Abstract:
Previous analysis of segment inventories (e.g., by Maddieson, 1984 and Lindblom and Maddieson, 1988) has shown that, across a large sample of languages, the segments that more commonly occur contrastively are those that are less complex on an intuitive scale of simplicity. But a segment type within any given language may occur in many or few words: segments that occur in more items may also be those that are in some sense more simple, or those that combine best with other segments in sequences. Within-language frequency of usage may therefore serve to show that segments of (near-) equal frequency in terms of mere membership in inventories in fact show different preference patterns. However, because there are many idiosyncratic properties of individual languages, it is important to look across a range of languages to reach general conclusions. The present paper reports on the relative frequency of the different onset consonants found in the universally present syllable type CV in a global sample of 31 languages (10 from Asia, 8 from the Americas, 6 from Africa, 3 from Europe, and 4 from the Pacific area). The counts are based on lexical frequency. [Work supported in part by NIH.]