Re: [tied] Re: *-tro-/*-tlo-

On 2007-06-21 20:46, stlatos wrote:

>> Not necessarily. The aspiration is older than the dissimilatory loss of
>> the first *l. At the time it arose, the *h2 was non-syllbic. One could
>> even propose *tl.h2-tlah2 > *tl.tHlah2 > *t&tHlah2, with a "schwa
>> secundum". The Greek word simply doesn't belong here.
>
> Are you saying h()>0 when t>th or not? You've said diff. things in
> the archives; I need to know your current rules and order to argue
> against this more effectively.

The rules are not mine; they are Olsen's. It seems that the laryngeal
was lost at least in some cases, e.g. in adjectives derived from stative
verb stems in *-eh1-, hence *-e-tHo- > Lat. -idus. There are also
isolated cases like Lat. stabulum < *sta-tHlom < *stah2-tlom, Gk.
á(w)etHlon < *h2wetHlom < *h2weh1-tlom, though in most cases the
lengthening effect of the laryngeal can be seen. I may have been
restored by analogy at a time when those nomina instrumenti (etc.) were
still trnsparently related to their verbs, or perhaps the loss was just
sporadic. We are dealing with a process whose effects have been largely
wiped out, so it's hard to say with certainty.

> More changes that show this is branch-specific not PIE include:
>
> * xY,s+u+ > eu-
>
> * xY,s+tlo+ > esthlos '~good'

I don't think it contains *-tlo-.

> * xr,+tro+ > arthron 'joint'

If it's the root found in the 'arm' word, it was *h2arh1-, conceivably
yielding *h2r.h1-trom > *h2r.tHrom > artHron.

Piotr