Studia Islandica - 01.06.1962, Qupperneq 183

181
syllables per phrase or sub-period; asyndeton; the use of tenses (pre-
terite and present respectively); the choice of verbs of saying; certain
epic formulas, etc. But unlike Wieselgren he comes to the result that
Egils saga on the points in question shows a striking similarity to
Snorri’s authentic writings. His conclusion runs thus: “on the whole
they have the same style, and the possibility that Snorri has written
Egils saga is now raised to a very high degree of probability” (92).
Van den Toom’s undertaking might seem foolhardy, in view of Nor-
dal’s convincing critique of Wieselgren’s linguistic study. He himself
has scarcely discussed the methodological aspects of his task. But in
fact Nordal’s argument about the manuscripts of Egils saga has not
by far the same fatal consequences for a positive outcome of the com-
parison (van den Toom) as for a negative one (Wieselgren). Perhaps
that seems at first sight a rather strange statement, but on second
thoughts it will become evident. Let us suppose that a comparison
between Snorri’s works and Egils saga had revealed such a close lin-
guistic affinity, that the assumption of a common authorship would
be a natural conclusion. Then the objection that the main manuscript
of the saga might differ considerably from the original, would miss
the point. For it would be a hazardous hypothesis that the copyists’
changes, perhaps through quite a line of manuscripts, should result
in a striking similarity between Egils saga and Snorri’s authentic writ-
ings. It is a more reasonable supposition that the similarity actually
depends on the fact that these works have the same author, and that
it would probably have proved still greater, if the original text of the
saga had been preserved.
Another point is that a linguistic-statistical comparison must be very
carefully executed and yield very clear-cut results, if it is to be con-
sidered as decisive in matters of authorship. Unfortunately, van den
Toorn’s investigation can hardly be said to satisfy such claims. The
present writer has not aimed here at a close critique of his material,
method and conclusions. But it ought to be remarked that, in spite of
his own just critique of Wieselgren, he exposes himself to much the
same grave objections. In fact some of his statistical figures lend them-
selves to interpretations which oppose his own. A few dubious points
have been discussed on pp. 10—13 in this paper.
A serious flaw in Wieselgren’s statistics was the want of testing
material outside Egils saga and Snorri. Van den Toorn has tried to
meet that deficiency, but not radically enough. Thus he considers none
of the other great family sagas, except occasionally Laxdœla saga —
and completely ignores Eyrbyggja saga, Njáls saga and Grettis saga.
Still more unsatisfactory, perhaps, is the circumstance that Snorri is
represented almost entirely by Ólafs saga Tryggvasonar (28 000 words)