Bibliotheca Arnamagnæana - 01.06.1999, Page 255

VIII “Vin5andin foma”
235
There is thus no instance of skal die vr- later than the year 1000, but there
are instances of r- already in the 9th century. Bragi as well as Egill used
word-forms both in vr- and in r-, and for the word (v)reidi the examples
doeument a period of overlap from the early 9th to the later lOth century.
It is probable that in this word in particular, people may have been aware
of the existence of both variants at the same time. It has often been
noticed that in Eddie poetry the word is particularly often found in
conjunction with the verb vega, so that vreidr vega looks like a poetic
formula, or in Mohr’s terminology, an “Erbzeile” or “Erbformel” (Mohr
1938: 218; cf. Jessen 1871: 29, referred to above).
These Eddie examples comprise the material used in this test.
In principle the test is very simple, but the actual occurrences may in-
volve problems of interpretation, which I shall illustrate by one ex-
ample, a long line in Solarljod, st. 26.1-2, without alliteration in its pre-
sent written form:
reidi verk
\>au Jtu un/j/7 hefr
(Skj. Al: 632)
The problem is that v- was lost also before -u and -o at an even earlier
stage than before -r.10 It might be restored by analogy in the 13th cen-
tury, however, a restoration which probably presupposes the identifica-
tion of /v/ with the voiced variant of /f/, which is frequent in Old Nor-
wegian and has become the rule in modern Norwegian, and sometimes
is found also in Old Icelandic from the late 13th century (Johannes L. L.
Johannsson 1924: 33). Consequently, this verse has four possible forms,
each belonging to a different period of time:
1) vreiøiverk
2) vrciSiverk
3) reiØiverk
4) reiØiverk
{rau ]ru vunnit hefr
{rau [ru unnit hefr
jrau Jru unnit hefr
Jrau {id vunnit hefr
(after the shift vu- > u-)
(after the shift vr- > r-)
(after the restoration u- > vu-)
At the first stage, the verse has perfeet alliteration in the form aa/ax. At
the second stage, however, the verse is either to be taken as lacking a
10 Cf. A. Noreen 1923: 169 [§ 235, 1. a].