Re: [tied] Re: Slavic endings

From: Piotr Gasiorowski
Message: 46064
Date: 2006-09-14

On 2006-09-14 11:18, stlatos wrote:

[citing my earlier posting]
> BTW, OE e:anian- < *auno:jan- 'bear young (of sheep and goats)',
> implying PGmc. *auna-, is often quoted (also by Pokorny) to justify
> the reconstruction of *gWH rather than *gW (the latter clearly
> supported by Greek). My private theory, however, is that *-kWn-,
> *-gWn- and *-gWHn- would have produced the same Germanic outcome,
> first merging as *-gWn- during the initial stage of Kluge's Law, then
> losing the velar closure and becoming *-wn- before the operation of
> nasal assimilation, thus escaping the further development to *-ggW-
> and (finally) *-kkW-. Cf. Goth. siuns 'sight' *seuni- < *sekW-n-i-
> (but not via _Vernerian_ *segWni-: Kluge's Law bleeds VL of its
> input!).

I have reconsidered the evidence and changed my mind about KL bleeding
VL. The Vernerian environment is necessary for Kluge's Law to apply (as
originally proposed by Kluge himself), so the part that can be called
"Prokosch's Law" (understood as the falling together of *-dn- with *-Dn-
from *-tn-' or *-dHn-) must follow Verner's Law, and yet be earlier than
the last stage of Grimm's Law (the devoicing of *b, *d, *g). The change
of *-gWn- > *-wn- must be ordered between Prokosch's Law and (revised)
Kluge's Law proper or Nasal Assimilation (operating only after short
vowels).

So here's my chronology:

(1) GL1 -- voiced aspirated stops become voiced fricatives (GLc),
voiceless stops become voiceless fricatives (GLa); the relative ordering
of the two components of GL1 seems to be undecidable.

(2) VL

(3) Prokosch's Law (PL): voiced non-sibilant fricatives become voiced
stops before *n in medial clusters.

(4) Voiced labiovelar simplification: -gWn- > -wn-.

(5) Kluge's Law (KL): *n is assimilated to a preceding voiced
stop (but only after a short nucleus).

(6) GL2 -- voiced stops become devoiced.

Examples:

*ligH-n-ah2- (accent irrelevant) > *liGno:- > (GL1) > *ligno:- (PL) >
*liggo:- (KL) > *likko:- (GL2)

*doik^-n-ó- > *daixná- (GL1) > *daiGna- (VL) > *daikna- (PL) > *taikna-
(GL2)

*swép-no- (OInd. svápna-) > *swéfna- (GL1); no other of the above
changes is applicable.

VL _must_ precede PL and GL2, but the following alternative orderings
are thinkable:

VL could be squeezed between GLa and GLc (in this order), or be an
allophonic rule within GLa (as suggested by Lass).

KL (with appropriate reformulations) could be later than GL2 if one
assumed *-dn- > *-tn- > *-tt- rather than *-dn- > *-dd- > *-tt-. It's
difficult to decide which ordering is more "natural", though my personal
preference is for the latter.

Kluge's original idea (*-Dn- > *-DD- > *-dd- > *-tt-) is untenable,
because *-Dn- is reflected as *-tn- after a long vowel or diphthong.

Stefan Schaffner's book on Verner's Law in Germanic nominal paradigms
(_Vernersche Gesetz..._ 2001, Innsbrucker Beiträge zur
Sprachwissenschaft) discusses a lot of Germanic material bearing out the
chronology suggested above.

Piotr