This review conflates phonetic and phonemic comparison in that most of the methodologies surveyed treat phonemes as idealized phonetic segments between which some phonetic or featural distance is to be computed. Regrettably I must leave unmentioned a large literature that compares phonemes or entire phonological systems on other bases (e.g. ‡).
1467-968X/asset/TRPS_left.gif?v=1&s=15d9e13f2d12b933569722e37dce715f19578740)
1467-968X/asset/TRPS_right.gif?v=1&s=3fbf5fe4ab30a37cf388608850965a6536ea016e)
Phonetic comparison algorithms†
Article first published online: 13 JUL 2005
DOI: 10.1111/j.1467-968X.2005.00153.x
Additional Information
How to Cite
Kessler, B. (2005), Phonetic comparison algorithms. Transactions of the Philological Society, 103: 243–260. doi: 10.1111/j.1467-968X.2005.00153.x
- †
Publication History
- Issue published online: 13 JUL 2005
- Article first published online: 13 JUL 2005
- Abstract
- References
- Cited By
Abstract
Appealing to phonetic similarity has traditionally been discouraged in linguistics, partly because it has been an ill-defined and subjective concept. But much research nowadays requires measures of similarity between words, from practical work in speech technology, information retrieval, and commercial branding to theoretical studies involving language comparison and history. Phonetic comparison algorithms are crucial to this work, enabling computer implementation as well as reliability and significance testing. But phonetic similarity is not a unitary concept. Various types of measures are discussed, with emphasis on those most appropriate for current and future work in historical linguistics.