Published online by Cambridge University Press: 22 April 2026
The verbal reduplicating prefix. If the verbal stem in Gothic contained an initial cluster of two consonants, only the initial consonant was transferred to the prefix, except for the indivisible clusters sk and st (no examples of sp are recorded). The few cases of these two types of transference preserved in Wulfila's Gothic are: type 1 fraisan ‘to attempt’: faífrais, gretan ‘to weep’: gaígrot, flokan ‘to lament’: faíflok, slepan ‘to sleep’: saíslep; type 2 skaidan ‘to divide’: skaískaiþ, ga-staldan ‘to acquire: ga-staístald. The digraph ƕ must have also represented two indivisible phonemes hw (and will here be so written for convenience), since hw, not h alone, was transferred to the prefix: hwopan ‘to boast': hwaíhwop. The indivisibility of the clusters sk and st may have been due to the fact that the IE stops *k and *t after s were not shifted, and hence could not be divorced from the preceding s. But this was not the case with the initial clusters fr, gr, fl, sl. To explain the loss of the second phonemes in the prefix we are justified in assuming that in pre-Gothic (i.e. in Proto-Germanic) these clusters were transferred to the prefix, but that the second phonemes r, l were later lost through dissimilation: *fraífrais > faífrais, etc. This loss brought the form of the prefix into the pattern of the more frequent type of reduplication with a single initial consonant; thus, faífrais like faífalþ (from falþan ‘to fold'), saíslep like saíso (from saian ‘to sow'), etc. One may object to this explanation on the ground that analogy alone could have produced the same result, since the great majority of reduplicating verbs in Gothic had a single initial consonant. Perhaps the safest assumption is that the leveling was brought about by a combination of phonetic change and analogy.
1 Geschichte der gotischen Sprache §233 footnote: 'Liegt hier der Rest einer Ableitungsart auf -anan vor?
2 Cf. the confusion between the present and past participle of strong verbs, due to the congruence of the suffix -ans (< -ands) of the present participle with -ans of the past participle; cf. Eph. 2.16 B af-slahans ‘having slain’ (apokteínas) = A af-slahands ‘slaying’.
3 Jellinek 198, §253: ‘Auffällig ist manniskodus* als Adjektivabstraktum (§215); ausserdem wäre der Ausdruck wörtlich übersetzt nicht anthrōpotēs, sondern ein *anthrōpinâtēs, was nicht existiert. Sollte humanitas als Vorbild gedient haben?‘
4 Cf. Lg. 21.97-8 (1945) and the literature there cited.
5 For the IE status in Sanskrit see Kluge, Nominale Stammbildungslehre der altgermanischen Dialekte §134: ‘manniskōdus “menschlichkeit” ist adjektivabstraktum (gebildet wie skr. jîvâtu f. “leben” und avest. ǧyâtu m. “leben”)‘.
6 See Streitberg, Gotisches Elementarbuch 5-6 §30.
7 Orthography and phonemes in Wulfila's Gothic, JEGP 49.225 (1950).
8 Not all forms are recorded, but they can be reconstructed on the pattern of ē- and jan -verbs.
9 Handbuch der vergleichenden gotischen Grammatik 244: ‘Got. hatjan kann auf *qəd-i̭ō zurückgehen ... Dass es erst aus hatan* umgebildet ist, kann man nicht beweisen.‘
10 OS hettian* may be postulated from the recorded nd-substantive hettiand.