Re: [tied] Re: Question on PIE Root h2ep- 'to grab'

On 2006-08-18 12:19, alexandru_mg3 wrote:

> But the problem is that with a root *h1ep- we still need to better
> account for:
>
> 1. the alternance a/e in Hittite forms epzi/ appanzi

This is exactly like *h1esti, *h1senti > eszi, asanzi. The initial *h1
was vocalised, giving Hitt. a-. There's no problem if we posit a regular
root present like *h1ép-ti, *h1p-énti.

> 2. The Latin api:scor, Latin aptus etc... as from *h1p- (that is
> not regular, as yourself have explained me) together with the need to
> explain the long e: in Latin co-e:pi: (see Lubotsky cognates on
> Leiden)

<aptus> may be somewhat irregular phonologically, but represents a
common Latin type of derivative with schwa secundum, i.e. a vowel which
doesn't belong to the regular system of PIE ablaut but is a substitute
for a zero grade (which would be impossible or clumsy in a given
formation). The same refers to the archaic verb apio: < *h1p-jé/o-, with
which <aptus> and <-e:pi:> were originally associated.

Lengthened-grade perfects are common in Latin:

lego: -- le:gi:
ago: -- e:gi:
sedeo: -- se:di:
capio: -- ce:pi:

etc.

Whatever the correct explanation of this length (and there is no
consensus on that yet), it must work for the whole bunch, not for -e:pi:
in isolation (BTW, *h1eh2p- does not account for the long /e:/ at all).

> 3. the long a: in Skt. a:pnoti, etc...

Perhaps Lubotsky is right about the secondary character of all IIr.
forms except for the perfect (originally *h1eh1(o)p- > *a:p-, with a
long vowel from the contracted reduplication), reinterpreted as a base
form. Another theoretically possible source of long *a: in Indo-Iranian
is the caus./iter. *h1op-éje/o-, with the root vowel lengthened by
Brugmann's Law (cf. Skt. a:páyati, Av. apaiieiti).

> 4. the Albanian jap/jep forms ('usually' the regular outputs of a
> accented short e) versus Albanian ap/(g.)nap forms
> (forcing 'everybody' starting with Pedersen to talk about an non-
> etymological "j")

It wouldn't be much trouble to derive the Albanian forms from *h1ep- and
*h1op-eje/o- if one insisted on the etymological connection despite the
semantic obstacles. The paradigm is clearly suppletive, which makes its
analysis difficult anyway.

> I mean: too many exceptions 'in the same time' related to the same
> supposed root *h1ep- => so maybe is better to think that there
> was 'something else' behind.

It's far less problematic than you originally made it look like.

> See Skt. a:p- (a:pnóti etc...) as one possible example of a possible
> *h1eh2p-

It could be a dozen other things just as well, so it can't be used as
evidence of *h2