Qualitative ablaut - case is even more closed

From: Glen Gordon
Message: 4903
Date: 2000-12-02

> >>Any n.pl. in *-@2 (Skt. -i, elswhere, including Hitt., -a).
> >
> >Examples?
>
>Lat. <nomina>. Sigh.

No need to sigh, Miguel. I wanted you to provide a clear example to discuss.
In the case of Latin /nomina/, how can we be sure that our /-a/ hasn't been
re-introduced. Are you asserting then that Latin /nomina/ directly comes
from IE *no:mn-x?

> >Now, let's see... First we have Hittite /nekuz/ which looks like *nekWt-
>to
> >me, rather than *nokWt- in postAnatolian languages. Could *nokWt- simply
>be
> >a later o-grade variant?
>
>Or <nekuz> could be analogical after oblique forms with *e (N.
>*n�kwts, G. *nekwt�s; like N. *w�dr, G. *wedn�s).

Maybe. If what you say is true, we should expect Late MidIE *nakWt- from an
a-grade verb. Hmm.

>No, the acc. is *podm (*pedm), with short *o (*e). There's >lengthened
>grade only in the nominative (*po:ds/*pe:ds).

Well, I'm just going by Mr Burrow, the Sanskrit guy, who describes *pod- as
a "root noun". Other root nouns are things like *re:g "chief". I was under
the impression that they are lengthened even in the accusative... but then,
come to think of it... The Early IE nominative would have been anciently
*rek: -> *re:k: (monosyllabic) while the accusative would have always had
been *rek:-em (disyllabic). I suppose there is no motivation to lengthen the
accusative if it was always disyllabic. Okay, fine.

>I see nothing about tonal accent here, and I see *a: > *o, as was my
>suggestion, so I would have to guess you're saying I'm right?

Erh, not quite. I don't think we envision the same pre-IE. As for tonal
accent, I told you that accented schwa becomes *e and unaccented schwa
becomes *o, except in conjugation. The previous stress accent was
responsible for the loss of the V in CVC before accented syllable, producing
CC, not tonal accent. Yet, IE must have had a tonal stage to explain the
agreement of Greek, Sanskrit and BaltoSlavic accent systems. No?

>Indeed not. This is what I claimed to begin with: *w�dr < **wa:dn,
>*wedn�s < **wa:dn-�:s.
>
>But whatever happened to *o < unstressed *e?

I'm not sure what you claimed anymore. I just remember nonsense about **mwe
and **swesw > *yus- and the like which seemed to show that your views on
early phonology haven't been fully worked out yet. I believe you mentioned
that some long vowel like *a: became *e which is ludicrous. It would be much
simpler to presume a short vowel begat short *e. Why do you want to always
multiply hypotheses?

So we now almost agree... Except, you are perverting the sound
correspondances I've been so careful to establish. Your MidIE
**wa:t:n/**wa:t:na:s, if it were ever a possible form, would yield IE
**wo:t:r/**wo:t:no:s which doesn't exist. How do you justify long vowels
that aren't attested?

It's far better explained with alternating short vowels *a and *e (schwa) on
the root *wat:-/*wet:- (later *wod-/*wed-) which is evocative of other
alternating *o/*e patterns, by the way.

>The pattern is wholly illusory. The "stative" has *a because of the
>colouring by *h2. The endings are, quite regularly:
>
>*-h2-e
>*-th2-e
>*-0-e

Hittite [-ha(ri), -ta(ri), -a(ri)], no? The 3ps may not have originally been
*-e. Afterall, there is clearly an automatic semantic link between the 3ps
ending (whether *-e or *-a) and the pronominal *e which can offer confusion.
There would be special motivation to alter the 3ps if 1ps *-xa became *-a in
laryngeal-losing languages. Thus, the 1ps ending would merge with the 3ps
ending unless something was altered either before or after this loss of
laryngeal. Certainly, later there would appear to be 3ps *-e.

You will have to deal with the fact that the supposed **-e for 1ps/2ps
endings is pure assumption, based only on the later 3ps ending. Logically,
it could also be *-a. Afterall, if it were *-a, we might better explain not
only the later e-grade non-stative and o-grade stative forms as due to
conjugational vowel harmony but we can also describe the mediopassive
endings as an analogical extension of the *-a terminating stative endings.
Note *-t + *-a + mediop.*-r becomes *-tar -> *-tor (labialisation of closed
syllable short *a).

>The *-e was not affected by zero-grade, even though it's unstressed,
>which must mean it's the thematic vowel. In the "active" thematic
>conjugation, the thematic vowel comes _before_ the personal endings:
>
>*-o-m
>*-e-s
>*-e-(t)

The thematic vowel is the real illusion. It is a meaningless filler for the
majority of conjugated IE verbs. The *-o- in the 1ps is because of the
labial colouring I mentioned, caused by surrounding labial phonemes (in this
case *-m). I've further precised on this sound change which occured BEFORE
the shift of most *a's to *o:

The rule is unaccented *eC{w} -> *aC{w}
[*C{w} is any labial phoneme like *m, *p, *b, *xW...]

Thus a vowel harmonious *-em first became *-am by the above rule, followed
by a change from *a to *o producing *-om

>It's not unusual for languages that have a perfective/imperfective
>(active/stative, etc.) contrast in the verb to show different
>constituent orders in the endings (e.g. Semitic with impf. prefix, >pf.
>postfix conjugation, Basque with ergative subject postfixed, but
>prefixed in the past tense, etc.).

I'm not interested in whether this is unusual or not. I'm only concerned
with finding the underlying pattern to this particular grammatical
phenomenon in IE. Semitic grammar is different from IE and surely arises
from different prehistoric processes. The e-grade/o-grade fiasco seems to
bring me directly to the concept of conjugational vowel harmony and explain
it with some reasonability. Do you have a better idea that I should
consider?

Anyways, before you send out another message, consider the following list of
changes (in chronological past-first order) that I believe must have existed
in Late IE from 5000 BCE onward. This is after the accent became "mobile"
via the loss of final vowels in MidIE and after the establishment of *e/zero
ablaut due to a strong stress accent:

- stress accent to tonal

- *-�x -> *-�:
*dwex "two" > *dwa:
*(e)kWtwex "eight" > *(e)kWtwa:

- e/o ablaut (unaccented *e > *o)
- resistance in vowel-harmonious conjugation
where e-grade is non-stative
and a-grade is stative.
- plural *-es remains by analogy
with *-mes & *-tes.

- *eC{w} -> *aC{w}

- shift of *a to *o
*�: -> *o:
*� -> *o
*aC -> *oC
*-a -> *-a (remains)

- *-o: -> *-o:u
- dual and locative in *-u arise
- new formations like *p�ku and *g�nu
- locative *-i from *ei "here"
- dative *-ei

- accent regularisation in athematic/thematic
- thematic gets initial accent
- athematic gets accent on
non-nominoaccusative case endings
- adjectives keep old accent

- *-om -> *-o:
- genitive in *-o: created (later instrumental)
- 1ps thematic *-o: from *-om

- *ego: replaces *meu
- meaning "I am here"
from *e-ge- (*e "this") plus them.1ps *-o:

- indicative *-i established from affixed *ei "here"

- 1ps them. n-i. *-o: reestablished as *-om


There. Isn't that beautiful?

- gLeN

_____________________________________________________________________________________
Get more from the Web. FREE MSN Explorer download : http://explorer.msn.com