Re: [tied] n/r (was: PIE suffix *-ro - 'similar-with')

----- Original Message -----
From: "P&G" <G&P@...>
To: <cybalist@yahoogroups.com>
Sent: Monday, January 09, 2006 2:53 AM
Subject: Re: [tied] n/r (was: PIE suffix *-ro - 'similar-with')


> >> > there is _no_ nasal infix. There is a nasal suffix which, under
> >> > certain
> >> > conditions, undergoes metathesis.
>
> >> What would you say these conditions are, under which a nasal suffix
> >> metathesises?
>
> > I do not know and I am not sure that we ever will.
>
> Aah. Then it sounds to me like a plausible hypothesis, but one without
> evidence. How would you distinguish a Cen-Co- root from a CeC-n-o or a
> C(e)nC-o? The only way I can think of, is to look for forms without
> the -n-, and alas, what we find in the evidence is an alternation between
> CeC and CenC, an infix.
>
> Peter

***
Patrick:

Truly, it is a difficult question to solve to everyone's satisfaction.

But, I would submit, unless we subscribe to Jens' infix, I think it would be
fair to say that infixes violate the Sprachgeist of PIE.

Aside from Jens', what other possible grammatical infix can be brought
forward for PIE?

Whereas, in a grammatical function, literally dozens of suffixes are easily
attested.

That distributional pattern is, perhaps not hard evidence, but it is
evidence unless one rejects the whole idea of statistical probability.

In other words, the *n of *Ce-n-Co could be a prefix but the probabilities
of expectation for this are so low that that we should regard it as a suffix
unless we have a compelling reason not to do so; also, the probabilities of
metathesis, though not as high, are higher than those for _any_ infix.

***