60 COMPARATIVE LANGUAGE 



In another direction, also, a study of the process of the spread 

 of linguistic changes, combined with a study of the mechanism 

 of dialect formation and early tribal migrations, would be of con- 

 siderable interest. Hirt, some time ago (IF, IK, 292), directed at- 

 tention to the similarities, both phonetic and morphological, in 

 neighboring but unrelated dialects. 1 "It is a well-known fact that 

 the same phonetic changes are met with in different but adjacent 

 dialectal areas. Most striking are such parallelisms in the languages 

 of the Balkan peninsula. Though much is uncertain, one fact is 

 plain, namely, that Rumanian, Albanian, and Bulgarian, three 

 fundamentally different languages, possess similar features which it 

 is hardly possible to ascribe to mere chance." It seems possible to 

 explain these similarities by assuming at the beginning a large 

 number of many small ethnic units of great mobility and only mod- 

 erate coherence. 2 These, moving with considerable ease 3 within a com- 

 paratively large geographical area, combined, often only temporarily, 

 with other units into larger bodies which may frequently enough 

 have employed a variety of heterogeneous dialects. These, accord- 

 ing to the degree of intensity of intercourse and according to the 

 duration of the union, could not help influencing each other. Finally, 

 a certain number of these units permanently consolidated, and, being 

 held together by a common material civilization, they began to form 

 a larger and more coherent unit, became more and more closely 

 knit as time went on, and in the same proportion in which the mem- 

 bers of this new body politic coalesced and began to feel their unity, 

 they were further and further separated from their neighbors, and 

 this contrast, which grew up on a political and economic basis, was 

 reflected in the independent development of the language which the 

 new group produced. Such must have been the process which gave 

 rise to definite dialects, 4 and this manner of forming them explains 

 why though in historical times we have clearly established dia- 

 lectal boundary zones we yet find surprising correspondences 

 between dialects which, in historical times, are completely inde- 

 pendent and distinct. They are due to the admixture of small roving 

 bodies of a different linguistic complexion which were themselves 

 absorbed by the larger mass, but which left a trace in the language of 

 those with whom they united. 



And, finally, if we maintain the distinction between primary and 

 secondary changes, we shall look for the causes of a change only 

 where that change is primary. It is, of course, true that all sec- 



1 On such similarities and their explanation see Kretschmer, Einleitung z. Gesch. 

 d. griech. Sprache, p. 98. 



* Edward Meyer, Gesch. d. Alterthums, n, p. 44; Kretschmer, loc. tit. p. 75. 



8 Note, e. a. the ease with which German tribes moved and combined, Lamp- 

 recht, Deut. Gesch. i (1891), p. 7; Goes. B. G. i, 31. 



4 Bremer in Paul's Grunariss, in, pp. 747, 763. 



