Bibliotheca Arnamagnæana - 01.06.1999, Blaðsíða 302

282
Part Two
which he drew the conclusion that, while the number of mentions of
08inn is greater than that of the other group in the heathen period, in the
Christian period it is the other way round.20 On the basis of his distorted
percentage it is possible to calculate the correct figures, however, and in
no period do they threaten the leading role of the OSinn-kennings.
Finally similar statistics for kennings involving goddesses rounds off
de Vries’s survey.
To assess the strength of his argument, I think it is fair to add up the
figures of the three sets of statistics, and for our purpose we still can
confine ourselves to the four periods 1050-1250. The summing up is
shown in Table 20 and the graph in Figure 2.
Table 20. Mythological kennings in skaldic poetry 1050-1250.
Period Total number of stanzas Stanzas with gods/ goddesses Stanzas without gods/ goddesses Percen- tage
1050-1100 199 20 179 10.05
1100-1150 252 29 223 11.51
1150-1200 712 101 611 14.19
1200-1250 493 41 452 8.32
Applying the same chi-square test to the total group as to the 05inn-
kennings, we find that this time the nuil hypothesis can be rejected- (p =
0.016). On the other hånd, this does not mean that de Vries’s hypothesis
is proved to be true. In faet a doser inspection of the figures shows that
the fali in the last period (1200-1250) counts for more than the preceding
rise, and we remember that the essence of the hypothesis is that there
was a heathen revival in the second part of the twelfth century. If this
20 “Tegenover 319 kenningen van Odin staan er slechts 138 van Baldr, Freyr en NjgrSr.
De percentgetallen zijn dus te herleiden tot zulke, waarbij van beide groepen in totaal
evenveel kenningen voorkomen. Om dit te bereiken heb ik die van de kleinste groep ver-
menigvuldigd met 2,3 [...] In den heidenschen tijd blijft het aantal der laatste kenningen
belangrijk beneden dat van Odin; in de periode daama ligt het er gewoonlijk boven” (de
Vries 1934a: 67-69).
Curiously enough, this conclusion was quoted without any reservation by Hempel
(1936: 116).