Re: [tied] *kap-

--- In cybalist@yahoogroups.com, "Patrick Ryan" <proto-language@...>
wrote:

> ***
> Patrick:
>
> No borrowing is ever out of the question completely as some future
> etymologist will, no doubt, exclaim when he sees what the Japanese

have made

> of 'McDonald's'; however, why should we consider a loan until all

ancestral

> possibilities have been thoroughly rejected? If I understand the

question,

> the major problem with *kap- for some is two voiceless stops in the

root, a

> PIE no-no.

Two voiceless stops isn't a problem. Two voiced stops is. So that's
not the issue here. There's two problems. First is that a single
reconstruction isn't possible. Some languages point to *kap, others to
*kabh, and others to *ghabh. It would be an odd coincidence if these
were separate roots that just happen to be very similar. Of course if
it's a loanword then we'd expect difficulty in reconstructing a single
form. The second problem is the occurance of *a. It's not supposed to
exist in PIE. Personally I have no problem with this because I think
PIE did have an *a. But if we take the view that every *a actually
comes from *H2e then we have a problem. We'd expect IIr. **khap in the
full grade or **k(h)ip in the zero grade. Neither of these appear in
Pokorny. Instead we have kapati, an expected development from *kapeti.

> Without boring you with details, I suspect strongly that the word
> should be reconstructed as **k(h)a(:)p- from a pre-PIE *kho?ap-,

which would

> radically change the root form. By itself, *kap- implies *kaHp- or

*k(h)ap-,

> leading to **ka:p- since *a cannot be maintained in PIE without having
> undergone (temporary) lengthening through either a laryngeal or lost

and

> compensated aspiration.

The IIr evidence seems to rule this out since it has no voiceless
aspirate. I also didn't see any form with *a:. If one doesn't believe
in PIE *a then this pretty much has to be a borrowing.