Studia Islandica - 01.06.1962, Qupperneq 193

191
a matter of course that the impression it made on him should have
left unmistakable traces in the pair word series. If Eyrbyggja had in-
stead been Egla’s next successor, the pair words between them would
have got a much higher frequency. As it is now, Laxdœla has come
first and in its tum given decisive and powerful impulses to the author
of Eyrbyggja. In the Eyrbyggja-table (XI) Egla has therefore simply
quit the field for Laxdœla. — Thus, the problem of the succession as
regards Laxdœla and Eyrbyggja seems to have been definitely solved.
These last tests of the pair word statistics has turned out to agree
surprisingly well with current views on the relations between certain
great sagas. When there is an obvious dependence, as in the case of
Njála and Laxdœla, the evidence of the pair words becomes very clear.
However, these series never reveal such a strong affinity between two
sagas as that between Egla and Snorri in the original series. And that
is as it should be, if the method used is to be proved as effective as
we hoped. In the relationship Snorri-Egla we have to do with some-
thing more than “influence”: common authorship.
The purpose of the verifying tests above has not been to make new
discoveries, rather the contrary. Since it is possible to postulate the
existence of certain commonly accepted facts, it seemed expedient to
test the effectiveness of the method by confronting these with the new
observations. The aim has simply been to utilize one more possibility
of testing the pair word statistics and thus corroborate the main result
of the present study: that Snorri is the author of Egla.
However, it must by this time have become perfectly clear that the
pair word method need by no means be some sort of subsidiary resort
only, suitable perhaps to confirm results already won in other ways.
In fact, it has definitively proved to be a very sensitive instrument
of research with its own inherent resources. This time it has only been
possible to exploit a small part of the material which the pair word
series and tables afford. Thus, for the sake of concentration and clarity,
the nature of the vocabulary listed, significant kinds of pair words
between different sagas, etc., have not been discussed at all. Various
other facts can be deduced from the material, completing and diffe-
rentiating the more summary testimony of the statistics. But the possi-
bilities of the statistics themselves are by no means exhausted. The pair
word series have been transposed into tables and statistics only so far
as this has been necessary for the argumentation of the main thesis.
I am convinced that a broad and well-organized research on the
basis of the method here applied would throw new light on many
otherwise inaccessible — perhaps unrecognized — problems in the
saga literature.