Km>gn (was: ficken)

From: stlatos
Message: 51782
Date: 2008-01-22

--- In, Piotr Gasiorowski <gpiotr@...> wrote:
> On 2008-01-21 01:00, stlatos wrote:
> > When I questioned other theories of his he and you gave different
> > answers. I'm not going to attempt to merge your answers in some way
> > instead of responding to each thing you say. For example, for words I
> > gave before that you said contained *u they were o-mo; neither you nor
> > Jens said they should be 0-mo within this theory and you even said
> > that *dhu:xmos < *dhouxmos was a possibility.
> Please consult his postings on this O-infix. He gives examples like
> *tog-áh2 vs. *bHug-áh2 and dHor-éje/o- vs. *k^ubH-éje/o-. He also
> explains the contrast between (older) *luk-éje-tor and (later,
> analogical) *louk-éje-ti.

I'm aware of it and responded to him in the past. My problems with
all aspects of it still remain. For example, you assume accent on a
separate syllable (since a C at that stage supposedly woudn't be
accented, part of the reason for the theory) but he just accounted for
contradictory examples (like tórmos) as 'accent retraction'
eliminating the most important reason to suppose C-infixation in his

> I do think something like *dHOwh2mó- > *dHu:mo- may account for the
> absence of laryngeal breaking in Greek, but that's my elaboration of
> Jens's theory. It applies only before laryngeals, with *Owh2/3 > *u:
> *Owh1 > *ou (as in *s(j)oumo- 'seam'). I don't even know if Jens would
> like it.

I'd be happy to hear any ideas. Anyone can respond to my theories,
you just happened to be the one to describe something I wanted to tie
into my recent message.

> > For this particular example, can you give one Latin noun in -gmus or
> > -gma or a reason why these in particular would be eliminated?
> Jens would predict *-mo- after a root like *tek- and *-no- after *pek-.
> That would make <pugnus> regular (phonetically, it had [-Nn-]).

The particular problem I wanted to talk about was Greek pugmé:. How
could that exist according to the timing of P-mn > P-n in PIE? The
evidence doesn't support mn. >> mos or any of the particulars.

> The
> levelling of the alternation to -n- in Latin after velars may of course
> be due to a language-specific change of *-gm- > *-Nm- > *-Nn-,

The timing requires, as I said, a change after a K stop. Since it
happens in similar rules in Gmc and some other branches it must be
part of a more extensive series of retroflexing (seen in *paL-x-mos >
pa:n.íH) which occurs in many different environments for other C also,
explaining much more than a few mo/no+ alternations.

If it happened after stop>+nasal/_nasal it's almost certain N, > eN
would have already occurred.

> but I
> wonder in what particular Latin words we would expect -gmus or -gma.

I'd expect SOME, at least one.

If you want to know what words I think had Km > Nn in Latin:

'striking, fist'
*pYugmó+ > L pugnus, pugna; G pugmé:

'sharp, point'
*xakYmó+ > L agna 'blade'; G akmé: '

'dripping, pool?'
*stagmó+ > L stagnum; G stagmos, stágma

Without such fairly close correspondences, but going on the model of
meaning for -mo- in G and L (for roots with other final C), probably
lignum, tignum, signum, re:gnum.

> The
> ideal candidates would be thematisations of neuters in *-m(e)n
> labial in the root. From *sek-, we get *sek-mn. > Lat. segmen, but this
> one accidentally fails to form a corresponding thematic noun, so we
> can't see what it would be.

What about signum?