L. In the new and important remarks which I am going to offer on the ancient languages and charaoters of Iran, I am sensible that you must give me credit for many assertions, which on this occasion it is impossible to prove; for I should ill deserve your indulgent attention, if I were to abuse it by repeating a dry list of detached words, and presenting you with a vocabulary instead of a dissertation; but, since I have no system to maintain, and have not suffered imagination to delude my judgment, since I have habituated myself to form opinions of men and things from evidence, which is the only solid basis of civil, as experiment is of natural, knowledge, and since I have maturely considered the questions which I mean to discuss, you will not, I am persuaded, suspect my testimony, or think that I go too far, when I assure you that I will assert nothing positively which I am not able satisfactorily to demonstrate. When Muhammed was born, and Anushiravan, whom he calls the Just King, sat on the throne of Persia, two languages appear to have been generally prevalent in the great empire of Iran; that of the Court, thence named Deri, which was only a refined and elegant dialect of the Parsi, so called from the province of which Shiraz is now the capital; and that of the learned, in which most books were composed, and which had the name of Pahlavi, either from the heroes who spoke it in former times, or from Pahlu, a tract of land, which included, we are told, some considerable cities of Irak. The ruder dialects of both were, and, I believe, still are spoken by the rustics in several provinces, and in many of them, as Herat, Zabul, Sistan, and others, distinct idioms were vernacular, as it happens in

every kingdom of great extent. Besides the Parsi and Pahlavi, a very ancient and abstruse tongue was known to the priests and philosophers, called the language of the Zend, because a book on religious and moral duties, which they held sacred, and which bore that name, had been written in it, while the Pazand, or comment on that work, was composed in Pahlavi, as a more popular idiom; but a learned follower of Zeratusht, named Bahman, who lately died at Calcutta, where he had lived with me as a Persian reader about three years, assured me that the letters of his prophet's book were properly called Zend, and the language Avesta, as the words of the Vedus are Sanscrit, and the characters Nagari; or as the old Sagas and poems of Iceland were expressed in Runic letters. Let us however, in compliance with custom, give the name of Zend to the sacred language of Persia, until we can find, as we shall very soon, a fitter appellation for it. The Zend and the old Pahlavi are almost extinct in Iran; for among six or seven thousand Gabrs, who reside chiefly at Yezd, and in Cirman, there are very few who can read Pahlavi; and scarce any who even boast of knowing the Zend; while the Parsi, which remains almost pure in the Shahnamah, has become by the intermixture of numberless Arabic words, and many imperceptible changes, a new language, exquisitely polished by a series of fine writers in prose and verse, and analogous to the different idioms gradually formed in Europe after the subversion of the Roman empire: but with modern Persian we have no concern in our present inquiry, which I confine to the ages that preceded the Mohammedan conquest. Having twice read the works of Firdausi with great

attention, since I applied myself to the study of old Indian literature, I can assure you with confidence, that hundreds of Parsi nouns are pure Sanscrit, with no other change than such as may be observed in the numerous bhashas, or vernacular dialects of India; that very many Persian imperatives are the roots of Sanscrit verbs; and that even the moods and tenses of the Persian verb substantive, which is the model of all the rest, are deducible from the Sanscrit by an easy and clear analogy: we may hence conclude, that the Parsi was derived, like the various Indian dialects, from the language of the Brahmans; and I must add, that in the pure Persian, I find no trace of any Arabian tongue, except what proceeded from the known intercourse between the Persians and Arabs, especially in the time of Bahram, who was educated in Arabia, and whose Arabic verses are still extant, together with his heroic line in Deri, which many suppose to be the first attempt at Persian versification in Arabian metre; but, without having recourse to other arguments, the composition of words, in which the genius of the Persian delights, and which that of the Arabic abhors, is a decisive proof that the Parsi sprang from an Indian, and not from an Arabian stock. Considering languages as mere instruments of knowledge, and having strong reasons to doubt the existence of genuine books in Zend or Pahlavi (especially since the well informed author of the Dabistan affirms the work of Zeratusht to have been lost, and its place supplied by a recent compilation), I had no inducement, though I had an opportunity, to learn what remains of those ancient languages; but I often conversed on them with my friend Bahman; and both of us were convinced, after

full consideration, that the Zend bore a strong resemblance to Sanscrit, and the Pahlavi to Arabic. He had at my request translated into Pahlavi the fine inscription exhibited in the Gulistan, on the diadem of Cyrus; and I had the patience to read the list of words from the Pazand in the appendix to the Farhangi Jehangiri. This examination gave me perfect conviction that the Pahlavi was a dialect of the Chaldaic, and of this curious fact I will exhibit a short proof. By the nature of the Chaldean tongue, most words ended in the first long vowel, like shemia, heaven; and that very word, unaltered in a single letter, we find in the Pazend, together with laila, night; meyd, water; nira, fire; matra, rain; and a multitude of others, all Arabic or Hebrew, with a Chaldean termination; so zamar, by a beautiful metaphor, from pruning trees, means in Hebrew to compose verses, and thence, by an easy transition, to sing them; and in Pahlavi we see the verb zamruniten, to sing, with its forms zamrunemi, I sing, and zamrunid, he sang; the verbal terminations of the Persian being added to the Chaldaic root. Now all those words are integral parts of the language, not adventitious to it like the Arabic nouns and verbals engrafted on modern Persian; and this distinction convinces me, that the dialect of the Gabrs, which they pretend to be that of Zeratusht, and of which Bahman gave me a variety of written specimens, is a late invention of their priests, or subsequent at least to the Muselman invasion; for, although it may be possible that a few of their sacred books were preserved, as he used to assert, in sheets of lead or copper at the bottom of wells near Yezd, yet, as the conquerors had not only a spiritual, but a poli

tical interest in persecuting a warlike, robust, and indignant race of irreconcileable, conquered subjects, a long time must have elapsed before the hidden scriptures could have been safely brought to light, and few who could perfectly understand them must then have remained; but, as they continued to profess among themselves the religion of their fore fathers, it then became expedient for the Mubeds to supply the lost or mutilated works of their legislator by new compositions, partly from their imperfect recollection, and partly from such moral and religious knowledge as they gleaned, most probably, among the Christians, with whom they had an interOne rule we may fairly establish in deciding the question, Whether the books of the modern Gabrs were anterior to the invasion of the Arabs ? When an Arabic noun occurs in them, changed only by the spirit of the Chaldean idiom, as werta for werd, a rose; daba for dhahab, gold; or deman for zeman, time, we may allow it to have been ancient Pahlavi; but when we meet with verbal nouns or infinitives, evidently formed by the rules of Arabian grammar, we may be sure that the phrases in which they occur are comparatively modern; and not a single passage which Bahman produced from the books of his religion would abide this test.

course.

We come now to the language of the Zend, and here I must impart a discovery which I lately made, and from which we may draw the most interesting consequences. M. Anquetil, who had the merit of undertaking a voyage to India in his earliest youth, with no other view than to recover writings of Zeratusht, and who would have acquired a brilliant reputation in France, if he had not sullied it by hi