> > The real problem here is that the words with /a/ seldom show the
> > "upper-class" variants with /e/ and /o/.
>
> Class VI 'draw' vs. class I 'drive', perhaps (all of class VI
> strong verbs are best explained as PPGmc -a-/-a:-/-a: -/-a- >
> PGmc -a-/-o:-/-o: -/-a-), Engl. grab vs OIc grípa, Engl. wag vs. OIc
> víkja "move"? Futher the OIc. class III verb exceptions gjalda
> "pay", gjalla "shout", hjálpa "help", skjálfa "tremble", skjalla
> "scold" with present root vowel /a(:)/ for /i/ (< PIE -é- before
> R). Other than that note the alternation -eu-/-u:- in most of the
> class II verbs vs. OE scu:fan, OHG su:fan; that alternation is of
> the same PIE/PPIE type as the -a-/-e- you were looking for. Note
> how large the -u:- subclass is in Dutch, as expected, I'm not
> convinced it grew later.
>
> Note also that some of the class II -u:C- subclass have -uCC-
> geminated counterparts: OHG su:f-, Engl. sup, sip, Schrijver's
> *dubb- etc, Sw class III dimpa, damp,
> http://ordnet. dk/ods/opslag? id=437209
>
> http://en.wikipedia .org/wiki/ Germanic_ strong_verb
>
And we certainly shouldn't forget the most typical of all, those from the whole *ka/unt- etc complex: Eng. hunt with original /u/ vs. Gothic hinþan, Sw hinna which are reorganized as a class III strong verb.
Torsten
Speaking of class variants, in rural US English, especially Southerm , there is a variant of /@/ that seems to be different from anything I've heard in any other dialect of English or other language, although I've heard something similar in some Scandinavian language, but with an "r" coloring. The sound really comes out in ""I'm going to hunt" > I'ma hunt" maybe /@m@ h#n?/ and seems to be contrast with /@/ that comes from stressed /ay, a/ and unstressed vowels. It's sort of a fronted or raised version of /@/