Studia Islandica - 01.06.1962, Side 186

184
terial has been excerpted and treated statistically for each category
separately. By this means one can check how far a certain affinity
between two texts is consistent throughout, in all categories. On pp.
20—23 the four categories are defined in more detail, and some pro-
blems connected with the collection and treatment of the material are
discussed.
3. Comments on the statistics. (Pp. 23—30). The statistics over
the pair words are presented in Tahles I—III (pp. 59—61). The cor-
responding material, i.e. the lists of the pair words themselves in the
various series and categories, is to be found in Appendices I—XII
(pp. 75—143). The page numbers in the appendices refer to the well-
known editions in Islenzk fomrit, published by Hið íslenzka fomrita-
félag (The Icelandic Ancient Literary Society), Reykjavik. The edi-
tions used are listed and some reflections on the manuscript tradition
and its bearings on the present research are to be found on pp. 24—25.
And now a general comment on the arrangement of Tables I—III.
In Table I Snorri A is compared with the five family sagas separately.
In Table II Snorri A is replaced by Snorri B. Table III, finally, is
arranged as Table I, with the difference that Egla has given place to
Snorri B. The purpose of this varying arrangement of the statistics
will appear in the discussion of each table separately. In the tables
there are two figure columns. The column to the left gives the abso-
lute numbers, that to the right the relative ones, adjusted according
to the varying length of the sagas. (Cf. the example given on p. 183
in this Summary.) Since Egla is used as the standard of comparison,
its figures are naturally the same in both columns, while the figures
for the other sagas are adjusted according to the volume of the texts
in question. Hence the adjusted column to the right gives the more
“correct” proportions. This correcting principle is also applied to Table
III, though Egla does not appear there. Of course, any one of the five
sagas could have been chosen as the standard of comparison. But as
we are primarily conceraed with Egla and its relations, it seemed most
convenient to choose that saga.
Otherwise the uniform arrangement of the tables should not need
further comments. We now turn to a closer examination of Tables
I—III.
Table I reveals that Egla has a very much larger number of pair
words with Snorri A than have any of the other sagas. Totally, for
all four categories, Egla presents a figure more than twice as high as
that for Laxdœla, the saga with the next highest figure (193:89). Lax-
dœla, Eyrbyggja and Grettla are much on the same level (89:80:80),
whereas Njála is far below them and does not reach to so much as a