Bibliotheca Arnamagnæana - 01.06.1999, Síða 219

and Introduction to Part Two
199
to declare as a matter of principle that the outcome of a translation will
have to be considered as a new text, and the same might be the case for
a translation from Proto-Norse to Old Norse. A consequence of this view
would be that it is not meaningful to assign a date prior to the period of
syncope to an Eddie text. In view of the abstraetness of our conception
of text this would be untimely, however, and would exelude important
contributions to the discussion.
To sum up: as a matter of principle, we shall make two methodologi-
cal decisions, which to a certain extent are arbitrary, but which are not,
as far as 1 can see, accessible to empirical refutation. They are thus un-
empirical decisions made in order to enable an empirical study of dat-
ing, based on the history of linguistics, stylistics etc.
1) Provided that the extant text is supposed to be some sort of recording
of an oral text and not composed by the scribe, this earlier text is
taken to be represented by the written text. They are taken to be the
same text, in a rather vague interpretation of ‘same’ and of ‘text’.
2) If there are not particularly strong arguments against it, the text is
considered as a unified whole.
The notion of dating
On the basis of these considerations it is possible to define our notion of
dating somewhat doser. The Edda consists of several collections of
written texts, all of which were committed to writing after 1200. On the
presumption that all or most of them have been orally transmitted before
being written down, the written texts are taken to represent the oral texts,
which have been composed in different periods of time. It is possible to
discuss the date of a text to the extent that we presuppose that its longer or
shorter history has left traces in the written form, which is accessible to
empirical study (cf. Schier’s points c-f). The immediate object of study is
these traces. A significant number of traces of early date are taken to
mean that the text is old, if this is not refuted by stronger arguments. If
there are no such traces, the text is considered to be of a later date.
Erik Noreen has remarked that a young poem may contain archaic
features, as is, in his view, the case for Lokasenna. He may be right, but
in the absence of more specific arguments, old-fashioned features will