Re: [tied] Germanic prefixes and Verner's Law [was: German "ge-" be

08-08-03 14:57, Miguel Carrasquer wrote:

> The question probably can't
> be resolved to everyone's satisfaction, and it doesn't really matter
> whether the development was first Grimm[1], then Verner:
>
> Grimm Verner
> *p ----> *P -----> *f
> \
> *bh ----> *B -----> *B ~ *b,
>
> as you claim, or first Verner, then Grimm[1]:
>
> Verner Grimm
> *ph ----> *ph -----> *f
> \
> *b ----> *b -----> *B ~ *b,
>
> as I would maintain. We cannot tell the difference. The fricative
> formulation has the advantage that it can be formulated as a law applying
> to voiceless fricatives, while in my version it applies to voiceless
> consonants in general (whether stops or the fricative *s), or, in case
> Grimm[2] came before Verner, more specifically to non-"glottalized" (or
> whatever distinguished post Grimm[2] *p, *t, *k from pre-Grimm[1] *f, *þ,
> *x) voiceless consonants.

While I agree that neither scenario is impossible, I nevertheless prefer
the first solution as the null hypothesis for the following reasons:

(1) Everybody agrees that "voiceless fricatives" constitute a natural
class. It's less certain if "non-glottalised voiced consonants" (or
something similar) do. The fact that *s was affected is, in my opinion,
a very important argument in favour of the view that Verner's Law
operated on the fricative output of Grimm's Law.

(2) The first scenario makes no assumptions about the ordering of the
component shifts of Grimm's Law and the phonetic nature of the pre-Gmc.
stop series or their post-Grimm reflexes. The second scenario has to
explain why, if *t etc. were voiced before becoming fricatives, they
merged with *dH rather than *d. This requires extra assumptions and
makes the formulation of Verner's Law unnecessarily dependent on the
validity of those assumptions, against the spirit of Ockham's principle.

>
> My reasons for thinking Verner came before Grimm are the following:
>
> - Verner depends on the position of the PIE accent, so it must predate or
> be contemporary with the shift to initial accent in Germanic. That's
> pretty old, judging by the pretty advanced state of erosion of final
> syllables in the oldest Germanic.

What oldest Germanic? Was the erosion really so advanced in early Runic?
I'd say that ca. AD 200-300 such syllables were stil more or less
intact, at least in Northwest Germanic. As for East Germanic, Wulfila's
Gothic was spoken in the 4th. century. If Verner's Law operated before
100 BC (the earliest Latin loans and the younger Celtic ones show
neither Grimmian nor Vernerian modifications) and the stress shift was
not much later, there was a gap of at least 300-400 years between that
and Wulfila. Judging from the observable tempo of erosion in Runic and
in late Old English, the time was quite sufficient.

> - There is no evidence that Grimm[1] (the fricativization of PIE *p, *t, *k
> to *f, *þ, *x) is particularly old. In fact, judging by such loans from
> Celtic as wolk- > *walh- and others, chances are that it's relatively
> recent... the fricativization, that is: the shift from PIE *p, *t, *k to
> PGmc. aspirated *ph, *th, *kh is, _necessarily_ as old as Grimm[2] [*b, *d,
> *g -> *p, *t, *k] and therefore roughly of the same antiquity as Verner.

See above. Recent as it may be, the fricativisation is at any rate older
than Latin loans in Germanic (or "Kaiser" would be "Heiser").

Piotr