--- In
cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
> I don't follow. -i is *always* the 1sg. middle past ending.
> It has nothing to do with the accent or the Ablaut. Why
> isn't the 2sg. *á-mam.s-ti then?
Sure, but the alternation between -i and -a as the ending of the 1sg
middle must have some basis, and unaccented -i vs. accented -á can
really be found. The propagation of -i to serve also in some places
where it cannot be original should not prevent us from spotting its
point of departure. Surely áduhi cannot reflect an original "*dhugh-
&2", nor ábhare an original "*bhér-a-&2". Of course one would have
liked to see a parallel alternation in the 2sg, but there just is
less allomorphy than the rules prescribe throughout the language
(perhaps throughout any language).
> >> >Phonologically the added vowel of the perfect could well be
the
> >> >thematic vowel, but that of the middle voice cannot.
> >>
> >> Why? As I said above, the vowel and the accent and the
> >> Ablaut and everything behave consistently with a thematic
> >> (-0-é-) paradigm.
> >
> >Because it has a zero-grade alternant which is something we do
not
> >have with the thematic vowel.
>
> What zero grade alternant? Do you mean zero grade of the
> root, or of the "thematic" vowel?
The "it" is the final vowel of the endings of the perfect and the
middle voice. The vowel ablauts in the middle voice, but not in the
perfect. Since the thematic vowel is ablaut-immune it cannot be
identified with the vowel of the middle-voice endings, whereas it
could conceivably be the final vowel of the perfect. But if you are
right when you say it also ablauts in the perfect, then that cannot
be the thematic vowel either.
> >> >I completely fail to se what an object marker would be doing
here.
> >>
> >> Marking a third person object, what else?
> >
> >Something which the perfect really denotes. Something like "...
and
> >still do" would be nice. It is not consistently transitive, not
even
> >typically so. And objects are not known to be marked this way in
> >this particular language, so it does not really fit anything.
>
> The premise is that the thematic vowel in the active
> paradigms (where it doesn't mark the subjunctive) _also_
> marked a 3rd. person direct object. As I have already
> admitted, this is not provable for the active.
And it is distinctly strange for the middle voice.
> Still, *if*
> this particular language (some stage of pre-PIE) marked the
> object in this way, then I see nothing amiss with it marking
> the object in the perfect/middle in a similar (but slightly
> different) way.
Sure, and if it was something else it marked, then there would be
nothing strange about having it do the same elsewhere. That's a pure
tautology.
> >> Neither the perfect nor the middle have exclusively
> >> intransitive semantics in PIE. Generalization of
> >> incorporated 3rd. p. object to all forms would not be a
> >> surprising development.
> >
> >I would indeed be surprised. You don't prove anything by playing
> >blasé.
>
> You've been talking with Glen too much. Let's stick to the
> factual arguments.
This is not personal: If I fail to understand something, I would
like to have it explained so that I do, not just having it stamped
as not surprising to those who know better but won't or can't
explain it.
> >The forms in /-e:-/ are really unproblematic, for that was the
form
> >of the subjunctive of the thematic conjugation which is what this
is
> >expected to reflect. The 1sg -am is a bit odd, but the inherited
> >form would be in -o:, i.e. identical with the indicative, so a
> >transfer of the form -am from the subjunctive type feram is
easily
> >motivated.
>
> And where does _that_ come from?
You shoot that out as if it were a central point in the argument
which it isn't. All it takes is that the a-subjunctive exists. I see
a few possibilities:
1. The type fuam fua:s fuat, in origin the injunctive of the aorist
stem *bhwaH- (*bhuH-), which was used as a subjunctive, caused
reanalysis of the segment /-a:-/ as a subjunctive morpheme. Against
it may be objected that we would like to use the same reanalysis to
explain the /-a:-/ of eram, era:s, erat as a *preterite* preterite
morpheme segmented off from the aor.inj.
2. It was the thematic subjunctive of H2-final roots (thus
Rix/Meiser). I fail to see a distributional support for this.
3. It was the same form as the Celtic a:-subjunctive, OIr. beraid ~
Lat. ferat. The Celtic formation however has now acquired an
explanation which is not easily transferred to Italic. Since the a:-
sbj. alternates with the s-sbj. according to root structure, it is
attractive to derive them both from *-&1s-e/o- ~ *-H1s-e/o-, which
would be parallel with the type in *-&1s-ye/o- ~ *-H1s-ye/o-. Sihler
does this (basically preceded by Thurneysen); McCone spells out the
same alternation, but will have it explained by analogy. An
intelligent rescue operation was worked out by my student Anders
Jørgensen who asked the question, What would it take? If OIr. beraid
is from *bher-&1s-ye-ti via *beraseti > *bera:Ti in Irish, then a
parallel explanation of Italic *fera:t (or perhaps *feraet) would
have to have passed through *ferayet with assimilation of *-sy- to *-
yy- and subsequent simplification before loss of single *-y- between
vowels. It seems that other examples of *-sy- in Latin, which show a
result /-yy-/ as eius, cuius, have the cluster after the first vowel
of the word, so that it is imaginable that the cluster was reduced
after non-first vowel in which case the forms can be regular and
congruent with those of other branches. Why can't I get the gifted
young scholar to publish this ingenious analysis? I think it
deserves everybody's attention even if it should turn out to be
wrong. I deem it quite likely.
Jens