Further Old Norse Secondary Formations

Abstract

Certain Old Norse secondary formations were discussed in Lg. 29.26-33 (1953). Here I shall continue the discussion by adding examples which at the time of my previous article I had not yet investigated. 1. THE LENGTHENING OF THE VOWEL e IN OLD ICELANDIC eta > eta 'TO EAT'. Initial short vowels sporadically suffered lengthening in Old West Norse, more often in Old Norwegian than in Old Icelandic.' This lengthening represents the beginning of a trend which later extended to a large number of words. In the few examples recorded in Old Norwegian and Old Icelandic there is no plausible explanation of the lengthening except as due to the initial position of the vowel with accentual variation, a purely phonetic process. Old Norwegian af > df 'of', at > dt 'at', akr > dkr 'field' are on a level with Old Icelandic ek > &k1 'ego', eta > eta 'to eat'. Noreen, nevertheless, assumes that the lengthening of the initial vowel e > e in eta > eta does not represent a phonetic process but was probably of analogic origin, the long vowel e following the pattern of the long vowel d in the anomalous preterit singular form it, a type of borrowed lengthening.2 This assumption is unfounded, primarily because Noreen has disregarded the lengthening of the initial vowel e > e in words which do not belong to a verbal ablaut system, such as the pronoun ek > ek, which cannot be of analogic origin. There is therefore no reason to assume that the forms ek and eta are on different levels. Furthermore, Noreen has not justified his assumed type of borrowed lengthening. He has not answered the question WHY this borrowing should have occurred, but has evidently based his assumption upon the fact that the vowels in eta and dt are both long. But this fact is no evidence that one vowel (6) became long because the other (d) was long. If Noreen's assumption were correct, we might expect the non-initial vowel e in the type vega 'to slay', fregna 'to ask' (of the same ablaut series) to have likewise suffered lengthening in imitation of the long vowel in the preterit singular vd and frd. That this never occurred is evidence that the vowel in eta was lengthened because of its initial position, parallel to ek > ek. 2. Fiandr, NOMINATIVE-ACCUSATIVE PLURAL OF fiandi 'ENEMY'. The substantive fiandi represents the present participle of the verb fia 'to hate', substantivized as a consonantal nd-stem. The nom.-acc. plural has the ending *-iR. We should therefore have expected, instead of fiandr, a form *fiendr (< *fiandiR) with i-umlaut of the vowel *a in the suffix. It is rightly assumed that the vowel a in fi-a-ndr was borrowed from the singular paradigm fi-a-nd-i, -a.3 But this is the only recorded example of an Old Icelandic nd-stem with such borrowing.