Re: [tied] Re: Eggs from birds and swift horses (was: the palatal s

On Sun, 15 Feb 2004 14:52:48 +0100 (CET), Jens Elmegaard Rasmussen
<jer@...> wrote:

>>>--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
>
>>
>>>I do not
>>>think the verbal desinences contained the w-element of the personal
>>> pronouns. If you write it into the original forms, you have to
>>>construe a whole arsenal of additional rules to get it out.
>>
>> I think the verbal desinences *did* contain the w-element of the
>> personal pronouns.  Thats why the 2sg. ending is *-s (< **-tw) and not
>> *-t, as we would otherwise have had.
>
>I know you do, that's part of the arsenal I was referring to. I respect it
>in so far as yours is not an analysis without basis in profound thoughts.
>Still, there are other s/t cases and I don't think their combined evidence
>adds up to /tw/.
>
>The relation between *-s and *-t is not so easy; many cases of *-s are
>rather plainly from older *-t, or from a common source of *-s/*-t-. There
>are not many cases of IE *-t, and if the 3sg active *-t is connected with
>the agent-noun suffix of Ved. is.u.bhr.'-t- 'arrow-carrier', then it has
>an allomorph with *-n- and even *-nt-, which opens a whole new ballgame.
>These are facts that ought to be considered, and they do not lead to /tw/
>as the origin of the second person marker in any obvious way.

Final *-t occurs in the 3sg. verbal ending (as well as, obviously, in the
3pl. *-en-t). It also occurs in the ablative ending *-ot and the
instrumental ending *-ét, although the latter became *-éh1 (except in
Hitt.). It also occurs as the final element of the stem in a number of
(neuter) nouns, such as *melit and *h2alut. Sanskrit adds *-t to root
nouns ending in *-i, *-u and *-r. (I'm not sure therefore if -bhr.-t can be
described as an agent noun: in other cases (jit "conquering", s'rut
"hearing", kr.t "making"), the meaning is that of a verbal noun), but I'm
not sure whether this can be derived from PIE itself. Then there are also
the nouns *ye:kWr.t "liver" (Skt. yakr.t, Arm. leard), *k^okWr.t (Skt.
s'akr.t) "excrement", *sneh1wr.t "nerve" (Arm. neard), with final *-rt (<
**-nt). I believe the pronominal neuter suffix *-d is also derived from
*-t, with the same voicing that we see in nom.sg.anim. *-z. I might be
overlooking something, but I'd say there is no lack of evidence for *-t
(unlike, for instance, for *-p or *-k(^)).

>>>> At the risk of repeating myself, I prefer to explain the *-s as a
>>>third
>>>> person ending, derived either from [the nominative **su of] *s(w)e
>>>(which
>>>> was a 3rd. person pronoun before it became a reflexive) or the
>>>> demonstrative *so. 
>>>
>>>If the point is that it makes sense as a reflexive, it specifically
>>> does *not* have nominative function.
>>
>> The point was that pre-PIE had matching 1, 2 and 3rd personal pronouns
>> **mu, **tu and **su (cf. Proto-Uralic *mV, *tV, *sV).  The third
>> person pronoun later changed into a reflexive pronoun (cf. Old Tamil
>> ta:n_ 3rd. person pronoun -> Mod Tamil ta:n_ reflexive pronoun), and
>> lost its nominative.  The reflexive accusative remains parallel to the
>> 1/2 pronouns (*me < *mwe [or *mme], *twe, *swe).
>
>Later than what? The sibilant is reflexive in Eskimo-Aleut, so that must
>be its original function.

Why? The sibilant is 3rd. person non-reflexive in Afro-Asiatic (*suwa,
*siya), Kartvelian (3rd. person S marker *-s), Uralic (*sV, poss.
*-sa/*-sen), Altaic (Turk. 3rd. S-marker/possessive *-sI, *-sen).

>It is as if you are trying to show me that a
>rfl.pron. cannot become a 3rd person pronoun, but it can.

Actually, I was trying to show that a 3rd.person pronoun can become a
reflexive.

>The possessive
>of the reflexive has turned into possessive of 3sg masc. in German sein,
>and into possessive of 3rd person sg. in general in French son. For the
>non-possessive Spanish se is well on its way to replace le under special
>consitions.

Except that *that* se is from *lle.

>So if the s-pronoun of Uralic is connected with the IE
>reflexive, the original function was reflexive, since that function
>reappears in the next branch, Esk.-Al.

Maybe the next branch is Altaic.

>>>Tocharian has many verbal stems that can only have arisen in the
>>> sigmatic aorist. And Hittite at least has some.
>>
>> According to Adams, the s-aorist survives in Tocharian only in the
>> Class III preterites.  Those are precisely the ones I meant, which
>> have -s(a) as
>> a third person singular desinence.  I reject the notion that they are
>> survivals of the s-aorist: they represent the seed from which the
>> s-aorist
>> grew.  In fact, Tocharian itself created an s-middle by extending the
>> /s/ (augmented with -a:-) throughout the paradigm in the middles of
>> the Class III verbs.
>
>> I don't think the thematic -se/so- presents are relevant in this
>> context. They are thematic, they're presents, and they're hard to tell
>> apart (at least in Toch A) from *sk^e-presents.
>
>They fall into place within the system we find in the other languages if
>the se/so-presents are aorist subjunctives. They have the exact shape of
>such forms, and a change from aorist subjunctive to present indicative
>is very common indeed, perhaps it is even the most common source of the
>e/o-presents.

Adams rather connects them with desideratives outside of Tocharian, not
with s-aorist subjunctives. The s-aorist subjunctives he associates with
imperative class III (s-imperatives).

>You do not know that "Tocharian itself created an s-middle"; the type
>corresponds to the middle voice of the s-aorist elsewhere, so what is the
>grand demonstration of a Tocharian innovation based on?

On Tocharian itself. There is no /s/ in the active (except 3sg.).

>>>The sigmatic aorist
>>>is not an outgrowth of the root aorist, but the regular aorist of
>>> verbs
>>> forming an sk-present.
>>
>> In Hittite at least (and, according to Adams, also in a pre-stage of
>> Tocharian, see "Tocharian", p. 76), *every* verb could potentially
>> form a sk-present.
>
>That looks more like an innovation, given the clearly lexicalized status of
>the present formations in the other languages, frequently matching those of
>Tocharian. In that system, lexicalized sk-presents are combined with
>lexicalized s-aorists often enough to represent a rule. In Tocharian
>itself, s- and sk-presents quite generally form prt.III, which may here be
>taken to reflect the s-aorist (though it also continues the perfect).

I'm not convinced Hittite is an innovation. Adams makes a plausible case
for Tocharian having inherited exactly the same system (where *all* verbs
could in principle make an iterative), with the iteratives subsequently
becoming causatives, unless they had previously displaced the present stem
(which then in turn became a subjunctive or an imperative).

>>>That makes the -s- of the aorist a
>>>suffix, not a desinence, from the beginning.
>>
>> Suffixes can become desinences and desinences can become suffixes. 
>
>But how would you get a desinential *-s into the position it occupies in
>the sk-presents?

I don't. I'm not convinced that I have to.

>Why would the putative desince have spread to all forms
>of the derived durative companion before it was accepted as part of the
>stem in the aorist itself? That is so peculiar that it demands a strict
>separation of the s-aorist from the sk-presents. And that just seems
>impossible, given such pairs as
>
>*pré:k^-s- : *pr.k^-sk^é-ti
>*H2é:y-s- : *H2is-sk^é-ti
>*g^né:H3-s- : *g^n.H3-sk^é-ti
>*yé:m-s- : *ym.-sk^é-ti
>*g^é:r&2-s- : Gk. ge:rásko:
>stative *-eH1-s- vs. *-eH1-sk^e/o-

That looks good. But if *-sk^é/ó- could be added to *any* verb, as in
Hittite (and pre-Tocharian in Adam's analysis), and if the s-aorist is the
productive aorist (as opposed to non-productive root aorists), I wonder if
I could make a similar list juxtaposing, s-aorists and, say, n-presents?
(I may try that later, now I have no time).



=======================
Miguel Carrasquer Vidal
mcv@...