Íslenskt mál og almenn málfræði - 01.01.2003, Síða 107
A Note on nokkurr
105
nokk- example. Text n2, which has the highest proportion of ‘i’
sPellings, has besides a single nakkvat no other forms of this pro-
noun.
Of the 13 exceptional forms two have ‘o’. It is very unlikely that
ihese should be interpreted as /0/, as the spelling ‘o’ for /0/ is extreme-
rare in S. This ‘o’ can stand for either /o/ or /q/. These two forms
are found in texts k and qn. Text k has further only a nakkvarr and two
nakkvat forms. Text qn has further a nekkverr and a nqkkvi form. It is
Possible to interpret both forms with ‘o’ as forms of the nakkvarr par-
adigm.
Altogether 9 of the 13 exceptional forms occur in the two younger
texts e and nn
Concluding we can say that in S nakkvat (22 times) is more fre-
9uent than nekkvat (10 times), and that nekkverr is the main paradigm
with 70 forms, whereas nakkvarr (twice), nakkverr (twice) and
nekkvarr (once) are exceptional. The 11 n0kk- forms are concentrated
m two younger texts. Of a potential nokk- or ngkk- paradigm only two
forms are found, which can also be interpreted as belonging to the
nakkvarr paradigm. By far the largest part of S is completely regular
and has only nekkverr in combination with either nekkvat or nakkvat.
3- Alexanders saga
*n his footnote 26 Hreinn Benediktsson describes his normalization of
ttle forms in Grágás and Alexanders saga. He remarks there that
n°kkurr is in these texts usually spelled nockorr, but nQckorr in about
of the cases in Grágás, and one nöckorr in Alexanders saga. The
tatter example does, however, not occur in AM 519 a 4° . It is found
m AM 226 fol., which Finnur Jónsson used to fill a lacuna in AM 519
a- The spelling with ‘o’ does not necessarily imply the phoneme /o/,
as /ö/ in AM 519 a is in the vast majority of cases written ‘o’, although
/ö/ from original /0/ shows far more often spellings indicating that the
v°wel results from umlaut (‘0’, ‘q’ ‘0’, ‘au’, or ‘ey’).
Where the paradigms for the Homily Book can be easily supple-
niented with frequency numbers by consulting Larsson’s Ordförrádet,