*gwh in Gmc.

[Douglas G Kilday]

[ moderator edited ]

Hans-Werner Hatting (6 Feb 2001) wrote:

>MCV wrote:

>>My original examples were: "liver", "four", "-leven, -lve", "oven",
>>"wolf", "leave"(?), "sieve"(?).  There's a labial in "wolf".

>There is also a labial in _four_ (PIE *kwetwor-). But I take Your point.
>Labials in the neighbourhood alone are not sufficient as an explanation, as
>they don't account for Your other examples.
>So, what are our choices?

>1. To accept these as cases of „untriggered sporadic sound change“, which
>is of course not satisfying;

>2. To try to extend the triggers for a sporadic sound change *kw > *p. One
>obvious candidate would be /l/. This would still leave „oven“ and
>„sieve“ unaccounted for. As I am without any library for the time being,
>what are the etymologies proposed for these words?

"oven" < OE <ofen> < PGmc *ufna- < PIE *aukwna-
cf. Lat. <o:lla> 'cook-pot' < <aulla> < PIE *aukwsla-

"sieve" < OE <sife> < PGmc *sif- < PIE *seikw- 'to flow'
cf. Lat. <siccus> 'dry' < PIE *sik(w)ko- 'flowed-out, dried-out'

>3. We could assume substrate influences or a dialect mixture in Germanic or,
>in other words, a mixing of features from neighbouring dialects, like in,
>e.g., the German dialect of Cologne, where we generally have the development
>/t/ > /ts/, /s/ (e.g. _zick_ /tsik/ „time“, NHG „Zeit“, but /t/ is
>kept in some function words like _et_ „it“, _dat_ „that“). But
>substrate and dialect influences are, of course, something of a „magic
>wand“, if there is no further evidence for their existence.

We use dialect-mixing to account for Lat. <lupus>, <bos>, etc. so what we
need here are "p-Gmc." and "q-Gmc." dialects. The ancestor of all attested
Germanic would be, like Latin, primarily a "q" dialect with admixture of
some words from "p" dialects. The problems are that numerals are seldom
borrowed between dialects and there is no known prehistoric Germanic
parallel to the spread of Latin, which resulted in the extinction of the "p"
dialects.

A variation on this theme is a three-stage model for the
Indo-Europeanization of the pre-Germanic population. In this scenario, the
first stage of contact between pre-Germans and IE-speakers resulted in the
borrowing of a few IE words into pre-Germanic. These few words, belonging to
a small set of categories, were not enough to influence pre-Gmc. phonology,
which lacked labiovelars and replaced them with (labio-)labials. The second
stage of more intense contact brought a large influx of IE words in which
the distinction between labials and labiovelars could not be ignored, so
pre-Germanic acquired the labiovelars along with the words. In the third
stage, the grammar was largely Indo-Europeanized under extensive mixing of
populations, and pre-Gmc. became Proto-Gmc.

This hypothesis requires justifying the early borrowing of words having Gmc.
labials for IE labiovelars, particularly the numerals. Many aboriginal
languages have quite lengthy numerals, which is no problem as long as
transactions are carried out by on-site negotiation with visible goods.
Short numerals have an advantage only for counting, and counting is only
useful when society has adopted the concept of "price" in terms of units of
currency (such as cattle). It is unlikely that IE-speakers introduced
herding to the pre-Germans, but they may have introduced currency and
counting, and pre-Germans would have quickly learned the convenience of the
short numerals. "Leave" is used in forming "eleven" and "twelve" and was
probably borrowed along with the numerals, since it is very useful in
transactions.

"Oven" and "sieve" might represent two of the useful novelties which
IE-speaking merchants traded to the pre-Germans. Awhile back I suggested
that "warm", "snow", and "Niere" might have had *bhw originally. In the
current hypothesis, PIE *ghw became pre-Gmc. *bhw in the first stage of
borrowing, but remained *ghw in the second stage. From *bhw came *vw and
then *w, from *ghw came *Gw and then *G/g except before *t, where it was
assimilated to *x/h. "Niere" and "liver" then belong together as
organ-names; perhaps the IE-speakers sought the organs of certain animals
for ritual or medicinal purposes. "Warm" could have been borrowed along with
"oven". I can't explain why "snow" would be borrowed into pre-Gmc. unless it
figured in compounds denoting novel products. "Wolf" is controversial; both
labial and velar are found in ON <ulfr> m., <ylgr> f. from Proto-Gmc.
*wulfaz m., *wulgi' f. In this hypothesis the IE root might have been
borrowed twice: once in the "first stage" when IE-speakers were trading for
skins, and *wlkw- became *wlpw-, then again in the "third stage" without
labialization of the stop when a feminine form was required.

It should be noted that the specific words mentioned do not constitute *all*
the first-stage borrowings in this hypothesis, but only the ones with
labiovelars in PIE. All the numerals into the hundreds, not just "four",
"five", "eleven", and "twelve", were presumably borrowed in the first stage,
as well as other organ-names and terms for products.

>4. We reconstruct a new series of phonemes for PIE, as has been proposed.

>My problem with approach no. 4 is that I don't know of any evidence for such a
>series other than from Germanic. If we assume that Germanic branched off
>earliest (a problematic assumption in itself), we would not expect such
>evidence, but in the scenario Douglas Kilday describes, we would expect some
>traces of the /pw/ series in Anatolian. And, I don't want to repeat myself,
>but I think the sound change /pw/ > /kw/ is not trivial – I would expect
>different outcomes in different branches of IE languages, not a simple split
>into a language keeping the series distinct, and the other ones merging them.

Yes, these are strong objections to the scenario I described earlier,
particularly the difficulty with /pw/ > /kw/. The new hypothesis avoids this
at the expense of introducing fresh assumptions about the origin of Germanic
and its position within IE, to which I expect further objections to be
raised.

DGK