[tied] Re: Allofamy, allofams

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2006-07-05 16:06, tgpedersen wrote:
>
> > The three verbs I know of with supposedly Verner variation
> > *-Vng-/*-Vnx- -> *-Vng-/*-V:x- are
> > *gang-/*ga:-
> > http://www.angelfire.com/rant/tgpedersen/Hng.html
> > *hang-/*ha:-
> > http://www.angelfire.com/rant/tgpedersen/Hng.html
> > *fang-/*fa:-
> > http://www.angelfire.com/rant/tgpedersen/HbHpHg.html
>
> Verner's law applies in the last two (PGmc. *xanx-/*xanG- and
> *fanx-/*fanG-), but not in the first pair! PGmc. *Ge:- didn't have

a

> final velar (never mind the misleading spelling of Mod.Ger. gehen,

where

> the <h> is orthographic, not etymological!).
>
> PGmc. *n before a velar fricative was lost in all Germanic

languages but

> it caused the nasalisation of the preceding vowel. The resulting

*a~:

> was still nasal in Anglo-Frisian, producing long /o:/ just like the
> outcome of *ans (OE go:s), *amf (OE so:ft), *anþ (OE to:þ). This is

why

> we have OE ho:n < *xanxan- (pp. hangen), fo:n < *fanxan- (pp.

fangen),

> but ga:n (pp. ga:n, no Verner). The *x in *xanxan, *fanxan was lost

in

> post-AFris. times and only when followed by a vowel (cf. 2sg.

fe:hst

> 'thou catchest' but gæ:st 'thou goest').
>

If these three glosses existed in Germanic when Verner applied, that
is correct. But ex hypothese (meo) they are allofams of something
else, thus potentially loans into Proto-Germanic. Eg. from a language
which had variation in finals *-Vn,x-/*-Vn,g- -> *-V:-/*-Vn,-, such
as many Sino-Tibetan languages do. Now suppose "go" was borrowed in
the form *ga:-/*gan,- but "hang" (originally the same word)
and "catch" were borrowed in the earlier form *hanx-/*han,- and
*fanx-/*fan, (before Grimm, obviously), we'd get the development you
described.

BTW I wondered whether ON gangari might be a Germanic calque on the
Nordwestblock(?) loan *hanx-ist-/*han,-ist-. That would match up the
Germanic (ultimately Latin) actor suffix -ari- with Nordwestblock(?)
-ist- which therefore also would be an actor suffix, which would be
OK with its use in NWBlock placenames. Or does someone have a better
Germanic sense for *-ist-?


Torsten