential statistical learning rapidly adapts to the predominant syntactic structure of the native language. Such adaptation may facilitate subsequent learning by highlighting statistical structures that are likely to be informative in the native linguistic environment. The study points to a possible developmental trajectory of grammar in preverbal children, and has implications for reassessing the scope and power of statistical learning. Usage-based accounts of language acquisition may appear limited as they rely on accumulated piece-meal knowledge that is local and strongly lexically-based. However, because infants in our experiment were exposed to novel sequences of nonsense syllables, their emergent prior biases for statistical groupings cannot be explained by familiarity with previously known words in their language. Instead, our results suggest that infants' sequential learning mechanisms adapt to the input broadly at a systemic level. By virtue of this adaptation, sequential learning mechanisms may be responsible for biasing toward general tendencies in the input that learners can capitalize on to arrange and parse novel word order sequences. These results are consistent with the hypothesis that general inductive biases emerge as the interaction between experience with the environment (linguistic in this case) and learning mechanisms. Such approach usefully points to ways to overcome the traditional nativist-empiricist debate.