[tied] Re: Lith. žinóti - why not a root g^neHH-?

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2009-10-13 19:19, alexandru_mg3 wrote:
>
> > a) Could you explain what is the distinction between 'based-on' and
> > 'influence' here:
> >
> > for 'influence' --> I could understand 'analogy'
> >
> > But how 'to base' the *R(z)-sk- verbal formations on acrostatic stem?
>
> *R(e)-sk^e- presents are sometimes paired with *R(e:)-s- stems (not
> necessarily aorists; there were also marginally preserved s-presents).
> Cf. Lat. pa:sco: < *pah2-sk^e/o- and pa:stor < *pah2s-tor-. It has been
> suggested that *-sk^e- presents are ancient (pre-PIE) *-je/o-
> derivatives of *-s- stems, though the phonological details of the
> development are problematic. According to Jasanoff, the original form of
> the sigmatic suffix was *-h1s- (as in desideratives and futures). If so,
> perhaps we could posit a pre-PIE metathetic develoment like *-hsj- >
> *-shj- > *-sç- > *sk^-.
>
> > Supposing that we accept R(e:)-sk-, R(e)-sk- forms on PIE times :
> > another questions is : could we have 'in the same time' both of them :
> > *g^ne:h3-ske/o (> njoh) and *g^neh3-ske/o ((g)no:sco:)?
>
> We could have *g^ne:h3-s-/*g^noh2-s- and *g^noh3-sk^e/o-, with
> *g^ne:h3-sk^e/o- as an analogical byform. Why not? Real languages often
> tolerate competing forms "in the same time".
>
> > No link to this: but I see no reason, not to have *g^nh3-ske/o-
> > formation here too: but the reality is quite opposite to this : no form
> > *g^nh3-ske/o- is attested in the derived languages...
>
> Gk. gnó:sko: is ambiguous; it could reflect the zero grade, but I agree
> that the bulk of the evidence points to a full grade.
>
> > b) what is your opinion about žinóti ? I mean the source of a: inside?
>
> It looks to me like an original nasal-infixed present, *g^n.-ne-h3-,
> reworked under the analogical influence of the far more numerous verb
> stems in *-ne-h2-.
>
> > c) do you think that CEHH- roots are theoretically possible?
>
> I can't think of any certain examples. PIE had rather strong
> morpheme-structure constraints allowing only a few types of clusters in
> this position. If Jasanoff is right, *g^ne(:)h3-s- could be rewritten as
> *g^ne(:)h3-h1s-, but here of course there is a morphological boundary
> separating the laryngeals and the second one wouldn't have been realised
> phonetically.
>
> Piotr
>

Thanks, Piotr.

Only for the sake of the argumentation:

WHAT-IF:
a) the root was *g^neh3h1- and
b) one of the laryngeals was 'sometimes' vocalized and 'sometimes' lost in different contexts

Some Results:

I) *g^nh3.h1-sk^e/o- (so a regular R(z)-sk^e/o-) would have been given Grk. gnó:sko: (h3. > o) and maybe even the Latin (g)no:sco: (h3. > o)
The same *g^nh3.h1-sk^e/o- would have been given regularly Alb. njoh < *gna:sk- < *gnah1sk- (with h3. > a)

II) If h3. > a in Baltic: *g^n.-n-h3.h1-ti- would have been given regularly žin-nah1-ti- > žina:-ti > žinoti


III) For any o-grade formation I expect that *g^noh3h1- > *g^noh3- (based on the lost of laryngeals in o-grades)

IV) For Skt. maybe we need to consider that h1 was lost in some other contexts than o-grades?.....

Seems that a lot of 'exceptions' could be explain in this way...

Marius