Preliminary Lexicostatistics as a Basis for Language Classification: a New Approach
George Starostin (Russian State University for the Humanities / Russian Presidential Academy of National Economy and Public Administration,
Journal of Language Relationship, № 3, 2010 - p.79-116
The article discusses the basic methodology that underlies the construction of a global lexicostatistical database for all of the world’s languages, currently one of the main tasks of the Evolution of Human Languages project at the Santa Fe Institute. The author presents several important modifications of the traditional lexicostatistical procedure, such as: replacing the traditional 100­item wordlist with a more compact list of 50 “ultra-stable” items; use of low-level protolanguage reconstructions as primary construction nodes; a combination of the comparative-historical method and principles of phonetic similarity as the basis for the cognate scoring procedure; and, most importantly, a heavy emphasis on semantic precision and severe restrictions on the use of synonyms.
Keywords: lexicostatistics, taxonomy, comparative method, language relationship, semantic reconstruction, Swadesh wordlist