Page:Language and the Study of Language.djvu/204

182 the community which laid the basis of any actually existing language or family of languages, we must admit the existence of some differences between the idioms of its individual members, or families. And if we suppose such a community to be dispersed into the smallest possible fragments, and each fragment to become the progenitor of a separate community, it might be said with a kind of truth that the languages of these later communities began their history with dialectic differences already developed. The more widely extended, too, the original community before its dispersion, and the more marked the local differences, not inconsistent with mutual intelligibility, existing in its speech, the more capital, so to speak, would each portion have, on which to commence its farther accumulation of dialectic variations. But these original dialectic differences would themselves be the result of previous growth, and they would be of quite insignificant amount, as having been able to consist at the outset with unity of speech; they might be undistinguishable even by the closest analysis among the peculiarities of idiom which should have arisen later; and it would be the grossest error to maintain either that these last were original and primitive, or that they grew out of and were caused by the first slight varieties: we should rather say, with entire truth, that the later dialects had grown by gradual divergence out of a single homogeneous language.

In an uncultured community, the value of such minor discordances of usage as may exist, and do always exist, among those who yet, as being able to communicate freely with one another, are to be regarded as speaking the same tongue, is at its maximum. The first effect of the cultivation of a language, as we style it, is to wipe out this class of differences, extending the area and perfecting the degree of linguistic uniformity. And its work is accomplished, first as last, whether the scale of variation over which its influence bears sway be less or greater, by selection, not by fusion. The varying usages of different individuals and localities are not averaged, but the usages of one part of the community are set up as a norm, to which those of the rest shall be conformed, and from which farther variation shall be