PROBLEMS IN COMPARATIVE PHILOLOGY 61 



ondary changes must have originated as primary changes, and as 

 such they are the direct result of certain forces. But as secondary, 

 that is, adopted changes, they appear where these forces never 

 existed. Not everywhere, therefore, where a certain change is ob- 

 servable may we expect to find the causes also to which it is due; 

 such generative forces can only be discovered where the change is 

 primary. It is wrong to infer that the mere use of a certain syn- 

 tactical form is prima facie evidence of a given mental attitude. 

 As soon as any syntactical phenomenon, such as the order of words 

 in a sentence, has become habitual, it is vain to seek for the causes 

 which lead a given speaker to arrange his words in the accepted 

 order. 1 And Siitterlin (Das Wesen der sprachlichen Gebilde, p. 11) 

 very properly points out that for the modern naive French speaker 

 the analytic il a aime is as much a unit as was the synthetic amavit 

 for a Roman. "At the time when the phrase il a aime was first 

 created, the single elements were still comparatively clearly felt; 

 but after it had once become habitual [that is, when uttered by those 

 who simply imitated it] it was fused into one whole. As a matter, of 

 fact, the uneducated Frenchman has no idea whether he pronounces 

 one word or three." In a similar way we may speak of the grouping 

 and moulding of a compound concept in the sentence only in those 

 cases where the process is really one of original analysis, but not in 

 those cases where we have a repetition of an analysis already made 

 and cast into linguistic form. 2 A good portion of the ordinary talk of 

 many persons is undoubtedly of this second, mechanical type. 3 



I pass over the problem of the origin of Indo-European inflection 

 whch has been discussed in its various bearings in two very recent 

 papers by Delbriick (Einleitung in das Studium der indogermanischen 

 Sprachen, 4th ed. p. 127) and Hirt (IF, xvn, 36). The latter especially, 

 collecting the scattered results of previous investigations (including 

 his own concerning Indo-European gradation and dissyllabic bases), 

 makes four points perfectly clear: namely, that the inflectional 

 system of the Indo-European languages was preceded by an inflec- 

 tionless period, traces of which are not at all rare in the historical 

 forms. Second, that the distinction of verbal and nominal inflection 

 is not original and that the whole sentence-architecture of the Indo- 

 European, with its characteristic division into subject and verbal 

 predicate, is a secondary growth. Third, all the tense-formatives 

 do not originally refer to time, but to the kind of verbal action. 

 Their tense-force is secondary throughout. And, fourth, that a 

 certain number of what used to be considered suffixes (but not all) 

 are not external accretions, but are the final syllables of a base, 



1 Wundt, Vdlkerpsychologie, Die Sprache, n, p. 365 ff. 



2 Compare Jerusalem's discussion of the " Ennnerungsurtheil " in his Urtheils- 

 f unction (1895). 



3 Compare Howells's Lady of the Aroostook, pp. 106 and 215. 



