PROBLEMS IN COMPARATIVE PHILOLOGY 57 



guage. Instead of attaching to these constructed forms the historical 

 value which they possessed in Schleicher's eyes, I should rather 

 regard them as algebraic grammatical notations * in which the com- 

 parison of parallel forms of two or more languages may be most 

 easily and conveniently summarized and expressed. The construc- 

 tion of a form like kmtdm means that Greek C-KOTOV, Latin centum, 

 German hund, Sanskrit catam, etc., should be grouped together. It 

 is in other words the common denominator of these forms. The 

 constructed m is a convenient symbol to mark the fact that Greek 

 a, Latin en, German un, Sanskrit a, etc., are to be paralleled and are 

 alike in so far as they are weak grades, a fact which for some does not 

 lie at all close to the surface and, indeed, is brought to light by the 

 comparison of cognates only. The signs which go to make up the 

 alphabet of the Indo-European are the symbolical expressions of 

 grammatical parallelisms rather than representatives of the historical 

 sources from which the sounds of the concrete languages are descended. 

 But even those who would grant historical reality to the formal 

 reconstructions of Indo-European words will hardly go so far as to 

 extend it to the semantic 2 reconstructions dealing with the force 

 and meaning of Indo-European cases, modes, and tenses. So long 

 as it was believed that from the very beginning the mode and tense 

 formatives were charged with a definite modal and temporal meaning 

 inherent in them, a formal reconstruction of the formative carried 

 with it semantic reconstruction also. But all recent investigations 

 (they have just been summarized and extended by Hirt 3 ) uniformly 

 tend to show that there was, generally speaking, no such inherent 

 meaning in these formatives. What we call the modal or the tense- 

 system of a language is the result of a very gradual development in 

 which old formal material has been adapted to certain semantic 

 uses. Witness, for instance, the use made of thematic (asa-ti) and 

 unthematic (as-ti) forms for the differentiation of subjunctive and 

 indicative, or the turning of the s-formative into a tense-sign. If, as 

 seems incontestable, the tense-system of the Indo-European lan- 

 guages is by no means primitive, but a secondary structure, into 

 which material of a previous period was built by charging old forms 

 with new meaning, 4 it is not necessary to assume that this new 

 system was uniformly worked out in what is called Indo-European 

 times, and the attempt to construct universally accepted Indo- 

 European meanings from which, by loss or addition, those of the 



1 This statement agrees with Delbriick, Einleitung in das Sprachstudium (1880) 

 p. 52. Einleitung in das Studium der indogermanischen Sprachen (1904), p. 126. 



2 For a fuller discussion of this see E. P. Morris's and the author's paper in the 

 Harvard Studies in Classical Philology for 1905. 



3 In the seventeenth vol. of the Indogermanische Forschungen. 



4 Compare also the acquired modal force of the augmentless preterites, Thurn- 

 eysen, Bezz. Beitr. vin, p. 282. 



