This study investigates how Korean L2 speakers produce word-final alveolars placed in the word-boundary palatalization environment in English. The purpose of our research is to see whether they show production variation according to lexical factors su...
This study investigates how Korean L2 speakers produce word-final alveolars placed in the word-boundary palatalization environment in English. The purpose of our research is to see whether they show production variation according to lexical factors such as word frequency and wordhood and phonological factors such as word-final alveolar types, number of syllables of the target words, number of word-final codas, etc. Four possible pronunciation variants were identified for word-final alveolars (canonical, palatalization, deletion, and wrong pronunciation). First, the results showed that like previous studies on L1 production, phonological variation was found for Korean L2 speakers. Specifically, canonical variants were predominant, and then the realization of palatalization was also common, whereas mispronounced variants such as deletion and wrong pronunciation were quite rare. Second, word frequency affected the likelihood of palatalization similar to native speakers of English. Palatalization was found more in high-frequency words than in low-frequency words. Third, wordhood affected the likelihood of palatalization. Fourth, production patterns as well as the likelihood of palatalization were affected by word-final alveolar types. Finally, backness of the vowels preceding word-final alveolars affected the occurrences of palatalized variants. That is, word-final alveolars underwent palatalization immediately following front vowels more frequently than back vowels. These findings suggest that production variability is observed even for L2 speakers as well as for L1 speakers. Furthermore, they provide additional support for the claim that frequency may be encoded in word representation of L2 speakers’ mental lexicon. Finally, they revealed that many phonological factors contribute to variability of production of categorical phonological rules by L2 speakers, indicating that listeners might consider lexical and phonological factors to recover speakers’ intended words.