Re: [cybalist] SV: The long awaited athematic answer to the athemat

----- Original Message -----

From: Urban Lindqvist

To: cybalist@egroups.com

Sent: Thursday, April 13, 2000 7:20 PM

Subject: [cybalist] SV: The long awaited athematic answer to the athematic question... Oy veh.

> From: Glen Gordon <glengordon01@...>
>
> PS: The Narten present couldn't have existed in IE unless we're
>     talking about laryngeals. I recall mention
>     of /stau'ti/ supposedly coming from *ste:u-? Why? As far as T.
>     Burrow in "The Sanskrit Language" conveys, the accentuation on
>     -u was the norm for athematic strong forms of stems ending in
>     -u before consonant initial suffixes like the 3rd person
>     singular -ti. Don't we find this same accentuation in duals
>     and in the word for "cow"? So... why isn't this just *steu-?

As usual I don't follow. I don't have Burrow's book available, so I can't check what he writes, and I don't understand in what way the accentuation is on the u.
However: *steu- > sto- before consonant, *steu- > stav- before vowel. There are "athematic strong forms of stems ending in -u before consonant initial suffixes" where we find the reflex -o-, such as reduplicating yuyoti and juhoti. If Burrow's rule only applies to root presents, it must be an Indo-Iranian rule, since we have sta:umi in Old Avestan (as opposed to Young Av. staomi).
Furthermore, since there are weak forms with full grade (stave, not *stuve), it's more natural to assign the verb to a known category than to assume a non-existent one (as far as I know), root presents with full grade in both strong and weak forms. For there are other Narten-presents, not only from roots ending in -u. Cf. sing. ta:s.t.i /Av. ta:šti (< *ta:ks.t.i) vs. pl. taks.ati 'fashion(s)' (compensatory lengthening is not an option here: cf. imperfect ata.s.ta).
I can find only three canditates for *-eu- > -au- before consonantal ending:

rauti 'cries', pl. ruvanti. Does not qualify, because it's a set.-root and in any case secondary (and late: bra:hman.as), since there's no trace of the final laryngeal (*h3reuH-; cf. Ved. aor. ara:vi:t, Greek o:ru:omai).

yauti 'unites'. From Atharvaveda on. In Rigveda we only find yuvati.

ks.naumi 'sharpens'. Rigvedic. The weak stem ks.nu- in Vedic could just as well be secondary; in Iranian we find full grade in weak stems: Old Persian 1 sg. a:-xšnauvaiy, Avestan 2 sg. opt. xšn@...:ša:.

Any others?

By the way:

> The Narten present couldn't have existed in IE unless we're talking about laryngeals.

Why not?

Urban

---

I don't understand the accentuation argument either.

 

One could adduce tauti (~tavi:ti) 'is mighty' to your list as another candidate that doesn't count, being a seT root. Let me point out that root presents in *-ei do not exhibit any unusual behaviour in the sg. with regard to Old Indic accent or quantity (*kei-toi > śete, *ei-ti > eti, *kþei-ti (Brugmannian) > kšeti); it would be strange if "Burrow treatment" were restricted to *-eu roots.

 

Glen seems to assume that only length caused by phonological processes (compensatory lengthening, coalescence) counts as genuinely IE. However, whether one likes it or not, there are clear instances of length playing a purely morphological role (as e.g. in vrddhied adjectives and probably in sigmatic aorists; I'd go as far as to question the "compensatory" account of vowel quantity in the nom. sg. of root nouns, at any rate in the form proposed by Szemerényi). No prestidigitation involving a laryngeal will work for *steu- (cf. stu-t- 'praise'), so classing it as a Narten present alongside similarly behaving stems like *tekþ- seems more reasonable than inventing a different ad hoc explanation in each case.

 

Piotr