in Sanskrit veda, the non-reduplicated perfect of the root vid, the English to wit, to know. Histōr, therefore, meant originally a knower, or a finder, historia, knowledge. Beyond the root vid we cannot go, nor can we tell why vid means to see, or to find, or to know. Nor should we gain much if from vid we appealed to the preposition vi, which means asunder, and might be supposed to have imparted to vid the power of dividing, singling out, perceiving (dis-cerno).1 It is true there is the same similarity of meaning in the Hebrew preposition bîn, between, and the verb bîn, to know, but why bîn should mean between is again a question which we cannot hope to clear up by mere etymological analysis.

All that we can safely maintain with regard to the nature of the Aryan roots is this, that they have definite forms and definite meanings. However chaotic the origin of language may by some scholars be supposed to have been, certain it is that here, as in all other subjects of physical research, we must attempt to draw a line which may separate the Chaos from the Kosmos. When the Aryan languages began to assume their individuality, their roots had become typical, both in form and meaning. They were no longer mere interjections with varying and indeterminate vowels, with consonants floating about from guttural to labial contact, and uncertain between surd, sonant, or aspirated enunciation. Nor were they the expressions of mere impressions of the moment, of single, abrupt states of feeling that had no reference to other sensations of a similar or dissimilar character. Language, if it then

1 On the supposed original connection between vi and dvi, see Pott, Etym. Unters. i. 705. Lectures, First Series, p. 44.

deserved that name, may at one time have been in that chaotic condition; nay, there are some small portions in almost every language which seem to date from that lowest epoch. Interjections, though they cannot be treated as parts of speech, are nevertheless ingredients of our conversation; so are the clicks of the Bushmen and Hottentots, which have been well described as remnants of animal speech. Again, there are in many languages words, if we may call them so, consisting of mere imitations of the cries of animals or the sounds of nature, and some of them have been carried along by the stream of language into the current of nouns and verbs.

It is this class of words which the Greeks meant when they spoke of onomatopeia. But do not let us suppose that because onomatopœia means making of words, the Greeks supposed all words to owe their origin to onomatopeia, or imitation of sound. Nothing would have been more remote from their minds. By onomatopeia they meant to designate not real words, but made, artificial, imitative words, words that any one could make at a moment's notice. Even the earliest of Greek philosophers had seen enough of language to know that the key to its mysteries could not be bought so cheaply. When Aristotle 1 calls words imitations (mimémata), he does not mean those downright imitations, as when we call a cow a moo, or a dog a bow-wow. His statements and those of Plato 2 on language must be read in connection with the statements of earlier

1 Rhet. iii. 1. τὰ γὰρ ὀνόματα μιμήματά ἐστιν, ὑπῆρξε δὲ καὶ ἡ φωνὴ πάντων μιμητικότατον τῶν μορίων ἡμῖν.

2 Plato, Cratylus. 423 Β. όνομα ἄρα ἐστίν, ὡς ἔοικε, μίμημα φωνῇ ἐκείνου 8 μιμεῖται καὶ ὀνομάζει ὁ μιμούμενος τῇ φωνῇ, ὅταν μιμῆται.

philosophers, such as Pythagoras (540-510), Heraclitus (503), Democritus (430-410), and others, that we may see how much had been achieved before them, how many guesses on language had been made and refuted before they in turn pronounced their verdict. Although we possess but scant, abrupt, and oracular sayings which are ascribed to those early sages, yet these are sufficient to show that they had pierced through the surface of language, and that the real difficulties of the origin of speech had not escaped their notice. When we translate the enigmatic and poetical utterances of Heraclitus into our modern, dry, and definite phraseology, we can hardly do them justice. Perfect as they are when seen in their dark shrines, they crumble to dust as soon as they are touched by the bright rays of our modern philosophy. Yet if we can descend ourselves into the dark catacombs of ancient thought, we feel that we are there in the presence of men who, if they lived with us and could but speak our language, would be looked upon as giants. They certainly had this one advantage over us, that their eyes had not been dimmed by the dust raised in the wars of words that have been going on since their time for more than two thousand years. When we are told that the principal difference of opinion that separated the philosophers of old with regard to the nature and origin of language is expressed by the two words phýsei and thései, "naturally" and "artificially," we learn very little from such general terms. We must know the history of those words, which were watch-words in every school of philosophy, before they dwindled down to mere technical terms.

With the later sophists thései, "artificially," or the still earlier nómó, "conventionally," meant no longer what they meant with the fathers of Greek philosophy; nay, they sometimes assumed the very opposite meaning. A sophist like Hermogenes, in order to prove that language existed conventionally, maintained that an apple might have been called a plum, and a plum an apple, if people had only agreed to do so.1 Another 2 pointed in triumph to his slave, to whom he had actually given a new name, by calling him "Yet," in order to prove that any word might be significative. Nor were the arguments in favor of the natural origin of language of a better kind, when the efficacy of curses was quoted to show that words endowed with such powers could not have a merely human or conventional origin.3

Such was not the reasoning of Heraclitus or Democritus. The language in which they spoke, the whole world of thought in which they lived, did not allow them to discuss the nature and origin of language after the fashion of these sophists, nor after our own fashion. They had to speak in parables, in

1 Lersch, Sprachphilosophie der Allen, i. p. 28. Ammonius Hermias ad Aristot. de Interpr. p. 25 Α. Οἱ μὲν οὕτω τὸ θέσει λέγουσιν ὡς ἐξὸν ὁμοῦν τῶν ἀνθρώπων ἔκαστον τῶν πραγμάτων ὀνομάζειν ὅ- ω ἂν ἐ ένῃ ὀνόματι, και θάπερ Ερμογένης ἠξίου. . . . Οἱ δὲ οὐχ οὕτως, ἀλλὰ τίθεσθαι μὲν τὰ ὀνό ματα ὑπὸ μόνου τοῦ ὀνομαθέτου, τοῦτον δὲ εἶναι τὸν ἐπιστήμονα τῆς φύσεως τῶν πραγμάτων, οἰκεῖον τῇ ἑκάστου τῶν ὄντων φύσει ἐπιφημίζοντα ὄνομα, ἢ τὸν ὑπηρετούμενον τῷ ἐπιστήμονι.

2 l. c. i. 42. Ammonius Hermias ad Aristot. de Interpret. p. 103. Εἰ δὲ ταῦτα ὀρθῶς λέγεται, δῆλον ὡς οὐκ ἀποδεξόμεθα τὸ διαλεκτικὸν Διόδωρον πᾶσαν οἰόμενον φωνὴν σημαντικὴν εἶναι, καὶ πρὸς πίστιν τούτου καλέσαντα τῶν ἑαυτοῦ τινὰ οἰκετῶν τῷ συλλογιστικῷ συνδέσμῳ 'Αλλά μην καὶ ἄλλον ἄλλῳ συνδέσμῳ· ποίαν γὰρ ἔξουσιν αἱ τοιαῦται φωναὶ σημασίαν φύσεώς τινος ἢ ἐνεργείας ἢ πάθους, καθάπερ τὰ ῥήματα χαλεπὸν καὶ πλάσαι. 8 Lersch, p. 44.

full, weighty, suggestive poetry, poetry that cannot be translated without an anachronism. We must take their words, such as they are, with all their vagueness and all their depth, but we must not judge them by these words as if these words were spoken by ourselves. The oracle on languages which is ascribed to Heraclitus was certainly his own. Commentators may have spoiled, but they could not have invented it. Heraclitus held that words exist naturally, but he did not confine himself to that technical phraseology. Words, he said,1 are like the shadows of things, like the pictures of trees and mountains reflected in the river, like our own images when we look into a mirror. This sounds like Heraclitus; his sentences are always like nuggets of gold, to use his own simile,2 without any of the rubbish through which philosophers have to dig before they can bring to light solid truth. He is likewise reported to have said, that to use any words except those supplied by nature for each thing, was not to speak, but only to make a noise. What Heraclitus meant by his simile, or by the word "nature," if he used it, we cannot know definitely; but we know, at all events, what he did not mean, namely, that man imposed what names he pleased on the objects around him. To have perceived that at that time, to have given any thought to that problem in the days when Heraclitus lived, stamps him once for all as a philosopher, ignorant though he may have been of all the rules of our logic, and our rhetoric, and our

1 Lersch, l. c. i. 11. Ammonius ad Arist. de Interpret. p. 24 B, ed. Ald. Bernays, Neue Bruchstücke des Heraclitus von Ephesus, Rheinisches Museum für Philologie, x. p. 242. χρυσὸν οἱ διζήμενοι γῆν πολλὴν ὀρῦσε oovoi kaì eúpíokovoi kiyor. Clemens Stromat. iv. 2, p. 565 P.