Re: [tied] 1sg. -o: [was Re: IE Thematic Vowel Rule]

Miguel Carrasquer wrote:

> So that would be the "double thematic" subjunctive of
> thematic verbs, which I suppose would have spread from the
> thematic subj. to the athematic subj. as well as to the
> thematic present indicative?

Sorry about my careless wording. As Mr. Jingle would have put it, "No
time -- much to say -- post in haste -- repent at leisure." Of course I
mean the "superthematic" subjunctive of thematic verbs and the
replacement of the expected **-o-m(i) (in both the thematic present
indicative and the subjunctive of athematic verbs) by that form.

> My solution is that the consonant in the 1st. person was *mW
> (and *tW > *sW in the 2sg., from earlier agglutinated
> pronouns *mu and *tu, respectively), which explains 1pl.
> Hitt. -wen(i), 1sg. Luwian -(a)wi and PIE 1du.
> *-wah2/*-wh2a(s). The present thematic ending *-o-mW-i
> became *-omWu > *-owu > *-o:u (like u-stem loc.sg. *-ew-i >
> *-ew-u > *-e:u), which would explain the Tocharian B 1sg.
> thematic ending -eu (< *-o:u), with irregular reduction to
> *-o: everywhere else. The Tocharian B 1sg. subjunctive -u
> must come from *-ou, which is harder to explain (*-omW
> should give (and gives) *-om, not *-ou).
>
> Neither your theory nor mine adequately explains why the
> Lithuanian 1sg. ending -ù is formerly acute (*-uó before
> Leskien's law). A double thematic *-oo(m) would inevitably
> produce a circumflex, and so would my *-o:(u), judging by
> akmuõ < *-o:(n), etc.
>
> As far as I can tell, only *-oh3 could have produced the
> required acute.

But what if the loss of *m was pre-PIE in the sense that at the time of
the IE breakup a plain long *-o: was the only realisation of the ending?
There would have been a contrast between this *-o: and later
contractions, or forms retaining a final consonant (even if the latter
could be dropped in sandhi). One could compare the 1sg. ending with the
animate thematic nom./acc.du. *-o:, if from analogical *-o-e. The Baltic
development is the same.

One question that should be asked at this point is why the early loss of
*-m after a long vowel was favoured in the 1sg. while a final *-m
survived in such cases as *d(H)(e)g^Ho:m or the gen.pl. *-(o-)-o:m . One
possible factor is the fact that verbs typically occupied the
clause-final position in PIE, and so could be subject to special
treatment by rules of sentence-level phonetics. A possible parallel may
be offered by Gk. do: 'house', which is always verse-final in Homer.

Almost needless to add, I think the variation in the 1sg. pronoun,
*h1eg^om ~ *h1eg^o: merely echoes that of the verb endings.

Piotr