Re: [tied] Re: Pretonic laryngeals in roots

From: Piotr Gasiorowski
Message: 47900
Date: 2007-03-16

On 2007-03-16 13:31, Sean Whalen wrote:

> These are not from PIE but the result of similar
> changes in a few branches. Even Iranian and Indic
> aren't exactly the same. H3 (xW) can cause this, too
> (* pYròxW >> prathamá-).

Iranian (MPer.) also occasionally shows forms with -tH- in words related
to <pratHamá->. I don't think this aspiration can be blamed on *h3, as
the word can be derived with far less trouble from *pro- (cf. comp.
pra-tara-, Av. Gk. pró-teros). It may be due to sporadic contamination
(e.g. of *pratama- with pratHa- 'spreading out').

> In some languages both the x() before and C
> following are important:
>
> Latin t>tH / xY(+syl) _ r
>
> * t.èr.-xY-tró+ > terebra
>
> * kYrìxYtró+ > cri:brum

The *h1 is not syllabic in the latter, and <terebra> is a secondary,
analogical form, considering its gender (an original neuter collective).
The older forms must have been *tér&1-trom (> Gk. téretron) and
*tr.h1-tHráh2, hence the levelled-out *ter&1-tHrah2 > terebra. There is
no aspiration e.g. in *hár&3-trom > ara:trum (with secondary /a:/, cf
Gk. árotron), where the laryngeal was syllabic.

> However, -idus is not from *-e-h1- + -to-. There's
> no reason for the h1 to disappear

Why not? It coalesces with the *t into an aspirate. This does not happen
when *h1 is part of the root (perhaps because of analogical
restoration), but in a string of suffixes there's nothing to protect it
from full coalescence.

> and there are
> cognates in other languages (where there's no tH>d).

vi:vidus : ji:vatHa-

> I even say that tH > T > f > v > b between e and o.
> The descriptions of the restrictions on dH > b are
> based on a faulty understanding of the data ...

Well, this is a rather bold statement.