Cross-linguistic variation in language similarity classification.
MaryAnn Walter · The Journal of the Acoustical Society of America · 2009
This study aims at identifying factors that make language sound structures seem more or less similar to English, and how those similarity judgments change according to the listener’s native language. Listeners from four different native language groups (English, Mandarin Chinese, Korean, and Turkish) sorted a group of 17 genetically and geographically diverse languages in terms of their sound-based distance from English. Placements of individual languages were analyzed, as well as groupings of similarly ranked languages and correlations among overall ranking structures of the different groups. Bilingual listeners exhibit more variability in their rankings than monolingual English speakers, rank their own language as less similar to English than other groups do, and rank languages of neighboring groups the least similar to it of all. Ranking correlations between language groups are significant, varying somewhat in magnitude depending on geographical proximity and typological/genetic relatedness of the listener group languages. This reflects the presence of consistent groupings within ranking structures for all language groups, which depend on sound-based factors such as the presence of perceptually salient speech sounds. These results enable predictions about relative intelligibility among international English users, native and non-native. [Work supported by a Mellon Postdoctoral Fellowship awarded by the Woodrow Wilson Foundation.]