> > So
> > PIE *stad- ->
> > PPGerm pres.
> > '0sg' *sta:i,
> > (loss of auslaut consonant with compensatory lengthening)
> > 3pl *stand-énti
> > (*-d´- -> *-nd´-)
The four main forms of the first five classes of Germanic, eg.
Gothic, strong verbs can be derived from PIE forms with ablaut
*< e, o, zero, zero>, ie.
Class I: verbs in -Vi-
Class II: verbs in -Vu-
Class III: verbs in -VRC-
Class IV: verbs in -VR-,
with *CeC- -> *Ce:- in the third form
Class V: verbs in -VC-,
with *CeC- -> *Ce:- in the third form and
preservation of the vowel of the first form (ie. *e) in the fourth.
But the sixth class goes like this in PIE:
*<o, o:, o:, o> or
*<a, a:, a:, a> or any combination of the two, BTW, since
PIE *o -> PGerm. *a and PIE *a: -> PGerm. *o:
Jasanoff has pointed out the importance of verbs with o-grade in
the present stem: they can't be fitted into the usual system in
which the present stem is assumed to have e-grade (at least in
the sg., the pl. should have zero grade if the Sanskrit pattern
applies, and it might, cf. the zero grade presents Goth. ga-lukan,
not **ga-liukan, Goth. trudan (<- *turdan), not **tridan, WGerm.
koman, not **kweman, OSw. sova, not **swefa, all presumably
generalized from the otherwise lost zero grade present pl.).
He uses from Gothic the verbs malan and faran, but doesn't mention
that they have a whole class to themselves in Germanic. The PIE
ancestor of malan, *melx- has e-grade present forms in other
languages. Since Hittite has an o-grade hi-conjugation malli
(3sg) he sets up a hi-conjugation paradigm for the present of
*melx-, with o-grade sg., and e-grade pl.
Perhaps there are other ways of explaining the o-grade of some
present stems. Note the full grade of the ppp. of the class VI,
normally we expect zero grade there because of the pretonic
position of the root vowel. Therefore the stress must have been
somewhere else. In fact none of the four forms have e-grade,
which is what we'd expect if the root vowel was stressed.
According to Jasanoff, Jens has proposed to explain the o-grade
presents as de-reduplicated intensive reduplicated presents
with vocalism *CeCo-. But suppose that instead of a reduplication
prefix we proposed a preverb/adverb as prefix? To some degree they
are equivalent, Latin verbs with reduplication typically lose
the reduplication, when prefixed with a preverb. In Lithuanian,
non-acute verbs come in two types, according to Stang, those that
retract the stress onto a prefix/preverb, and those that don't.
If we assume a similar pattern for PIE, we could explain why some
Germanic verbs follow the *< o, o:, o:, o> and some follow the
*< e, o, zero, zero> pattern. Note BTW that eg in Russian there
are verbs that stress the prefix in some forms but not in others,
eg.
prodát', prodám etc, but
pródal, prodalá etc
pródan, prodaná etc
A verb type like that in PIE would explain the isolated *o of the
pret.sg. of the *< e, o, zero, zero> inflection.
If this stress-alternation affected consonants too, there might
even have been an alternation in a single verb like *-bhér-,
*´-por-. Note that for some reason the verb, *bher- is suppletive,
ie. is replaced by some other verb, in Greek, Latin, Old Irish and
Tocharian, in the forms that would otherwise have had *-bhor-.
The various o-grades, which arose in prefixed verbs, were then
transferred in some cases to the unprefixed verbs (cf. Russian
l'gat', rvat' with the rare anlaut sequences, which are probably
also the result of 'de-prefixing')
I should point out that in Skt., a preverb, the standard place of
which is immediately before the verb, might be stressed, but in
main clauses the verb itself might not. This might be interpreted
as the presence of preverb-stressed verbs in Skt. too.
Torsten