Re: Ir. cas(s) and IE models (was: Ligurian)

--- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
<bhrihstlobhrouzghdhroy@...> wrote:

>
> > AFAIK, Irish is a Celtic language, so as a matter of internal
> > *coherence* we should first reconstruct a Proto-Celtic etymology

before

> > doing that for "PIE".
>
> I'd have expected such a statement by a Neogrammarian. Such
> statements always imply models. In my model (as in yours, I suppose),
> PIE had split into hundreds of palaeodialects;
>

But in my model the real PIE (I'd rather call it paleo-IE) has little to
do with Neogrammarian's "PIE" (notice my use of quotation marks).
Neither I think there were so many paleo-dialects.

> some of them later
> coalesced into Irish IE, which in turn came to be part of

Proto-Celtic.

>

Irish (actually Goidelic) must be diachronically younger than
Proto-Celtic, as otherwise we'll have the same "anomaly" than Alinei's
Romance being older than Latin.

> Your objection, on the contrary, implies that PIE first split into
> Proto-Celtic and then into Goedelic and so on.
>

See above.

> > In the case of your proposed etymology, we lack
> > evidence a labiovelar in Celtic.
>
> No, because we don't know a priori the meaning of Cassi-. Both you
> and I adhere to Birkhan's proposal, but this is crucially based on an
> Irish-Germanic comparison. Within Celtic, we can suppose, but not
> prove, that cas and cassi- are cognate. It's just like conn and
> penno-: they may be connected (Pederson), but they don't necessarily
> need.
>

Not exactly. I was speaking about the absence of cognate P-Celtic forms
with the expected reflex, not the existence of a countervidence.

> > According to Schrijver (quited by De Vaan), after a labial consonant
> > Proto-Italic /o/ was unrounded to /a/, but only in *open* syllables
> > (e.g. *mori > mari), so *kWos-lo- would give regularly Latin co:lum,

but

> > not qua:lus, which must be either a loanword or have a different
> > etymology (De Vaan chose the latter).
>
> Be careful: I've written *kwö-, with Schwa secundum
>

Using non-std notations can led to *confusion*. Anyway, your
reconstruction looks like an ad-hoc one.