>On Sun, 4 Jan 2004, Tim May wrote:
> I've heard a recording of someone saying vprtskvni, and it sounds to
> me as if - phonetically - the r is syllabic and there are epenthetic
> schwas after the [k] and both [v]s.  Or maybe there's a schwa before
> the r as well, I don't know.  I can't find this online any more, but I
> can mail you the .wav file if you like.
> I've read that the formidable clusters in Salishan languages (like the
> Nuxalk example above) _don't_ contain epenthetics, and that these
> languages sound much more consonantal than Georgian.  I can't find any
> recordings of these, though.

I'm afraid I don't have any soundfiles, but I can pint you to Marika
Butskhrikidze's online dissertation 'the consonant phonotactics of
Georgian' which can be found at:

each of the chapters is in a separate pdf file, chapters 3 & 4 are the two
that focus on the actual language data.

A somewhat technical paper on harmonic clusters can be found at:


with commentary by Bruce Hayes at:

Hayes' commentary is interesting because he theorizes that the voiced (and
presumably voiceless & ejective) harmonic clusters, /dg, bg/ etc, may be
velaric egressive, given that the anterior stop has more articulatory
intensity than a corresponding word juncture sequence /d#g/.

I'll have to dig out my copy of Nater's Bella Coola/Nuxalk grammar, but if
memory serves, clusters with voiceless fricatives have no epenthetic
vowels, but those with sonorants vary between having syllabic sonorants and
epenthetic vowels. Can't recall what stop clusters do, but I think they
have epenthetic vowels as well (though weak, and, I think, possibly