lahata, to exercise the tongue.

lagana, to lick.

lassa, to lick.

lasaba, to lick.

lâsa, to taste.

lahasa, to lick.

lisn, the tongue.

lasa'ha, to be maligned.

The same root exists as "lih" in Sanskrit, as λeixo in Greek, as "laigon" in Gothic, as "ligh" in Celtic, and in Latin "lingua." Again, with the frequent transition of l, the dental semi-vowel, into d, the dental media, we find, corresponding to the Latin lingua, or dingua, the Gothic "tuggo," and the English "tongue." That the word "glossary" should have grown out of this root LAK, may seem startling; still there is not a link wanting to connect the two words either in their form or in their meaning. Turning to the Turanian languages, we find the Finnic lakkia, to lick, though it may be doubted whether Mangu leke, to polish, Finnic laaha, the same, or Finnic lau, to speak, could safely be referred to the same source.

§ 7. The three different Directions of Grammar, Turanian, Semitic, and Arian, represented by the three Sons of Feridun, Tur, Silim, and Irij.

A. Tur.

As we have thus been carried back to times when we see the three principal tongues, which we may represent as the three sons of Feridún, as not yet separated, it may be of interest to catch at least one glimpse of them as they are leaving their common home and starting off in different directions. What they carried away from home were roots and pronouns. Two of them, Silim and Irij, seem both to have held the secret how a root could be divided and changed so that it might be used as a subject or as a predicate. Túr also may have known it; but he either forgot it, or he did not like to tamper with those sacred relics which he had carried away from his father's house. Under his care they remained the same, without addition or diminution; and when they had to be used, they were only set and framed like precious jewels, but neither divided nor polished down. Now there were at least four things which Tur had to express with his roots and pronouns. If he possessed a

* Conf. pag. 286.

root for cutting, he wanted to say, I cut (present); I cut (past); cutter, i. e. knife; and my cutter, i. e. my knife. These four little phrases were indispensable for him if he wished to get on in the world. As long as he was alone with his family and children, he no doubt could make them understand by some expressive accent when ngò tà (moi battre) meant “I beat,” and when ngô là meant my stick" (moi-bâton). What followed would generally remove all uncertainty, if it existed; for ngo.ta.ni, I-strike-thou (moi battre vous), could only mean "I strike thee." Again, as he could express to-day by "this light," and yesterday by "that light," perhaps his wife and children were not slow in understanding when he said kin-tien ngo.ta, this day I strike, i. e. I strike now (tout à l'heure moi battre); or tso-tien ngo.ta, that day I strike, i. e. I struck (jadis moi battre).*

All this may seem so natural, as far as construction goes, that at first one hardly discovers any thing peculiar in these different modes of expression. Still, in the construction of these two expressions, ngo.ta, I beat, and ngo-ta, my stick, there is something so individual and peculiar, that neither Silim nor Irij could imitate it. This is the liberty of putting the predicate first in one sentence and last in another. Silim could say ngo.ta, I beat (é.qtol), but never ngo-ta, my stick. He would have to put the predicate last in both phrases, and say ta-ngo, stick of me, like e.qtol, I-striking. Irij again, at least in his early youth, could say ngo-ta, my stick (mad-danda), but never ngo.ta, I-striking. Instead of this he had to say striking-I (tudâ.mi). This peculiarity by which Tur put the predicate sometimes first, sometimes last, may originally have been involuntary. As his roots were not yet distinguished as nominal and verbal, as subjective and predicative, his ngo.ta, I strike, may not have been meant for I striking, but, like ngo-ta, my stick, for my-striking. Still we shall see that, among his descendants, even after they had learned to distinguish between nominal and verbal roots, and between subjective and pre

6

* "Qu'un étranger me dise, Moi avoir soif, moi vouloir boire, moi désirer manger,' je comprends ce langage; mais je ne puis m'empêcher de sentir que c'est un langage sans vie, sans nerf, sans liaison. Pourquoi ? Parce que l'âme du discours, la force unitive, le nœud de la proposition, l'essence du jugement, le verbe en un mot s'y fait désirer, malgré la présence de l'infinitif.”—Dissertation Critique, par l'Abbé Darrigol, p. 97.

dicative pronominal affixes, some retained the power of putting the subject first as well as last; such as agha-m, father of me, i. e. my father, and sanî.bin, knowing-I, i. e. I know. This applies, however, only to the Nor-Western descendants of Tûr; his other descendants place the predicate first always.

B. Silim.

Silim, as we saw, started from home fully aware that his roots might be made to answer two purposes. He therefore divided his roots into simple nouns and fuller verbs; also, he kept one set of his pronouns, which had already grown and multiplied around him, for his verbs, and another for his nouns. He had only one difficulty, which, with all his acuteness, he could not overcome: he could never think a predicate without first having thought his subject. Therefore he could say wrath (of) God, and wrath (of) me, but not God (s) wrath, and my-wrath. He also could say beating (of or to) me, i. e. I did beat, and I-beating, i. e. I beat, but not beating-I, i. e. I beat. The opportunity, however, which he had of forming at least these two verbal compounds, beating (of) me, and I-beating, was not lost by Silim; and as he found it essential to make his friends understand either that he had paid or that he meant to pay, he took the first form, paying (of) me, i. e. paying (belonging to me, or possessed and had by me), in the sense of the preterite, while the mere assertion of I-paying was left to answer the purpose of a present or a future payment.

C. Irij.

The mind of Irij was more comprehensive than that of Silim. He was able to think, as it were by one grasp, ideas such as "goldpiece," "God's love," &c., and he expressed them by a compound word, in which the predicate being second in thought, and therefore more present to his mind, came first in language. Now, as he could say God's love, μпrр-o-woλıç, father-land, Mahâ-râga, always putting the predicate first, he could also say, I-love, I-wife, but only in the

* 'INTоTÓτаμоs, which is generally mentioned as an exception, is only a literal translation of an Egyptian word. On the difference between 'Avdpópiλos and Þíλavδρος, Τιμόθεος and Θεότιμος, Δωρόθεος and Θεόδωρος, see Pott, Personennamen, p. 88.

sense of my love, and my wife, because his first word is always the predicate. South of the "Snows" his descendants retained this manner of expression for many centuries. They said, mat-putra, tvat-putra, asmad-putra, my-son, thy-son, our-son. Their Northern brethren, however, found it more expedient to express the predicative nature of these pronouns more distinctly than could be done by mere position. They therefore formed an independent predicative form, whether genitive or adjective. This they were able to handle with greater freedom, so that they might now say TéкVOV Eμòv as well as Eμoù TÉкVOV. As to his verbal compounds, Irij had two ways opened before him, only just in the contrary direction to those of Silim. He could say loving-I, i. e. I love; and he did say so, after his verbal base had been qualified by reduplication or by Vikaranas. This compound phrase, however, was a mere predication, and could therefore hardly be restricted to any point in time, whether past, present, or future. It simply asserted a quality or an action. How then could Irij express his preterite? As he had as yet no auxiliary possessive verb, like the "habere" and "tenere" of his descendants, he could only use his possessive pronouns. But his possessive pronouns he could only use before a verbal base, while he was accustomed to mark all other formal changes at the end of words. Silim, when he found himself in the same dilemma, simply divided his pronouns in two, and put half before and half after the verb.* Irij had to do the same; but as he was putting his pronoun before the word, trying to pronounce ma-gâ, my-going, i. e. I went, the pronouns were so strongly attracted towards the end of the root, that all that remained in the place originally intended for the whole predicative pronoun was not even a distinctive consonant, as in Hebrew, but only a

noun.

* Ewald (§ 152.) explains the formation of the Hebrew Aorist in the following manner :- "The prefixes had to be pronounced as short as possible: one consonant, not even followed by a vowel, was all that remained of the prefixed proThis consonant happened to be the same for several persons; confusion would inevitably have arisen, unless, by a very natural expedient, the pronominal prefix had been divided, so that the characteristic letters only remained as prefix, while the rest were thrown towards the end of the word. The pronoun of the second person sing. fem. being a-tin, atin was divided into at + in. At was shortened into t and prefixed, while in was suffixed, thus giving ti.qtel.î(n), thou (woman) killest.

strongly accented vowel, common to all these pronominal prefixes, and now called the augment; while the consonants, without their final vowels, were suffixed and placed at the end of the root. Thus, if there was a root lip, to write or paint, it could first be raised to a verbal base by reduplication. This verbal base lilep, writing, followed by predicative suffixes, would then give an aoristic compound, lilep-a, writing-I, I write, lilep-itha, writing-thou, thou writest. If afterwards a new and more actual verbal base was produced by the insertion of a nasal, such as limp, then, by the addition of predicative suffixes, limpāmi, limpasi, limpati, might be formed; and as these forms would express the present act of I am actually writing, the old present lilepa would in time take the sense of a perfect, I have written. The same root lip, however, being used as a subject, and not as a predicate, participating, therefore, more in the nature of a substantive than of an adjective, would, if preceded by possessive pronouns, express my-writing, i. e. writing belonging to me, i. e. I wrote, and thus á-lip-am (instead of ma-lip) would form the simplest and most primitive Arian preterite.

D. The Descendants of Tur divided according to their Employment of the Pronominal Affixes.

We have still to see how Tur proceeded in his verbal formations, as it is not likely that he could be satisfied with the Chinese juxtaposition of pronouns and words. Some of his descendants in Bhota and Bhotánta introduced formal elements to indicate the predicative or verbal nature of their roots; they formed their verbal bases, as we saw, by reduplication. They also used formal elements to indicate the predicative nature of their pronouns, and thus formed genitives, or pronominal adjectives. In Chinese already we have ngo-tisin, my heart. In ngo-ti-sin, ti, though originally it may have been a pronoun, cannot be compared with the Hebrew aser, or the Ethiopic za (masc.) and enta (fem.). In the Ethiopic mazmor za Dâwith, za is the masculine demonstrative or relative pronoun, referring to mazmor. It means the psalm which (to) David. But the Chinese min-li or min-ti-li expresses not the people which (is) power, i. e. the people of power, but people's power, where people's is the predicate, and therefore to be expressed either as the first