Roger Blench
Kay Williamson Educational Foundation
<r.blench@odi.org.uk>
Are there four additionalunrecognised branches of Austroasiatic?
Austroasiatic is usually considered to have twelvebranches and these are on the whole well characterised, although the internalclassification of the phylum remains controversial. This paper evaluates thepossibility that there were originally four other branches, whose existence cannow only be inferred from residual lexicon in languages which are currentlyconsidered non-Austroasiatic. These four hypothetical branches are;
a) The language of the Shompen. Eitherunclassified or considered a language isolate, Shompen has a number of cognateswith mainland Austronesian which are not shared with other Nicobareselanguages. This raises the possibility that Shompen represents a separate andpresumably earlier migration to the Nicobars and that apparent similaritieswith Nicobarese may be due to borrowing.
b) The languages of coastalVietnam priorto the Chamic migrations. Paul Sidwell has observed that a significantpercentage of the lexicon of Chamic is neither Austronesian nor knownAustroasiatic. It may be that the languages assimilated by the incoming Chamicspeakers represented either a quite unknown phylum or else were an unrecognisedbranch of Austroasiatic.
c) Acehnese. Thurgood and Sidwell treatedAcehnese as related to the Chamic languages, and it certainly has a significantstratum of Chamic lexicon. However, it also has cognates with mainstreamAustroasiatic and vocabulary with no clear etymology. It is therefore possible,as Diffolth (p.c.) has argued, that Acehnese represents a residualAustroasiatic language that has come under heavy Chamic influence.
d) Bornean substrate languages. SomeAustronesianists (Adelaar in particular) have argued that unusual phonologicalfeatures ofBorneo languages such as Bidayuhpoint to a possible Austroasiatic substrate. Other linguists, including RobertBlust, claim that these features can be explained by processes internal toAustronesian. However, cultural evidence for mainland presence onBorneo is extremely strong, and such early contact isquite likely.
All these hypotheses remain to be tested; the paperevaluates the case for each proposal. In some cases, the poor quality of thedata (e.g. Shompen) may mean that they are presently undecidable. Additionaltypes of data, in particular cultural and archaeological, provide support forsome hypotheses and these may prove valuable in interpreting the strictlylinguistic findings.