Rit (Vísindafélag Íslendinga) - 01.06.1970, Síða 139

i37
well-definable relations—though by no means those of a one-
to-one analogue—to the phonological structure of language.
If so, these relations could become the object of a formal
analysis, and such an analysis would be a prerequisite for the
utilization of poetic texts as linguistic evidence. In the case of
skaldic prosody there is, I think, clear evidence to indicate
that such a special study would be well worth while. And it
might then, indeed, turn out that modern skaldic metrics—as
distinct from old skaldic prosody—has been under undue
classical influence, i.e., has been ‘a carry-over from Latin’ of
a kind.
In summary, all the earliest sources of evidence require the
positing of a separate stage of development, different from the
following stage, as outlined in §3.5; and they are, in turn,
satisfactorily accounted for by the structural characteristics of
this stage, without recourse to ad-hoc assumptions. That the
next manuscript to mark the distinction long vs. short—the
GkS 2087 40, more than a century later—does in fact denote
the vowel quantity in á, nú, búa, etc., quite differently (viz.,
by the accent mark) then implies a new stage of development,
with a different quantitative structure; this relates to the fact
that, in the intervening period, certain vocalic changes occurred
which may well have led to, and were indeed almost bound
to result in, a reorganization of the quantitative structure and
which represented the transition from the earliest to the
following stage of development (ibid. 55-60).
(2) The absence of the so-called younger a-umlaut, e.g.,
in landum, as opposed to its presence in Igndum, is, according
to Einar Haugen, ‘obviously ... a simple case of analogical
simplification within the paradigm.’ There can, of course,
be no dispute that such morphophonemic reorganization
eventually played a decisive role in the development of the
paradigms involved; cf., e.g., Norw. barn, plur. barn, butDan.
barn, plur. bsrn. The earliest (East) Norwegian examples of
such clear analogical forms date from the second half of the
thirteenth century. Thus, what is in dispute is not whether
L