Innate Biases in Language Learning (2014)

Undergraduate: Brandon Prickett

Faculty Advisor: Elliott Moreton
Department: Linguistics

---

Languages have rules about which sounds can come next to other sounds within words. For example, in English, the word ¿bnick¿ would not sound as good as the word ¿blick¿, even though neither is an actual word. This preference is a result of constraints in English. Hayes and White (2013) explored how well people learn natural and unnatural constraints in their language. Natural constraints are motivated by what sound combinations are physically easier to produce. If Language is an entirely learned skill, like riding a bike, then different types of rules in languages should be equally learnable. However, Hayes and White¿s findings suggest that natural constraints could be learned more easily by speakers of English. This implies that there could be an innate bias causing people to prefer these particular rules. The current study seeks to explore other variables that might have contributed to Hayes and White¿s findings besides the naturalness of their constraints. Past work in Linguistics and Psychology has shown that people learn simpler patterns better than they learn complex ones, even patterns that involve sound patterns in made-up languages (Shepherd et al 1961, Pycha et al 2003). Online experiments were used to test complexity as a factor that might be confounding Hayes and White¿s findings. The results suggest that naturalness does people¿s ability to learn real world sound patterns and that complexity does not.