Re: [tied] Thoughts on the existence of *H1

----- Original Message -----
From: "Glen Gordon" <glengordon01@...>
To: <cybalist@yahoogroups.com>
Sent: Wednesday, September 12, 2001 11:03 PM
Subject: Re: [tied] Thoughts on the existence of *H1


> Damn you and your clever devil's advocate skills!
>
> Okay, take a hypothetical *-CrH1C- then. The pronunciation
could
> have been *[-C&r&C-] since a syllabic *r equals *[&r] and
*H1
> theoretically equals *[&] for the most part.

<schnip>

> Alternatively, we might say that the "epenthesised vowel"
was
> there already. Then we can do away with a senseless *H1
altogether, thereby
> simplifying our theory while maintaining an equally
> acceptable chain of changes:
>
> *C&r&C- (-CrH1C-) > *-Cr&:C- > *-Cra:C-


Okay, but why is *r syllabic in the first place? In *-CrC-
it's the interconsonantal position that makes it syllabic,
but if *H1 is really a vowel, *r is prevocalic and should be
non-syllabic. Something like *-eCr&Ce- should be syllabified
like this:

*-e.Cr&.Ce-

See? There is no motivation for compensatory lengthening.
*h1 behaves like a consonant also when it comes to
conditioning syllable structure. With a consonantal *h1
(say, [h]) we get acceptable scenarios:

*-e.Crh.Ce- > EITHER *-e.Cr&h.Ce- > *-e.Crah.Ce- >
*-e.Cra:.Ce-
OR *-e.C&rh.Ce- with branch-specific continuations
(involving compensatory lengthening e.g. in Balto-Slavic and
Indic).

> And what of the interpretation that stems like *g^enH1-
are in fact
> simply disyllabic stems (ie: *g^ene-)? The added
advantage: IE
> looks more like an everyday language with both
monosyllabic AND
> disyllabic roots at its disposal just like almost every
other
> human language on the planet.

"The root" is not really a phonetic concept, because it is a
bound entity, like quarks. Apart from cases of
zero-derivation, it is not pronounced in isolation. The root
is the central morpheme of a derived word. Roots, if we
consider their representation in the human mental lexicon,
are unsyllabified strings of segments. A root might well
look like this: {mgreskiawwwts}, if a child could find
enough evidence for internalising it in this way in the
process of language acquisition. Most languages prefer
shorter roots since the various meanings worth representing
can be encoded much more economically (especially if you
have a largish phoneme inventory at your disposal). But a
word -- a phonetic string that can be uttered in
isolation -- must meet language-specific pronounceability
criteria. Structural constraints apply to complete words
with all the affixes and inflections attached. Roots are
manipulated in the process, often losing or gaining a
segment. Constraints demanding that the "root" part of the
string should overlap at least one or _exactly_ one syllable
peak are not rare cross-linguistically. They help listeners
to parse the phonetic input into morphemes.

Of course we have disyllabic _stems_, e.g. *bHere- (the root
*bHer- plus thematic *-e-), and a root may likewise have a
disyllabic surface realisation in some cases. For example, a
root noun derived from {CReC} may have the Nom.sg.
*C&Re:C(-s).

Finally, it is hard to doubt the consonantal character of
the other laryngeals, especially *h2. How would you explain
the strange phonological conspiracy that makes *h2 and *h1
to have the same effect in virtually all environments, if
*h2 is a consonant but *h1 is a vowel? What I mean is that
e.g. -CRHC- and -CNHC- sequences develop in the same way
(with the same resyllabifications and compensatory
processes), no matter which "laryngeal" is involved. Either
we are dealing with some kind of truly miraculous
coincidence, or the "laryngeals" form a natural class.

Piotr