Voiceless Labiovelar Stop in Celtic and Its Role in the Classification of Celtic Languages
Tatyana Mikhailova (Lomonosov Moscow State University, tamih.msu@mail.ru)
Journal of Language Relationship, № 2, 2009 - p.79-90
The most obvious phonological distinction between the two main branches of Celtic languages (Goidelic and Brittonic) involves the outcome of the unvoiced labiovelar */kw/. In Goidelic it was delabialized in most environments and merged with /k/, but in Brittonic it became fully labialized and yielded /p/. This basic difference, proposed by John Rhys in 1882, has given rise to the unfortunate terms “P-Celtic” (Welsh, Cumbric, Cornish and Breton) and “Q-Celtic” (Irish, Scottish and Manx). The Continental Celtic languages (Gaulish, Lepontic and Celtiberian) show some fluctuations in regard to this sound, and it is highly probable that they represented allophonic variants of Proto-Celtic *qw for a very long time (in spite of the depth of divergence of Proto-Celtic — appr. 1200 BC according to glottochronological calculations). Toponymic and ethnonymic data of later periods (400 BC — 400 AD) demonstrate that the variants of this sound were acoustically similar to the speakers, especially within the zone of their prominent contacts — West of Scotland and North-East of Ireland. We therefore presume that this most important element of Celtic historical phonology cannot serve as a distinctive principle in the construction of the tree diagram of Celtic languages and only makes more obscure the scheme of their development.
Keywords: Celtic languages, comparative phonology