This fact, as a bare undeniable fact, should have startled the Darwinians, even as it startled the venerable Darwin, when I simply set the facts before him, and he immediately drew the necessary consequences. Of any danger there could be no fear. The facts are there and show us the right path. And it is not only simple facts, but the consequences of preëxisting conditions which render every so-called transition from animal to man absolutely unthinkable. Language—as ethnologists should have learned—has neither originated from artificial signs, nor from imitation of sounds. That we can communicate with signs without saying a word, that we even now use signs in our speech, is best learned in southern races, and in such pantomimes as _L’enfant prodigue_. We have long known that imitations of sound exist in greater or lesser numbers in every language, and how far they can reach has probably never been shown in such detail as by myself.(49) But that our Aryan tongues, and also the Semitic, and all others that have been studied scientifically, originated from roots, is now generally known and recognised. That these roots may in remote times have contained an element of imitation, we may readily concede, for it is really self-evident; only we should not from the beginning bar our way by conceiving them as mere imitations of sound. If this were so, the problem of language would long since have been solved, and the first formation of ideas would require no further reflection. It must be conceded on the other side that the origin of roots still contains much that is obscure, and that even Noiré’s _clamor concomitans_ does not explain every case. Only it is firmly established that a scientific analysis of language leaves a certain number of roots which are not mere sound-imitations, such as “bow wow,” or “moo moo.” There are people who have taken much pains to discover whether the roots ever had an independent existence, or if they have merely been scientifically abstracted, or shelled out of the words in which they occur. These are vain questions, for we can never of course come at the matter historically, and the attempt to prove the necessity of the one or the other view is a useless undertaking. It appears to be the most reasonable plan to assume for the Aryan languages a period that approaches the Chinese, in which roots had the same sound and the same form as the corresponding noun, adjective, and verb. Even in Sanskrit roots appear at times still unchanged, although it is quite right that as soon as they take on grammatical functions, they should no longer be called roots. Much may be said in favour of both views, without arriving one step nearer our goal. If we now only remember that the whole Sanskrit language has been reduced to 121 primitive ideas, and that the roots denoting these (which are of course much more numerous) are not imitations of sound in the strict sense of the word, but sounds about whose origin we