Onset-to-Onset Probability and Gradient Acceptability in Korean

Bi-phone probability, or the relative frequency with which two segmentally adjacent phonemes co-occur in a language, has effectively explained the influence of speakers’ sensitivity to the phonotactic probability of sound sequences on lexical processing. In this paper, we argue that speakers are also sensitive to onset-to-onset probability in their native language, and therefore bi-phone probability alone is not a sufficient estimate of sequential probability since onset-to-onset probability reflects how often two onsets that are segmentally non-adjacent co-occur in the language. To support our argument, we present an experiment which shows that adult Korean speakers are sensitive to onset-to-onset probability in their language and that their sensitivity is manifested in their gradient acceptability judgment of non-words.