summary on 'cotton'

[Laurent Sagart]

Greetings to all:


This is a summary of responses received after I posted a query on this list concerning one of Dempwolff's reconstructions: *kapas 'cotton'; I asked whether the Oceanic forms adduced by Dempwolff in support of his reconstruction (Fidji kava 'string', Tonga, Futuna kafa 'string', Samoa ava 'string' ) were convincing. I received replies from Robert Blust, Paul Geraghty, Frederic Plessis and Stan Starosta. I thank all of them. A couple of references on cotton in the Pacific were provided:


Johnson, Rubellite K., and Bryce G. Decker.  1980.  Implications of the Distribution of Native Names for Cotton(gossypium spp.) in the Indo-Pacific.  249-307.


Robert Langdon (1982) New World cotton as a clue to the Polynesian past. In: Siikala, Jukka (ed.). 1982. Oceanic Studies: Essays in honour of Aarne A. Koskinen Suomen Antropologinen Seura (The Finnish Anthropological Society). Helsinki.


Robert Blust indicated that he thought the Oceanic forms cited by Dempwolff are not clearly related, and the forms in insular SEA could very well represent a loan distribution. Paul Geraghty pointed out that the Proto Polynesian reconstruction is *kafa and it only applies to sennit, ie coconut-fibre cord. Fijian kava means 'roll (of sennit)'. He suggested that a MP word reconstructed by Blust: *ka(ng)ba(Ct) 'to tie, wrap up' (#184 in Blust's Austronesian Etymologies of 1971, OL IX, 2:127):) on the basis of forms in Malay, Maranao and Ilocano, might be related. 


I personally find Geraghty's suggestion attractive. The semantics of 'bind' > 'sennit string' seem to me preferable to Dempwolff's 'cotton' > 'sennit string' (name of instrument out of corresponding verb, a well-established semantic process, whereas the evolution 'cotton' to 'sennit string' seems to require the not obvious assumption that strings were made out of cotton fibre at some point in Austronesian history. At the very least, Blust's *ka(ng)ba(Ct) 'tie' looks like a viable alternative to Dempwolff's etymology. 

L. Sagart
CRLAO
54 Bd Raspail
75270 Paris cedex 06
France
Tel: +33 1 49 54 24 18
Fax: +33 1 49 54 26 71

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/an-lang/attachments/20011112/fb957ba9/attachment.htm>