Bibliotheca Arnamagnæana - 01.06.1959, Page 268
254
proach of the pagans, w. 1002-48 (shortened); vv. 1396-1411; shorten-
ing of vv. 1802-50 (return of Charlemagne); shortening of w. 1974-88,
2083-2114, and 2134-43. In the last three cases the shortening has led to
errors in the translation which the translator would not have committed
if he had had the full text of these passages before him.
There remains a group of variant readings where either the translator
or some French remanieur may be responsible: the addition after v. 112
(description of the chessboards); Ganelon’s oath on the pagan “Bible” in
v. 611; the variants in vv. 1776-79, 2200, 2503-06; the shortening of vv.
520-62, 1152-87; the omission of vv. 170-76, 889, 1470-86, 2193-99,
2203-06, 2404—09, 2488-95, 2512-24, and prohably the additions after
vv. 855 and 999.
It will be seen that this last group is of far less importance than the
others. This is indeed the real reason why it is so difficult to decide
whether this group of changes must be attributed to the translator or to
the French remanieur. As for the variants which are certainly changes
made by the translator, they have all proved to belong to the same groups
as those discussed in the previous chapter: errors, as in vv. 1324, 1605,
2528-62, and 841-45; phrases, summaries, etc. in vv. 348, 453, 456, 670-
71, 1918, 2209, 2233-45, and in a few other cases.
There is throughout the Runzivals j>åttr a marked contrast between on
the one hånd the errors and ahsurdities in the translation of certain pas-
sages, and on the other a number of omissions and changes carried out by
someone who knew the poem well. It is scarcely possible that a person
who has shown himself ignorant of French literary and cultural traditions
should be able to revise the poem, and since we know that the errors were
committed by the translator, it is unlikely that he was also responsible for
the sweeping, but generally consistent, shortening and changes made in the
Kms version of the poem.
If the arguments put forward in this chapter are accepted, we have, in
Kms, a fairly complete but slightly ahridged translation of an old variant
version of the Chanson de Roland. There is nothing intrinsically improbable
or even surprising in this assumption: the Kms version must have been
created in the 12th ecntury, possibly even before 1150, at a time when local
variants of chansons de geste were more common than in the 13th century.
The generally accepted stemma of the MSS and versions of the Chanson
de Roland, proposed by Theodor Muller and vigorously and successfully
defended by Bédier, makes Kms a memher of the a group: