Íslenskt mál og almenn málfræði - 2020, Síða 324

is still a question, however, if this justifies Þorgeir’s solution of “removing the
weight distinction” (p. 147); this would make the existence of Verschleifung (res-
olution) in odd verses (pp. 155–57) a puzzle.
To substantiate his new finding, Þorgeir is in need of an analytical tool to dis-
tinguish between those syllables that may fill the dips in trochaic verses of
kviðuháttr, and those that may not. For this purpose he invents the terms “re -
stricted vowels” (Old Norse unstressed i, a and u, on p. 42) and “restricted sylla-
bles”, which contain these vowels only, besides “full syllables”, which can be both
light and heavy (p. 96, cf. the table on p. 106).2 His choice of terminology could,
however, be questioned. For instance, it could be regarded as unfortunate that
“light” and “restricted” syllables, which share the same property of filling dips or
unstressed positions in the metre, are at the highest level assigned to different
groups of syllables. In Þorgeir’s system, verses such as Arbj 3.8 heim um sóttan
and Nóregs konungatal (Nkt) 32.2 land með hringum are scanned differently, as
sxsx vs. svsx (pp. 125, 130), where s = “strong”, v = “weak” and x = “restricted”.3
The only reason for this difference is the restriction against function words (such
as the preposition með) in dips in the oldest kviðuháttr poems from the tenth cen-
tury. Þorgeir even changes his notation as he proceeds: whereas Arbj 5.2 tryggt at
líta is scanned sxsx (p. 130), with a “restricted” infinitive marker, Sturla Þórðar -
son’s Hákonarkviða (c. 1264; Hkv) 41.2 öll at líta is scanned svsx (p. 126), with a
“weak” infinitive marker, apparently because the metre at the time of Hkv had
blurred the distinction between “restricted” and “weak” syllables.
In section 5.2, on syllable weight and Craigie’s law, Þorgeir discusses the dif-
ferent frameworks for metrical scansion applied by Kari Ellen Gade (1995) and
Kristján Árnason (1991), and takes side with Kristján’s “simplification” of the
Sievers system from five to three rhythmic types, which essentially represent the
Sievers types A, C and D (example verses on p. 105). For all its simplicity, Krist -
ján’s classification fails to explain some crucial features of Old Norse poetry,
most importantly the fact that a compound may fill the three last metrical posi-
tions of a dróttkvætt verse, e.g., raddkleif at Þórleifi (Þjóð Haustl 1.4), or lǫgsóta ver-
fótum (Eyv Lv 13.2). These (and other examples) have to be scanned according to
Sievers’ type E, with a stressed fourth position, since otherwise the compound
would be assigned a stress pattern that is highly anomalous from a Germanic per-
spective (unstressed–stressed–unstressed). That this metrical pattern (Sievers’
type E) exists is confirmed by dróttkvætt verses like rauð fnýsti ben blóði (Hornkl
Gldr 2.5) and gerum þar fyr sjǫt sólar (Eg Lv 6.7), with alliteration on both fourth
and fifth positions, following the principle “no alliteration without stress” (see
the discussion of this principle below).
Klaus Johan Myrvoll324
2 The term “restricted syllables” is used already on p. 58, without any explanation or
definition.
3 One may ask why “weak” is not abbreviated w rather than v, but this is immaterial.