--- In
cybalist@yahoogroups.com, "Tom Brophey" <TBrophey@...> wrote:
>
> What sort of argument would you accept as evidence? How about this:
> If there were still three laryngeals, then under zero-grade *ah1,
> *ah2, and *ah3 all have the same reflex of /i/ in Vedic. This
> implies that in that time frame *h1, *h2, and *h3 had all changed
to
> be i-colored.
No, it does not presuppose any i-coloration of the consonantal
alternants. If the propvowels accompanying interconsonantal
laryngeals were ultrashort, i.e. shwa-like, they could easily merge
if they were originally different (as I believe they were) and
subsequently change into a more chromatic vowel. The merger of the
shwas into /a/ in most other branches would demand an "a-colored"
pronunciation of the laryngeals and thus make it impossible to
reconstruct a common protolanguage.
> It seems improbable that they would all make the same
> change. And if they did and were so similar, how did they maintain
> separate identities?
Why is that improbable? All three IE k's have become /k/ in Old
Irish. All three Sanskrit sibilants are /s/ in any varieties of
Modern Indic. All short vowels have merged in Latin internal open
syllables. Is it also improbable that the /h-/ of Greek hépomai,
hêpar, and hésperos has three different origins? They were close,
yet different, phonological change can be very subtle.
> Do I infer correctly from your statement that you would agree that
> there is no evidence in Indo-Iranian for multiple laryngeals?
There is plain evidence for a different treatment of /t/ contiguous
with H1/2 and H3: only H1/2 aspirate. There is also the difference
that H3 voices in píbati. There is the evidence that H1 does not
produce a word-internal shwa after a stop while H2 certainly does.
But all this is common to IE as a whole, so it does not constitute
evidence of a difference between the surviving laryngeals in post-
PIE Indo-Iranian. But what would the product of merger be? The only
phonetically probable thing I can suggest is a break, a pause in the
word where the laryngeal once stood and there is now a minute of
silence in its memory. I know that's not what I said, but perhaps it
should have been.
> > The three Greek colours of syllabic resonants followed by
> > laryngeals present the same oppositions of coloration as the
> > laryngeals had when they coloured adjacent /e/ in a prestage of
> > PIE, That certainly indicates that the laryngeals were still
there
> > in the relevant post-PIE linguistic stage when the specifically
> > Greek sonorant colorations were effected.
>
> As you know, Patrick has proposed a hypothesis that laryngeals did
> not color adjacent vowels, but rather the vowel color was an
> inherent quality of the vowels. One line of evidence against this
> hypothesis is the Greek reflexes of zero-grade with an adjacent
> laryngeal. (That is what "the three Greek colours of syllabic
> resonants followed by laryngeals" refers to, is it not?) What
> other lines of evidence are there against this hypothesis?
There certainly are the surviving differentiated consonantal
reflexes of Anatolian. For Tocharian, K.T.Schmidt has discovered
quite a number of cases of H2 > k in clusters with sonants. There is
also a delightful Armenian rule found by Olsen: between vowels,
rH1/3 > r, but rH2 > rr (the "strong r"). There is the nice little
case story of Germanic quick: PGmc. *kwikwaz from IE *gWiH3wós has
assimilated /H3w/ to the initial /gw/, thereby revealing the
phonetic affinity between /H3/ and the voiced labiovelar that was
expected anyway. For Celtic, Hamp has pointed to an almost pervasive
loss of /H3/ before labiovelars. Since these are all separate events
restricted to specific branches, they presuppose a differentiated
pronunciation of the ocnsonantal laryngeals at a time postdating the
separation of the branches.
Opposed to this, I can see no evidence in favour of the idea you are
so bravely fighting for. Why this spin?
Jens