On Wed, 18 Feb 2004, Miguel Carrasquer wrote:
> On Tue, 17 Feb 2004 21:55:23 +0000, elmeras2000 <jer@...> wrote:
>
> >> The question was: if we
> >> "know" that the PEA 2pl. ending *-ci (containing the PEA phoneme
> >*/c/,
> >> which is different from */t/, or */d/) derives from something with
> >**t
> >> higher up (we can get this result by internal reconstruction,
> >and/or by
> >> comparison with Uralic), then is there a reason *not* to think
> >that other
> >> items containing */c/, such as the reflexive suffix *-c (> -ni),
> >*also*
> >> derive from something with **t "higher up"?
> >
> >Yes, the compelling reason is that //t// and //c// do not show up
> >as /c/ on the same level. You are mixing French and Latin.
>
> You'll have to explain it a bit more clearly for me to understand.
>
> The ordinary reflexive (abslutive sg.) is -ni < *-n~ < *-c, so because of
> its position in the Auslaut, the /c/ is difficult to compare directly
> with
> the /c/ in the 2pl. suffix -/ci/.
No, because of the position, the reflexive -ni from *-ñ < *-c is
very easy to compare with 2sg -n from *-t.
The dual forms are less clear, but still in final analysis safe enough.
The CYup. forms of the inergative are -t&k/z&k of both 2nd person and
reflexive, but the ergative contrasts -pt&k and -m&k, pointing to a weaker
consonant in the reflexive than in the 2nd person.
The contrast is repeated in the rest of the system:
2pl ie. -ci/-zi, erg. -pci
refl.pl.ie. -t&ng/-z&ng, erg. -m&ng
not a good case for identifying 2nd person and reflexive.
It goes on with plural posession:
2du.ie.pl. -t&k (*-d-t-g)
2du erg.pl. -pt&k (*-d-m-t-g)
refl.du.ie.pl. -t&k (*-d-c-g)
refl.du.erg.pl. -m&k (*-d-m-c-g)
2pl.ie.pl. -ci (*-d-t-d)
2pl.erg.pl. -pci (*-d-m-t-d)
refl.pl.ie.pl. -t&ng (*-d-c-d)
refl.pl.erg.pl. -m&ng (*-d-m-c-d)
Aleut does not have much, but it does show where the ng-forms belong:
2du.ie.sg. -dix (E -t&k/-z&k : *-t-g)
2du.erg.sg. -mdix (E -pt&k : *-m-t-g)
refl.du.ie.sg -dix (E -t&k/-z&k): *-c-g)
refl.du.erg.sg. -max (E -m&k) : *-m-c-g)
2pl.ie.sg. -ci (E -ci/-zi) : *-t-d)
2pl.erg.sg. -mci (E -pci) : *-m-t-d)
refl.pl.ie.sg. -dis (E -t&ng/-z&ng) : *-c-d)
refl.pl.erg.pl. -mang (E -m&ng) : *-m-c-d)
The lenition product of /t/ is expteced to be -d-, but we find -z- in
Eskimo. However in the reflexive it *is* expected to be -z-. Conversely,
in Aleut we find -d- for both, which is expected only in the 2nd person.
The two voiced spirants must have been so close that they could be
confused. In other cases the lenition product of -t- is in fact -d- in
Eskimo (e.g. ptc. -tuq/-duq, trans. -taR-/-daR-), and of course -t- and
-c- are opposed to each other in their unchanged form, and also when
geminated.
This should make it clear that *-t- and *-c- are opposed to each other on
all levels. There are minor changes due to articulatory contact and
phonetic similarity, but nothing that really remains enigmatic.
> We have, however, the reflexive dual
> and
> plural, which you said were -z&k ~ -t&k and -z&ng ~ -t&ng in C. Yupik, so
> I
> assume *-c&g and *-c&d in PEskimo. I don't know what the forms with
> -/t/-
> are. The 2pl. suffix in C. Yupik is -zi, I think, from *-ci.
>
> So as far as I can see, the reflexive is *-c- in PEsk, -z- in C. Yupik,
> and
> the 2pl is *-ci in PEsk., -zi in C. Yupik. I think I'm comparing Latin
> with Latin and French with French. On the other hand, it worries me that
> you gave Aleut abs.pl.refl. -din/-dis, whereas the 2pl. in Aleut is -c^i,
> so there's a difference there.
The C. Yupik reflex of PEsk -c- is -c- (its 2pl.ie.sg. is -ci/-zi, with
inherited lenition after vowel). Like some other dialects, this one has
both, -c- and -z-, for it inherited both from an already-existing
alternation which is also reflected in Aleut.
Jens