an Indian cannot say 'I smoke,' without using such a number of concrete pictures, that his immensely long word to represent that monosyllabic action means, I breathe the vapour of a fire of herb which burns in a stone bowl wedged into a pierced stone.** Imagine such a process of word-formation as this applied to a cultivated or literary language: imagine that instead of the word inapplicabilities, we were obliged to use a sort of printer's-hyphen word like

the plural-condition-of-not-being-able-to-fold-onething-into-another;' imagine this, I say, and then see the immense victory which has been achieved by the Aryan race, in adopting inflectional synthesis as the basis of their grammatical structure. Or, again, take such a word as 6 recommence;' which of us in ordinarily using that very common word is ever likely to be troubled by remembering that it is composed of the six elements, re-cum-in-it-i-a-re, viz., 'to be going into again with;' or which of us ever dreamt till recently that the latter re in that form, the re in which all Latin inflections end, is a demonstrable relic of esse (Sanskrit as), the infinitival form of the verb to be?' And yet, be it observed, that although thousands of such synthetic forms could be selected from English, English has become far less purely inflectional than any other language of the whole Indo-European, or Aryan family. It has discarded its inflections unsparingly;

* Voy, dans les grands déserts du Nouveau-Monde, p. 392.

indeed, so unsparingly that barely a dozen of them in the whole language are left in common use, and the others it has so completely pared down that they are unrecognisable to any eye but that of the philologian. Who, for instance, would imagine that the d in the word had is the relic of an auxiliary verb which once possessed, in a single tense of the middle voice, no less than eight inflections? Yet even a few terminations of this kind furnish an undoubted proof that the fundamental idea and structure of English, as of all the Indo-European languages, is (in spite of its immense development of the analytical process) synthetic and inflectional. It may have happened to some of my audience to have had the rare pleasure of hearing Professor Huxley lecture, and show how the idea of the structure of the entire animal kingdom may be represented by a vertebra with two lateral processes; and how, in a lobster, for instance, every single function of every single articulation, from the jaw to the flipper, is provided for by a modification of this single structure. The demonstration is wonderful, but it is not, I think, more wonderful

than the demonstration of the manner in which man has moulded the faculty of language into a thousand different forms, which yet retain their own marked individuality, and has, without any violence or discontinuity of development, made the simplest pronouns express the most complex multiplicity of conceptions and relations.

2. But this similarity of grammatical structure in all Aryan languages is accompanied by an ultimate identity in the vast majority of roots.

It is now a matter of simple notoriety that not merely in sounds and letters, but in fundamental radical structure-and not only in words which might conceivably have been borrowed from obvious natural sounds, but in words deduced through a long series of imaginative metaphors or fanciful analogies *—the vocabulary of any single Aryan

* These cautions are exceedingly necessary. The hurry of etymologists has led them to see signs of relationship between languages radically distinct simply by virtue of a few onomatopoetic, or even purely illusory, homonyms; yet it does not need much acumen to see that avy has no connection with the German Auge, or laus with the German Laus! Lars, says Professor Whitney, has as much to do with laird as it has with deputy-sheriff. People have compared the Polynesian mati, an eye, with the modern Greek, but the word in modern Greek is simply a contraction of ouμáriov; and the Hebrew kophar with the English cover, although cover is a corruption of conoperire. Klaproth even compared the Japanese ta tchin with the English teaching! If these accidental phonetic coincidences or apparent coincidences of written words were worth noticing, one might connect the Chinese uhr with the German Ohr, or the Jenisei eg with the English egg, or the Galla aba with the Welsh afon, or the Karib alaiba with the Gothic hlaifs (loaf). Pott has some

language, in spite of the effacing influences of time, and the disturbing elements of foreign admixture, stands in a very close relation to the vocabularies of all the rest. The numerals, the pronouns, the most ordinary and essential verbs, the words for all the commonest relationships, for the parts of the body, for nearly all the domestic animals, for the most necessary cereals, and the most familiar metals, are substantially the same in all the languages of this great family. That such is the fact may be seen by any one who will take the trouble to examine a few comparative lists; but it may be more interesting to observe that even when the words in several branches are different, the roots of them all are to be found in the family possession; and that very often when the words are as absolutely unlike each other as they can possibly be, they can yet be deduced, through easy stages of differentiation, from a common original stock. As an instance of the first fact, take the word 'horse.' In Persian it is fâl, in French cheval, in German Pferd, in Anglo-Saxon wicg, in Polish koń; and these words have no connection with each other: yet there is not, I believe, one of them which is not excellent remarks on this subject in Zeitschrift der deutschen morg. Ges. ix. 405 fg. 1855) See, too, Max Müller in Bunsen, Philos. of Hist. i. 356: I believe that there is hardly a word in any language to which, making the usual allowance for change of form and meaning, some other word might not be found almost identical.' One of the first scholars to state this principle clearly was Job Ludolf, in the Preface to his Ethiopic Grammar and Lexicon, 1702.

F

traceable to a Sanskrit root. The root horse' (if it be not, as I myself believe, ultimately an onomatopoeia) may allude to the spirit of the animal (karasa,' what passion'*); cheval, the Latin caballus, is from an old root, capala, swift. Wicg is probably from vága, rapidity; koń, from çôn, to be red' (or bay-coloured), and so on: the simple explanation being that in the Aryan mother-speech the animal had different names, derived from different attributes, and in the struggle for existence which takes place among words no less than among living organisms, the effects of accident caused one form or other to prevail.†

The radical changes which are effected in the

* M. Pictet's remark (Orig. indo.-eur. i. 349) that the word 'cherry' has apparently the same origin ('où cependant le mot rasa a le sens de suc'), will not add to the probability of this derivation. And yet it is certain that language does seize on the most marvellous analogies. 'What the German philosopher described as the relation of a cow to a comet,' says Mr. E. B. Tylor, 'is sufficient, and more than sufficient, to the language-maker-both have tails.' How remote, for instance, is the apparent connection between flies and musquets, or varnish and the golden hair of an Egyptian queen? Who would have expected that the word money derives its origin from a temple of the goddess Juno, or that treacle has anything to do with wild beasts? or that the names naphtha and emery are traceable to legends about Jeremiah and King Solomon? Yet so it is.

It is often curious to watch the effect of this struggle for existence; for instance, gold in Greek is xpuσós; in Latin it is aurum; yet such a word as Onσavpós, 'a treasure,' shows that xpvods is a usurper, and has expelled its rival form which is triumphant in the sister tongue. Similarly, such a compound as Boúkoλos shows that there was once in Greek, no less than in Latin, a verb colo, I tend; but in Greek it has been ousted by νέμω.