Re: Laryngeals revisited

--- In cybalist@yahoogroups.com, "C. Darwin Goranson"
<cdog_squirrel@...> wrote:

> > > >h1 was earlier /h/ and then /?/.
> > > >h2 was earlier /x/ and then /h/.
> > >
> > > Wouldn't this have led to a collapseof h1 and h2?
> >
> > I don't think so. It seems that h1 became /?/ before h2

became /h/.

> > In fact, h2 may not have become /h/ until sometime after IE began

to

> > break apart.
>
> In that case, then there must be a need to create LEVELS of
> Proto-Indo-European based on approximate dates. Level 1 or PIE-1
> could be the earliest traceable state of PIE, Level 2 or PIE-2
> would be slightly later, until PIE-n (n being some number), the
> latest possible state of PIE before break-up. This is just an idea
> for practicality, but when being able to trace back to several
> degrees of origin, wouldn't it be more useful?

Ideally, the thing to do is recognize relative or absolute stages of
development, where each stage is marked by a change in the
phonological and morphological rules of the language.

For laryngeals, I've come up with four relative stages of development:

Stage 1 - h1 = /h/, h2 = /x/, h3 = /¿/; there is also /?/, but only
as a consonantal anticipation of phonemically vowel-initial words
under sandhi-less conditions.

Stage 2 - h1 > /?/, merges with original /?/; h3 > /xW/. (These
changes do not necessarily happen at the same time.)

Stage 3 - h2 > /h/, h3 > /hW/; h1 may have disappeared at this time.

Stage 4 - h2, h3 > /ø/.

Stage 4 and possibly Stage 3 happened after IE began to break apart.

> > It seems like there may have been anticipatory rounding (not an
> > uncommon phenomenon) in vowels preceding labialized phonemes in
> > closed syllables. Admittedly, we do not see this in the
> > traditionally-reconstructed proto-language, and nor do we see it
> > in the daughter languages. However, it would help to explain
> > some inconsistencies in root-noun vocalism, such as *(H)
> > re:gs 'king, ruler' vs. *(H)o:kWs 'eye' and *wo:kWs 'word, voice'
> > (however cf. Greek (w)epos < *wékWos) without needing laryngeals.
>
> Are there any current examples of this kind of rounding? Say, in
> English or German or other modern IE languages?

English back vowels retain their quality both before and after labial
glides, e.g.:

'water' (otherwise would be pronounced as 'waiter'),
'what' (otherwise as with the /&/ in 'cat'),
'war' (otherwise as with the /a/ in 'car'),
'wolf' (otherwise as with the /O/ ~ /A/ in 'pot')
'draw' (otherwise as with the /au/ in 'drought'),
'law' (" "),
'saw' (" "),
etc.

English /l/ also preserves back vowels -- compare 'talk' and 'walk'
vs. 'tack' and 'whack'.

- Rob