Bibliotheca Arnamagnæana - 01.06.1999, Blaðsíða 259

VIII “VinSandin foma”
239
would lose its alliteration if vrata were read instead of rata.'6 Grandt-
vig’s reasoning implies that st. 106.2 originally had as main stave a
word-form beginning in v- alliterating with vrata, and that this form has
been exchanged for one in r- in order to restore the metrics after vrata
had lost its v-. But the test as such presupposes that no general metrical
restoration has taken place, since if that were the case, the forms in vr-
would be irrecoverable. Although Grundtvig’s argument is sound in
principle, it is invalid in this particular case, and Jessen’s view remains
the more acceptable one.
The cases quoted so far demonstrate that before the examples can be
used as evidence, a critical examination on several levels is required,
from the point of view of textual criticism, of etymology and of metrics.
Not much needs to be said on textual criticism, except that conjectu-
ral examples must naturally be discarded, unless they are warranted by
solid arguments independent of the metrical problems involved in the
test.
From the point of view of etymology, safe examples with vr- have to
be distinguished from more hypothetical instances. The following Eddie
words seem to have an undisputed etymology involving older forms in
vr-: reidr, reidi, reini, reka, rengja. Also Rindr has, in spite of Bugge’s
objections (1867: 137), a relatively certain etymology in vr-; both Brate
(1913: 110) and Lundberg (1913: 51-52) relate it to a root *wr-/wer-,
even if they differ on the meaning of the (homonymous) root.17 Rata in
the meaning ‘travel’ has also an undisputed etymology in vr- as far as I
know (Germanic *wraton), but the same written form may represent
(h)rata in the meaning ‘collapse’ (Gripisspd 36.3). Rati is by most
scholars taken to derive from a form in vr- (Germanic root *wrat-, cf.
Lidén 1929: 220-21), but Gering (1927: 127) takes it to be a cognate
form with lat. rodo, giving the meaning ‘rodent’ to 05inn’s auger. Rata-
toskr should probably follow rata. Rgskr probably has an etymology im-
plying vr-, cognate with Gothic gawrisqan - ‘bear fruit’ (de Vries 1962:
458; Feist 1939: 213), but it seems that a possible relation to Gothic
16 Cf. the diplomatic edition by Ludvig Wimmer and Finnur Jonsson (1891: 7.2 and
11.4-5).
17 Loewenthal 1931: 317 holds that Rindr is cognate with Serbian rudina, ‘lawn’, but this
word also seems to be developed from a root in wr-, wer- being lost in protoslavonic
(Petersson 1916: 142—43).