The idea that languages without vowels exist is an enduring urban myth, the linguistic equivalent of the crocodile in the sewer or the poodle in the microwave. Those who have made a careful study of allegedly vowelless languages now maintain that there simply isn't any such thing. One of the most actively discussed cases was the phonological system of Kabardian, one of the Northwest Caucasian (NWC) or Abxaz-Adyghean languages. After a lively exchange of articles between Aert Kuipers, Morris Halle and others, over the possibility of an adequate phonologization of NWC languages which made no use of vowel phonemes, more recent analyses (e.g. those of John Colarusso in Canada, Catherine Paris in Paris, and of the majority of indigenous specialists in the Caucasus itself) require at least two, sometimes three vowel phonemes -- or more accurately, clusters of vocalic features, to generate a surprisingly rich inventory of phonetic vowels. (Interested readers should consult Colarusso's recently-published Grammar of the Kabardian Language). By the way, I have heard several NWC languages spoken, including a bit of Kabardian, and none of them would give a naive listener the impression of vowellessness. They have nothing to compare to those celebrated Bella Coola or Wishram jawbreakers, which sound a bit as though the speaker was trying to whisper and eat granola at the same time.

At the same time, as I read through some of the contributions to the Great Kabardian Debate, I was struck by just how close people like Kuipers got to their goal of eliminating vowels from the phonology. The arguments often hinged on the analysis of a handful of minimal pairs, which required an opposition of height or length. Were these lexical items -- some of them borrowings from Turkish or Arabic -- to be magically erased from the vocabulary, would we have to admit Kuiper's phonologization as adequate, and therefore a (phonemically) vowelless language as compatible with universal constraints on human phonological systems? I honestly don't know the answer, nor do I know enough NWC to attempt a thought-experiment along these lines (My field experience is to the southwest of NWC territory, in the Kartvelian-speaking highlands of Georgia). As some of you may know, Winfried Lehmann maintains in all seriousness that an anterior stage of Proto-Indo-European was phonemically vowelless, though equipped with a feature of sonority which triggered a vocalic realization of sonant phonemes. To sum up, I for one would prefer that we not rule out the POSSIBILITY of vowellessness, despite the absence of an attested example. A cautionary tale: A well-known syntactician once gave a paper demonstrating that languages with OSV and OVS word order were excluded by universal grammar. A missionary linguist, recently returned from South America, raised his hand and politely explained that such languages did indeed exist, but that all known examples were spoken by small indigenous communities in a little-explored region of northern Amazonia. One can only speculate on what linguistic diversity might have vanished without a trace in the Americas, Oceania -- or Europe, for that matter.

Hope this helps!

**************************************************************
Kevin Tuite                         514-343-6514      (bureau)
D=E9partement d'anthropologie         514-343-2494 (t=E9l=E9copieur)
Universit=E9 de Montr=E9al
C.P. 6128, succursale centre-ville
Montr=E9al, Qu=E9bec H3C 3J7      tuitekj@mistral.ere.umontreal.ca
**************************************************************