> >In other words, with some words, you'll have to resort to
> >'expressiveness' to explain the gemination, which is no explanation
> >at all.
>
> Why not? In many languages, "expressive" formnations do
> have their own peculiar phonology and phonotactics, and
> follow different historical developments.
What is 'expressive'? What does it express? It sounds to me like
someone is playing on the word's connotations of 'hypochoristic' and
'diminutive' but doesn't want to say it straight out, since that would
provide an actual criterion for evaluating the use of that epithet, by
which it would surely fail. Those supposed 'expressive' forms have
nothing semantic in common.
Anyway, I thought I'd reconstruct a typical verb of the geminate
language. Let's take dump/duck etc, Schrijver's example:
pres.
*dúbo:
*dúges
*dúget
...
*dungént/dúmbont
pret.pass.3sg
*dups
I've assumed the language is IE, and that it was absorbed in Germanic
pre-Grimm. Further that PIE had allophones *ke/*po (and *ku) for the
phoneme *kW (and correspondingly for the other labiovelats),
corresponding to some pre-stage of Greek, and that the various
languages generalized one the various allophones.
The preterite I reconstruct based the 'subitive s' in some German
verbs: plumpsen etc, and as a starting point of a later s-aorist. It
would have been constructed mana kartam - style: by-me (was) dumped, etc
Now, interestingly, this verb group has a Swedish cognate which goes
by the 1st strong conj.: dimpa, damp, etc "fall down".
I've proposed earlier to emend the ablaut scheme from
e, o, zero
ei, oi, i
eu, ou, u
to
e, o, zero
ei, ei, i
ou, ou, u
with later analogical remodelling to the first scheme. That way we
take account of the fact that even though PPIE had the three vowels i,
a, u it seems in the old scheme like verbs always had the vowel /a/.
But it occurred to me that the VRC scheme, that of the Germanic first
strong conj., needed to be adapted too, from
eRC, oRC, R.C
so that it was spread more evenly, so to speak, in PPIE on the three
vowels (Miguel did something similar by letting all PPIE i, a, u short
vowels -> schwa, IIRC), so to
aRC, aRC, aRC
iRC, iRC, iRC
uRC, uRC, uRC
which became
eRC, oRC, aRC/iRC/uRC/R.C
by analogy
which apparently happened to the Swedish verb.
Torsten