Jens:
>I would take it up as a good suggestion if it had not been disproved
>over and over again. But yes, most ablaut was caused by the accent;
>and yes, unaccented short vowels used to be deleted. Only, that does
>not mean that all unaccented short vowels are of a younger make,
Strangely, you are only serving to back yourself _into_ my theories
since I do _not_ think that _all_ unaccented short vowels were
deleted! Please note the "paradigmatic resistance", as we seen in
*pedos (not **pdos), that I just grilled Miguel with! It is the
fact that we never see asyllabic stems in paradigm-derived forms
that shows us that, indeed, the *e here is probably from *& which
failed to disappear in Mid IE *p&d-as& despite being unstressed.
However, thematic vowels are just different from all this. Starting
with what we know about them...
Thematic vowels are found in both verbs and nouns. We see it just
as well in *bHer-e-ti and *bHer-o-mes as we do in *to-m and
*ekw-o-syo. Due to this, despite Miguel's pleas, thematic vowels
do not seem to have any inheirant meaning attributed to them.
There is no evidence that they are meaningful morphemes. Therefore,
the reason why they present themselves in so many differing words
must involve a trivial, subconscious process in the language,
rather than a conscious and meaningful one.
The thematic vowel alternation *e/*o appears to be best explained
by originally being the same vowel with trivial vowel lengthening
before voiced segments (something we see in a number of other
languages). The lengthened counterpart must have become *o and the
short counterpart became *e over time. At any rate, they were once
one vowel, to be certain.
So, they are not a morpheme and they were a single vowel, which we
may write as *&. Thus for thematic words like *ekwo-, the
antecedent form appears to be *ekw&-. The funny thing about these
thematics however is that they both have this "most peculiar" vowel
AND there accent has been secondarily regularized to the initial
syllable. Funny hunh? We know that it has been regularized for the
simple fact that it does not exhibit the more unintuitive pattern
seen in athematics like *kwon-. Surely, we would not pick the
acrostatic pattern as representative of anything ancient because
it is too regular to be real! (Or perhaps, Jens, you would for sheer
arguments sake...)
Given this, thematics and their thematic vowels in both verbs and
nouns show nothing but innovation. They do not exhibit the more
ancient pattern seen elsewhere and cannot possibly represent it.
So, we must suspect that the thematic vowel is also an innovation.
To conclude that it is most ancient ignores all the facts.
As I've said, this innovation is the misanalysis of genitival stems
as "thematic" stems plus a nominative *-s, which is why we still
have adjectives with final accent. This is because the acrostatic
regularization only operated on nouns and verbs.
>I have since realized I overlooked the obvious, namely its position:
>the them.vow. is the only vowel occurring in stem-final position. I
>am not sure what that exactly means, but it apparently imparted a
>special kind of resistence on the vowel.
It apparently changed the accent too :P Or rather, the more efficent
solution: The thematic vowel and all words formed by it are an
innovation of the Late IE period AFTER the loss of unstressed schwa
that had triggered zero-grading in late Mid IE.
>Anyway, it is a descriptive fact that vowels in this position alternate in
>a way all their own, and totally uninfluenced by the
>accent. [...] The thematic vowel shows alternations governed by the
>phonetic property of the following segment [...]. Since the VERY
>SPECIAL status of the thematic vowel must have its phonetic
>justification VERY FAR back in prehistory,
Correction: Since the "very special" status of the thematic vowel
doesn't operate under the earliest rules we know of in preIE, and
because it is otherwise surrounded by clear innovations (such as
the acrostatic accent) there is no logical reason to assume that it
is ancient.
>Therefore, the thematic type is not younger than the ablaut.
>If it were, we would find the same e/o alternation depending on the
>voicing of the following segment with other vowels also, which we
>do not.
No, the thematic vowel was not THE unaccented vowel of Late IE by
the time e/o alternation came into being. Miguel is helping me
piece together in my head the syllabic rules of each period with
his enqueries but perhaps I can show better by examples rather than
trying to come up with exact wording that covers every detail.
It does make sense, because only *& has observably lengthened before
voiced segments as can be seen by its results. However unaccented
vowels other than *& do not appear to be lengthened to any
appreciable degree to have formed the same alternation. So we see
the following happening (after acrostatic regularization):
mLIE PIE
*dekm *dekm 'ten' (unaccented *m)
*kunas *kunos 'of the dog' (unaccented *u)
*pedas *pedos 'of the foot' (unaccented *e)
*kW&s-y&: *kWesyo 'of what' (unaccented *&)
*bHer&nt *bHer&:nt *bHeronti 'they carry' (unaccented *&)
*ekw&m *ekw&:m *ekwom 'horse [acc]' (unaccented *&)
Notice how nothing but unaccented *& is lengthened before voiced
segments. Again, unaccented *e in *pedas results from the Mid IE
paradigmatic resistance rule which preserves the vowel of
paradigm-derived forms even when the root is unaccented. So this
totally covers all bases without logical conflict.
- gLeN
_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail