22-08-03 22:35, etherman23 wrote:
> Actually I admit to one (possibly two in pre-Anatolian PIE) because
> the evidence points to one (e.g. Armenian h).
Actually, Armenian /h-/ (as oppsed to Anatolian aitches) isn't
particularly good evidence. It can reflect initial *s or *p, and there
is some irregular variation between /h/ and zero in Armenian. It's hard
to guarantee that initial /h-/ isn't simply a prothetic onset-filler in
some cases (similarly in Albanian, pace Eric Hamp).
> However I reject the
> notion that all vowels are derivable from /e/, that the laryngeal had
> a syllabic variant, and that the laryngeal had numerous coloring
> qualities.
How do you reject such notions? By just declaring your rejection?
Incidentally, it's a crude simplification to say that "all vowels derive
from /e/", and it's plainly incorrect to say that the laryngeal theory
assumes "numerous colouring qualities". Its mainstream version only
claims that *h2 colours an adjacent *e to *[a] and that *h3 colours an
adjacent *e to [o]. Does "numerous" include "two" in plain English?
> These are all part and parcel of the laryngeal theory.
> There are numerous instances in Anatolian where a laryngeal should
> exist according to theory but doesn't appear.
It's relatively well established that *h2 and *h3 survive as Hittite <h>
(<-hh->) in certain positions (e.g. word-initially, unless syllabic),
while *h1 is lost in all positions (in Hittite and the rest of
Anatolian). I'm not aware of any serious counterexamples.
> Any language can be reconstructed with one vowel and a sufficiently
> large number of coefficients sonantiques that disappear while
> coloring that vowel. That doesn't mean that all languages have only
> one vowel! So we need to have some direct evidence for them. Armenian
> and Anatolian provide such evidence for at least one, but no more
> than two. And these two branches aren't even all that consistent with
> each other, so we may be talking about parallel innovations.
The first two sentences of the above paragraph contain a straw-man
argument. PIE vowel qualities are not exclusively due to laryngeal
colouring. Regular ablaut, for example, has nothing to do with the
laryngeals. Only *a may owe its phonemicisation to *h2 (unless there are
PIE *a's of other origin -- and there probably are).
The laryngeals were not introduced in order to make reconstructed PIE a
language with one vowel, but in order to integrate some seemingly
irregular types of ablaut with the ordinary pattern of *e/*o/zero. In
other words, they helped to make sense of some unexpected vowel
qualities and of alternations involving long vowels, reconstructed
schwas and "long" syllabic liquids and nasals. It has since been found
that the laryngeals produced a number of other effects as well. The
collective weight of such indirect evidence is highly satisfactory,
because it has the form of arguments starting from different (and
independent) premises and converging on a uniform solution. Of course
it's nice to have direct reflexes of some of the laryngeals in
Anatolian, but the direct evidence isn't as crucial as you make it seem.
Piotr