Niuja Waurda

[llama_nom]

> > 
> > > *muldigs (A) "dusty"
> > 
> > I think there may be a tendency, not perfectly adhered to, to use the
> > suffix -eig- after a long root, -ig- after a short stem.  Hope I've
> > got that the right way round...
> 
> I don't remember exactly. I was assuming long stem = short suffix. The
> example might be hairdeis vs. harjis < xerdijiz, xarijiz.

I was thinking specifically of this variation in the adjectival suffix
-eig-.  'gabig' gets 18 hits at Project Wulfila (including the verbs:
gabigjan, gabignan), 'gabeig' just 8.  Compare the examples with long
roots in Wright 394: anda-nemeigs, ansteis, hroþeigs, laiseigs,
listeigs, mahteigs, þiuþeigs, us-beisneigs, waurstweigs; but then
there are two attestations of 'sineigs' with a short root.  Apart from
this, 'i' for long 'ei' appears "nut ganz vereinzelt und fehlerhaft"
(Braune/Helm 1956, 17.3).  So even if there is no such rule generally
applied, the long -eig- is the more normal.  I can't remember where I
read this theory that the variation was a tendency to reduce the
ending after a short root (could it have been Streitberg's
Elementarbuch?), but it seems credible, and would presumably have been
motivated in the same way as Sievers' Law, the process whereby
Germanic -j- became -ij- after a long stressed syllable, which is what
accounts for the examples you mention:

*herdjaz > *herdijaz (syllabification of -j- after long root) > Go.
hairdeis.

*harjaz > *harjaz (no syllabification after a short root) > Go. harjis.

Another parallel is the tendency for derivative suffixes to carry a
heavier stress after a long root in Old English verse.  Thus in
Beowulf: Hwílum cyninges þegn 867b, which scans well as Sievers type B
with "resolution" (where a short stressed syllable counts together
with a following unstressed syllable as metrically equivalent to one
long stressed syllable) in the first lift: x x ú (x) x /, alliteration
on 'c'.  Similarly Beowulf 87a, Sievers type A: þráge geþolode = / x x
ú (x) x, with resolution on the second lift.  But with a long root,
it's possible to count the (originally) long suffix (medially and very
occasionally finally) as metrically equivalent to a secondary stress,
e.g. Beowulf 166b, Sievers type D: Heorot eardode = ú (x) / \ x.  In
final position, Beowulf 1112b, Sievers type A: æþeling manig = ú (x) \
ú x -- here the use of a secondary stress in the first dip allows the
short syllable /mo/ to count as the second lift, and blocks resolution.

LN

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/gothic-l/attachments/20070417/5b5bfcf8/attachment.htm>