PROBLEMS IN GERMANIC PHILOLOGY 293 



in the case of the latter vowels, and I consider it therefore advisable 

 first to examine the relation between u and o before proceeding 

 further with the i and e. Germanic u is of twofold origin, because it 

 corresponds (1) to Indo-European u, e. g. in the preterite plural and 

 past participle of the second ablaut-series, i. e. in forms like Goth. 

 budun, budans. Also in words like juk "yoke," ufar "over," etc. 

 (2) Another u is developed before or after syllabic liquids or corre- 

 sponds to an Indo-European weak vowel. Here belong the u found 

 in the third and fourth ablaut-series, e. g. Goth, hulpun, hulpans, 

 O. H. G. wirtun (= Mod. Ger. wurderi), Goth, numans ; and the u 

 of nouns like Goth, wulfs "wolf," hund "hundred," guma "man," 

 and many others. 



Both varieties are treated in the Germanic languages in exactly 

 the same manner. In Gothic they both remain u, except before 

 r and h, where they are changed (or "broken" according to Grimm's 

 terminology) to a vowel which was probably pronounced o, but, 

 curiously enough, is spelled au, e. g. Goth, waur^un (=Mod. Ger. 

 ivurderi), tauhun (=Mod. Ger. zogen). In Old High German both 

 are treated in accordance with Holtzmann's rule, e. g. zugun, gi- 

 zogan; wurtun, gi-wortan. 



The twofold origin then of the Germanic u stands in no connection 

 with its twofold form (u and o) in the Germanic languages. And 

 there is no reason why we should not regard the u of Goth, juk as 

 identical with that of Lat. jugum, Greek vyov, Sanskr., yugam. 

 The o, therefore, of O. H. G. joch must be regarded as a later sub- 

 stitute for the Gothic vowel. 



As regards the relation between u and o in Germanic, the opinion 

 advocated here is probably the one which is at present shared by 

 the majority of scholars. It is true that here and there we still meet 

 with the doctrine of an Early Germanic o, preserved in West Germanic 

 and changed in Gothic to u. This doctrine, however, finds no support 

 in the u- and o- vowels themselves, but is based on the parallelism 

 of the i- and e- vowels; the supposition being that if 0. H. G. 

 e is older than Goth, i, it would seem rational to regard O. H. G. 



as older than Goth. u. I am for my part willing to admit that 

 the vowels e and o in West Germanic are strictly parallel. But the 

 conclusion I would draw from this fact is a different one. If it is 

 possible to regard West Germanic o as a more recent vowel than the 

 Goth, u, the problem may be solved by regarding West Germanic 

 e as more recent than the Goth. i. Let us return then to the vowels 



1 and e, in order to see whether the latter theory can be sustained. 



Like Goth, u, Goth, i also has a double origin; for it corresponds 

 (1) to Indo-Eur. i in the preterite plural and past participle of 

 the first ablaut-series, e. g. Goth, bitun, bitans, and in words like 

 i-s "he" = Lat. is, fisks = Lat. piscis, widuwo Lat. vidua; (2) to 



