'ice' and 'pot' (RE: epenthetic glide.)

[Koontz John E]

On Tue, 24 Jun 2003, Rankin, Robert L wrote:
> In the other cases like 'ice', 'kettle', etc. it may help to look at
> the Ofo cognate to see if there was an initial syllable.

For 'ice' the only SE cognate is Tutelo noNxi.

For 'pot' (under 'kettle')  Biloxi and Tutelo have various longer forms in
initial yes- and the CSD offers PSE *yes-.

The Dhegiha forms are PDh *Ree'ghe OP ne'ghe, Kaw c^ee'ghe, Os ce'ghe.
For Winnebago-Chiwere PWC *Ree'xe, IO de'xe, Wi dee'x.  Presumably all
these vowels should have been heard long, though the sources don't always
support that.

Dakotan has c^hegha, of course, and Mandan we'rex(e) (without or without
the e-absolute marker).  CH is PCH *wira'xa, Cr bila'xa, Hi wira'xa.

The CSD offers *yeSE as a reconstruction, where S indicates fricative
symbolism grades, and E signifies that the final vowel is the e ~ a set of
disputed nature.

If PS were actually *wVyeS-, maybe *wiyes-, that might explain a lot.  We
don't know alot about PMV *py (*wy) < PS *wVy-, but there is this.  The
inflection of *DEM=...ye 'to think' (Da epc^a 'I think', a "defective"
stem in traditional temrinology - with only this first person; OP
ebdhe(gaN)) shows that Dakota py for *wy and Dhegiha has *bdh.  A few noun
sets with *py and *ky show that Dakota tends to lose initial p in pc^, so
PrePDa *wyegha might well come out pc^hegha ~ c^hegha.  Dhegiha has bdh
for *pr in verb inflection and verb stems, but *R in nouns in initial
position, as in ne 'lake' vs. Da (Te) ble.  If Dh treats *py as *pr in all
contexts, then PrePDH *wyeghe would come out *Reghe.  Since WC works like
Dh wrt *wr (and, I think also *wy), it has *Rexe.  CH has forms like PCH
*wiraxa, where it looks like x has lowered preceding e.  Maybe SE lacks
wi- before *y?

This doesn't help with Dhegiha wi- < *w-yi A1-P2 in transitive verbs.  It
looks like those should have *Ri-, which they don't.

This argument is not entirely new, as the Bob's discussions show.  I think
this might be one of those cases where the editors didn't quite get their
ideas into the reconstruction.

Anyway, this is certainly a case where it pays to notice that *R reflexes
are sometimes fairly late developments of *r (or *y) in clusters.

JEK