PROBLEMS IN GERMANIC PHILOLOGY 299 



to be remnants from the Primitive Teutonic period appeared to yield 

 more naturally to a different explanation, until -gastiR remained as the 

 only instance in which there is at least the semblance of greater 

 antiquity than in the corresponding Gothic form. I have no doubt, 

 then, that in this case, too, appearances are deceptive; and whether 

 or not we are able to explain the -4- with certainty: we are hardly 

 warranted in ascribing, on account of this form, a Primitive Teutonic 

 character to the language of the Runic inscriptions. 



For the present I would suggest the following explanation. It is 

 perhaps not necessary to identify -gastiR with Goth, gasts. We must 

 remember that -gastiR only occurs as the second member of com- 

 pounds, and that compounds belonging to i-stems sometimes follow 

 the ja-declension. To be sure, Gothic has no compounds in -gasteis, 

 but it has, e. g. the adjectives af-haimeis and ana-haimeis belonging 

 to the t-stem haims. If this explanation be correct, Runic -gastiR 

 would be in Latin not hostis, but * -hostius. 



The next form to claim our attention is the accusative horna, which 

 I have translated by the plural "horns," while it is generally regarded 

 as a singular, corresponding to Gothic haurn. But we must remember 

 that near Gallehus not one, but two, golden horns were found. Both 

 were adorned with figures in the same style, and evidently belonged 

 together, but only one of them had an inscription. Hence the inscrip- 

 tion was probably meant for both horns. There may have been three 

 horns originally, intended perhaps to decorate the wall of a festival 

 hall in such a way that the larger one with the inscription hung 

 in the middle, and a slightly smaller one on each side. But even if the 

 number amounted from the outset to not more than two, the artist 

 might have spoken of them as " the horns." There is no reason, then, 

 to regard horna as a pre-Gothic form, since the nominative-accusative 

 plural of haurn in Gothic is haurna. 



Finally we have to examine the preterite tawido, a form which in my 

 opinion sheds a clearer light on the language of these inscriptions than 

 any one of the words with which we have concerned ourselves so far. 

 The current interpretation is that the form tawido, which finds a 

 parallel in faihido and similar Runic preterites, appears as a very old 

 form, not only as compared with the Old Norse weak preterites in -8a, 

 but also as compared with the corresponding Gothic forms. As the 

 ending of the first person singular of the Gothic weak preterite is -da 

 (or in other cases -\>a and -to) , this view may seem at the first glance 

 unobjectionable. And yet it betrays clearly its origin from a time 

 when the comparative study of the Germanic languages stood in its 

 first tentative stages. For the ending -8a of Literary Old Norse cannot 

 be identified with the Gothic ending -da, because the latter would 

 have to appear in Old Norse as -3i. This may not only be inferred 

 from cases like Goth. hana=O. N. hani, Goth. fadar=*0. N. fatir, 



