Íslenskt mál og almenn málfræði - 2020, Page 334

rics and linguistics call for response from me. I see his other two parts as good
additions to my thesis, in addition to issues mentioned by Driscoll.
In the metrical part Myrvoll finds many faults with the notation that I use and
prefers the traditional notational system of Sievers. He begins by noting a change
in my notations, before and after AD 1000. To address his objections, I must
introduce my notation. It is traditional to group lines (verses) of an Old Norse
poem into rhythmic groups and to produce tables of lines of the same rhythmic
type. I used a simple method for doing so, based only on observable features; syl-
lable types and the location of nominals. For Egill’s poetry I find it remarkable and
fortunate that it is possible to use the same labels as Sievers did and to distinguish
between all his rhythmic types, by taking note of these observable features. This
was not my primary goal, but I imagined that it would please those used to the
Sievers types that I grouped together the same lines as they would, even if I did
not refer to all the features of the Sievers types, such as a secondary stress and the
location of alliteration. This works well for all kviðuháttr poetry before AD 1000.
After that, changes take place that require more factors (such as word divisions) to
be included if the same distinctions are to be kept (this is not addressed in the the-
sis). As an example, I can label the trochaic type in Arinbjarnarkviða as A1, like
Sievers did, although I denote it by sxsx. Here s stands for a strong syllable that
must have a heavy syllable type, and x can only stand for an unstressed syllable
(having a vowel in the a-i-u system described by Hreinn Benediktsson (1972) in
his edition on the First Grammatical Treatise). This holds true until AD 1000 when
the second position in a trochaic line became capable of holding any type of sylla-
ble, but it continued not to be strong. To denote such a position I use the symbol
v. For that reason I denote trochaic lines in the young poems Nkt and Hkv by svsx
to allow for all possibilities. My change in notation is thus due to a change in the
meter and should not be objectionable.
In my opionion, the above illustrates a flaw with the Sievers system. It is too
admissive. It describes kviðuháttr before and after AD 1000 in the same way.
Myrvoll wishes to defend the Sievers system against the simplified system
that Kristján Árnason (1991) introduced for dróttkvætt and against the notation
that I introduced for kviðuháttr. In both instances, his arguments can be turned
against the older system. His first argument regards an interesting feature of
dróttkvætt, namely that a compound may fill the three last metrical positions of a
dróttkvætt verse. Myrvoll asserts that the Sievers system explains this, but it only
allows it. It is better to allow such oddities to stand out; there may be an inter-
esting linguistic explanation for them, as I hope to be able to demonstrate for this
particular phenomenon at a future date. The same goes for the alliteration on the
antepenult, on which I included some discussion in my thesis. Myrvoll claims I
am wrong when I say that the Sievers types and Craigie’s law directly conflict
with each other. The Sievers type in question prescribes a heavy syllable for
nouns where Craigie’s law prescribes a light syllable. This cannot be reconciled.
Þorgeir Sigurðsson334