Bibliotheca Arnamagnæana - 01.06.1979, Page 132

118
Besides its use for æ and æ, p stands for e (Jré.ssa 2v30, and XXXIII béjoit
2v27, véll 2v21, jjétta 2vl6 etc.) and, although there are no instances in.XB,
also for é (XXXIII grét 2v8, mér 2r29 etc.). The digraph ie also occurs once
in hieåaN 2v 15, a word which had once been pronounced with a short e that
later underwent lengthening. In this case ie presumably stands for é, since é
was occasionally written ie, particularly after h, even in this early period.
The diphthong ei is written with and without the superscript accent and
subscript hook: heims lr26, heilags lr25, méinlayst lr5, néitem lr5 etc.
The graph sequence ei, marked and unmarked, stands for i in himinn
presumably as a result of confusion with heimr: heimneskra 1 r26, héimiNS
[?] lrl, but cf. himna XXXIII 2v6.
The diphthong ey is so spelt in XB: fa/roneyte 2v22, treystiz 2v25, deyia
lrl8, but once éy (déyia lrl6). In XXXIII, other spellings also occur: æy
(æyj^emårk lr5, cf. eyjjemorkiNe 2v2) and y (hyrf)e 2r20).
The diphthong au is written with the graph sequence av (saclavs/r
1 r 18, bavj} 2v 11, mandavjrr 2vl3, |)av 2vl6), except for one occurrence of
ay (méinlayst lr5) and one of æy in XXXIII (ræynir lr20). (Unger,
doubtless mistakenly, corrects to ræyn(d)ir). The ligature æ does not
stand for the diphthong in this hånd.
Having examined the vowel graph system of hånd B, one may consider
the question of the use of the accent and subscript hook. In faet, there
appears to be no pattern in the use of marked graphs and graph clusters.
With only two exceptions, the marked graphs are used also in their
unmarked forms to represent the same phonemes, and with equally few
exceptions the marked graphs and graph sequences are used for more than
one phoneme.
The accent mark, apart from its use in combination with the subscript
hook is employed only with monophthongs in X and indicates vowel
length in about 2/3 of its occurrences: vårt lr4, for 2v23, oarga 2v27, dyr
2v27, håtiå 2vl 1, stigaz 1 r21 etc., but lvnd lr3, ségia 2vl3, lands 2vl7 etc. A
high proportion of these instances involve monosyllabic words, but as
examples above show, by no means all.
In unstressed syllables, e is very mueh more frequent than i (the ratio is
about 5:1). However, the distribution is not random; e occurs normally in
word final position (e.g. sigrhelge 2vll, kvNe 2vl3, enge 2vl3 etc.), and i
elsewhere (e.g. allir lr2, eggiat>ir lr2, farhs 2vl0 etc.). In XXXIII, e and i
occur with approximately equal frequency in the distributional pattern
just described as a result of the reduced use of the er abbreviation sign. Of