Misperceptions of foreign language phonology.

Abstract

For many years it has been assumed that when hearing familiar sounds in unfamiliar combinations, listeners will perceive the sounds accurately. In the recent years, this assumption has been challenged [Halle et al. (1998) and Berent et al. (2006)]. The present study investigates what listeners actually hear when presented with familiar consonants in unfamiliar combinations. A group of monolingual native English speakers was asked to transcribe Russian words containing only consonants attested in English but presented in a two-, three-, and four-consonant combinations (CC, CCC, and CCCC clusters) which are not legal in English. The results indicate that the accuracy of response and the nature of errors depend on the type of cluster. CC clusters (e.g., /pn/) were most often misperceived as containing a vowel between the two consonants, and perception accuracy was well-explained by phonological principles, such as sonority. On the other hand, CCC (e.g., /vzb/) and CCCC (e.g., /fstr/) cluster transcriptions contained relatively few vowel insertions, but many deletions and substitutions and their accuracy were better explained by acoustic factors, such as voicing of the consonants. These results suggest that speech perception is influenced by both phonological and acoustic factors.