Bibliotheca Arnamagnæana - 01.06.1959, Qupperneq 113

99
b) The passages where the Swedish and the Icelandic versions agree
in every detail.
However we can be fairly content if there is general agreement between
S and a reading common to all the Icelandic MSS, even if S has Swedish
instead of Norse words and expressions. When there is a difference be-
tween a and Bb, the reading supported by 5 must always be that of the
original translation, except in cases where a correction may have been
made by two authors independently. This rule does not apply to omis-
sions ; all versions tend to shorten the text and two scribes may easily have
omitted the same words, especially in a text of this kind, where omissions
suggest themselves on practically every page. In general the Swedish text
is not sufficiently close to the Norse to make it possible for us to declare
with absolute certainty that the actual wording of the Icelandic MSS is
that of the original translation.
A critical text must be based on a, as Unger’s edition is, supplemented
and corrected with the help of Bb and S. With the exceptions just men-
tioned, this text will be that of the Icelandic version from which a and Bb
are descended, which can be approximately dated to the end of the 13th
or the beginning of the 14th century. Mistakes and omissions in the
original Icelandic version may occasionally be corrected with the help of
S or D, or both. When a, Bh, S, and D all have different readings, the
French MSS may sometimes help us to decide which MS has preserved
the reading of the original Norwegian translation. But the difference be-
tween the Kms and the French MSS is frequently considerable, and in
such cases our decision must be based on a thorough study of the way in
which the Icelandic editors and scribes have treated their source32.
32 Cp. the case of the “camels” and “mules” above. It is often dangerous to try
to correct the main MS, as is shown by Unger’s treatment of the saga text which
corresponds to O v. 1297:
E Gualter fiet un paien, Estorgans,
where the other French versions have Otes instead of Gualter, a has an obviously
corrupt reading:
SiSan rei3 fram Estorgant i moti heidingja cinum ok hjo til hans-
while Bb has the correct version:
SiSan reiS fram af Frankismonnum Hatun sterki, ok honum moti Estorgant
af heiSingja lidi, ok hoggr Hatun til hins heiSna me3 sverSi sinu, = S.
Unger, instead of adopting the Bb reading, has corrected the a reading in accord-
ance with O: SiSan reiS fram Valteri i moti Estorgant (p. 51019). M. Aebischer’s
note, Rol. Bor. p. 170, is misleading; there is no parenthese carrée lacking in the
edition: the same bracket applies to notes 13 and 14.
7*