The Sacred Fire

In the old Iranian cosmology fire was reckoned the seventh creation, present as a life-force within the other six, "distributed in all", and so animating the whole world^[1]; Zoroaster, Boyce writes, deepened this inheritance by apprehending fire as the creation of Aša Vahišta, "Best Righteousness", and as the instrument of God's judgement at the last day, when a flood of molten metal would test all mankind.^[2] Fire thus carried a moral charge no other element did. It was bound up with truth by an ancient judicial practice: one accused of lying or of breaking a contract could be put to an ordeal by fire (passing through flame, or having molten metal poured on the bare breast) and was held innocent only if he survived; the mildest form had him swear an oath while drinking a draught of sulphur (whence New Persian sowgand xordan, "to take an oath", literally "to consume sulphur"), a fiery substance thought to burn a perjurer inwardly. "Fire thus acquired an association with truth," Boyce concludes, "and hence with aša."^[2][1]

The prophet made the daily discipline of the faith turn on it. His followers were to pray always in the presence of fire: a terrestrial fire, hearth or ritual, or else the sun or moon on high. The Gāthās state the ideal in Zoroaster's own voice, and Boyce quotes the line:

"At the offering made in reverence (to fire) I shall think of truth (aša) to the utmost of my power." (Gāthās, Yasna 43.9, cited in Boyce, "Ātaš")^[3]

Fire, then, is not adored as a deity; it is the pure thing before which one turns the mind toward truth. That distinction (icon, not idol) is one the Zoroastrians themselves have always insisted on, repudiating the outsiders' label of "fire-worshippers" on the ground that the flame simply helps fix the thoughts on God and aša.

For the Achaemenid period the plainest witness to how the fire and the gods were approached is Herodotus, writing in the mid fifth century, and his report is of an emphatic absence of temples:

"The Persians have no images of the gods, no temples nor altars, and consider the use of them a sign of folly." (Herodotus, Histories 1.131, trans. Rawlinson)^[4]

Instead, he says, they go up to the tops of the mountains and there sacrifice, calling the whole vault of heaven Zeus, and honouring the sun and moon, the earth, fire, water and the winds. Nor could the rite be performed without a priest:

"A Magian man stands by them and chants over them a theogony (for of this nature they say that their incantation is), seeing that without a Magian it is not lawful for them to make sacrifices." (Herodotus, Histories 1.132, trans. Macaulay)^[4]

Herodotus knew that the Persians venerated fire greatly (he mentions it again at 3.16, where Cambyses' wish to scourge the embalmed body of Amasis offends both Persian and Egyptian custom)^[5], yet, as Boyce stresses, he does not single them out as fire-worshippers in any remarkable way, and he knows of no temples of any kind among them. Four centuries later Strabo, drawing on Hellenistic sources, gives the same open-air picture:

"Now the Persians do not erect statues or altars, but offer sacrifice on a high place, regarding the heavens as Zeus; and they also worship Helius, whom they call Mithras, and Selene and Aphrodite, and fire and earth and winds and water." (Strabo, Geography 15.3.13, trans. Jones)^[6]

Darius's own inscriptions mention āyadanā, "places of worship", which the usurper Gaumāta is said to have destroyed and which Darius restored (Behistun, DB §14). But archaeologists have found no remains that can be read as early-Achaemenid temples, and Boyce takes the āyadana to have been simply sacred places of regular resort, which fits Herodotus exactly.^[7][8] The worship was of the open air; the fire it centred on was tended not behind walls but on a stand raised where it could be seen.

The exact word is essential here, because the single object most often misnamed in accounts of Persian religion is the stepped stone stand on which the fire was set. Western scholars have habitually called it a "fire-altar", but Boyce insists the word is a misnomer that blurs its real significance.^[9][8] Many ancient faiths had altars on which a fire was kindled to consume an offering; the Zoroastrian fire-holder (Pahlavi ātaš-dān) is a different thing: not a hearth for burning up a sacrifice, but a raised repository designed to hold and sustain a wood-fire, its offerings made to the fire rather than on it.

The diagnostic feature is the top. A true fire-holder has a hollow top, cut to carry the deep bed of hot ash a continually burning wood-fire needs. Fragments of two or three such holders, finely wrought in white stone, were found at Pasargadae and are assigned on technical grounds to the time of Cyrus the Great (sixth century BCE): a three-stepped top and a three-stepped base joined by a slender rectangular shaft, the top hollowed for the ash.^[10][8] These, Boyce writes, are the first of a long line of such objects, and it is possible that one of them held the Great King's own hearth-fire, exalted into a dynastic fire that burned as long as he reigned.

Two contrasts sharpen the point. First, the pair of tall plinths that still stand on the Pasargadae plain, often lumped in with the fire-holders, are not ātaš-dāns at all: they lack a deep fire-bowl, and excavation showed they always stood unsheltered in the open, so they cannot have served a cult of ever-burning fire. What they were instead is unsettled: Pierre Amiet, re-examining Stronach’s plans, argues that the two cubes are not even a matched pair, since they were built one after another and do not align, and proposes that each was the stone base of a tower whose upper storey of perishable timber has vanished, a smaller forerunner of the Zendan at Pasargadae and the Ka’ba-ye Zardosht at Naqsh-e Rostam rather than an altar of any kind.^[11] On no reading are they fire-holders. Second, an object that looks like a prototype of the form was excavated in a Median building of about the eighth century at Tepe Nush-e Jan near Hamadān: a mud-brick stand, plastered white to look like stone, with a stepped top. But its top holds only a shallow bowl, charred round the rim, too small to sustain an ever-burning wood-fire; "what rites this Median altar served remain unknown."^[9] The shallow-topped Median stand and the hollow-topped Persian holder are two different things, and only the second is a Zoroastrian fire-holder.

The tombs and the Pasargadae holders are monumental and few. A far larger and humbler body of evidence has come into focus only recently: the seal impressions on the Persepolis administrative tablets. The Fortification archive (509–493 BCE) and the Treasury archive (492–457 BCE) preserve dozens of small scenes in which a figure stands before a stepped stand carrying fire, and Mark Garrison’s catalogue of them assembles seventy-four seals bearing eighty-five examples of what the literature has long called “fire altars”, most published for the first time.^[12] They fall into two forms. One is the familiar stepped stand, a pillar or conical base rising to a two- or three-stepped top; the other, which Garrison calls the “tower structure”, is a rectangular pillar with recessed panels and a crenellated or V-shaped crown. Only the first ever carries a visible fire.

The archive shifts the question from the monuments to the everyday practice of the heartland, and Garrison’s reading of it cuts against Boyce at the decisive point. He interprets the scenes alongside the Elamite texts of the same archive, which name a crowded pantheon of Elamite and Iranian gods among whom Auramazdā holds no privileged place, and he prefers “Mazdaism”, the worship of Auramazdā, to the loaded “Zoroastrianism”.^[12] Where Boyce made the fire-holder’s flame a pure icon, kindled solely as a focus for prayer and never a hearth for an offering, the seal scenes show animals led to slaughter and liquids poured onto the flame, and Garrison concludes that the fire here is not iconic at all:

“The fires on the stepped structures do not appear to be iconic; on the contrary, the fires seem quite functional, a conduit for the consumption of liquid and animal sacrifices intended for a deity.” (Garrison, The Ritual Landscape at Persepolis)^[12]

If he is right, the very object Boyce set apart from the ordinary altar was, in the early Achaemenid heartland, doing an altar’s ordinary work of receiving sacrifice. The deity to whom it was offered the seals almost never name, the one divine sign being a crescent on a single sealing, and the archive’s texts leave open any of a dozen recipients. The seals thus confirm the open-air stepped stand as a real fixture of early Persian cult while unsettling the confident reading of its fire as the pure Zoroastrian flame.

One seal stands apart. A chalcedony cylinder in Paris, cut with the Aramaic legend zrtwštrš, “Zoroastrian”, shows two priests in tiara and mouth-veil at a three-stepped stand beneath the winged figure; one holds the barsom, the other tips a ladle of fat or oil into the fire. Albert de Jong takes it for the one Achaemenid-era image of the ātaš-zōhr, the offering made to the fire, and it matches Strabo’s rite closely.^[13] But it is a single seal of uncertain date, and it cannot by itself fix what the many unlabelled scenes meant.

The fullest Achaemenid image of the rite is carved above the tomb of Darius at Naqsh-e Rostam, and reproduced over the tombs of his successors. The king stands on a stepped platform, one hand raised in reverence, before a fire-holder of exactly the Pasargadae profile (a three-stepped top and base, the shaft panelled) on which a pyramid of flames leaps up.^[14] Overhead floats the figure in the winged ring, long identified with Ahura Mazdā, a reading now strongly contested in favour of the royal khvarnah;^[15] the debate is set out at The Winged Symbol (the figure in the winged ring). On Darius's tomb it seems to carry a double sense, khvarnah and the sun, with the Akkadian crescent-moon disc set behind it. Since Zoroastrian custom lets prayer be said before a terrestrial fire or facing the sun or moon, Boyce reads the sculpture as showing the king at prayer "according to orthodox Zoroastrian prescriptions", before the fire and beneath the lights of heaven at once.^[8]

The fire-holder's form is remarkably stable. The Pasargadae holders of Cyrus's day, the ātaš-dāns on the tombs of Darius and every king after him, and later Parthian and Sasanian representations on seals, carvings and coins all share the stepped profile; the Sasanian kings would set a fire-holder with a royal diadem tied round its shaft on the reverse of their coins, the fire declared the king's own. The continuity of the image across the tombs is itself, for Boyce, an argument: the later Achaemenids are known to have been Zoroastrians, and a change of faith in any reign would surely have altered so charged a funerary scene, so the unbroken repetition tells for the Zoroastrianism of the whole line, a maximalist reading, and one other scholars resist (see Ahura Mazdā on the wider debate).

Behind the fire on the stand lay a fire of state. Boyce reconstructed a dynastic fire kindled at a king’s accession and burning as long as he reigned, the exalted hearth-fire of the Great King himself, and thought one of the Pasargadae holders might have carried it.^[8] Classical writers knew such a fire as perpetual: Quintus Curtius calls the fire the Magi tended before the king aeternum, “eternal”.^[16] Its clearest trace is a death-rite. When Hephaestion died, Diodorus reports, Alexander ordered the sacred fires of Asia put out, a thing the Persians did only at the death of the king, so that the extinguished royal fire marked the close of a reign and was rekindled for the successor.^[16] Briant reads the episode as a genuine survival: the Achaemenid king had a fire bound to his person and quenched at his death.^[16] The dynastic fire is the one point at which the later temple institution and the early kingship touch. An ever-burning royal fire, tended by Magi, is credible for the Achaemenid court long before the graded temple fires that would one day house such flames were built.

The enclosed fire-temple (a roofed sanctuary housing a consecrated fire kept perpetually burning, tended by priests at the appointed hours) is the picture most people carry of Zoroastrian worship, and it is the wrong picture for Darius's day. Herodotus, Boyce notes, knows of no temples among the Persians; no certain remains of any fire-temple have been found from Achaemenid times; and the temple cult of fire "seems to have been first instituted in the later Achaemenid period."^[2] She links its rise to a specific innovation: the introduction under Artaxerxes II (405–359 BCE) of a temple image-cult of the goddess Anāhitā, to which the orthodox are thought to have answered by founding temples with no man-made image but a consecrated fire, "the only icon permissable for a true follower of Zoroaster", the fire being an image no idolater could pervert.^[8] The oldest undisputed remains of a Zoroastrian fire-temple are Parthian, on the Kuh-e Khwaja in Sistan; the great named fires and the whole apparatus of graded temple fires (Ātaš Bahrām, Ādurān, Dādgāh) belong to the Parthian and, above all, the Sasanian centuries. The temple cult was in essence an extension of the hearth-fire cult, the fire "enthroned" on its stand and still a wood-fire fed the traditional offerings five times a day, but as an institution it postdates the early kings.

The dating does not rest on Boyce alone. Albert de Jong, whose reading of the Greek and Latin evidence is guarded against projecting the later faith backwards, arrives at the same result from the sceptical side. Herodotus’ flat denial that the Persians had temples or altars, de Jong argues, is simply correct, for what Herodotus denies is the Greek bōmós, the sacrificial altar that carries a fire and burns the victim’s portions on it, and “nothing remotely resembling the bomos has been found so far”.^[13] The cult statue, and with it the temple, he places with the Anāhitā cult under Artaxerxes II, after Herodotus wrote; even later Greek authors, he notes, rarely call a Persian fire-temple a naós, since to a Greek a temple without a cult-image was scarcely a temple.^[13] Where Strabo does describe roofed pyraitheia, de Jong shows he means enclosed buildings with a fire kept burning under a heap of ash, and that the scene belongs to the Parthian-era communities of Cappadocia.^[13] The maximalist Boyce and the minimalist de Jong, at odds over method and continuity, converge on this: the enclosed temple fire is a later thing.

The anachronism has a documented false trail, and it shows how such errors enter the record. It was once argued, from the Persepolis Elamite tablets, that there were nineteen fire-temples in Pārs in the early Achaemenid period (W. Hinz and H. Koch). The claim rested on reading the tablet-word haturmakša as "fire-priest". But the contexts, as edited by R. T. Hallock, show the man at work as a secular middleman in food commodities; the nineteen fire-temples dissolve on a corrected reading, deductions Boyce calls "made in defiance of the evidence".^[17][8] The archive scholarship has since refined the verdict without reviving the temples: Wouter Henkelman upholds the priestly etymology behind the word (Old Persian ātṛvaxša-, "fire-fanner", an auxiliary fire-priest, the spelling objection answered), while confirming that most of its bearers appear in the tablets as administrators of cereals and fruit, only a handful as officiants.^[18] It is a caution that an institution's name in a much later source is no warrant for its existence in an earlier one.

The fire was served by the Magi, the hereditary priestly class that Greek writers associated especially with the Medes. Their office was to perform the sacrifices and tend the fire, to chant the sacred recitations that were the living memory of an unwritten Avesta, to read dreams and omens, and to officiate at the rites of death. Herodotus makes the priest indispensable to any offering (1.132, above); Strabo, describing the Persian fire-rite, gives the ritual purity in exact and vivid detail:

"To fire they offer sacrifice by adding dry wood without the bark and by placing fat on top of it; and then they pour oil upon it and light it below, not blowing with their breath, but fanning it; and those who blow the fire with their breath or put anything dead or filthy upon it are put to death." (Strabo, Geography 15.3.14, trans. Jones)^[6]

Every element there is a guard against defilement: barkless wood, no breath on the flame, nothing dead or unclean brought near it. Describing the fire-sanctuaries the Magi kept in Cappadocia in his own day, Strabo adds the detail of the priest's dress that later tradition names the padām, the mouth-and-nose veil worn lest a breath reach the fire, and the bundle of ritual rods, the barsom (see the barsom plaque):

"In the midst of these there is an altar, on which there is a large quantity of ashes and where the Magi keep the fire ever burning. And there, entering daily, they make incantations for about an hour, holding before the fire their bundles of rods and wearing round their heads high turbans of felt, which reach down over their cheeks far enough to cover their lips." (Strabo, Geography 15.3.15, trans. Jones)^[6]

This is a Parthian-era scene of a diaspora community under alien rule rather than a snapshot of Darius's Persis, and it describes a temple fire the early period had not yet built. Strabo is a witness to be handled with the same care as any late source. But the ritual grammar it preserves, the strict guarding of the flame's purity, reaches back to the reverence Herodotus already saw.

The reverence for fire was one face of a wider horror of pollution that governed how the Persians treated all the good creations: fire, water, earth. To befoul any of them was a sin, and Herodotus preserves the everyday scruples: the Persians will not spit or make water or wash their hands in a river, but hold running water in great honour. The gravest problem this created was the disposal of the dead, for a corpse was the archetype of pollution and could be given neither to the fire, which it would defile, nor, ideally, to the earth or water. The later, "classical" Zoroastrian solution was exposure: the body laid on a raised place to be stripped by birds and sun, the dried bones only then gathered away, so that no good creation was contaminated. Strabo records the practice already in its double form:

"They smear the bodies of the dead with wax before they bury them, though they do not bury the Magi but leave their bodies to be eaten by birds." (Strabo, Geography 15.3.20, trans. Jones)^[6]

Here is one of the period's genuine puzzles, and it resists tidying. It is not clear why the Achaemenid kings built tombs at all, given how defilement of fire and earth was dreaded (Cyrus his gabled house at Pasargadae, Darius and his successors their chambers cut high in the cliff at Naqsh-e Rostam), sealing embalmed bodies in stone rather than exposing them. Boyce's answer is that this focuses too narrowly: the undoubtedly Zoroastrian Sasanian kings did the same, and Cyrus' sepulchre seems to have set a precedent of royal entombment carried out with the utmost care to keep the corpse from touching the good creations, stone being reckoned impermeable to pollution: the chamber thick-walled, raised on a stepped plinth well above the earth, with a stone floor and a double stone roof.^[8] The rock-cut royal tomb is a king's exception within the purity system rather than a breach of it; and the very tension between the kings' tombs and the exposure the faith prescribed is itself evidence that Achaemenid practice was not yet the classical Zoroastrianism of a thousand years later. The fire the king honours on his tomb-relief, and the sealed chamber beneath it, belong to the same creed of purity, and to a stage of it still short of the dakhma.

Gathered together, the strands make the fire's meaning plain. It was the purest of the visible creations and the icon of the truth the whole Persian moral order turned on: the arta the king swore by, the aša against the Lie. What set the Achaemenid fire apart was its bond with the throne: the flame on the royal tomb, and the dynastic fire kindled at an accession and quenched at the king's death, made it the hinge between arta and kingship, the point where the order of the cosmos met the legitimacy of the Great King.