Íslenzk tunga - 01.01.1959, Qupperneq 34

32
PETER G. FOOTE
It is obviously extremely difficult to apply mechanical analysis of
this kind to anything so human and variable as language and ortho-
graphy in a manuscript-age. We have at least three scribal ,layers‘ to
deal with, and it would be too much to expect sharply defined
differences as soon as a new part begins. It appears from the above
that a habit apparently imposed by a more regular form in one part
may be only gradually discarded, and that concentrations of dis-
crepant forms may be found in part II, where a scribe, at some stage
in the transmission, appears to have fallen into a certain habit and
then out of it again with equal suddenness. From the above examples,
however, it cannot be seen that such blocks of discrepant fonns
appear within the same close limits, and thus they cannot be regarded
as indicating further divisions in part II (such as might be ascribed
to the work of different scribes or the use of different exemplars),
although a more extensive and closer investigation of the manuscript
may lead to different conclusions.17 The differences between I and
II as a whole, on the other hand, appear in general to be well marked,
though not all the examples given above can be equally significant
(braut, for example, is obviously suspect).
The figures for the subordinate parts of I are included for what
they are worth. In some instances they seem sufficient at any rate to
support the conclusion that one would draw from the of/um figures,
viz. that the ultimate provenance of I (a) was in some way different
from that of I (b—c).
These observations may contribute towards a solution of the
combination of I and II, although there are many imponderables.
A reasonable inference from the differences noted above would be
that I and II were not originally parts of the same work. Ortho-
graphic features common to both parts must of course have been
imposed at every stage, from the time of the combination of the two
parts down to the writing of 291; some of them may have already
17 It is presumably only a coincidence, for example, that the reappearance of
ek 7820 corresponds so closely to the first appearance of brot 7823, since the
former disappears again after 9930, while the latter is last found at 12712.