[In the essay below, Boyce examines the way in which the veneration of fire, which served an important role in Zoroaster's teachings, was transformed by his followers to a ritual temple cult.]

The iconoclastic movement in Christianity has been carefully studied, as has Islamic iconomachy, but the origins of both still present problems; and in investigating these consideration should certainly be given to the fact that Zoroastrianism, ancient and until the 9th century a.d. immensely influential, had an iconoclastic movement which preceded both, and which may well have played a part in inspiring them. Zoroastrian iconoclasm has been ignored for various reasons. The history of the faith is poorly documented for all periods before the 17th century a.d., and has to be pieced together (as far as this is at all possible) from sparse and diverse sources. It is easy, therefore, to overlook whole strands in its composition. Moreover, the assumption that the cult of temple fires was original to it, and remained its sole form of public worship, has obscured this particular issue. That such an assumption has been generally made is in itself a tribute to the success of the Zoroastrian iconoclasts, who triumphed so completely that in the end fire was the sole icon in the temples of their faith, and they and their coreligionists became known to the world at large simply as ‘fire-worshippers’.

The fact is that, though veneration of fire is very ancient among the Iranians, and was of supreme importance in Zoroaster's teachings, the cult of temple fires appears to have been unknown in early Zoroastrianism.¹ Indo-Iranian religion had taken shape during millennia of nomadic wanderings on the Central Asian steppes, and its cult was therefore materially very simple, without temples, altars or statues. The Iranians, like the Vedic Indians, held tenaciously to this tradition. The essence of Zoroastrian devotional life was worship of Ahura Mazdā, the Creator, in the presence of his own creations, namely the sky, water, earth, plants, animals, man and fire. The last, held to be the all-pervading element which gave life and warmth to the rest, was represented visibly both by the sun on high and by fire on the domestic hearth, which from time immemorial was tended with reverent care and never allowed to go out. In Zoroaster's teachings fire was linked with Aša, the yazata of righteousness and good order; and his followers were enjoined to pray either at their hearths or in the open, turned towards the sun, so that they had fire always before them to help fix their thoughts on righteousness.

This tradition of worship under the sky or in the home was continued evidently during the early Achaemenian period. The great sanctuary at Zela in Asia Minor, founded, it is said, in thanksgiving in the 6th century b.c., consisted of an artificial mound raised on the plain so that men could go up to offer their veneration there;² and at Pasargadae two massive plinths still stand in the open, one with steps leading up it; and it has been suggested that these were built so that the king, mounting upon the one, could fix his eyes on fire set on the other and thus pray in fitting manner before a great assembly.³ Still in the mid-5th century b.c. Herodotus records that ‘as to the usages of the Persians … it is not their custom to make and set up statues and temples and altars’.⁴ Instead they climbed high into the mountains to offer sacrifice there. The Western Iranians were exposed, however, to strong influences from their alien subjects and neighbours—Elamites, Babylonians, Assyrians, Mannai and others—all of whom used statues and altars in their worship. Near Hamadan, in Medean territory, a curious tower-like structure has been excavated, thought to belong to the 8th century b.c., and in it was found an altar, about waist-high, with broad, stepped top and shallow bowl, in which fire was evidently sometimes kindled, for traces of burning remain.⁵ Moreover, in the carvvings set above the tombs of Darius the Great and his descendants each king is represented as standing on a three-stepped dais, facing fire burning on a three-stepped ‘altar’.⁶ In the light both of Herodotus' report, and the absence of temple ruins at Pasargadae and Persepolis,⁷ it is possible that these ‘altars’ bore occasional fires only (like, perhaps, the second plinth at Pasargadae), placed upon them for the public performance of royal acts of devotion. (Such ‘altars’, called simply ‘fire-holders’,⁸ are still to be found in the outer rooms of all old fire temples in the Yazdī area.) Later usage suggests that such fires were either kindled when needed, or created from embers brought from the nearest hearth fire. Another possibility is that each fire upon a funerary monument was the king's personal fire, that is, his hearth fire, elevated thus to burn in a manner fitting to royal dignity, and dying when he died.

The oldest temple ruin as yet to be found in Zoroastrian Persia is one excavated at the Achaemenian capital of Susa.⁹ This has been attributed, on the evidence of architectural detail, to the reign of Artaxerxes II Memnon (404-359)—the very monarch who is reported to have imposed an image-cult generally upon his subjects.¹⁰ He was much attached, we are told, to Anaïtis, an alien fertility goddess whose cult had already been adopted by Western Iranians at the time when Herodotus wrote. She had become assimilated, it seems, to the Iranian river-goddess ^*Harahvatī Arsdvī Sūrā,¹¹ who came to be known thereafter as Arsdvī Sūrā Anāhitā; and at some point in his long reign Artaxerxes is said to have given orders that statues to her should be erected in many of the chief places of his empire, including Medean Ecbatana (Hamadan), at or near Persepolis itself,¹² and in Bactria in the remote north-east, a noted Zoroastrian stronghold. The cult thus dictatorially established was fostered evidently with the utmost lavishness. Verses in the hymn to Arsdvī Sūrā Anāhitā are held to describe one such statue, and they present the goddess as wearing golden shoes and earrings, a precious necklace and jewel-encrusted mantle, with a radiate crown upon her head.¹³ Splendid temples were evidently built to house these costly images, and later that at Ecbatana (tiled, it is said, with silver and with gold-plated columns) was ruthlessly plundered for its wealth by Macedonian soldiery.¹⁴

The king's power was absolute in Achaemenian Persia; and it is natural that there should have been men, both priest and lay, who were ready to bow to Artaxerxes' will and do their utmost to please him. The verses incorporated in Arsdvī's hymn illustrate this conformity; and possibly the Persian word for image, *uzdaēsa (Middle Persian uzdēs) was coined at this time to justify the new cult. It seems to mean, like Greek ειxών, a ‘showing forth, representation’;¹⁵ and perhaps in evolving it Zoroastrian iconophils sought to characterize their new images as a semblance of the divine only, in whose presence men should pray in order to direct their thoughts to what lay beyond, rather than idols, to be worshipped for themselves. Yet however scholar-priests of the royal party may have argued the matter, the introduction of an image cult must have shocked the orthodox profoundly; for by it a man-made statue was substituted for the living icon of fire, the creation of Ahura Mazdā, which had been enjoined as qibla by the prophet himself.¹⁶ There was, moreover, a serious doctrinal consideration involved. Zoroastrian theologians taught that originally Ahura Mazdā had made his creation in spirit-form only, mēnōgīhā as it was expressed in Middle Persian.¹⁷ His Adversary, Anra Mainyu, countered with an evil creation, also intangible; and thereafter Ahura Mazdā by a mighty exertion of power enabled his mēnōg creation to ‘put on appearances’, that is, to take physical (gētīg) forms.¹⁸ This second stage was beyond Anra Mainyu's capacity, and so the powers of evil have no material bodies of their own, but steal shapes to inhabit in the furtherance of wickedness. To the orthodox, therefore, an image maker was guilty both of impiety, in seeking to perform the act of creation himself (the prerogative of God the Creator), and also of rash folly, since he had fashioned an empty form which a daēva or evil being could enter to misappropriate the worship intended for the divinity, and grow stronger thereby. Hence in surviving Zoroastrian works temples with statues in them are referred to as the ‘abode of dēvs’ (nišēmag ī dēwān),¹⁹ and the term ‘image worship’ (uzdēs-parastagīh) has for a synonym ‘demon worship’ (dēwīzagīh).²⁰

In these circumstances one can safely assume that an impulse towards iconoclasm sprang into being among Zoroastrians with the setting up of the first statues to Arsdvī Sūrā Anāhitā in the 4th century b.c.; and there may well have been unsung martyrs then in this cause. Royal patronage brought it about, however, that the image cult was firmly implanted; and the energies of the orthodox seem to have been turned therefore into another channel, that of instituting the veneration of fire, the true Zoroastrian icon, as a rival temple cult. The origin of the movement cannot be closely dated; but since temple fires were still unknown, it seems, in the mid-5th century, but were widely attested after the downfall of the Achaemenians—all across their former dominions from Parthia to Asia Minor (then no longer an Iranian possession), it is a reasonable assumption that the cult was instituted in late Achaemenian times, probably very soon after that of images.²¹ It appears, therefore, that whereas at the beginning of the Achaemenian period the Zoroastrians had no sacred buildings for public worship, by the end of it they had temples of two kinds, the one sheltering images, the other sacred fires.

This state of affairs evidently continued all through the Parthian period. Strabo records that in his day the Persians in Cappadocia maintained both ‘holy places of the Persian gods’, and also fire-sanctuaries, pyraithoi.²² In the latter, he says, stood altars bearing a great heap of ashes, on which the fire was kept ever alight; and in connection with one of the former he speaks of a wooden image, which on occasion was carried in procession. It appears, not surprisingly, that the Iranians had different names for these two kinds of shrines. The Parthians themselves seem to have called fire-temples *ātarōšan, a word meaning perhaps ‘place of burning fire’,²³ and known from Armenian atrušan (for Armenia was steeped in the Zoroastrian culture of the Arsacids). A shrine to a yazata or divinity was called a *bagin, a term derived from older *bagina, and meaning ‘(a place) belonging to the gods’.²⁴ This term occurs in Sogdian as faγn ‘temple’; and as a Parthian loanword in Armenian it was used in the singular for an altar set before an image (as in the phrase ‘to the altar of Anāhīt's image’ bagnin anahatakan patkerin²⁵), and in the plural for a temple.²⁶Bagnapet (another Parthian loanword in Armenian) was the title of the chief priest of such a temple, and has its equivalents in other Middle Iranian languages (MPersian bašnbed, Sogdian faγnpat),²⁷ and is attested at Mathurā as bakanapati.²⁸ Wherever details occur, they show this group of words to have been associated with the cult of images, not fires. Thus in a Manichaean Middle Persian fragment there is a reference to uzdēsān, bašnbedān ‘images (and) masters of image-temples’,²⁹ and a Sogdian text contains a description of golden images, jewel-adorned, within a faγn.³⁰ The words bagin, bagnapet, or their equivalents, are not attested in later Zoroastrian usage, and presumably they ceased to be current when the iconoclastic movement finally triumphed during the Sasanian epoch.

In Seleucid and early Parthian times, strong Hellenic influences in Iran must have encouraged an increased use of statues by the Zoroastrians. For these periods, as for the Achaemenian epoch, there is pitifully little internal evidence, and most data derive from lands on the borders of the Parthian Empire. In Zoroastrian Armenia, for instance, we learn that there were temples ‘where is sculptured … Aramazd’;³¹ and in others stood Anāhīt's image.³² There was a famous golden statue of this yazat at Erez, which was carried off by one of Mark Antony's soldiers in 36 b.c.;³³ and there are references to offerings made to her there.³⁴ Strabo writes of statues to Anāhīt in Cappadocia also, and possibly to Vohu Manah.³⁵ The temple in Ecbatana, built, it seems, to house one of the Anāhitā statues set up by Artaxerxes II, was pillaged by the Seleucid Antiochus III in 209 b.c.;³⁶ but thereafter it was restored once more, for Isidore of Charax records that sacrifices were continually offered there in his own day (sometime, that is, between 27 b.c. and a.d. 77).³⁷ In Armenia statues are further recorded to Mihr (Mithra),³⁸ Tīr³⁹ and Vahaghn (Vsrsthraghna, yazata of Victory).⁴⁰ The Greeks equated the last-named with their own Herakles, whose cult-name, Kallinikos, ‘Victorious’ must have helped the identification.⁴¹ The Iranian Vahaghn/Varahrān was patron-divinity of travellers; and beside the ancient highway which passes by Bisutun, near Hamadan, there is a little shrine to Herakles Kallinikos, with a carving in high relief of the god, and an inscription in both Greek and Aramaic, showing the meeting of the two cultures, Hellenic and Iranian.⁴² The inscription tells that the shrine was made in the year 164 of the Seleucid era.

A number of other Herakles shrines and statues are known from Parthian Iran,⁴³ and the god is generally shown naked in the Greek style. The Iranians, with their long tradition of worship without images, had not even by this time, it is evident, created an iconography of their own. The first Anaïtis/Anāhitā statues of the Achaemenian period were presumably Semitic in inspiration; and subsequently the Armenians declared that all their icons were made by Greeks, ‘for no one in Armenia knew how to make statues’.⁴⁴ A fine bronze head of a goddess has been found near Erzinjan, which is thought to be from a statue, larger than life-size, of Anāhīt,⁴⁵ and this is certainly indistinguishable from a Greek Aphrodite. The work subsequently of zealous Christians in Armenia, and ardent iconoclasts in Iran itself has insured that very little of such statuary survives. The evidence of the Parthian coins, and those of the Kushans in the east, as well as of the sculptures of Nimrud Dagh in Armenia, combines to show, however, that in the post-Hellenic period the Iranians grew accustomed to having the yazatas of their faith identified with gods of the Greek pantheon, and represented plastically in the same way; and the archaeological and literary evidence attests that there were shrines where these representations took the form of cult-images, within the framework of Zoroastrian worship.

The use of images seems to have become widespread during the Parthian period in the home also. Thus Josephus tells how the widow of a Parthian nobleman, having been made captive, ‘took along the ancestral images of the gods belonging to her husband and herself—for it is the custom among all the people in that country to have objects of worship in their houses and to take them along when going abroad …’.⁴⁶ At first, he says, she performed the due rites before these secretly, so they were evidently small objects, which could be honoured unobtrusively. How reliable Josephus is in his statement that possession of such household images was general there is no means of testing; but there is archaeological and literary evidence to show that, under Hellenistic influence, images came to be used in the universally popular cult of the dead. Excavations in Old Nisa,⁴⁷ the ancestral capital of the Arsacids, have uncovered two halls in which were many statues of men and women in Parthian dress, some larger than life-size, others small. These were made of clay, painted, and realistically modelled; and it is suggested that the halls in which they stood were shrines where rituals for the souls of the dead were performed, the statues being fashioned in honour of individuals to receive the offerings. Greek influence is apparent in the craftsmanship, and was presumably responsible for inspiring the practice. The use of images in the cult of the dead is recorded also in Zoroastrian Armenia;⁴⁸ and it appears to be attested in the remains of a shrine at Shami, a village of Khuzistan in south-western Iran.⁴⁹ Here there came to light in the 1930's a damaged but still splendid statue in bronze, life-sized, of a nobleman wearing Parthian dress.⁵⁰ Excavation of the mound where it had lain uncovered a brick platform which had apparently been partly roofed over to protect cult images—for the remains of other statues were found there, in bronze and marble, some big, some small, as at Nisa.⁵¹ There was also a square imagebase, and before it a small, elegant altar of Hellenic type.⁵² All the statues had been broken into pieces, and the shrine itself burnt over them. Whether this was the work of iconoclasts, or simply the result of local feuding, there is no means of knowing; but the fact that the custom of making images of the dead was wholly unknown in later times shows that this too must have roused the wrath of those opposed to icons, and so in course of time have been suppressed. The practice was possibly considered a little less wicked than that of making images for the divine beings, since such statues were no more than reproductions of the physical forms which men had once possessed. Nevertheless the departed soul belongs wholly to the mēnōg state, and to fashion anew a physical form for it, before the Last Judgment and the resurrection of the body, may well have been held by the orthodox to be both impious and rash (again as creating an empty abode for dēvs). There was also the potentially corrupting power of all icons to attract worship to themselves, as is indicated by the epithet by which Agathangelos stigmatised the Zoroastrians of Armenia, urvapast ‘soul-worshippers’.⁵³ St. Gregory alludes moreover to their habit of prostrating themselves before the images of the dead.⁵⁴

If the cult of images increased during the Parthian period, so too, evidently, did the rival one of sacred fires; so much so that when Ardašir Pāpakān overthrew the Arsacids and established the Sasanian Empire, in about a.d. 224, one of his first acts was to suppress the many fires which had been founded by local rulers,⁵⁵ since these evidently provided a cultic focus for dynastic claims. Subsequent developments show, however, that this was a purely political measure, for Ardašir and his successors distinguished themselves both by founding many new sacred fires, and by giving full support to the iconoclastic movement, which now became triumphant. It seems likely that this movement had already begun to gather strength in the latter part of the Arsacid period, though the indications are necessarily slight. Valakhš (Vologeses) I (c. A.D. 51-80) replaced representations of yazads on the reverse of his coins with a burning fire; and if he was also the Valakhš who commanded his subjects to gather up and preserve the Zoroastrian holy works, he may well have been moved to do this through well-instructed orthodoxy.⁵⁶ The Parthian Empire was, however, a loosely-knit confederacy rather than a firmly controlled state, and whatever position the later Arsacids themselves took up in this controversial matter, there is no evidence that they sought to impose a uniformity of observance on their Zoroastrian subjects throughout the land.

This, however, is what their successors, the Sasanians, did. It is probable that the Persian priests, favoured naturally by the new Persian dynasty, were burning to show that they were superior in zeal and orthodoxy to the Parthian priesthood which had so long had the chief voice in the community. They attacked heresy, set up an inquisition to deal with nonconformists and apostates, and (it seems) took stern measures to root out the cult of images, replacing this wherever possible with that of fires. Greek influence had been as strong, however, in Persia as in Parthia, and the rockcarvings of the Sasanians show that these monarchs made no objection to representation as such of the yazads of their faith.⁵⁷ Thus in the investiture-scenes of Ardašīr I Ohrmazd is portrayed as a noble bearded figure in Persian dress, with turreted crown and barsom-rods in his left hand;⁵⁸ and Varahrān appears, as in Parthian days, as a naked Herakles with club.⁵⁹ In the rock-carvings of later kings are shown Mihr with radiate crown,⁶⁰ and Anāhīd holding a tilted jug from which the waters flow.⁶¹ Such representations were set even on the walls of fire temples, in painting, or stucco in high relief;⁶² for portrayals of this kind evidently did not offend Zoroastrian iconoclasts as did free-standing images. Presumably, not being fully-fashioned forms or objects of cult, they were not regarded as potential homes for dēvs, nor yet as presumptuous imitations of the works of God.⁶³

The indications are, however, that, with regard to free-standing images, the Sasanians were active iconoclasts before ever they rose to imperial power. The family had the hereditary care of a great temple to Anāhīd at Istakhr (a town lying between Achaemenian Pasargadae and Persepolis, which was to become their dynastic capital). This temple was probably an ancient foundation, and evidently it housed originally a statue to the yazad, for the Muslim historian Mas'ūdī learnt in the 9th century a.d. that it had once been an ‘idol temple’, but that the idols had been removed and fire installed in their place.⁶⁴ Tabari (who drew on Sasanian sources) calls the Istakhr temple ‘the temple of the fire of Anāhīd’;⁶⁵ and the destruction of the statue there probably took place before Ardašīr seized power, or at the latest during the first decades of his dynasty's rule, for his grand-daughter, the Queen of queens of his son Šābuhr I, bore the name Ādur-Anāhīd ‘Anāhīd of the Fire’,⁶⁶ being named evidently for the patron yazad of the family. It is quite possible that the Istakhr image was the first of Anāhīd's statues to be overthrown; and thereafter, it seems, the divinity as she was venerated at this shrine was known by the distinctive appellation ‘Anāhīd of the Fire’. (Other shrines to her were naturally made by springs and streams, and the name Āb-Nāhīd, ‘(A)nāhīd of the Water’ is still commonly given to girls by Zoroastrians of the Yazdi area, in the north of Pars.⁶⁷) It is unlikely that it will ever be possible to date at all closely the establishment of the Istakhr fire; but the probability seems that in general the active phase of Zoroastrian iconoclasm had its beginnings in the first century a.d., at a time when Hellenic influences were waning and there appears to have been a stirring of orthodox zeal in Parthian overlord and Persian vassal alike.

That Ardašir's forbears were already convinced iconoclasts is suggested also by the fact that this king is known to have begun the destruction of images during his campaigns of conquest—although again we are dependent for evidence on the border-land of Armenia. Here he is said, on mastering the country, to have shattered statues of the dead, and to have set a sacred fire in the temple of Ohrmazd at Pakaran⁶⁸ (presumably in place of the image there). Even at this relatively late period only scraps of evidence survive for tracing developments within Iran itself; and in the main we hear more about the positive encouragement given by the kings and their chief priests to the founding of sacred fires than we do of the overthrow of images. Throughout the Sasanian period, it seems, the propaganda was broadcast that ‘the Varahrān fire represents goodness, and images are its adversary’ (ātakhš ī warahrān wehīh, ud uzdēs pityārag);⁶⁹ and whereas, as we have seen, the image shrine was characterized as ‘the abode of dēvs’, the ‘house of fire’ was called ‘the abode of yazads’, and it was said that the divine beings gathered there thrice daily (at the times when the devout should say their prayers in the presence of sun or fire), leaving gifts of ‘virtue and righteousness’.⁷⁰ One of the most active founders of sacred fires during the early years of the dynasty was the priest Kirdēr, who held office during five reigns, and rose to great power and wealth. In his inscriptions he lays proud claim to founding many Varahrān fires, a work, he says, of benefit to Ohrmazd and the yazads, whereby water, fire and cattle were also deeply satisfied;⁷¹ and he further states that he had brought it about that ‘images were destroyed and the haunts of demons laid waste, and the place and abode of the yazads [i.e. fire sanctuaries] were established’ (uzdēs gugānīh ud gilistag ī dēwān wišōbīh ud yazdān gāh ud nišēm āgīriy).⁷²

Despite the strength of the Sasanian monarchy, and the zeal from the outset of the Persian priests, many generations evidently lived and died before the long-established use of images was wholly suppressed. Thus of the two relevant cases recorded in the Sasanian law-book (the Mādigān ī Hazār Dādestān) one took place as late as the 6th century a.d., during the reign of Khosrau Anōširvān (531-579).⁷³ The two men concerned in it, named Kaka and Ādurtōhm, owned a piece of land in common on which they had ‘a house as an image-shrine’ (khānag pad uzdēsčār). The priests ordered the image to be removed from it, and set in its place an Ādurōg. This was a sacred fire of a minor grade, which could be tended by a layman with the same rites and respect as a household fire.⁷⁴ After this fire had burnt for a while in the former uzdēsčār (evidently to drive out the dēvs and purify the place) the Dīvān ī Kerdagān or Ministry of (Religious) Works was prepared to take it back into its own keeping—for plainly the maintenance of even the humblest fire, with its need for fuel and regular tending, was more costly than that of a statue, and could not safely be imposed on the unwilling. However, the two men petitioned to be allowed to keep the fire, undertaking to endow it with the land on which the image-shrine had stood; and they built it a fitting sanctuary, in which it was installed with due ceremony. In the other recorded case (which has no indication of date) a judge had had an image removed from an image-shrine (uzdēs kadag), and later a man other than the original owner of the shrine installed an Ādurōg there.⁷⁵ These cases show that a law must have been passed under the Sasanians forbidding the veneration of images, although there is no suggestion that those involved were punished except by the removal of the statue. When it was not possible (for expense or other reasons) to replace it by a sacred fire there must have been a grave sense of loss for the worshippers, and a danger to faith; but the priestly iconoclasts stated firmly that ‘when the worship of images is ended, little departs with it of belief in the spiritual beings’ ([ka] uzdēs parastišnīh be absīhēd, mēnōg warrawišnīh andak abāg be šawēd).⁷⁶ The resulting patterns of public worship can still be seen today in Yazd and its surrounding villages. There each place has its fire temple or temples, and also shrines to individual yazads, notably Mihr, Bahrām (Vsrsthraghna), Tīr, Aštād and Srōš. These shrines are regularly visited by the devout, who go there to pray, to take solemn vows, or to make acts of contrition or thanksgiving to the divinity concerned; and on the feast day of the yazad a communal act of worship is performed. But these shrines are empty, except for a pillar on which fire is kindled on each separate occasion. Incense is burnt, candles are lit, and other offerings are made; but there is no icon now at whose feet to lay them, the yazad being once more present only as an invisible spirit, as in the early days of the faith—though the alien practice of building a shrine for him, a ^*bagin or ^*bašn, has thus persisted.⁷⁷

Although there is the law-case to show that the image cult had not been entirely eradicated everywhere even by the 6th century, nevertheless iconoclasm must have won its main victories long before then. Ardašīr I is known to have imposed a number of new measures on the Zoroastrians who had come under his rule, and to have waded through blood to enforce them;⁷⁸ and the likelihood is that the law against images was one of these, for the iconoclastic campaign must have been begun early in the Sasanian period for the image cult to be so thoroughly obliterated that hardly any reference to it was made by Muslim historians (several of whom were themselves Persians of Zoroastrian stock). When one of them does mention that once there were images in Zoroastrian shrines, it is only to speak of this as something very remote and far-off.⁷⁹ A faint survival of the old icon cult seems to have persisted, however, in the Yazdī region, in connection with the worship of the much-loved Mihr; for in some Muslim villages there the face of the sun-god with radiate crown is still embroidered in traditional designs.⁸⁰ Since, however, the Arabs know the sun as female, the moon as male, this portrait is called that of Khorshīd Khānom, the ‘Lady Sun’. Such pictures would not, it is clear, have offended the Zoroastrian iconoclasts; but they are wholly in breach of later Muslim edicts on the subject, and thus illustrate the stubbornness of the devotion to icons, once their use had become thoroughly established in Persia.

Although the Zoroastrian iconoclasts were victorious in the end, their battle had evidently been hard-fought and prolonged, lasting over 800 years and probably indeed longer, for it is likely that much controversy preceded the edict by which Artaxerxes II imposed an image cult on the whole community. The iconophils had the initial advantage of support from the all-powerful throne; and the cult had a strong stimulus subsequently from the inspiration provided by Greek craftsmen, who created works of noble beauty. Yet fire is itself one of the most beautiful of icons, and to pray in its presence was to follow the example of the prophet, as well as to maintain an age-old tradition of Iranian worship—considerations which gave orthodoxy the strength to triumph in the end. During the long struggle which it had to wage, however, its theological weapons must have become well sharpened and its doctrines ever more clearly defined. Throughout the period of controversy Zoroastrianism was temporally immensely powerful, as the state religion of three successive empires; and it had moreover the authority conferred by lofty ethical teachings and a clearly defined dogmatic system.⁸¹ Its influence is already acknowledged in the transmission to other faiths of fundamental doctrines concerning the existence of God and the Evil One, the individual and last judgments, resurrection of the body, and life everlasting; and it would be strange if it had not contributed also to the debate on the propriety of making representations of the divine, which was a problem that exercised the minds of Greek philosophers and Jewish prophets, the early Buddhists, Christian priests and Muslim theologians. Controversy raged about this matter, in fact, over the whole area at whose centre lies Iran; and it is only the existence of so many blank pages in Zoroastrian history that has prevented the realisation that it was a burning issue for the Zoroastrians also.

The question of direct influence is naturally one which can only be approached with great caution; but it seems probable that it was exerted on at least two faiths, Christianity and Islam. In the case of the former, Armenia provided a channel for the transmission of Zoroastrian ideas. Under the Parthians, and governed by a cadet branch of the royal Arsacid family, this country had been predominantly Zoroastrian by profession, and the fact that the image cult was well established there means that the iconoclasts had abundant occasion to raise their voices, even if in vain. In a.d. 301, not long after the overthrow of the Arsacids, the Armenian king Tiridates III embraced Christianity (partly, it is thought, out of hostility to the Sasanian regime); and there ensued a general overthrowing of Zoroastrian images and a setting up of Christian ones instead. The Armenian Christian church never officially opposed the veneration of images;⁸² but it is very likely that earlier controversy on this matter (stimulated by Ardašīr's ruthless iconoclasm) continued among the Christians of the land,⁸³ whose links were now westward with Byzantium, where Christian iconoclasm was subsequently to spring into being.⁸⁴ Meantime Islam had been born, whose followers in its early years showed no aversion to representational art.⁸⁵ It was only in the 9th century a.d., when through massive conversions, Iran had come to play a leading part in the Muslim community, that Islamic doctrines took shape in this respect; and their bases—usurpation of the prerogative of the Creator, the wickedness of making shapes to be inhabited by evil powers—are precisely those which appear to have been established centuries earlier by Zoroastrian divines. The Muslim theologians carried their own laws to a logical extreme in forbidding any representations whatsoever. For the Zoroastrian, however, moderation is a virtue, and the Persians evidently kept their own iconoclasm, though strictly enforced, within well-defined bounds.⁸⁶ This limitation was also no doubt due to the fact that they, like the Christian iconoclasts, had a long tradition of the use of images to contend against, and could only win their war by fighting it intensively on a narrower front. Within the Iranian community it seems to have been the Persians who both instituted the image cult and finally brought it to an end; but during the intervening centuries it evidently affected the Zoroastrians far and wide.

Notes

1. This was argued forcefully by S. Wikander, Feuerpriester in Kleinasien und Iran, Lund 1946, 56 ff.; but he obscured a sound case by postulating that a temple cult of ever-burning fire had existed independently of Zoroastrianism and before that faith arose (a supposition unsupported by evidence); and that this cult was adopted into Zoroastrianism in the 4th century b.c. as a part of the worship of Arsdvī Sūrā Anāhitā. Since this divinity is a yazatā of water, the unlikelihood of such a supposition was apparent. [Note: the Avestan term yazata, fem. yazatā, Middle Iranian yazat/yazad, ‘being worthy of worship’ is kept throughout this article rather than being rendered by some imperfect equivalent which would obscure the characteristic Zoroastrian doctrine that all beneficent divine beings were created by Ahura Mazdā (who in the beginning alone was), in order to help and serve him in his task of redeeming the world. Having been created, they are to be worshipped in their own right, although always as subordinate to him. The Zoroastrian yazata is thus both more than an angel, and different in his station from the independent god of a pagan pantheon.]

2. See Strabo, XI.8.4.512.

3. See D. Stronach, ‘Urartian and Achaemenian tower temples’, JNES 26, 1967, 287; for a detailed account of the plinths see Stronach, Iran III, 1965, 24-27 with Pl. VII.

4. I.131.

5. See M. Roaf and D. Stronach, ‘Tepe Nūsh-i Jān, 1970: second interim report’, Iran XI, 1973, 132-38 with Pl. VI-VIII. The shallowness of the bowl in the altar top makes it impossible that ever-burning fire should have been maintained there, for this requires a deep bowl of hot ashes to sustain it. The excavators, although acknowledging this fact, nevertheless call the altar a ‘fire altar’, and the building containing it a ‘fire temple’.

6. These carvings are superbly reproduced by E. F. Schmidt, Persepolis III, Chicago 1970.

7. The so-called ^*Frātadāra temple by Persepolis, attributed by G. Widengren (Die Religionen Irans, Stuttgart 1965, 131, 358) to the Achaemenian period, is in fact later. See, with full bibliography, K. Schippmann, Die iranischen Feuerheiligtümer, Berlin 1971, 177-85.

8. Either āδokhš (the old Zoroastrian term), or kalak, a common Persian word for ‘brazier’.

9. See M. Dieulafoy, L’acropole de Suse, Paris 1893, 411 ff.; K. Erdmann, Das iranische Feuerheiligtum, Leipzig 1941, 15-16; Schippmann, op. cit., 266-74.—The term ‘Persia’ is used throughout the present article in its restricted meaning of Pārs (the present Iranian province of Fārs).

10. Berossus, fragment apud Agathias II.24, Dindorf, Historici graeci minores, II.221; Clemens Alexandrinus, Protrepticus, V. 65.3. On the form of the goddess' name there see Wikander, Feuerpriester, 61 n. 2.

11. See H. Lommel, ‘Anahita-Sarasvati’ Asiatica, Festschrift F. Weller, Leipzig 1954, 405-413.

12. On the force of the phrase εν πέρsαιs (in the citations from Berossus) see G. Hoffmann, Auszüge aus syrischen Akten persischer Märtyrer, Leipzig 1880, repr. Liechtenstein 1966, 137; Wikander, op. cit., 65. If a statue to the goddess were in fact erected at Persepolis itself, this would again pose a problem with the lack of identifiable temple ruins there; but perhaps it was the famous temple to Anāhitā at Istakhr nearby (a site still unexcavated) which was founded by Artaxerxes. On this temple see further below.

13. Yašt 5.126-8. It has been suggested that it was ‘Anāhitā’ with her eight-rayed crown who was represented on the coins of Demetrius I of Bactria, see P. Gardner, Catalogue of coins in the British Museum: Greek and Scythic kings of Bactria and India, 1886, Pl. III. 1; W. W. Tarn, The Greeks in Bactria and India, 115, cf. 135.

14. Polybius X.27. On this incident see E. R. Bevan, The House of Seleucus, London 1902, repr. 1969, II 18.

15. From the root ‘show’, Skt. diś-, Av. daēs-, which occurs with the same preverb in Khotanese uysdīśś- ‘expound, declare’, see R. Emmerick, Saka Grammatical Studies, London 1968, 16. Since the existence of a Zoroastrian image-cult was not formerly recognized, it used to be held that Persian uzdēs meant ‘heathen idol’, and the word was accordingly understood to derive from the base daēs- ‘form, shape’, and was interpreted as ‘out-form’, i.e. ‘monstrous thing’. See P. Horn, Grundriss der neupersischen Etymologie, Strassburg 1893, 295; W. B. Henning, ZII, IX, 1933, 225.15; H. S. Nyberg, Hilfsbuch des Pehlevi, Uppsala 1931, II, 230.

16. Cf. his words in one of his own hymns, Yasna 43.9; ‘Then indeed at the gift of veneration to thy fire truly shall I think of righteousness (aša-) to the utmost of my power’.

17. Zoroastrian theological utterances survive only in works compiled in the Sasanian period; but these clearly had a long tradition behind them, going back in essentials to the teachings of the prophet himself, see H. Lommel, Die Religion Zarathustras nach dem Awesta dargestellt, Tübingen 1930, passim; H. S. Nyberg, Die Religionen des alten Iran, deutsch von H. H. Schaeder, Leipzig 1938, Ch. 8.

18. Dādestān ī dīnīg (ed. T. D. Anklesaria) Purs. XXX. 5; text with transl. by H. W. Bailey, Zoroastrian problems in the ninth-century books, Oxford 1943, 112. On mēnōg/gētīg see most recently S. Shaked, ‘The notions mēnōg and gētīg in the Pahlavi texts and their relation to eschatology’, Acta Orientalia XXXIII, 1971, 59-107.

19. See Zand ī Vohuman Yašt (ed. and transl. by B. T. Anklesaria) VII. 37 (where the text has the late form uzdēstčār for ‘image shrine’). On similar beliefs among neo-Platonists and Christians see E. Bevan, Holy Images, London 1940, 91-3. They are strongly held also by Muslims.

20. Contrast Pahl. Vd. I 9 with Iranian Bundahišn (ed. T. D. Anklesaria), 206.15.

21. The dating is that suggested by Wikander, op. cit., although his interpretation of the development is different.

22. XV.3.14. That Strabo made this distinction has been stressed by O. Reuther, A survey of Persian art (ed. A. U. Pope), I, 1938, 559; A. Godard, Athār-é Irān III, 1938, 19.

23. See Wikander, op. cit., 98, 219; E. Benveniste, JA, 1964, 57.

24. See W. B. Henning, BSOS, VIII, 1936, 583-5; BSOAS, XXVIII, 1965, 250 f.

25. See apud M.-L. Chaumont, JA, 1965, 174.

26. See H. Hübschmann, Armenische Grammatik, I 114.85.

27. See Henning, BSOAS, XII, 1948, 602 n. 3.

28. See H. W. Bailey, BSOAS, XIV, 1952, 420 f.

29. See F. C. Andreas-W. Henning, Mitteliranische Manichaica aus Chinesisch-Turkestan II (SPAW, Phil.-hist. Klass, VII, 1933), 311 (M 219 R 17-18).

30. See Henning, BSOS, VIII, 1936, 584-5 (M 5731 = T II D 117b V 11 ff.).

31. See S. der Nersessian, ‘Une apologie des images du septième siècle’, Byzantion, XVII, 1944-5, 63, and cf. Agathangelos, CIX.133 (V. Langlois, I 167).

32. Nersessian, art. cit., 64.

33. Pliny, Natural History, XXIII.4.24.

34. See the passages brought together by M.-L. Chaumont, JA, 1965, 167-81.

35. XV.3.15.

36. Polybius X.27 (see above, n. 14).

37. Parthian Stations, 6.

38. Agathangelos, CX.134 (Langlois, I 168).

39. See most recently W. Eilers, Semiramis (Sb. Österreichische Ak. der Wissenschaften, 274 Bd., 2 Abh.), Vienna 1971, 43-4.

40. See G. Dumézil, RHR, CXVII, 1938, 152-69; J. de Menasce, RHR, CXXXIII, 1948, 1-18; E. Benveniste, The Persian religion according to the chief Greek texts, Paris 1929, 64-6. The identity of the Iranian divinity of whom a wooden statue existed at a Cappadocian shrine, and to whom Strabo (XV. 3.14) refers as ‘Omanos’, remains doubtful. He is widely taken to be Vohu Manah, but this is by no means certain.

41. There are a number of traces of the worship of this popular Greek god in Iranian territory, and small terracotta figurines of him have been found in abundance in the ruins of Seleucia on the Tigris, see W. von Ingen, Figurines from Seleucia on the Tigris, Ann Arbor 1939, 106-8 with pl. XVIII.

42. For the Greek inscription see L. Robert, Gnomon, XXXV, 1963, 76. For knowledge of the (unfinished) Aramaic version I am indebted to the kindness of my colleague, Dr. A. D. H. Bivar. The Iranians still at this time used Aramaic for written records and documents, as under the Achaemenians.

43. For the statues see R. N. Frye, The heritage of Persia, London 1962, 156, with Pl. 68-71, and 87 (from Commagene). For the shrine at Masjed-i Suleiman in Khuzistan see R. Ghirshman, Comptes rendus de l’Académie des Inscriptions et Belles-Lettres, 1969, 493, Schippmann, Feuerheiligtümer, 249; and for that on Mt. Karafto in East Kurdistan Aurel Stein, Old Routes of Western Iran, London 1940, 324-46. Herakles is equated with Varahrān in the monument on Nimrud Dagh.

44. See Nersessian, Byzantion XVII, 75.

45. See M. H. Ananikian, Armenian Mythology, Boston 1925, Pl. III opp. p. 26.

46. Antiquities, XVIII.344.

47. See the reports by V. Masson, G. A. Pugachenkova and G. A. Koshelenko, detailed references apud G. Frumkin, Archaeology in Soviet Central Asia (Handbuch der Orientalistik VII. ed. J. E. van Lohuizen-de Leeuw), Leiden 1970, 144-6.

48. See, e.g., Moses Khorenaci, II.40 (Langlois II 101).

49. On this site, with references to earlier literature, see Schippmann, Feuerheiligtümer, 227 ff. As it is unique in character among known Iranian sanctuaries, there are naturally divergent opinions about the cult to which it was devoted.

50. See Aurel Stein, Old routes of Western Iran, 130-2.

51. See A. Godard. “Les statues parthes de Shami’, Athār-é Irā II, 1937, 285-305.

52. See Stein, op. cit., 154.

53. See Ananikian, Arm. Mythology, 94.

54. See Nersessian, Byzantion XVII, 61.

55. See the Tansar Nāme (Letter of Tansar), ed. M. Minovi, Tehran 1932, 22, transl. M. Boyce, Rome 1968, 47.

56. The absurd but often repeated statement that there was no such thing as Zoroastrian orthodoxy before the Sasanians is a tribute to the propaganda of the Sasanian priesthood, who to increase their own authority attributed confusion and ignorance to their predecessors. See in more detail Boyce, A history of Zoroastrianism (Handbuch der Orientalistik, I, ed. B. Spuler), Leiden, Vol. II, Ch. 3 (in the press).

57. See W. Hinz, Altiranische Funde und Forschungen, Berlin 1969, for admirable photographs of these carvings.

58. See ibid., p. 123 ff. with plates 57, 60.

59. See ibid., p. 123 with Plates 57, 59.

60. See, e.g., E. Herzfeld, Am Tor von Asien, Berlin 1920, Pl. XXIX; A. U. Pope (ed.), Survey of Persian art, Pl. 160b.

61. See Herzfeld, op. cit., 92-3, with Pl. XLIV; Pope, Survey, Pl. 160 a. Anāhīd is also represented in the investiture scene of Narseh at Naqš-i Rustam, see Pope, Survey, Pl. 157b.

62. See the description of wall-carvings in the ruins of the great fire temple at Istakhr given by Mas‘ūdī, Les Prairies d’Or § 1403 (ed. Ch. Pellat, Vol. II, Paris 1965). The walls of the ‘palace’ besides the fire temple on the Kūh-i Khwāja in Seistān were richly decorated with paintings, which included representations, in Hellenistic style, of divine beings; for descriptions and bibliography see Schippmann, Feuerheiligtümer, 57-70. Even more strikingly, fragments of human figures, life-size or a little larger, and in very high relief, belonging, it seems, to a stucco frieze, were found within what was probably the fire-sanctuary itself (room PD) at Takht-i Suleiman, see D. Huff, Iran IX, 1971, 181-182, and further Boyce, A History of Zoroastrianism, II. Ch. 4.

63. On the apparent anomaly of the setting up of royal cult-statues by the Sasanians themselves see Boyce, op. cit., II, Ch. 3.

64. Loc. cit. He states that ‘the queen Humāy, daughter of Bahman, son of Isfandiyār’ was responsible for removing the images and transforming the shrine into a fire temple; but this transposing of the event into the legendary past is naturally not to be taken seriously, for the Zoroastrians (who have no historical tradition) tend to connect anything remote in time with the Kayanian dynasty who helped the prophet establish the faith.

65. This is rendered by Nöldeke (Tabarī, 17) a little freely as ‘the fire temple of Anāhīd’.

66. See Šābuhr's Parthian inscription on the Ka‘ba-yi Zardušt, l. 18. (His marriage with his daughter was a highly meritorious one according to the ancient Zoroastrian law of xvaētvadatha.)

67. See J. S. Sorushian, Farhang-e Behdinān, Tehran 1956, 201 (under Āb-Nahīr).

68. Moses Khorenaci, II.77 (Langlois II 119).

69. Dīnkard, ed. D. M. Madan, Bombay 1911, 551.13-15. The term pityārag is a theological one, meaning something evil brought into being by Ahriman in deliberate opposition to something good created by Ohrmazd.—A number of Pahlavi passages concerning uzdēs were collected by A. V. W. Jackson, ‘Allusions in Pahlavi literature to the abomination of idol-worship’, Sir Jamsetjee Jejeebhoy Madressa Jubilee Vol., ed. J. J. Modi, Bombay 1914, 274-85.

70. Sāyest nē-šāyest (ed. F. M. Kotwal, Copenhagen 1969) XX.1.

71. Ka‘ba-yi Zardušt ll. 9, 10 (facsimile ed. by W. B. Henning, Corpus Inscriptionum Iranicarum, Part III, Vol. II, Portfolio III (London 1963), Pl. LXXII-LXXV; transcription and translation by Ph. Gignoux, JA, 1968, 394-5.

72. KKZl. 10; facsimile, Pl. LXXII, LXXIII; translation, Gignoux, loc. cit. On the verbal construction (with imperfect passives) see W. B. Henning, Handbuch der Orientalistik I (ed. B. Spuler), IV.1, 102.

73. Mādigān ī Hazār Dādestān, Part II (ed. T. D. Anklesaria) Bombay 1912, 37.2-8; transl. J. de Menasce, Feux et fondations pieuses dans le droit sassanide, Paris 1966, 25; Boyce, BSOAS, XXXI, 1968, 63-4 (both then assuming that the case involved infidel idol-worshippers).

74. See Boyce, ‘On the sacred fires of the Zoroastrians’, BSOAS XXXI, 1968, 52-68; ‘On the Zoroastrian temple cult of fire’, JAOS, (in the press).

75. MHD, Part II, ed. J. J. Modi, Poona 1901, 94.3-6; transl. Menasce, op. cit., 31; Boyce, BSOAS, XXXI, 64. The Middle Persian term uzdēs kadag, lit. ‘image house’, corresponds exactly to MP. ātakhš kadag (Pers. āteš kade) ‘fire house’, one of the standard names for a fire temple.

76. Dīnkard, ed. Madan, 553.16-17; cf. ibid., 551.17-19.

77. Partly, probably, to secure a measure of immunity for them from violation, the Zoroastrians have come to call these shrines by Muslim terms, i.e. Pīr-i Mihrīzed etc., or generally, ma‘bad. Although the fire temples have all been rebuilt, some very pleasingly, since the second half of the 19th century, as have certain much-loved mountain shrines (places of general pilgrimage), the village sanctuaries mostly remain as they were during the years of oppression, humble mud-brick buildings given their aura of sanctity only through the devotions of centuries which have been paid there. Their inconspicuousness has meant that their existence has hitherto been largely overlooked by non-Zoroastrians.

78. See Tansar Name, text 14, transl. 39.

79. E.g. Mas‘ūdī, loc. cit. (nn. 62, 64, above).

80. For knowledge of this I am indebted to the kindness of Khanom Ferangis Shahrokh, who has a striking collection of such old embroideries.

81. The writer finds it impossible to agree with those scholars who interpret this great faith as the product of compromise and confusion. On the contrary, the fundamental doctrines taught by Zoroaster appear to have been maintained with admirable strictness by his followers down to the 19th century, when the sudden impact of European ideas and modern science had a cataclysmic effect on their theology.

82. See S. der Nersessian, Byzantion XVII, 67, with n. 37.

83. Later in the 4th century there was some return to Zoroastrianism, and Faustus of Byzantium (transl. Langlois I 295) relates that ‘a number of statues … were erected which were openly venerated’.

84. For some recent studies in this field see G. B. Ladner, ‘The concept of the image in the Greek fathers and the Byzantine iconoclastic controversy’, Dumbarton Oaks Papers, VII, 1953, 1-33; E. Kitzinger, ‘The cult of images before iconoclasm’, ibid., VIII, 1954, 85-150; M. V. Anastos, ‘The ethical theory of images formulated by the iconoclasts’, ibid., 155-160; A. Grabar, L’iconoclasme byzantin, dossier archéologique, Paris 1957. (I am grateful to Miss Helen Potamianos for kindly drawing my attention to these.)

85. See E. C. Dodd, ‘The image of the word; notes on the religious iconography of Islam’, Berytus XVIII, 1969, 35-61.

86. In general it was sculpture in the round which roused the wrath of iconoclasts; but whereas the Zoroastrians showed toleration for religious carvings in high relief, Christians after them were opposed to these also (see Bevan, Holy Images, 148).