*-iH_2 [was Re: Domesticating the Horse]

[Richard M. Alderson III]

On Wed, 1 Mar 2000, Pat Ryan (proto-language at email.msn.com), following up a
query about a form in a Stirling post, wrote:

>>> By the way, Pokorny lists *u.lkwi: for 'female wolf'.  Where do you get
>>> *ulkwiha?

> <MCV>
>> For <ha> read <h2> (or <A>).

> <PR>
> Well, yes, we could *assume* that it is a case of utilizing both formants in
> combination, but how do we know?

By "both formants", I'm assuming you mean your notion of *y and *H_2 as
feminine formations.  I won't address that.

> Could it not equally well be *u.lkwiH1- or *u.lkwiH3-?

> And why would not *ulkwiH(2)e develop into *u.l/wl.kwia:  ?

Last question first:  Vowels are only lengthened with the loss of a *following*
laryngeal, so if the proto-form were ?*ul{k^w}i{H_2}e, the outcome would have a
short vowel *a, not a long *a:.  However, the suffix is *-i{H_2}, with no
trailing *-e.

For the remainder of this post, I am going to write just the numeric portion of
the TeX constructs {H_1}, {H_2}, and {H_3} to represent the e-, a-, and o-
coloring laryngeals.  This should make things easier to read without added
ambiguity.

The feminine-forming suffix *-i2 is reconstructed based on Skt. -i:, Gk. -ia,
etc.; Latin has extended the form with an additional suffix, -g-, as in
_stri:x, stri:gis_.  The Greek outcome is especially interesting, since this
is one of the few places where the laryngeals do not undergo fully parallel
developments: We find *i1 > _i:_, *i2 > _ia_, and *i3 > _iO:_.

This last fact, of course, answers the question of reconstructing ?*ul{k^w}i1
or ?*ul{k^w}i3:  Were the suffix not *-i2, we would have a different outcome
in Greek.

								Rich Alderson