56 COMPARATIVE LANGUAGE 



of dialects. 1 What else does the great diversity of the Italic dialects, 

 with their marked divergence in the most common words as well as 

 in the grammatical material, indicate but that the hordes and tribes 

 which invaded Italy were far from uniform, and that the linguistic 

 unification accompanied the political and economic consolidation of 

 Italy under the leadership of Rome? If we once admit that the Indo- 

 European ethnic group long before the opening of history inhabited, 

 and developed in, a large area embracing Middle and Eastern Europe 

 and reaching far into Western Asia, then the assumption of a well- 

 rounded and evenly developed grammatical structure 2 becomes as 

 impossible as that of a uniform culture. It is not necessary to assume 

 that all Indo-Europeans once possessed knowledge of, and terms for, 

 agriculture, and that the absence of such terms among the Eastern 

 branch is due to loss. Are we not justified in seeing here primitive 

 differences? Exactly 3 so it seems to me unnecessary to assume a 

 fully developed and generally accepted differentiation of optative 

 and subjunctive throughout all Indo-European territory. Is it not 

 possible that the Italic tribes, for instance, did not fuse what was 

 originally distinct, but represent a section and stage which never 

 utilized the ie : I etc. forms for the purpose of differentiating between 

 wish (optative) and will (subjunctive)? And may it not be just as 

 incorrect to speak of the meaning of a common Indo-European 

 optative as it would be to speak of a common Indo-European agricul- 

 ture? In other words, are not many supposed losses and fusions in 

 reality rather primitive and original local absences and primitive 

 and original local failures to differentiate? It seems to me that con- 

 siderations like these must have been the cause which have led, in 

 recent standard works on comparative syntax, to the substitution of 

 " Gebrauchssphaere " for the older "Grundbedeutung." The latter 

 implied unity, local uniformity; the former puts in its place multi- 

 plicity, and thus allows for primitive local differences which we may 

 find continued in the historical languages. 



I am not here attacking the starred, constructed forms of our 

 comparative grammars, the value of which no sane scholar under- 

 rates. What I try to combat is the belief that these constructed 

 forms can be utilized for historical inferences. Since the method by 

 which they are produced is purely logical (namely, a summation of 

 correspondences and an elimination of differences), their character 

 is essentially unhistoric. But this unhistoric quality in no way im- 

 pairs their value as aids in the grammatical study of a given lan- 



1 Heyck, Histor. Zeitsch., vol. 85, p. 68; Wrede, ibid., vol. 87, p. 39; Koegel, 

 Gotting. Gel. Am., 1897, p. 648. 



2 A. Ludwig wrote in 1867: " Auch die Ursprache in der Zeit ihrer (nattirlich 

 immer nur relativen) Vollkommenheit bot keine vollkommerie Einheit." (Si- 

 tzungsberichte d. phil. hist. Cl. d. Kais. Akad. d. Wiss. Wien. 1867, vol. 55, p. 134.) 



* The following paragraph has been elaborated more fully by E. P. Morris and 

 the author in the Harvard Studies in Classical Philology for 1905. 



