Genesis of Germanic

In case the Shibbolethisation process I suggested really works, one
could propose the genesis of Germanic thus:

PIE (or perhaps later?) has a rule
{Th, T', T}H ->{T#, Th, T'}, where T = any plain stop (= {p, t, k})
and T# = {f, thorn, x}
and Th = aspirated stop

thus in stems of the form

T-H-C-

you have

Nom. Sg. TeHC- > Te:C-
Acc. Sg. THeC- > T'eC-
Gen. Sg- THC- >? TeC-

in T'-H-C-

Acc. Sg. T'HeC- > TheC-

in Th-H-C-

Acc. Sg. ThHeC- > T#eC-

In general, you could call the whole syatem "the left shift rule"
(this is what happens when you mix linguistics with your computer
science education).

Germanic generalizes the Acc.Sg. forms. Other IE branches generalize
the Nom.Sg. forms.

A phonological shibboleth relationship develops (sociologically)
between the Proto-Germanic speakers and some other tribe in the area.
The preferred variants in the T#/Th, Th/T', T'/T shibboleth pairs
therefore spread everywhere in Proto-Germanic. The "wrong" halfs are
purged, but a few roots with the "wrong" half survive, giving rise to
later "Wechselformen" à la Møller (French <maire>/<majeur>).

Then Th -> T, T -> D (= {b, d, g}) in Proto-Germanic.

At some later time, stress is shifted to the first syllable (due to a
Baltic Finnic substratum?). Perhaps this is what triggered the
confusion in the first place?

(And I left Verner out in this first round. It all got rather
complicated).

I would suspect that the first Proto-Germanic speakers were the
Bastarni (cf Du. <verbasteren> 'to degenerate'), and the language was
spoken as a trade language in the south-bound river systems of Russia
and the Ukraine (all the d-n- rivers). In relation to "Wechselformen"
(note above), observe that you have Nom.Sg. forms for gods and
peoples (Da:nu, Tuatha Dé Danaan, Danes?) and Oblique forms for
rivers and places (Tanais, Tanakvisl, Borys-tenes, Tannis bugt,
Pytheas' <Tanais> at Jutland), cf. the place name <den Haag>, one of
the only survivors from a purge of the Masc.Obl.article <den> in
Dutch.

Torsten