Gripla - 2020, Side 20

19
emphasized what he read as the fragmentary nature of the C-redaction
by removing the three þættir and printing them after the saga’s main text,
as well as by printing Þórarins þáttr ofsa as a text entirely separate from
Ljósvetninga saga.43 In addition, his critical apparatus and overall argumen-
tation gives the impression that chapters 22–31, which can only be found
in the C-redaction, were a part of the A-redaction’ – despite the fact that,
as discussed above, codicological evidence indicates that this part would
not have been included in 561.
The Svart á hvítu edition of the saga from 1986 was used as the
basis of our stylometric analysis.44 This edition published the A- and
C-redactions separately, by which – unlike Björn Sigfússon’s íslenzk fornrit
edition – it highlighted the A-redaction’s fragmented nature. In the Svart
á hvítu edition segment that is titled “Ljósvetninga saga (C-gerð),” i.e. the
C-redaction, there are several readings where the C-redaction manuscript
readings are indeed preferred, but in most places the edition keeps the una-
mended A-redaction reading over that of the C-redaction, thereby under-
playing the differences in phrasing throughout the parallel text. From the
viewpoint of a stylometric analysis where the choice of words is key, we
have therefore opted to revise their reading of the parallel chapters in the
C-redaction, based on the readings in AM 485 4to.45 We also confirmed
their version of the A-redaction against the manuscript.
All of these editors of Ljósvetninga saga made decisions that influ-
enced the text’s reception: Þorgeir Guðmundsson and Þorsteinn Helgason
ignored the significant variance evident in the A-redaction; Guðmundur
Þorláksson created a composite text, providing misleading episode titles
A STYLOMETRIC ANALYSIS OF LJóSVETNINGA SAGA
43 Ljósvetninga saga, ed. Björn Sigfússon, 143. Despite this, the three first þættir still influence
Björn’s chapter count of the C-redaction but not that of the A-redaction, creating a
some what disorienting effect which further strengthens his representation of an ‘eclectic
C-redaction’.
44 Íslendingasögur og þættir, II, ed. Bragi Halldórsson et al.
45 The choice of AM 485 4to as the basis for our revisions of the Svart á hvítu text is
justified in Tirosh, “On the Receiving End,” 36-50. There it is argued that of the earliest
extant paper manuscripts, AM 485 4to reflects the most faithful (though certainly not
perfect) transmission of AM 162 c fol. In a text-sensitive study like the one conducted
here, the manuscript chosen by Björn Sigfússon for his íslenzk fornrit edition JS 624
4to is problematic due to its addition of too many words and clauses for the purpose of
clarification and creating a more streamlined narrative.