PROBLEMS IN GERMANIC PHILOLOGY 295 



u (instead of o) is clearly due to the plural bi-durfun and the subjunc- 

 tive bi-durfi. The i, therefore, of the infinitive wizzan may be ex- 

 plained as due to the plural wizzun and the subjunctive wizzi, an 

 explanation which we shall naturally also apply to the participles 

 wizzanti and gi-wizzan. Granted that gi-wizzan is due to the analogy 

 of the plural wizzun and the subjunctive wizzi, the explanation of the 

 participle gi-bizzan can no longer be doubtful. Its i is due to the 

 analogy of the plural bizzun and the subjunctive bizzi. 



But why should in the first ablaut-series the participle gi-bizzan 

 adopt the vowel of the preterite plural bizzun, while in the second 

 ablaut-series the participle gi-wortan is not influenced by the plural 

 wurtun? I believe that I am able to give a satisfactory answer to this 

 question, and I think that the answer, if it be correct, will prove to 

 be of some interest, as it would lead to the result that to a certain 

 extent analogical changes depend on definite laws just as much as 

 phonetic changes. 



We are concerned in the case under discussion with a law which 

 regulates the ablaut in West Germanic (or more precisely in Norse 

 and West Germanic) in such a way as to require a harmony or an 

 equal balance between the grade of the infinitive or the present tense 

 and that of the past participle. We may say in a general way that 

 Holtzmann's rule does not affect the past participle, unless it can 

 affect also the infinitive and certain forms of the present tense. To 

 be sure, if we adopt this wording we shall have to except the cases 

 in which the infinitive and present are formed with the suffix -/- 

 (e. g. O. H. G. sitzan, "to sit"), since these particular presents 

 have no effect on the past participle. We have to remember, how- 

 ever, that the suffix j is found only with verbs whose preterite 

 plural has the vowel e or o, so that an analogical influence of this 

 vowel on that of the past participle would be out of question. But 

 in order to avoid exceptions of this kind it is preferable to restrict 

 our law to those ablaut-series in which we have in Gothic (and in 

 Primitive Teutonic) one and the same vowel in the preterite plural 

 and the past participle, i. e. to the first, second f and third series. 

 The law then may be formulated thus: (t Wherever the preterite 

 plural and the past participle had originally one and the same vowel 

 admitting of the application of Holtzmann's rule, this rule has taken 

 effect only when it could also affect the infinitive and the plural of 

 the present tense. If, however, the present tense shows in the plural 

 the same vowel as in the singular and in the infinitive, the past 

 participle will retain the vowel of the preterite plural and not be 

 modified by Holtzmann's rule." 



From our point of view, then, we shall divide the first three ablaut- 

 series into two groups, which may be distinguished as simple and 

 complex forms of ablaut. 



