the number of its words and the precifion of its phrases; but it is equally true and wonderful, that it bears not the leaft resemblance, either in words or the structure of them, to the Sanfcrit, or great parent of the Indian dialects; of which diffimilarity I will mention two remarkable inftances: the Sanferit, like the Greek, Perfian, and Germon, delights in compounds, but, in a much higher degree, and indeed to such excess, that I could produce words of more than twenty fyllables, not formed ludicrously, like that by which the buffoon in ARISTOPHANES defcribes a feast, but with perfect seriousness, on the most folemn occafions, and in the most elegant works; while the Arabick, on the other hand, and all its fifter dialects, abhor the compofition of words, and invariably exprefs very complex ideas by circumlocution; fo that, if a compound word be found in any genuine language of the Arabian Peninsula, (zenmerdab for inftance, which occurs in the Hamásab) it may at once be pronounced an exotick. Again; it is the genius of the Sanferit, and other languages of the fame stock, that the roots of verbs be almoft universally biliteral, fo that five and twenty hundred such roots might be formed by the compofition of the fifty Indian letters; but the Arabick roots are as univerfally triliteral, fo that the compofation of the twenty-eight Arabian letters would

give near two and twenty thousand elements of the language: and this will demonftrate the furprising extent of it; for, although great numbers of its roots are confeffedly loft, and fome, perhaps, were never in use, yet, if we fuppofe ten thousand of them (without rekoning quadriliterals) to exift, and each of them to admit only five variations, one with another, in forming derivative nouns, even then a perfect Arabick dictionary ought to contain fifty thousand words, each of which may receive a multitude of changes by the rules of grammar. The derivatives in Sanfcrit are confiderably more numerous: but a farther comparison between the two languages is here unneceffary; fince, in whatever light we view them, they seem totally diftinct, and must have been invented by two different races of men; nor do I recollect a fingle word in common between them, except Suruj, the plural of Siraj, meaning both a lamp and the fun, the Sanfcrit name of which is, in Bengal, pronounced Súrja; and even this refemblance may be purely accidental. We may easily believe with the Hindus, that not even INDRA himfelf and his heavenly bands, much less any mortal, ever comprehended in his mind fuch an ocean of words as their facred language contains, and with the Arabs, that no man uninfpired was ever a complete master of Arabick: in fact no

perfon, I believe, now living in Europe or Afia, can read without study an hundred couplets together in any collection of ancient Arabian poems; and we are told, that the great author of the Kámùs learned by accident from the mouth of a child, in a village of Arabia, the meaning of three words which he had long fought in vain from grammarians, and from books, of the highest reputation. alone, that a knowledge of these two venerable languages can be acquired; and, with moderate attention, enough of them both may be known, to delight and instruct us in an infinite degree: I conclude this head with remarking, that the nature of the Ethiopick dialect seems to prove an early establishment of the Arabs in part of Ethiopia, from which they were afterwards expelled, and attacked even in their own country by the Abyffinians, who had been invited over as auxiliaries against the tyrant of Yemen about a century before the birth of MUHAMMED.

It is by approximation

Of the characters, in which the old compofitions of Arabia were written, we know but little; except that the Koran originally appeared in those of Cufab, from which the modern Arabian letters, with all their elegant variations, were derived, and which unquestionably had a common origin with the Hebrew or Chaldaick; but, as to the Himyarick letters, or those which

we see mentioned by the name of Almufnad, we are still in total darkness; the traveller NIEBUHR having been unfortunately prevented from visiting fome ancient monuments in Yemen, which are faid to have infcriptions on them: if those letters bear a strong resemblance to the Nágarì, and if a story current in India be true, that some Hindu merchants heard the Sanfcrit language spoken in Arabia the Happy, we might be confirmed in our opinion, that an intercourfe formerly fubfifted between the two nations of oppofite coafts, but should have no reason to believe, that they fprang from the fame immediate ftock. The first fyllable of Hamyar, as many Europeans write it, might perhaps induce an Etymologist to derive the Arabs of Yemen from the great ancestor of the Indians; but we muft obferve, that Himyar is the proper appellation of those Arabs; and many reafons concur to prove, that the word is purely Arabick: the fimilarity of fome proper names on the borders of India to those of Arabia, as the river Arabius, a place called Araba, a people named Aribes or Arabies, and another called Sabai, is indeed remarkable, and may hereafter furnish me with obfervations of fome importance, but not at all inconfiftent with my present ideas.

II. It is generally afferted, that the old religion of the Arabs was entirely Sabian; but I can

offer fo little accurate information concerning the Sabian faith, or even the meaning of the word, that I dare not yet fpeak on the subject with confidence. This at leaft is certain, that the people of Yemen very foon fell into the common, but fatal, errour of adoring the Sun and the Firmament; for even the third in defcent from YOKTAN, who was confequently as old as NAHOR, took the furname of ABDUSHAMS, or Servant of the Sun; and his family, we are affured, paid particular honours to that luminary: other tribes worshipped the planets and fixed ftars; but the religion of the poets at least seems to have been pure Theism; and this we know with certainty, because we have Arabian verses of unfufpected antiquity, which contain pious and elevated sentiments on the goodness and justice, the power and omniprefence, of ALLAH, or THE GOD. If an inscription, said to have been found on marble in Yemen, be authentick, the ancient inhabitants of that country preferved the religion of EBER, and profeffed a belief in miracles and a future ftate.

We are alfo told, that a ftrong resemblance may be found between the religions of the pagan Arabs and the Hindus; but, though this may be true, yet an agreement in worshipping the fun and ftars will not prove an affinity between the two nations: the powers of God represented as