46 COMPARATIVE LANGUAGE 



thing has been so systematically attacked in efforts to crush nation- 

 ality, from the time of the Assyrian kings, one of whom boasts of 

 having carried into captivity the subjects of the four quarters of the 

 world and made them of one speech, 1 down to the present day, when 

 the "language question" is a burning problem in nearly every 

 European state. 



Recent years have seen much discussion of the nature and origin 

 of dialects, the result of which only emphasizes how inseparable 

 are the linguistic and historical aspects of the question. In the first 

 place, " dialect" is not a purely linguistic concept. It does not, any 

 more than the term "language," represent a definable degree of 

 speech- variation (or speech-unity, according to the point of view). 

 It rests upon a combination of linguistic and historical elements. It 

 is true that in place of the popular conception of a dialect, a precise 

 and purely linguistic definition has been suggested, 2 but it is one that 

 yields a concept too limited in scope to be usable, and if adopted 

 would only necessitate the invention of a new term as elastic as the 



" The subjects of the four quarters (of the world, speaking) strange languages 

 and varied dialects, inhabitants of mountain and plain, over whom the warrior 

 of the gods, lord of all, rules, whom I had carried into captivity in the name of 

 Ashur, my lord, with my powerful staff, I made of one speech and settled them 

 therein." From the cylinder inscription of Sargon (722-705 B. c.), in Assyrian 

 and Babylonian Literature. Selected Translations, R. F. Harper. 



- Oertel, Lectures on the Study of Language, p. 92 ff . Strictly, there is always 

 some variation between the speech of any two persons and even between two 

 utterances of the same person, so that, objectively considered, the only absolute 

 dialectal unit is the momentary utterance of a single individual. To call this 

 a dialect would obviously be absurd. But such variations may be top minute to be 

 noticed, so that subjectively they do not exist. Accordingly it is proposed to 

 make the test subjective instead of objective. " A dialectal unit is constituted by 

 the speech of all those persons in whose utterances variations are not sensibly 

 perceived or attended to. Subjective uniformity makes the dialect," is Oertel's 

 thesis. Higher groups he would classify as dialect-family, language, and language- 

 family, emphasizing that these represent only ideal types in contrast to the con- 

 crete type represented by a dialect as defined. It may be admitted that in this 

 way one can make of dialect a concrete and purely linguistic concept, and one 

 that is somewhat more comprehensive than that obtained objectively. But it is 

 still a too limited concept to which to restrict the term dialect. We could not speak 

 of the dialect of a single town, so long as it included, as often, perceptible varia- 

 tions in the speech of different classes. Its speech-form would rather be a dialect- 

 family. Or, waiving the matter of variation within a single town, we could speak 

 for example, to illustrate from Greek dialectology, of the speech of Tegea in Arcadia 

 as a dialect, but what we commonly call the Arcadian dialect would be a dialect- 

 family, what we commonly call the Arcado-Cyprian dialect-family would be a 

 language, and what we commonly call the Greek language would be a language- 

 family. The fact is, of course, that we cannot have a complete set of terms of 

 absolute value for all degrees of even perceptible variation, and if all but one 

 must necessarily be ideal types, not to be defined precisely, what is the advantage 

 of making an absolute concrete type of this one? Yet we had no right perhaps to 

 illustrate from the Greek, for it is obvious that the term dialect as defined cannot 

 be properly applied to any phase of speech no longer extant. For it is only in the 

 case of living speech that it is possible to take testimony as to what variations 

 are perceptible and so secure the subjective test. 



I see no objection to the continued employment of the term dialect, as of dialect- 

 family, language, etc., in its present elastic sense, its special application being 

 shown by the context. 



