peopled and civilised by Asia, and in Asia, if any where, we shall probably find it. If I turn to the Sanskrit, one of the very oldest languages of Asia, I find that the word ka is a noun substantive signifying both the sun and fire, and am consequently led to suspect that the Greek verb kaio, I burn, is compounded of the Sanskrit root ka, fire, and ego, I, the Personal Pronoun; but as I discover that the Pronoun I, in the dialect of Boeotia, one of the oldest of Greece, was expressed, not by Ego, but Io, by joining the latter word to the Sanskrit root ka, my analysis of kaio becomes complete, and I now know, not only that it signifies I burn, but why and how it signifies I burn. Kaio is explained in Latin by the words uro and cremo, and the Latin language is said to have been chiefly formed from Æolic Greek. Are the Latin words simple or compound, and if the latter, whence derived? I find in Hebrew the word aor, or aur, signifying fire, and suppose the Latin uro to be formed by the elision of the initial a of aur, and the addition of Ego, or Io (Baotian), contracted to o; I also find in Coptic the word chrom, or krom, fire, and suppose cremo to be formed in the same way. To try another instance from the Sanskrit. In Greek we find the word oikos, a house, which I have little doubt was derived from the Sanskrit okas, but in Greek we also find the verb oiko (oik@), I dwell; and how was that formed? I believe by the addition of the pronoun Ego, or rather Io, to the simpler form of the Sanskrit oka, a house. But though the Sanskrit and Greek mutually illustrate each other, and we commonly find the roots of the latter in the former, this is not invariably the case; and we sometimes find the simple root of the verb in Greek in the shape of a noun substantive. For instance daio in Greek signifies I burn, which I should confidently say was derived from the Sanskrit dah, reduce to ashes, did I not find in Greek the word dai a torch, which with the addition of final omega forms the verb. But have we made all the use that we can do of the established fact that the first Person of the Greek verb consists of a simple invariable root, joined to a personal pronoun, and secondly that that personal pronoun is also the Verb Substantive To

be? Can we discover any thing from the analogy of the Sanskrit, a language cognate with the Greek and Latin, and probably the mother of both? The 13th volume of the Edinburgh Review, in its notice of Wilkins's Sanskrit grammar, gives the conjugations of the Verb Substantive asmi, sum, and the verb jivami, vivo; and if we look at both attentively we shall discover that the latter is compounded of the Sanskrit noun jiv, or jiva, life, which never varies, joined to the persons of the verb astun, to be, unchanged except by dropping the letter S, and writing ami instead of asmi. The verb sebami is conjugated in the 33d volume of the Edinburgh Review, in the notice of two works of Bopp, to which the same remark applies, and we may also observe an extraordinarily close resemblance between the Sanskrit Active, and the Greek Middle Voice. But that Middle Voice appears to be clearly the Sanskrit root seb, joined to the persons or terminations of the Verb Substantive To be. But in Greek sebo and sebomai have precisely the same signification, colo, adoro; how are we to account for this? The root of both words is clearly the Sanskrit seb, worship, and we may write them in Greek thus:

Zaß-Ew, the obsolete form of eimi, sum, by contraction σß, literally, worship I am, or worship I.

Zéß-ouai, contracted from eomai, another of the obsolete forms of eimi, literally worship I am, or worship I.

But as the obsolete forms of eimi perform so important a part in the conjugation of the Greek verb, it will be necessary, before we proceed farther, to endeavour to ascertain what they were.

XLI. On the Auxiliary Verb To be.

The persons of the Present Tense of this verb are so much alike in all the principal languages of the Sanskrit branch of the great Indo-European family, as irresistibly to suggest the idea of a common origin among all the people that use it. Our examples comprise the Sanskrit, which appears to be the basis of most of the languages still used in the vast continent of India, of three forms of the Persic, of the Slavonic

and Lithuanian, which compose the speech of the principal nations of the North of Europe, of the Greek and Latin, which, with some modifications, still continue to be spoken in the South, and of the Moso-Gothic, which is the oldest written form of the Anglo-Saxon, German, and English.

Here we may remark that the Sanskrit verb asmi, and more especially if we restore what Dr. Prichard regards as its oldest plural forms, asmus, astha, asanti, corresponds almost letter for letter with the oldest form of the Greek soμì, and that the same verb, as it appears as an auxiliary, and is joined in the process of conjugating to other roots, drops its s for the most part, and approximates to the common Greek form of eiuì; as jiv-ami, I live, seb-ami, I venerate. We recognise the Latin sum in the Persic shum, the Latin sumus, in the Persic shuim, and the Latin sunt in the Persic shund, and may remark that both the old Slavonic and Lithuanian, or Lettish, strikingly resemble the Sanskrit and oldest Greek; while the persons in the singular number of the Moso-Gothic approximate very closely to the first Persic form am, ai, ast. The verb substantive appears to have experienced many changes in most languages, and chiefly in the shape of contractions, as that which is most used is naturally most worn; but in addition to this common source of change, the Auxiliary Verb, To be, in Greek, seems to have undergone other corruptions, from the circumstance of there existing three or four other verbs in the language so

* Prichard on the Eastern Origin of the Celtic Nations, p. 97.

closely resembling it, that their tenses have unquestionably been confounded, and we are under the necessity of casting a glance on them before we proceed with our arduous attempt to restore the obsolete parts of eiuí. These verbs are εἶμι and ἵημι eo, vado, ἵημι mitto, ἵεμαι cupio, ήμαι sedeo, elpai indutus sum; and as the Port Royal appears to have paid more attention to these verbs than any other grammar, I shall depend chiefly on its authority.

1. Aorist,

'EIMI, Sum, I am.

From w, with a smooth breathing comes tipi acuted.

Jor, Poetical.

ov, Ionic (Eustathius), also an Imperfect.

From comes pa, which, with a rough breathing, expresses desire. From e to sit, comes likewise pai.

Such are the verbs so closely resembling sipì, to be, in their form, that I entertain not a doubt that many tenses are arranged under them, in all existing grammars and lexicons, which in an early period of the Greek language, perhaps anterior to the general use of alphabetical writing, certainly formed part of the Auxiliary Verb. It may be necessary, however, to premise a few words on the polemical part of my subject, before I enter on that which is purely didactic.

The article, in the 33rd volume of the Edinburgh Review, noticing two works of Bopp, to which I have had occasion to allude more than once, says, the First Preterite (Imperfect) is formed in Sanskrit by prefixing A to the root, as the Imperfect in Greek is by the augment. Thus the first person in Sanskrit is Asebam, in Greek σεßov (p. 433). But the First Preterite in Sanskrit is not formed merely by prefixing A to the root, but also by changing Ami, the termination of the Present, into Am, nor is the Imperfect in Greek formed by merely prefixing the augment ɛ, but also by changing the w of the Present into ov, as Sebami, Asebam, Sanskrit; σéßw, eσeßov, Greek. The root is the same in both instances, and the additions to that root are in Sanskrit Aam, and in Greek Eov. But in Greek čov is a distinct tense of the verb To be, and we find it arranged in the Eton Grammar among the dialects of v, eram. In both languages, therefore, the Imperfect is formed by prefixing A or E to the root, and adding Am, or On, as a termination. But ov, the Imperfect of sipì, sum, forms žov, ees, en, according to the Eton Grammar; the real model of the terminations of this tense, therefore, appears to be lov, the Second Aorist of eiui, vado, which, in the form of lov, was formerly one of the tenses of the Auxiliary Verb To be; and if we prefix its ɛ to the root TUTT, and add its termination ov, we shall have ĕ-TUTT-OV, the Imperfect of TUTT, letter for letter throughout all the Persons. If we now turn to -TUTTT-ÓμNV, the Imperfect of the Passive Voice, as the addition to the root is souny, the ɛ being prefixed as an augment, and the oμny annexed as a termination, we can hardly doubt that that word is ἐόμην, the Imperfect of the obsolete ἔομαι, which itself in the slightly contracted form of opal forms the Present Tense, Passive Voice τύπτομαι. The First Future of all Greek verbs in w, except liquids in λ, u, v, p, of the fifth conjugation, terminates in σω, έσω, οι ησω if we resolve the penultimate double letters and έ into their elements πs and xs; and here it appears to me that there cannot be a doubt that the formative is low, the Future of the obsolete ew, sum, added to the root, and generally contracted to ow, as from TUT, TÚT-σw,