Sacred Texts  Zoroastrianism  Index  Previous  Next 
Buy this Book at

The Zend Avesta, Part I (SBE04), James Darmesteter, tr. [1880], at

p. xxx



§ 1. The collection of Zend fragments, known as the Zend-Avesta 1, is divided, in its usual form, into two parts.

The first part, or the Avesta properly so called, contains the Vendîdâd, the Vispêrad, and the Yasna. The Vendîdâd is a compilation of religious laws and of mythical tales; the Vispêrad is a collection of litanies for the sacrifice; and the Yasna is composed of litanies of the same kind and of five hymns or Gâthas written in a special dialect, older than the general language of the Avesta.

These three books are found in manuscripts in two different forms: either each by itself, in which case they are generally accompanied by a Pahlavi translation; or the three mingled together according to the requirements of the liturgy, as they are not each recited separately in their entirety, but the chapters of the different books are intermingled; and in this case the collection is called the Vendîdâd Sâdah or 'Vendîdâd pure,' as it exhibits the original text alone, without a translation.

The second part, generally known as the Khorda Avesta or 'Small Avesta,' is composed of short prayers which are recited not only by the priests, but by all the faithful, at certain moments of the day, month, or year, and in presence of the different elements; these prayers are the five Gâh, the thirty formulas of the Sîrôzah, the three Âfrigân, and the six Nyâyis. But it is also usual to include in the Khorda Avesta, although forming no real part of it, the Yasts or hymns of praise and glorification to the several

p. xxxi

[paragraph continues] Izads, and a number of fragments, the most important of which is the Hadhôkht Nosk.

§ 2. That the extent of the sacred literature of Mazdeism was formerly much greater than it is now, appears not only from internal evidence, that is, from the fragmentary character of the book, but is also proved by historical evidence. In the first place, the Arab conquest proved fatal to the religious literature of the Sassanian ages, a great part of which was either destroyed by the fanaticism of the conquerors and the new converts, or lost during the long exodus of the Parsis. Thus the Pahlavi translation of the Vendîdâd, which was not finished before the latter end of the Sassanian dynasty, contains not a few Zend quotations from books which are no longer in existence; other quotations, as remarkable in their importance as in their contents, are to be found in Pahlavi and Parsi tracts, like the Nîrangistân and the Aogemaidê. The Bundahis contains much matter which is not spoken of in the existing Avesta, but which is very likely to have been taken from Zend books which were still in the hands of its compiler. It is a tradition with the Parsis, that the Yasts were originally thirty in number, there having been one for each of the thirty Izads who preside over the thirty days of the month; yet there are only eighteen still extant.

The cause that preserved the Avesta is obvious; taken as a whole, it does not profess to be a religious encyclopedia, but only a liturgical collection, and it bears more likeness to a Prayer Book than to the Bible. It can be readily conceived that the Vendîdâd Sâdah, which had to be recited every day, would be more carefully preserved than the Yasts, which are generally recited once a month; and these again more carefully than other books, which, however sacred they might be, were not used in the performance of worship. Many texts, no doubt, were lost in consequence of the Arab conquest, but mostly such as would have more importance in the eyes of the theologian than in those of the priest. We have a fair specimen of what these lost texts may have been in the few non-liturgical fragments which we still possess, such as the Vistâsp Yast and

p. xxxii

the blessing of Zoroaster upon King Vistâsp, which belong to, the old epic cycle of Iran, and the Hadhôkht Nosk, which treats of the fate of the soul after death.

§ 3. But if we have lost much of the Sassanian sacred literature, Sassanian Persia herself, if we may trust Parsi tradition, had lost still more of the original books. The primitive Avesta, as revealed by Ormazd to Zoroaster and by Zoroaster to Vistâsp, king of Bactria, was supposed to have been composed of twenty-one Nosks or Books, the greater part of which was burnt by Iskander the Rûmi (Alexander the Great). After his death the priests of the Zoroastrian religion met together, and by collecting the various fragments that had escaped the ravages of the war and others that they knew by heart, they formed the present collection, which is a very small part of the original book, as out of the twenty-one Nosks there was only one that was preserved in its entirety, the Vendîdâd 1.

This tradition is very old, and may be traced back from the present period even to Sassanian times 2. It involves the assumption that the Avesta is the remnant of the sacred literature of Persia under the last Achæmenian kings. To ascertain whether this inference is correct, and to what extent it may be so, we must first try to define, as. accurately as we can, the exact time at which the collection, now in existence, was formed.

§ 4. The Ravâet quoted above states that it was formed 'after the death of Iskander,' which expression is rather vague, and may as well mean 'centuries after his death' as 'immediately after his death.' It is, in fact, hardly to be doubted that the latter was really what the writer meant; yet, as the date of that Ravâet is very recent, we had better look for older and more precise traditions. We find such a one in the Dînkart, a Pahlavi book which enjoys great authority with the Parsis of our days, and which, although it contains many things of late origin 3, also comprises many

p. xxxiii

old and valuable traditions. According to a proclamation, ascribed to Khosrav Anôsharvân (531-579), the collection of the Avesta fragments was begun in the reign of the last Arsacides, and was finished under Shapûr II (309-380). King Valkash (Vologeses), it is said, first ordered all the fragments of the Avesta which might have escaped the ravages of Iskander, or been preserved by oral tradition, to be searched for and collected together. The first Sassanian king, Ardeshîr Bâbagân, made the Avesta the sacred book of Iran, and Mazdeism the state religion: at last, Âdarbâd under Shapûr II, purified the Avesta and fixed the number of the Nasks, and Shapûr proclaimed to the heterodox 1: 'Now that we have recognised the law of the world here below, they shall not allow the infidelity of any one whatever 2, as I shall strive that it may be so 3.'

§ 5. The authenticity of this record has been called in question, chiefly, I think, on account of the part that it ascribes to an Arsacide prince, which seems hardly to agree with the ideas generally entertained about the character of the Sassanian revolution 4. Most Parsi and Muhammedan writers agree that it was the Sassanian dynasty which raised the Zoroastrian religion from the state of humiliation into which the Greek invasion had made it sink, and, while it gave the signal for a revival of the old national spirit, made Mazdeism one of the corner stones of the new establishment 5. Therefore it seems strange to hear that the first step taken to make Mazdeism a state religion was taken by one of those very Philhellenic Parthian princes, who were so imbued with Greek ideas and manners. Yet this is the

p. xxxiv

very reason why we ought to feel some hesitation in rejecting this document, and its being at variance with the general Parsi view speaks rather for its authenticity; for as it was the general post-Sassanian tradition that the restoration of Mazdeism was the work of the first Sassanian kings, no Parsi would ever have thought of making them share what was in his eyes their first and best title of honour with any of the despised princes of the Parthian dynasty.

§ 6. It is difficult, of course, to prove directly the authenticity of this record, the more so as we do not even know who was the king alluded to. There were, in fact, four kings at least who bore the name of Valkhash: the most celebrated and best known of the four was Vologeses 1, the contemporary of Nero. Now that Zoroastrianism prevailed with him, or at least with members of his family, we see from the conduct of his brother Tiridates, who was a Magian (Magus) 2; and by this term we must not understand a magician 3, but a priest, and one of the Zoroastrian religion. That he was a priest appears from Tacitus’ testimony 4; that he was a Zoroastrian is shown by his scruples about the worship of the elements. When he came from Asia to Rome to receive the crown of Armenia at the hands of Nero, he wanted not to come by sea, but rode along the coasts, 5, because the Magi were forbidden to defile the sea 6. This is quite in the spirit of later Zoroastrianism, and savours much of Mazdeism. That Vologeses himself shared the religious scruples of his brother appears from his answer to Nero,

p. xxxv

who insisted upon his coming to Rome also: 'Come yourself, it is easier for you to cross such immensity of sea 1.'

§ 7. Thus we hear on one hand from the Parsis that the first collection of the Avesta was made by an Arsacide named Vologeses; and we hear, on the other hand, from a quite independent source, that an Arsacide named Vologeses behaved himself as a follower of the Avesta might have done. In all this there is no evidence that it is Vologeses I who is mentioned in the Dînkart, much less that he was really the first editor of the Avesta; but it shows at all events that the first attempt to recover the sacred literature of Iran might very well have been made by an Arsacide, and that we may trust, in this matter, to a document which has been written perhaps by a Sassanian king, but, at any rate, in a Sassanian spirit. In fact, in the struggle between Ardavan and Ardeshîr, there was no religious interest at stake, but only a political one; and we are expressly told by Hamza that between Ardeshîr and his adversaries there was perfect accordance in religious matters 2. It can, therefore, be fairly admitted that even in the time and at the court of the Philhellenic Parthians a Zoroastrian movement may have originated, and that there came a time when they perceived that a national religion is a part of national life. It was the merit of the Sassanides that they saw the drift of this idea which they had the good fortune to carry out; and this would not be the only instance, in the history of the world, of an idea being sown by one party and its advantages reaped by their adversaries.

p. xxxvi

§ 8. Another presumptive evidence of the groundwork of the Avesta being anterior to the age of the Sassanians is given by the language in which it is written. That language not only was not, but had never been, the national language of Persia. It is indeed closely connected with the ancient Persian, as found in the cuneiform inscriptions of the Achæmenian kings, from which modern Persian is derived; but the relations between ancient Persian and Zend are of such a kind that neither language can be conceived as being derived from the other; they are not one and the same language in two different stages of its development, but two independent dialects in nearly the same stage, which is a proof that they did not belong to the same country, and, therefore, that Zend was not the language of Persia. Now the language used in Persia after the death of Alexander, under the Arsacides and Sassanides, that is, during the period in which the Avesta must have been edited, was Pahlavi, which is not derived from Zend, but from ancient Persian, being the middle dialect between ancient and modern Persian. Therefore, if the Sassanian kings had conceived the project of having religious books of their own written and composed, it is not likely that they would have had them written in an old foreign dialect, but in the old national language, the more so, because, owing both to their origin and their policy, they were bound to be the representatives of the genuine old Persian tradition. Therefore, if they adopted Zend as the language of religion, it must have been because it was already so when they appeared, that is to say, because the only remnants of sacred literature then extant were written in Zend, and the editors of the Avesta had Zend writings before them.

This does not, of course, prove that all we find in the Avesta is pre-Sassanian, and that the editors did not compose new Zend texts. Although Zend was not only a dead language, but also a foreign one, it was. not an unknown language: that it was well understood by the learned class, the priests, appears from the Pahlavi translation, which was made by them, and which, the deeper

p. xxxvii

one enters into the meaning of the text, has the fuller justice done to its merits. The earliest date that can be ascribed to that translation, in its present form, is the last century of the Sassanian dynasty, as it contains an allusion to the death of the heresiarch Mazdak, the son of Bâmdâd 1, who was put to death in the beginning of the reign of Khosrav Anôsharvân (about 531). Now the ability to translate a dead language is a good test of the ability to write in it, and in the question of the age of the Zend texts the possibility of new ones having been composed by the editors cannot be excluded à priori. Nay, we shall see further on that there are passages in these texts which look very modern, and may have been written at the time when the book took its last and definitive form. But whatever may be the proportion of the new texts to the old ones (which I believe to be very small), it is quite certain that the bulk of the Avesta is pre-Sassanian.

§ 9. The date assigned by the Dînkart to the final edition of the Avesta and to its promulgation as the sacred law of the nation, agrees with what we know of the religious state of Iran in the times of Shapûr II. Mazdeism had just been threatened with destruction by a new religion sprung from itself, the religion of Mânî, which for a while numbered a king amongst its followers (Shapûr I, 240-270). Mazdeism was shaken for a long time, and when Mânî was put to death, his work did not perish with him. In the Kissah-i Sangâh, Zoroaster is introduced prophesying that the holy religion will be overthrown three times and restored three times; overthrown the first time by Iskander, it will be restored by Ardeshîr; overthrown again, it will be restored by, Shapûr II and Âdarbâd Mahraspand; and, lastly, it will be overthrown by the Arabs and restored at the end of time by Soshyos. The Parsi traditions about Âdarbâd, although they are mixed with much fable, allow some historical truth to show itself. He was a holy man under Shapûr II, who, as there were many religions and heresies in Iran and the true religion

p. xxxviii

was falling into oblivion, restored it through a miracle, as he gave a sign of its truth by allowing melted brass to be poured on his breast, without his being injured. Setting aside the miracle, which is most probably borrowed from the legend of Zoroaster, this account receives its true interpretation from the passages in the Kissah-i Sangâh and the Dînkart, which imply that Âdarbâd restored Mazdeism, which had been shaken by the Manichean heresy, and that in order to settle it upon a solid and lasting base, he gave a definitive form to the religious book of Iran and closed the Holy Writ. And even nowadays the Parsi, while reciting the Patet, acknowledges Âdarbâd as the third founder of the Avesta; the first being Zoroaster, who received it from Ormazd; the second Gâmâsp, who received it from Zoroaster; and the third Âdarbâd, who taught it and restored it to its purity.

Therefore, so far as we can trust to inferences that rest upon such scanty and vague testimonies, it seems likely that the Avesta took its definitive form from the hands of Âdarbâd Mahraspand, under King Shapûr II, in consequence of the dangers with which Mânî's heresy had threatened the national religion. As the death of Mânî and the first persecution of his followers took place some thirty years before Shapûr's accession to the throne, it may be presumed that the last revision of the Avesta was made in the first years of the new reign, when the agitation aroused by Mânî's doctrines and imperfectly allayed by the persecution of his disciples had not yet subsided, and the old religion was still shaking on its base 1.

§ 10. It follows hence that Zend texts may have been composed even as late as the fourth century A.D. This is, of course, a mere theoretical possibility, for although the liturgical parts of the Yasna, the Vispêrad, the Sîrôzah, and

p. xxxix

the Khorda Avesta must be ascribed to a later time than the Gâthas, the Vendîdâd, and the Yasts, and may belong to some period of revision, they certainly do not belong to the period of this last revision. Âdarbâd was only the last editor of the Avesta, and it is likely, nay, it is beyond all question, that the doctors of the law, before his time, had tried to put the fragments in order, to connect them, and to fill up the gaps as far as the practical purposes of liturgy required it. Therefore instead of saying that there are parts of the Avesta that may belong to so late a period as the fourth century, it is more correct to say that no part of it can belong to a later date.

There are two passages in the Vendîdâd which seem to contain internal evidence of their date, and in both cases it points to Sassanian times, nay, the second of them points to the age of Manicheism. The first is found in the eighteenth Fargard (§ 10): Ahura Mazda, while cursing those who teach a wrong law, exclaims:

'And he who would set that man at liberty, when bound in prison, does no better deed than if he should flay a man alive and cut off his head.'

This anathema indicates a time when Mazdeism was a state religion and had to fight against heresy; it must, therefore, belong to Sassanian times. These lines are fully illustrated by a Parsi book of the same period 1, the Mainyô-i-Khard:

'Good government is that which maintains and orders the true law and custom of the city people and poor untroubled, and thrusts out improper law and custom; . . . and keeps in progress the worship of God, and duties, and good works; . . . and will resign the body, and that also which is its own life, for the sake of the good religion of the Mazdayasnians. And if there he any one who shall stay away from the way of God, then it orders him to return thereto, and makes him a prisoner, and brings him back to the way of God; and will bestow, from the wealth that is his, the share of God, and the worthy, and good works,

p. xl

and the poor; and will deliver up the body on account of the soul. A good king who is of that sort, is called like the Yazads and the Ameshâspeñds 1.'

What doctrines are alluded to by the Vendîdâd is not explained: it appears from the context that it had in view such sects as released the faithful from the yoke of religious practices, as it anathematizes, at the same time, those who have continued for three years without wearing the sacred girdle. We know too little of the Manichean liturgy to guess if the Manicheans are here alluded to: that Mânî should have rejected many Zoroastrian practices is not unlikely, as his aim was to found a universal religion. While he pushed to extremes several of the Zoroastrian tenets, especially those which had taken, or might receive, a moral or metaphysical meaning, he must have been very regardless of practices which could not be ennobled into moral symbolism. However it may be with regard to the foregoing passage, it is difficult not to see a direct allusion to Manicheism in lines like the following (IV, 47 seq.):

‘Verily I say it unto thee, O Spitama Zarathustra! the man who has a wife is far above him who begets no sons; he who keeps a house is far above him who has none; he who has children is far above the childless man, he who has riches is far above him who has none.

‘And of two men, he who fills himself with meat is filled with the good spirit much more than he who does not so; the latter is all but dead; the former is above him by the worth of an Asperena, by the worth of a sheep, by the worth of an ox, by the worth of a man.

'It is this man that can strive against the onsets of Astôvîdhôtu; that can strive against the self-moving arrow; that can strive against the winter fiend, with thinnest garment on; that can strive against the wicked tyrant and smite him on the head; it is this man that can strive against the ungodly Ashemaogha 2 who does not eat 3.'

p. xli

That this is a bit of religious polemics, and that it refers to definite doctrines and tenets which were held at the time when it was written, can hardly be doubted. It may remind one of the Christian doctrines; and, in fact, it was nearly in the same tone, and with the same expressions, that in the fifth century King Yazdgard branded the Christians in Armenia 4. But however eager the Christian propaganda may have been for a time in Persia, they never endangered the state religion. The real enemy was the heresy sprung from Mazdeism itself; and Christianity, coming from abroad, was more of a political than a religious foe. And, in point of fact, the description in the above passage agrees better with the Manichean doctrines than with the Christian 5. Like Mânî, Christian teachers held the single life holier than the state of matrimony, yet they had not forbidden marriage, which Mânî did; they put poor Lazarus above Dives, but they never forbade trade and husbandry, which Mânî did; and, lastly, they never prohibited the eating of flesh, which was one of the chief precepts of Mânî 6. We find, therefore, in this passage, an illustration, from the Avesta itself, of the celebrated doctrine of the three seals with which Mânî had sealed the bosom, the hand, and the mouth of his disciples (signaculum sinus, manus, oris) 6.}

p. xlii

§ 11. We must now go a step farther back, and try to solve the question whence came the original texts out of which the editors of the Avesta formed their collection. Setting aside the Dînkart, we have no oriental document to help us in tracing them through the age of the Arsacides, a complete historical desert, and we are driven for information to the classical writers who are, on this point, neither very clear nor always credible. The mention of books ascribed to Zoroaster occurs not seldom during that period, but very often it applies to Alexandrian and Gnostic apocrypha 1. Yet there are a few passages which make it pretty certain that there was a Mazdean literature in existence in those times. Pausanias, travelling through Lydia in the second century of our era, saw and heard Magian priests singing hymns from a book 2; whether these hymns were the same as the Gâthas, still extant, we cannot ascertain, but this shows that there were Gâthas. The existence of a Zoroastrian literature might be traced back as far as the third century before Christ, if Pliny could be credited when he says that Hermippus 3 had given an analysis of the books of Zoroaster, which are said to have amounted to 2,000,000 lines 4. For want of external evidence for ascertaining whether the original texts were already in existence in the later years of the Achæmenian dynasty, we must seek for internal evidence. A comparison between the ideas expressed in our texts and what we know of the ideas of Achæmenian Persia may perhaps lead to safer inferences.

§ 12. That all the Avesta ideas were already fully developed in the time, or, at least, at the end of the

p. xliii

[paragraph continues] Achæmenian dynasty, appears from the perfect accordance of the account of Mazdeism in Theopompos 1 with the data of the Zend books. All the main features of Mazdean belief, namely, the existence of two principles, a good and an evil one, Ormazd and Ahriman, the antithetical creations of the two supreme powers, the division of all the beings in nature into two corresponding classes, the limited duration of the world, the end of the struggle between Ormazd and Ahriman by the defeat and destruction of the evil principle, the resurrection of the dead, and the everlasting life, all these tenets of the Avesta had already been established at the time of Philip and Aristotle. Therefore we must admit that the religious literature then in existence, if there were any, must have differed but little, so far as its contents were concerned, from the Avesta; its extent was greater of course, and we have a proof of this in this very account of Theopompos, which gives us details nowhere to be found in the present texts, and yet the authenticity of which is made quite certain by comparative mythology 2. Therefore there is nothing that forbids us to believe, with the Parsis, that the fragments of which the Avesta is composed were already in existence before the Greek invasion 3.

§ 13. But it does not follow hence that the Achæmenian Avesta was the sacred book of the Achæmenians and of Persia, and it must not be forgotten that the account in Plutarch is not about the religion of Persia, but about the belief of the Magi and the lore of Zoroaster. Now if we consider that the two characteristic features of Avestean Magism are, so far as belief goes, the admission of two principles, and so far as practice is concerned, the prohibition of burying the dead, we find that there is no evidence

p. xliv

that Achæmenian Persia admitted the former, and there is evidence that she did not admit the latter. But, at the same time, it appears that both the belief and the practice were already in existence, though peculiar to one class, the sacerdotal class, the Magi.

The question whether the Achæmenian kings believed in dualism and knew of Ahriman, is not yet settled. Much stress has often been laid on the absence of the name of Ahriman in the religious formulae engraved by Darius and Xerxes on the rocks at Persepolis and Naqs-i Rustam 1. But it is never safe to draw wide conclusions from negative facts: Darius and Xerxes speak of Aurâmazda quite in the style of the Avesta, and their not speaking of Ahriman is no sufficient proof of their not knowing him; they did not intend to publish a complete creed, nor had they to inscribe articles of faith.

The account of the Persian religion in Herodotus also leaves, or seems to leave, Ahriman unnoticed. But it must be borne in mind that he does not expound the religious conceptions of the Persians, but only their religious customs; he describes their worship more than their dogmas, and not a single tenet is mentioned. He seems even not to know anything of Ormazd, who was, however, most certainly the most supreme god of Persia in his days; yet, in fact, he clearly alludes to Ormazd when he states that the Persians worship Zeus on the summits of mountains, and call by the name of Zeus the whole circle of the heavens, which exactly agrees with the character of Ormazd 2. In the same way the existence of Ahriman is indirectly pointed to by the duty enforced upon the faithful to persecute and kill noxious animals, as it was only on account of

p. xlv

their being creatures of the evil principle and incarnations if of it, that this custom was enjoined as a religious duty 1. It appears, it is true, from the words of Herodotus, that it was only a custom peculiar to the Magi 2; but is shows, at least, that the belief in Ahriman was already then in existence, and that dualism was constituted, at least, as a Magian article of faith.

If we pass now from dogma to practice, we find that the most important practice of the Avesta law was either disregarded by the Achæmenian kings, or unknown to them. According to the Avesta burying corpses in the earth is one of the most heinous sins that can be committed 3; we know that under the Sassanians a prime minister, Seoses, paid with his life for an infraction of that law 4. Corpses were to be laid down on the summits of mountains, there to be devoured by birds and dogs; the exposure of corpses, was the most striking practice of Mazdean profession, and its adoption was the sign of conversion 5. Now under the Achæmenian rule, not only the burial of the dead was not forbidden, but it was the general practice. Persians, says Herodotus, bury their dead in the earth, after having coated them with wax 6. But Herodotus, immediately after stating that the Persians inter their dead, adds that the Magi do not follow the general practice, but lay the corpses down on the ground, to be devoured by birds. So what became a law for all people, whether laymen or priests, under the rule of the Sassanians, was only the custom of the Achæmenians.

The obvious conclusion is that the ideas and customs which are found in the Avesta were already in existence under the Achæmenian kings; but that taken as a whole, they were not the general ideas and customs of the whole of Persia, but only of the sacerdotal caste 7. There were

p. xlvi

therefore, practically, two religions in Iran, the one for laymen and the other for priests. The Avesta was originally the sacred book only of the Magi, and the progress of the religious evolution was to extend to laymen what was the custom of the priests.

§ 14. We are now able to understand how it was that the sacred book of Persia was written in a non-Persian dialect: it had been written in the language of its composers, the Magi, who were not Persians. Between the priests and the people there was not only a difference of calling, but also a difference of race, as the sacerdotal caste came from a non-Persian province. What that province was we know both from Greek historians and from Parsi traditions.

All classical writers, from Herodotus down to Ammianus, agree in pointing to Media as the seat and native place of the Magi. 'In Media,' says Marcellinus (XXIII, 6), 'are the fertile fields of the Magi . . . (having been taught in the magic science by King Hystaspes) they handed it down to their posterity, and thus from Hystaspes to the present age an immense family was developed, hereditarily devoted to the worship of the gods. . . . In former times their number was very scanty . . ., but they grew up by and by into the number and name of a nation, and inhabiting towns without walls they were allowed to live according to their own laws, protected by religious awe.' Putting aside the legendary account of their origin, one sees from this passage that in the time of Marcellinus 1 (fourth cent. A.D.) there was in Media a tribe, called Magi, which had the hereditary privilege of providing Iran with priests. Strabo, writing three centuries before Marcellinus, considered the Magi as a sacerdotal tribe spread over the land 2 . Lastly, we see in Herodotus (III, 65) that the usurpation of the Magian Smerdis was interpreted

p. xlvii

by Cambyses, as an attempt of the Medes to recover the hegemony they had lost, and when we learn from Herodotus (I, 101) that the Medes were divided into several tribes, Busae, Paraetakeni, Strouchates, Arizanti, Budii, and Magi, without his making any remark on the last name, we can hardly have any doubt that the priests known as Magi belonged to the tribe of the Magi, that they were named after their origin, and that the account of Marcellinus may be correct even for so early a period as that of Herodotus.

§ 15. Parsi traditions agree with Greek testimonies.

That the priesthood was hereditary, we see from the statement in the Bundahis, that all the Maubeds are descendants from King Minochihr 1, and even nowadays the priesthood cannot extend beyond the priestly families; the son of a Dastur is not obliged to be a Dastur, but no one that is not the son of a Dastur can become one 2.

That they came from Media, we see from the traditions about the native place of Zoroaster, their chief and the founder of their religion. Although epic legends place the cradle of Mazdean power in Bactria, at the court of King Vistâsp, Bactria was only the first conquest of Zoroaster, it was neither his native place, nor the cradle of his religion. Although there are two different traditions on this point, both agree in pointing to Media; according to the one be was born in Rai, that is in Media, properly so called; according to the other he was born in Shîz, that is in Media Atropatene.

The former tradition seems to be the older; it is expressed directly in the Pahlavi Commentary to Vendîdâd I, 16 3; and there is in the Avesta itself (Yasna XIX, 18 (50)) a passage that either alludes to it or shows how it originated.

'How many masters are there?'

p. xlviii

'There are the master of the house, the lord of the borough, the lord of the town, the lord of the province, and the Zarathustra (the high-priest) as the fifth. So is it in all lands, except in the Zarathustrian realm; for there are there only four masters, in Ragha, the Zarathustrian city 1.'

'Who are they?'

'They are the master of the house, the lord of the borough, the lord of the town, and Zarathustra is the fourth 2.'

This amounts to saying that the high-priest, the Maubedân Maubed, held in Rai the position of the dahvyuma, or lord of the land, and was the chief magistrate. It may be suspected that this was the independent sacerdotal state which is spoken of in Marcellinus, and this suspicion is raised to a certain degree of probability by the following lines in Yaqût:

'Ustûnâwand, a celebrated fortress in the district of Danbawand, in the province of Rai. It is very old, and was strongly fortified. It is said to have been in existence more than 3000 years, and to have been the stronghold of the Masmoghân of the land during the times of paganism. This word, which designates the high-priest of Zoroastrian religion, is composed of mas, "great," and moghân, which means "magian." Khaled besieged it, and the power of the last of them 3.'

According to another tradition Zarathustra was born in Atropatene. The very same commentary which describes Ragha as being identical with Rai, and the native place of Zartust, also informs us that Ragha was brought by others

p. xlix

to be Atropatene. Traditions, of which unfortunately we have only late records, make him a native of Shîz, the capital of Atropatene 1: 'In Shîz is the fire temple of Azerekhsh, the most celebrated of the Pyraea of the Magi; in the days of the fire worship, the kings always came on foot, upon pilgrimage. The temple of Azerekhsh is ascribed to Zeratusht, the founder of the Magian religion, who went, it is said, from Shîz to the mountain of Sebîlân, and, after remaining there some time in retirement, returned with the Zend-Avesta, which, although written in the old Persian language, could not be understood without a commentary. After this he declared himself to be a prophet 2.'

Now we read in the Bundahis that Zartust founded his religion by offering a sacrifice in Irân Vêg (Airyanem Vaêgô) 3. Although this detail referred originally to the mythical character of Zoroaster, and Irân Vêg was primitively no real country, yet as it was afterwards identified with the basin of the Aras (Vanguhi Dâitya) 4, this identification is a proof that the cradle of the new religion was looked for on the banks of the Aras. In the Avesta itself we read that Zoroaster was born and received the law from Ormazd on a mountain, by the river Dareg5, a name which strikingly reminds one of the modern Darah river, which falls from the Sebîlân mount into the Aras.

To decide which of the two places, Rai or Atropatene, had the better claim to be called the native place of Zoroaster is of course impossible. The conflict of the two traditions must be interpreted as an indication that both places were important seats of the Magian worship. That both traditions may rely on the Avesta is perhaps a sign that the Avesta contains two series of documents, the one emanating from the Magi of Ragha, and the other from the

p. l

[paragraph continues] Magi of Atropatene 1. Which of the two places had the older claim is also a question hardly to be settled in the present state of our knowledge 2.

Whether Magism came from Ragha to Atropatene, or from Atropatene to Ragha, in either case it had its origin in Media 3. That Persia should have submitted in religious matters to a foreign tribe will surprise no one who thinks of the influence of the Etruscan augurs in Rome. The Magi might be hated as Medes, but they were respected and feared as priests. When political revolutions gave vent to national hate, the Persian might willingly indulge it, and revel in the blood of the foreign priest 4; yet whenever he had to invoke the favour of the gods, he was obliged to acknowledge that he could not do without the detested tribe, and that they alone knew how to make themselves beard by heaven 5. When and how the religious hegemony of Media arose we cannot say: it is but natural that Media 6,

p. li

having risen sooner to a high degree of civilisation, should have given to religion and worship a more systematic and elaborate form, and in religion, as in politics, the best organised power must sooner or later get the upper hand. It is likely that it began with the conquest of Media by Cyrus: Media capta ferum victorem cepit. . . . Cyrus is said to have introduced the Magian priesthood into Persia (Xenophon, Cyrop. VIII, I, 23), which agrees with the legend mentioned by Nikolaus that it was on the occasion of the miraculous escape of Crœsus that the Persians remembered the old λογία of Zoroaster forbidding the dead to be burnt.

The Medic origin of the Magi accounts for a fact which perplexes at first sight, namely, the absence of the name of the Magi from the book written by themselves 1; which is natural enough if the word Magu was not the name of the priest as a priest, but as a member of the tribe of the Magi. The proper word for a priest in the Avesta is Âthravan, literally, 'fire-man,' and that this was his name with the Persians too appears from the statement in Strabo (XV, 733) that the Magi are also called Πύραιθοι. It is easy to conceive that the Persians, especially in ordinary parlance, would rather designate their priests after their origin than after their functions 2; but the Magi themselves had no reason to follow the Persian custom, which was not always free from an implication of spite or scorn. The only passage into which the word found its way is just one that betrays the existence of this feeling: the enemy of the priests is

p. lii

not called, as would be expected, an Âthrava-tbis, 'a hater of the Âthravans' (cf. the Indian Brahma-dvish), but a Moghu-tbis, 'a hater of the Magi 1.' The name, it is true, became current in Pahlavi and modern Persian, but it was at a time when the old national quarrels between Media and Persia were quenched, and the word could no longer carry any offensive idea with it.

§ 16. The results of the foregoing research may be summed up as follows:--

The original texts of the Avesta were not written by Persians, as they are in a language which was not used in Persia, they prescribe certain customs which were unknown to Persia, and proscribe others which were current in Persia. They were written in Media, by the priests of Ragha and Atropatene, in the language of Media, and they exhibit the ideas of the sacerdotal class under the Achæmenian dynasty.

It does not necessarily follow from this, that the original fragments were already written at the time of Herodotus 2.

p. liii

[paragraph continues] But as the Magi of that time sang songs of their gods during sacrifice, it is very likely that there was already a sacred literature in existence. The very fact that no sacrifice could be performed without the assistance of the Magi makes it highly probable that they were in possession of rites, prayers, and hymns very well composed and arranged, and not unlike those of the Brahmans; their authority can only be accounted for by the power of a strongly defined ritual and liturgy. There must, therefore, have been a collection of formulae and hymns, and it is quite possible that Herodotus may have heard the Magi sing, in the fifth century B.C., the very same Gâthas which are sung nowadays by the Mobeds in Bombay. A part of the Avesta, the liturgical part, would therefore have been, in fact, a sacred book for the Persians. It had not been written by them, but it was sung for their benefit. That Zend hymns should have been sung before a Persian-speaking people is not stranger than Latin words being sung by Frenchmen, Germans, and Italians; the only difference being that, owing to the close affinity of Zend to Persian, the Persians may have been able to understand the prayers of their priests.

§ 17. It may, therefore, be fairly admitted that, on the whole, the present texts are derived from texts already existing under the Achæmenian kings. Some parts of the collection are undoubtedly older than others; thus, the Gâthas are certainly older than the rest of the Avesta, as they are often quoted and praised in the Yasna and the Vendîdâd; but it is scarcely possibly to go farther than a logical chronology. One might feel inclined, at first sight, to assign to a very recent date, perhaps to the last revision of the Avesta, those long enumerations of gods so symmetrically elaborated in the Yasna, Vispêrad, and Vendîdâd. But the Account of Mazdeism given by Plutarch shows that the

p. liv

work of co-ordination was already terminated at the end of the Achæmenian period, and there is no part of the Avesta which, so far as the matter is concerned, may not have been written in those times. Nay, the Greek accounts of that period present us, in some measure, with a later stage of thought, and are pervaded with a stronger sense of symmetry, than the Avesta itself. Such passages as the latter end of the Zamyâd Yast and Vendîdâd X, 9 seq. prove that, when they were composed, the seven Arch-Dêvs were not yet pointedly contrasted with the seven Amshaspands, and therefore those passages might have been written long before the time of Philip. The theory of time and space as first principles of the world, of which only the germs are found in the Avesta, was fully developed in the time of Eudemos, a disciple of Aristotle.

§ 18. To what extent the Magian dogmatical conceptions were admitted by the whole of the Iranian population, or how and by what process they spread among it, we cannot ascertain for want of documentary evidence. As regards their observances we are better instructed, and can form an idea of how far and in what particulars they differed from the other Iranians. The new principle they introduced, or, rather, developed into new consequences, was that of the purity of the elements. Fire, earth, and water had always been considered sacred things, and had received worship 1: the Magi drew from that principle the conclusion that burying the dead or burning the dead was defiling a god: as early as Herodotus they had already succeeded in preserving fire from that pollution, and cremation was a capital crime. The earth still continued to be defiled, notwithstanding the example they set; and it was only under the Sassanians, when Mazdeism became the religion of the state, that they won this point also.

The religious difference between the Persians and their Medic priests was therefore chiefly in observances. Out of the principles upon which the popular religion rested, the sacerdotal class drew by dint of logic, in a puritan spirit,

p. lv

the necessity of strict observances, the yoke of which was not willingly endured by the mass of the people. Many acts, insignificant in the eyes of the people, became repugnant to their consciences and their more refined logic. The people resisted, and for a time Magian observances were observed only by the Magi. The slow triumph of Magism can be dimly traced through the Achæmenian period. Introduced by Cyrus, it reigned supreme for a time with the Pseudo-Smerdis, and was checked by Darius 1. It seems to have resumed its progress under Xerxes; at least, it was reported that it was to carry out Magian principles that he destroyed the Greek temples, and that the first who wrote on the Zoroastrian lore was a Magian, named Osthanes, who had accompanied him to Greece 2. New progress marked the reign of Artaxerxes Longimanus. The epic history of Iran, as preserved in the Shah Nâmah, passes suddenly from the field of mythology to that of history with the reign of that king, which makes it likely that it was in his time that the legends of Media became national in Persia, and that his reign was an epoch in the political history of Magism 3. But the real victory was not won till six centuries later, when national interest required a national religion. Then, as happens in every revolution, the ultra party, that had pushed to the extreme the principles common to all, took the lead; the Magi ascended the throne with Ardeshîr, one of their pupils 4, and the Magian

p. lvi

observances became the law of all Iran. But their triumph was not to be a long one; their principles required an effort too continuous and too severe to be ever made by any but priests, who might concentrate all their faculties in watching whether they had not dropped a hair upon the ground. A working people could not be imprisoned in such a religion, though it might be pure and high in its ethics. The triumph of Islam was a deliverance for the consciences of many 1, and Magism, by enforcing its observances upon the nation, brought about the ruin of its dogmas, which were swept away at the same time: its triumph was the cause and signal of its fall 2.


xxx:1 A very improper designation, as Zend means 'a commentary or explanation,' and was applied only to explanatory texts, to the translations of the Avesta. Avesta (from the old Persian âbastâ, 'the law;' see Oppert, Journal Asiatique, 1872, Mars) is the proper name of the original texts. What it is customary to call, 'the Zend language' ought to be named, 'the Avesta language;' the Zend being no language at all; and, if the word be used as the designation of one, it can be rightly applied only to the Pahlavi. The expression 'Avesta and Zend' is often used in the Pahlavi commentary to designate 'the law with its traditional and revealed explanation.'

xxxii:1 Ravâet ap. Anquetil, Mémoires de l’Acad. des Inscr. et Belles-Lettres XXXVIII, 216; Spiegel, Zeitschrift der Deutschen Morgenländischen Gesellschaft IX, 174.

xxxii:2 J. Darmesteter, La légende d’Alexandre chez les Parses.

xxxii:3 We find in it a description of the four classes, which strikingly reminds p. xxxiii one of the Brahmanical account of the origin of the castes (Chap. XLII; cf. the first pages of the Shikan Gumânî), and which was certainly borrowed from India; whether at the time of the last Sassanians, when Persia learnt so much from India, or since the settlement of the Parsis in India, we are unable to decide: yet the former seems more probable.

xxxiii:1 Gvêt rastakân. We are indebted to Mr. West for the right translation this word.

xxxiii:2 Thus translated by West (Glossary of the Book of Ardâ Vîrâf, p. 27).

xxxiii:3 Haug, Essay on Pahlavi p. 145 seq., 149 seq.

xxxiii:4 Spiegel, Eranische Alterthumskunde III, 782, n. 1.

xxxiii:5 S. de Sacy, Mémoires sur quelques antiquités de la Perse. Cf. Masudi, 125. II, 125.

xxxiv:1 Perhaps five (see de Longpérier, Mémoire sur la Numismatique des Arsacides, p. 111).

xxxiv:2 'Magus ad eum Tiridates venerat' (Pliny, Nat. Hist. XXX, 6).

xxxiv:3 Pliny very often confounds Magism and Magia, Magians and Magicians. We know from Pliny, too, that Tiridates refused to initiate Nero into his art: but the cause was not, as he assumes, that it was 'a detestable, frivolous, and vain art,' but because Mazdean law forbids the holy knowledge to be revealed to laymen, much more to foreigners (Yast IV, 10; cf. Philostrati Vita Soph. I, 10).

xxxiv:4 'Nec recusaturum Tiridatem accipiendo diademati in urbem venire, nisi sacerdotii religione attineretur' (Ann. XV, 24).

xxxiv:5 He crossed only the Hellespont.

xxxiv:6 'Navigare noluerat quoniam inspuere in maria, aliisque mortalium necessitatibus violare naturam eam fas non putant' (Pliny, l. l. Cf. Introd. V, 8 seq.).

xxxv:1 Dio Cassius, LXIII, 4. The answer was mistaken for an insult by Nero, and, as it seems, by Dio himself In fact Vologeses remained to the last faithful to the memory of Nero (Suet. Nero, 57). What we know moreover of his personal character qualifies him for taking the initiative in a religious work. He seems to have been a man of contemplative mind rather than a man of action, which often excited the anger or scorn of his people against him; and he had the glory of breaking with the family policy of Parthian kings (Tacitus, Annales, XV, 1, 2). It was under his reign that the first interference of religion with politics, of which the history of Persia speaks, took place, as he was called by the people of Adiabene against their king Izates, who had become a Jew (Josephus, Antiq. XX, 4, 2).

xxxv:2 Hamzae Ispahensis Annales, ed. Gottwaldt, p. 31 (in the translation).

xxxvii:1 Vide infra, p. xli, note 3.

xxxviii:1 Shapûr II ascended the throne about 309 (before being born, as the tradition goes): and as he appears from the Dînkart to have taken a personal part in the work of Âdarbâd, the promulgation of the Avesta can hardly have taken place at an earlier date than 325-330. Âdarbâd and the Fathers at Nicaea lived and worked in the same age, and the Zoroastrian threats of the king of Iran and the Catholic anathemas of the Kaisar of Rûm may have been issued on the same day.

xxxix:1 See the book of the Mainyô-i-Khard, ed. West; Introduction, p. x seq.

xl:1 Chap. XV, 16 seq. as translated by West.

xl:2 Ashemaogha, 'the confounder of Asha' (see IV, 37), is the name of the fiends and of the heretics. The Parsis distinguish two sorts of Ashemaoghas, the deceiver and the deceived; the deceiver, while alive, is margarzân, p. xli 'worthy of death,' and after death is a darvand (a fiend, or one of the damned); the deceived one is only margarzân.

xl:3 The Pahlavi translation illustrates the words 'who does not eat' by the gloss, 'like Mazdak, son of Bâmdâd,' which proves that this part of the commentary is posterior to, or contemporary with the crushing of the Mazdakian sect (in the first years of Khosrav Anôsharvân, about 531). The words 'against the wicked tyrant' are explained by the gloss, 'like Zarvândâd;' may it not be Kobâd, the heretic king, or 'Yazdgard the sinner,' the scorner of the Magi?

xli:4 Elisaeus, pp. 29, 52. in the French translation by Garabed.

xli:5 At least with orthodox Christianity, which seems to have alone prevailed in Persia till the arrival of the Nestorians. The description would apply very well to certain gnostic sects, especially that of Cerdo and Marcio, which is no wonder as it was through that channel that Christianity became known to Mânî. Masudi makes Mânî a disciple of Kardûn (ed. B. de Meynard, II, 167), and the care which his biographer (ap. Flügel, Mânî, pp. 51, 85) takes to determine the length of time which intervened between Marcio and Mânî seems to betray some dim recollection of an historical connection between the two doctrines.

xli:6 The patriarch of Alexandria, Timotheus, allowed the other patriarchs, p. xlii bishops, and monks to eat meat on Sundays, in order to recognise those who belonged to the Manichean sect (Flügel, p. 279).

xlii:1 'Those who follow the heresy of Prodicus boast of possessing secret books of Zoroaster,' Clemens Alex. Stromata I. Cf. the ἀποκαλύψεις Ζωροάστρου forged by Adelphius or Aquilinus (ap. Porphyr. Vita Plotini, § 16).

xlii:2 Ἐπᾴδει δὲ ἐπιλεγόμενος ἐκ βιβλίου (V, 27, 3).

xlii:3 See Windischmann, Zoroastrische Studien, 288.

xlii:4 'Hermippus, qui de tota arte ea (magia) diligentissime scripsit et viciens centiens milia versuum a Zoroastre condita indicibus quoque voluminum ejus positis explanavit.' . . . (Hist. Nat. XXX, 1, 2). He had written a book περὶ μάγων (Diog. Laert. Prooem. 8).

xliii:1 In Plutarch, De Iside et Osiride, §§46-47.

xliii:2 Men, when raised from the dead, shall have no shadow any longer (μήτε σκιὰν ποιοῦντας). In India, gods have no shadows (Nalus); in Persia, Râshidaddîn was recognised to be a god from his producing no shadow (Guyard, Un grand maitre, des Assassins, Journal Asiatique, 1877, I, 392); the plant of eternal life, Haoma, has no shadow (Henry Lord).

xliii:3 Persian tradition cannot be much relied on, when it tries to go back beyond Alexander, and on that special point it seems to be more an inference of later ages, than a real tradition; but the inference happens to be right.

xliv:1 Professor Oppert thinks he has found in Darius' inscriptions an express mention of Ahriman (Le peuple et le langue des Mèdes, p. 199); yet the philological interpretation of the passage seems to me still to obscure to allow of any decisive opinion. Plutarch introduces Artaxerxes I speaking of Ἀρειμάνιος, but whether the king is made to speak the language of his own time, or that of Plutarch's time, is left doubtful. As to the allusions in Isaiah (xlv), they do not necessarily refer to dualism in particular, but to all religions not monotheistic. (Cf. Ormazd et Ahriman, §241.)

xliv:2 Vide infra, IV, 5.

xlv:1 Vide infra, IV, 35; cf. Fargard XIII, 5 seq.; XIV, 5.

xlv:2 Herod. I, 140.

xlv:3 Vide infra, V, 9.

xlv:4 Procopius, De Bello Persico, I, II.

xlv:5 Ibid. I, 12.

xlv:6 Herod. I. 140.

xlv:7 There are other features of the Avesta religion which appear to have been foreign to Persia, but are attributed to the Magi. The hvaêtvôdatha, the holiness of marriage between next of kin, even to incest, was unknown to p. xlvi Persia under Cambyses (Herod. III, 31), but it is highly praised in the Avesta, and was practised under the Sassanians (Agathias II, 31); in the times before the Sassanians it is mentioned only as a law of the Magi (Diog. Laert. Prooem. 6; Catullus, Carm. XC).

xlvi:1 Or of the historians from whom he copies. Still he seems to speak from contemporary evidence. Sozomenus (Hist. Eccles. II, 9) states that the care of worship belonged hereditarily to the Magi 'as to a sacerdotal race,' ὡ?'σπερ τι φῦλον ἱερατικόν.

xlvi:2 Τὸ τῶν Μάγων φῦλον (XV, 14).

xlvii:1 Bundahis 79, 13.

xlvii:2 Dosabhoy Framjee, The Parsees, &c. p. 277.

xlvii:3 'Ragha of the three races,' that is to say, Atropatene (vide infra); some say it is 'Rai.' It is 'of the three races' because the three classes, priests, warriors, husbandmen, 'were well organized there. Some say that Zartust was born there . . ., those three classes were born from him.' Cf. Bundahis 79, 15, and Farg. II, 43, n. 2. Rai is the Greek Ῥαγαί.

xlviii:1 Or possibly, 'in the Zarathustrian Ragha.'

xlviii:2 The Commentary has here: 'that is to say, he was the fourth master in his own land.'

Their spreading and wandering over Mazdean lands appears from Yasna XLII, 6 (XII, 34): 'We bless the coming of the Âthravans, who come from afar to bring holiness to countries;' cf. infra, p. lii, note 1, and Farg. XIII, 22.

xlviii:3 Dictionnaire géographique de la Perse, traduit par Barbier de Meynard, p. 33. Cf. Spiegel, Eranische Alterthumskunde III, 565. A dim recollection of this Magian dynasty seems to survive in the account ap. Diog. Laert. (Prooem. 2) that Zoroaster was followed by a long series of Magi, Osthanae Astrampsychi, and Pazatae, till the destruction of the Persian empire by Alexander.

xlix:1 The Persian Gazn, the Byzantine Gaza Ganzaka, the site of which was identified by Sir Henry Rawlinson with Takht i Suleiman (Memoir on the Site of the Atropatenian Ecbatana, in the journal of the Royal Geographical Society, X, 65).

xlix:2 Kazwini, and Rawlinson, l.c. p. 69.

xlix:3 Bund. 79, 12.

xlix:4 See Farg. I, p. 3.

xlix:5 See Farg. XIX, 4, 11.

l:1 This would be a principle of classification which unfortunately applies only to a small part of the Avesta.

l:2 Still, if we follow the direction of the Zoroastrian legend, Magism must have spread from west to east, from Atropatene to Ragha, from Ragha to Bactria; and Atropatene must thus have been the first cradle of Mazdeism. Its very name points to its sacred character; oriental writers, starting from the modern form of the name, Adarbîgân, interpret it as 'the seed of fire,' with an allusion to the numerous fire springs to be found there. Modern scholars have generally followed the historical etymology given by Strabo, who states that, after the death of Alexander, the satrap Atropates made himself an independent sovereign in his satrapy, which was named after him Atropatene. This looks like a Greek etymology (scarcely more to be trusted than the etymology of Ῥαγαί, from ῥήγνυμι), and it is hardly to be believed that the land should have lost its former name to take a new one from its king; it was not a new-fangled geographical division, like Lotharingia, and had lived a life of its own for a long time before. Its name Âtarpatakân seems to mean 'the land of the descent of fire,' as it was there that fire came down front heaven (cf. Ammianus l.c.)

l:3 The Pahlavi names of the cardinal points show that Media was the centre of orientation in Magian geography (Garrez, Journal Asiatique, 1869, II).

l:4 Magophonia (Herod. III, 79).

l:5 Ὡς ἀυτοὺς μόνους ἀκουομένουσ (Diog. Laert. Prooem.); cf. Herod. I, 132 Ammian. l. l.

l:6 An echo of the old political history of Media seems to linger in Yast V, 29, which shows Azi Dahâka reigning in Babylon (Bawru); as Azi, in his legendary character, represents the foreign invader, this passage can hardly be anything but a far remote echo of the struggles between Media and the Mesopotamian empires. The legend of Azi is localised only in Medic p. li lands: he addresses his prayers to Ahriman by the banks of the Sipît rût (Bundahis 52, 11), his adversary Ferîdûn is born in Ghilân, he is bound to Mount Damâvand (near Rai).

li:1 In their own language, the Zend; of which the modern representatives, if there be any left, should therefore be looked for in Atropatene or on the banks of the Caspian sea. The research is complicated by the growing intrusion of Persian words into the modern dialects, but as far as I can see from a very inadequate study of the matter, the dialect which exhibits most Zend features is the Talis dialect, on the southern bank of the Aras.

li:2 The Pahlavi has 'one who hates the Magu-men.' In the passage LIII (LII), 7, magéus is not a Magian, and it is translated by magi, 'holiness, godliness,' related to the Vedic magha. Afterwards the two words were confounded, whence came the Greek statement that μάγος means at the same time 'a priest' and 'a god' (Apollon. Tyan. Ep. XVII).

lii:1 A further echo of the anti-Magian feelings may be heard in Yasna IX, 24 (75): 'Haoma overthrew Keresâni, who rose up to seize royalty, and he said, "No longer shall henceforth the Âthravans go through the lands and teach at their will."' This is a curious instance of how easily legendary history may turn myths to its advantage, The struggle of Haoma against Keresâni is an old Indo-European myth, Keresâni being the same as the Vedic Krisânu, who wants to keep away Soma from the hands of men. His name becomes in the Avesta the name of an anti-Magian king it may be Darius, the usurper (?), and ten centuries later it was turned into an appellation of the Christian Kaisars of Rûm (Kalasyâk = *ἐκκλησια[κός]; Tarsâka).

lii:2 If the interpretation of the end of the Behistun inscription (preserved only in the Scythian version) as given by Professor Oppert be correct, Darius must have made a collection of religious texts known as Avesta, whence it would follow, with great probability, that the present Avesta proceeded from Darius. The translation of the celebrated scholar is as follows: 'J’ai fait une collection de textes (dippimas) ailleurs en langue arienne qui autrefois n’existait pas. Et j’ai fait un texte de la Loi (de l’Avesta; Haduk ukku) et un commentaire de la Loi, et la Bénédiction (la prière, le Zend) et les Traductions.' (Le peuple et la langue des Mèdes, pp. 155, 186.) The authority of Oppert is so great, and at the same time the passage is so obscure, that I hardly know if there be more temerity in rejecting his interpretation or in adopting it. Yet I beg to observe that the word dippimas is the usual Scythian transliteration of the Persian dipi, 'an inscription,' and there is no apparent reason for departing from that meaning in this passage; if the word translated 'la Loi,' ukku really represents here a Persian word Abasta, it need not denote the Avesta, the religious book, p. liii as in that case the word would most certainly not have been translated in the Scythian version, but only transliterated; the ideogram for 'Bénédiction, prière,' may refer to religious inscriptions like Persepolis I; the import of the whole passage would therefore be that Darius caused other inscriptions to be engraved, and wrote other edicts and religious formulae (the word, 'traductions' is only a guess).

liv:1 Cf. V, 8.

lv:1 Darius rebuilt the temples which the Magus Gaumata had destroyed (Behistun I, 63). The Magi, it is said, wanted the gods not to be imprisoned within four walls (Cic. de Legibus II, 10). Xerxes behaved himself as their disciple, at least in Greece. Still the Magi seem to have at last given way on that point to the Perso-Assyrian customs, and there were temples even under the Sassanians.

lv:2 Pliny, Hist. Nat., XXX, I, 8.

lv:3 Cf. Westergaard, Preface to the Zend-Avesta, p. 17. This agrees with what we know of the fondness of Artaxerxes for religious novelties, It was he who blended the worship of the Assyrian Anat-Mylitta with that of the Iranian Anâhita (the ascription of that innovation to Artaxerxes Mnemon by Clemens Alexandrinus (Stromata I) must rest on a clerical error, as in the time of Herodotus, who wrote under Longimanus, the worship of Mylitta had already been introduced into Persia (I, 131)).

lv:4 Agathias II, 26.

lvi:1 De Gobineau, Histoire des Perses, II, 632 seq.

lvi:2 We ought to discuss here the Scythian theory of Magism; but thus far we have been unable to find anywhere a clear and consistent account of its thesis and of its arguments. Nothing is known of any Scythian religion, and what is ascribed to a so-called Scythian influence, the worship of the elements, is one of the oldest and most essential features of the Aryan religions.

Next: Chapter IV. The Origin of the Avesta Religion