Learning nonadjacent dependencies in phonology: Transparent vowels in vowel harmony

Nonadjacent dependencies are an important part of the structure of language. While the majority of syntactic and phonological processes occur at a local domain, there are several processes that appear to apply at a distance, posing a challenge for theories of linguistic structure. This article addresses one of the most common nonadjacent phenomena in phonology: transparent vowels in vowel harmony. Vowel harmony occurs when adjacent vowels are required to share the same phonological feature value (e.g. V_+F C V_+F). However, transparent vowels create a second-order nonadjacent pattern because agreement between two vowels can ‘skip’ the transparent neutral vowel in addition to consonants (e.g. V_+F C V^T_–F C V_+F). Adults are shown to display initial learning biases against second-order nonadjacency in experiments that use an artificial grammar learning paradigm. Experiments 1–3 show that adult learners fail to learn the second-order long-distance dependency created by the transparent vowel (as compared to a control condition). In experiments 4–5, training in terms of overall exposure as well as the frequency of relevant transparent items was increased. With adequate exposure, learners reliably generalize to novel words containing transparent vowels. The experiments suggest that learners are sensitive to the structure of phonological representations, even when learning occurs at a relatively rapid pace.