114 PRINCIPLES OF LINGUISTIC SCIENCE. 



possible that an original human race ghould have separated into tribes before 

 the formation of any language so distinctly developed, and of such fixed forms, 

 as should leave traceable fragments in the later dialects of the sundered por- 

 tions. x\.moug all the varieties of human speech there are no differences which 

 are not fully explainable upon the hypothesis of unity of descent. 



That the linguistic student also cannot bear positive testimony in favor of 

 such descent is equally demonstrable, although not by so direct an argument. 

 There is here no theoretic impediment in the way, but a practical one. It might 

 be hoped that traces of an original unity would be discoverable in all parts of 

 human language ; only examination could show that such is not the case. But 

 investigation, however incomplete, has already gone far enough to leave no 

 reasonable expectation of making the discovery. 



The processes of linguistic change alter the constituent parts of language in 

 every manner and to every degree, producing not only utter difference between 

 words which were originally one, but also apparent correspondence between 

 those which are radically unconnected. There are no two languages on the 

 face of the earth between which a diligent search may not bring to light resem- 

 blances which are easily proved by a little historical study to be no signs of rela- 

 tionship, but only the result of accident. Now, the more remote the time of sepa- 

 ration of two related languages, the more numerous will be their differences, the 

 more scanty their resemblances ; hence, the more ambiguous will be the indica- 

 tions of their connexion ; until finally a point is reached where it is impossible 

 to decide whether apparent coincidences which we discover are genuine, or only 

 accidental, and evidence of nothing; and, in the comparison of languages, that 

 point is actually reached. When we come to hold together the forms of speech 

 belonging to different families, the evidence tails us. It is no longer of force to 

 prove anything to our satisfaction. The families are composed of such lan- 

 t^-uao-es as can be seen to have grown together out of the radical stage. If there 

 is community between them, it must lie in their roots alone ; and to give the 

 comparison this form is virtually to abandon it as hopeless. To trace out the 

 roots of any family, in their ultimate form and primitive signification, is a task 

 of the very gravest difficulty. By the help of the great variety and an- 

 tiquity of its dialects, and especially by the Sanscrit, the task can be somewhat 

 satisfactorily accomplished for the Indo-European tongue; but the Semitic 

 roots, as already explained, are of the most perplexingly developed form. Radi- 

 cal correspondences among the great branches of the Scythian family are hardly 

 sufficient to prove the ultimate relationship of those branches ; and to hope that, 

 in the blind confusion of Malay, African, and American dialects," linguistic 

 analysis will ever arrive at a confident recognition of their primitive germs, is 

 altogether futile. Accidental correspondences are, if anything, more likely to ap- 

 pear among roots than in the forms of developed speech. Authorities arc much 

 divided upon the question whether the Indo-European and Semitic families are 

 proved connected, with a decided preponderance of the best and safest opinions 

 on the negative side. If it may possibly be hoped that their connexion will 

 yet be established, with the help of evidence coming from outside of language, 

 the same hope cannot be entertained as to the connexion of either of these with 

 any other family, and yet less as to the inter-connexion of all the families. 



We come, finally, to consider the origin of language. We may claim that 

 the problem has been greatly simplified by what has already been proved as to 

 the history of speech. Did we find the latter everywhere and always a com- 

 pletely developed and complicated apparatus, we might be tempted to despair of 

 explaining its origin otherwise than by the simple hypothesis of a miraculous 

 agency. But we have seen that the wealth of the noblest tongues comes by 

 slow accumulation from an early poverty. We have only to satisfy ourselves how 

 men should have become possessed, at first, of the scanty and humble germs of 

 language. And, in the first place, there is no reason for supposing them 



