The Magi

The fullest ancient account is Herodotus', and he is explicit: the Magi were one of the six tribes of the Medes (1.101), a hereditary clan who held an influential place at the Median court as interpreters of dreams and readers of signs (1.107).^[1] Many classical writers followed him in taking the Magi for a Median people; Strabo (15.3.1) calls them instead a Persian tribe. The word magu- itself has no agreed Indo-European etymology, and that gap has kept the question open. Some scholars supposed it a pre-Iranian element absorbed into Median society; Émile Benveniste argued instead that it signified a member of a particular social class already in proto-Iranian, and kept that sense into the Avesta, while Giuseppe Messina and Ernst Herzfeld held the Magi a class or estate rather than a tribe at all.^[2]

The strongest single argument that the title was met as something foreign and specific is scribal. When the Elamite and Babylonian clerks copied Darius's Behistun text, they found no word of their own for the Magus of the Old Persian and simply transliterated the title. De Jong, weighing the same evidence, adds a philological caution that cuts against the ethnic reading: the Avestan cognate of magu- appears to mean "(member of a) tribe", so that the word may have been confused with the name of an actual tribe, and nothing in the Greek or the Iranian record requires the Magi to have been Medes at all.^[3] Against this, the whole later tradition treats them as a caste defined by function and descent, not by a homeland; and their bearers in the documents carry Iranian names wherever the empire's Iranians were stationed, from Persis to Babylonia to Egypt.

Dandamayev, in the reference treatment, declines to settle it: whether the sixth-century Persians already had priests of their own whom they called Magi, and whether the term also covered any Mede who performed priestly functions and so carried no ethnic sense at all, cannot now be recovered.^[4] The clearest fact about them is a negative one. In Media and Persia alike the Magi were the only recognised body of priests, though the old Elamite cults kept their own; de Jong's summary is that the Magi "certainly were the main (possibly the only) religious specialists in the realm."^[3] Whatever they were by origin, by the reign of Darius they were the priesthood of the Iranian west.

The surest Achaemenid evidence for the Magi is also the least literary. In the Elamite Fortification tablets from Persepolis the maku appear repeatedly, mostly as recipients of rations, grain, flour, rams, wine, beer and fruit, drawn from the royal warehouses for the requirements of the cult and for their own keep. They are found at Persepolis and throughout south-western Iran, and their personal names, where given, are Iranian. Their functions here are concrete. One document (PF 1955) has the Magus Ukpiš (Henkelman's Upiš) receiving 360 quarts of barley in four equal shares, as offering for the ceremony called lan, the most widespread type of sacrifice in the whole archive, for the Mišebaka ("All Gods"), and for rites at a named mountain and a named river, a libation-service to peak and stream that matches exactly what the Greeks report.^[5][6] (An older reading took the second beneficiary for Mithra; the re-edited text gives the "All Gods".)

The titles the tablets carry alongside maku show a priesthood already differentiated by office. De Jong sets out the set the tablets seem to distinguish, a Magus (magu-), a fire-priest (ātrvaxša-), a high-priest (āθravapati-) and a specialist in the hymns (framazdā), while stressing that the interpretation of the Elamite spellings remains unsettled.^[3] The archive's re-edition has since firmed the readings. Henkelman reports Hallock's own later collation of the "memoriser" title as pirramadda, Old Persian framātā "master, commander", so that makuš pirramadda is a "chief magos" (Rüdiger Schmitt compares the later mōbed, from magu-pati), while the fire-title ātṛvaxša- "fire-fanner" stands, its bearers appearing in the tablets mostly as administrators of grain and fruit rather than at the altar.^[6] That last point is a genuine correction of the older picture: it was once argued, from these very tablets, that there had been nineteen fire-temples in early Achaemenid Persis, a claim that dissolved once the priest-title behind it was seen to belong to a functionary who mostly moved commodities (see the sacred fire).

The tablets also show the Magi were not only cultic functionaries. The same men handled and disbursed grain, so that the title magus can even appear as the personal name of a grain-handler. Beside them the archive names another word for priest, Elamite šatin, more common than maku, whose bearers serve Iranian gods (Ahura Mazdā among them) but sometimes the supreme Elamite god Humban as well, so that the two terms may have overlapped, and Iranian and Elamite priesthoods worked side by side in the one administration. Here, in the driest of records, is the firmest ground the subject affords: from at least the reign of Darius I the Magi were the salaried priests of the Achaemenid crown, an institution of the working state rather than a Greek traveller's marvel.

The classical writers fasten above all on one thing: the Magus was indispensable to the rite. Herodotus gives the description that fixed the image for antiquity, and marks the necessity in a single clause:

"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 1.132, trans. Macaulay)^[1]

The act of sacrifice was performed by the layman himself; the Magus was the officiant who made it lawful and sang over it the birth of the gods. The prayer at such a sacrifice, Herodotus stresses, was never for the self alone:

"For himself alone separately the man who sacrifices may not request good things in his prayer, but he prays that it may be well with all the Persians and with the king." (Herodotus 1.132, trans. Macaulay)^[1]

De Jong, reading the whole passage closely, isolates just these features as the ones that struck Herodotus: that the rite was directed not to the good of the individual but to the community and the king, that a Magus had to be present, and that the flesh was not burned for the god but simply carried away.^[3] The chant is the detail that explains the institution. An unwritten Avesta lived in the memory of the priests, and the offering was valid only when a trained voice sang the proper words over it; this is why the archive pays a class of "memorisers", and why the priesthood was hereditary, since the liturgy passed from father to son by heart. The layman could kill the beast, but only the Magus knew the words that made the killing worship. That monopoly of the word, more than any doctrine, is what set the Magi at the centre of Persian religion.

Before the Magi were the priests of Persia they were, in the Greek picture, the diviners of Media, and the interpreting of dreams and omens stayed one of their chief offices. Herodotus has them at the Median court expounding the dreams of Astyages that foretold the rise of Cyrus (1.107, 1.120), and reading the meaning of an eclipse and of monstrous births; when their reading failed, Astyages impaled the Magi who had advised him. The office followed the priesthood into the Achaemenid court. Xerxes, hesitating over the invasion of Greece, is made to consult the Magi on a dream and a solar omen, which they interpret to his liking (7.19, 7.37). The divinatory art is bound up with the same expertise as the liturgy: a class trained to carry and construe a sacred lore reads the signs of the world as it recites the hymns, and the two together make the Magus the specialist the king keeps at his side. The Greek writers, who had their own seers and dream-readers, found this readily intelligible, and it is one of the points on which their reports are least likely to be pure invention.

The Magi also formed the young. Xenophon, writing an idealised life of the elder Cyrus, makes them the priests of the Persian kings, into whose hands the ceremonies of libation, incantation and sacrifice were given, whose instruction the kings followed in religious matters, and who took part in the coronation of each new king (Cyropaedia 8.3.11–12). He fixes the founding of the office on Cyrus himself:

And then the Magi were first charged with singing hymns to all the gods with the coming of every day, and he sacrificed every day to the gods the Magi named. ... Thus what was established then still endures even now in the court of whoever may be king.

Xenophon, Cyropaedia 8.1.23–24, trans. Ambler^[7]

The attribution to Cyrus is Xenophon's construction, not history; but the picture of a permanent priestly office at the heart of the court, chanting the daily hymns and prescribing the king's own worship, agrees with everything the documents show. Plato's circle knew the same tradition in sharper form: the heir to the Persian throne, they held, was given at fourteen to four royal tutors, of whom the wisest taught the mageia of Zoroaster son of Oromazes, that is the worship of the gods, along with the arts of kingship (Alcibiades I 121e–122a). The Magi, on this account, were the priests of the realm and also the guardians of the prince's education in religion and rule, a role that made them a standing presence beside the throne without ever making them a power that could command it.

The priesthood went to war with the king. The sacred fire was borne before the troops: Curtius Rufus describes Persian soldiers carrying the holy flame on silver altars at the head of the march, the Magi following behind and singing their ancient hymns (Historiae 3.3.9). On campaign the Magi did the work of divination and the management of the elements that the crossing of a hostile country required. As Xerxes' host moved into Europe, Herodotus has its Magi conduct a sequence of sacrifices aimed at the powers of the land it passed through: libations to the heroes at Ilium (7.43), and at the river Strymon a rite for good omens on the eve of the crossing:

"by that water the Magi slew white horses, offering thus sacrifice for good omens." (Herodotus 7.113, trans. Godley)^[8]

The white horse, the beast of the sun and of the gods, was the costliest of offerings; the crossing at the Edonian town of Nine Ways followed. The most vivid of these notices has the Magi turn their art against a storm off the Magnesian coast that was wrecking the fleet:

For the storm lasted for three days; and at last the Magians, by using victims and wizards' spells on the wind, and by sacrificing also to Thetis and the Nereids, did make it to cease on the fourth day, or mayhap it was not of their doing but of itself that it abated.

Herodotus 7.191, trans. Godley^[8]

Herodotus's dry "or mayhap it was of itself" is his own scepticism, not the Persians'; what the passage preserves is a real function of the priesthood, the ritual command of wind and water on which an army in the field depended, and the offerings to river, sea and heaven that the Fortification tablets record in their bare administrative form. The Magi of the campaign narratives and the Magi of the ration-lists are the same men, seen once from the Greek deck and once from the Persian warehouse.

Two practices caught the foreign eye above all, and both belong to the purity-religion the Magi kept. The first is the treatment of the dead. Herodotus reports that the Magi did not bury a body until it had been torn by a bird of prey or a dog, and that they, unlike other men, killed with their own hands everything that crawls and flies (1.140), a habit that fits the later Zoroastrian war on the xrafstra, the noxious creatures of the Hostile Spirit. Strabo makes the exposure of the priests' own bodies explicit:

"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 15.3.20, trans. Jones)^[9]

The logic is the purity of the good creations: a corpse must not be allowed to defile earth or fire or water, and so the Magi's own dead were given not to the ground but to the birds, the rite that later Zoroastrianism would formalise in the dakhma. The same scruple governs the smallest daily habit:

"Into a river they neither make water nor spit, neither do they wash their hands in it, nor allow any other to do these things, but they reverence rivers very greatly." (Herodotus 1.138, trans. Macaulay)^[1]

Boyce reads Herodotus' account of Persian practice in his own day (1.131f.) as the earliest description of the Zoroastrian purity laws in action, the killing of noxious creatures, the avoidance of polluting water, the exposure of the dead, and holds that the open-air sacrifices he reports accord closely with living Zoroastrian lay usage.^[10] Whether that continuity reaches all the way back to Zoroaster is the disputed matter (below); what is not disputed is that these were the observances the Magi enforced, and that they are already, in Herodotus, a coherent creed of purity and not a bundle of curiosities.

From the priests' own title comes a word with an extraordinary afterlife. Through the Greeks magos gave mageia, the arts of the Magi, and the word split. It kept its dignified sense, a wise man of the East, master of the gods' worship and of the stars; but it also acquired a contemptuous one, and was used for conjurers, sorcerers and quacks. The two senses did not arrive in that order. De Jong shows the derogatory sense was in fact old, present already in the fifth-century tragedians and perhaps earlier still in a fragment attributed to Heraclitus, of about 500 BCE, whose authenticity is debated, so that "Persian priest" and "magician" ran together in Greek almost from the first, and their entanglement is the central difficulty in reading any classical notice of the Magi: one can rarely be sure which the writer means.^[3] To Sophocles' Oedipus, calling Teiresias a scheming magos, the word is already an insult; to Plato it can name the true religion the Persian prince is taught (Alcibiades I 122a) or the fraud of a wandering beggar-priest. That double meaning is the root of the English "magic" and "magician".

Two further confusions thickened the Greek image. The first is the Greek habit of running the Magi together with the "Chaldeans", the Babylonian priest-astrologers; Dandamayev notes that the similarity of some of their activities made the two easy to conflate, and the association of the Magi with astrology and occult craft is largely a Hellenistic overlay on the Iranian priesthood.^[4] The second is the ethnographic report, going back to Xanthus of Lydia, among the earliest Greek writers to name the Magi, that they practised next-of-kin marriage, taking their mothers, sisters and daughters to wife. De Jong treats this as a real and central datum of the Greek tradition rather than mere slander: close-kin union (xwēdōdah) was, in the later Zoroastrian tradition, an honoured institution, and the Greek reports, however luridly framed, register a genuine Iranian practice they found monstrous.^[3] The Greek "Magus" is thus a composite: a true memory of a real priesthood, a garbling of it with Babylonian astrology, and a projection onto it of secret and forbidden arts. Disentangling the three is the first task of any modern account, and the reason the classical evidence must be handled with such care.

The dignified sense of the word had one afterlife more durable than any. In the opening of Matthew's gospel magoi come "from the east" to Jerusalem, following a star, and lay before the infant Jesus gold, frankincense and myrrh (Matthew 2:1–12).^[11] The evangelist uses the plain Greek word for the Persian priests, and the sense he intends is the honoured one: learned men of the East, readers of the heavens, whose recognition of a new king carries weight precisely because they are the wisdom of the great empires. Later Christian tradition would make them three, name them, and crown them kings, but the text says only magoi, and in that word the old Iranian priest-astrologer, stripped of the Greek suspicion of sorcery, entered the central story of another religion. Rendered in English as the "wise men" or, keeping the Greek, the Magi, they are the reason the word survives in common speech at all with its dignity intact, beside the "magic" that carries its disrepute.

The documents show the Magi wherever the empire's Iranians went. In Babylonia, where Persian and Median officials and soldiers were stationed, Magi came to perform the rites, and the local records name them, always with Iranian names, in administrative and legal capacities: a Magus checking a temple's flour-store at Uruk, another supervising workmen, a Magus named Zattumšu holding a field near Kish beside the lands of Marduk's own temple, a Magus among the witnesses to a business deal. In Egypt, an Aramaic deed of 434 BCE from the Jewish military colony at Elephantine lists two Magian priests, Mithrasarah and Tatt, among the witnesses when a Jewish colonist gave his wife half of his house.^[12] Their presence at Persepolis is confirmed in Aramaic as well as Elamite, one Aramaic docket on a Fortification tablet naming the mgw. Later, Arrian records that Magi were appointed to guard the tomb of Cyrus the Great at Pasargadae and to sacrifice a horse there each month (Anabasis 6.29), an endowed cult of the royal dead that, as Boyce notes, was maintained until Alexander came.^[10] The priesthood was an institution of the whole empire and no local curiosity, and most scholars read that portability, a class of specialists who could carry the cult to a garrison in Egypt or a temple-town in Babylonia, as the mark of a caste of function rather than a people of one land.

The Magi enter Achaemenid political history at its most violent hinge, and the episode is the single Old Persian attestation of the word. In 522 BCE, by Darius's account, a Magus named Gaumāta claimed to be Bardiya son of Cyrus, seized the royal power, and was overthrown by Darius and his six confederates. In the Babylonian version of the Behistun text he is called "a Mede, the Magu"; in the Elamite, again the title is simply transliterated.^[13] Herodotus tells the same story with the usurper under the name Smerdis, a Magus whom Cambyses had made steward of his household and who impersonated the murdered prince (3.61). Whether Gaumāta was a real pretender or a fiction with which Darius covered a coup against the true Bardiya is one of the deepest problems of the reign, and it is a problem about Darius rather than about the priesthood (see Bardiya and Gaumāta and the Accession of Darius).

What the episode shows about the Magi is narrower and firmer than the identity question, and it runs against the obvious reading. The Greek tradition holds that the killing of the pretender was remembered in an annual festival, the Magophonia, the "killing of the Magi", on which (Herodotus 3.79) no Magus might show himself abroad; the privileges the victorious Seven claimed for their houses, unannounced entry to the royal presence and the right to furnish the king's wives, were likewise made law (see the Seven).^[14] Yet whatever a Greek festival-name seems to imply, the episode does not mark a Persian hostility to the Magi as such. The priesthood continued as the crown's own priests without a break; a Magus who claimed the throne was an offence against the king, not the Magi against Persia. Under Xerxes, Dandamayev observes, the influence of the priesthood grew rather than shrank.^[4] The one Magus a Persian king ever condemned by name he condemned for reaching above his station, and the class he belonged to lost nothing by it.

The hardest question about the Magi is their relation to Zoroaster, and it is genuinely unsettled. The Greek writers make it reasonably clear that the later Magi were Zoroastrians: Aristotle, quoted by Diogenes Laertius, reports that the Magi taught two first principles, a good spirit called Zeus or Oromasdes and an evil one, Hades or Areimanios, the earliest outside statement of the dualism; and a chief Magus, Ostanes, travelled with Xerxes as a figure Greek tradition knew as a Zoroastrian high priest. But the religion of the earlier Magi is fought over. Jacques Duchesne-Guillemin thought it unlikely they followed Zoroaster at the outset, and held the supreme god of the Medes to have been Mithra, not Ahura Mazdā. Ilya Gershevitch made the sharpest case for a distinction, taking the early Magi not as adherents of any one religion but as professional priests, technical experts who served the cult of any Iranian god for pay and adopted Zoroaster as their prophet only in the first half of the fourth century BCE, becoming thereby the authors of the syncretism later called Mazdaism.^[2] R. C. Zaehner supposed rather that they took Zoroaster's teaching and bent it to the older popular notions of Iran.

On the other side de Jong, surveying the same Greek evidence, judges that the customs Herodotus ascribes to the Magi, the exposure of the dead and the killing of noxious creatures, are typical of evolved Zoroastrianism, so that the Magi of that account "are ... best regarded as Zoroastrian priests", even while other passages may preserve traces of non-Zoroastrian Magi whose religion can no longer be recovered.^[3] Mary Boyce pressed the strongest continuity case, that the early Achaemenids and their priests were already Zoroastrians in the broad old-Iranian mould from which the reform grew (see Ahura Mazdā and Zarathustra). Dandamayev's own conclusion is a deliberate agnosticism: it is as difficult to say when and how far the Magi adopted Zoroastrianism as it is to say whether the Achaemenid kings were Zoroastrians at all.^[4] In the polytheistic world of the time there were no fixed dogmas, and it was natural, he adds, that the Magi absorbed other religious ideas alongside the teaching of Zoroaster and became the makers of eclectic and syncretistic tendencies. That uncertainty has to stand. The Magi were the Mazdā-worshipping priests of the western Iranians; the precise shape of their creed in Darius's day is one of the things the evidence does not tell.

The texts describe the Magi; the material record shows them, though sparingly. On the Persepolis tablets a class of seal impressions depicts two priests flanking a fire-altar beneath the winged symbol, with the mortar and pestle of the haoma rite on a table between them, a scene Erich Schmidt read as the haoma ceremony itself, the priest at the mortar the havanana.^[15] The fire-holder they attend is the free-standing stepped stand of the open-air cult, and the bundle of rods the Iranian priest carried in worship, the barsom, is given a face in the same Achaemenid present by the gold votive plaques of the Oxus Treasure, which show a figure in Median dress holding just such a sheaf of twigs. Beyond Persis, a relief of the fifth century BCE from Daskyleion in Asia Minor shows two Magi at a fire-altar in tunic and trousers, their noses and mouths covered by the ritual mouth-veil (paitidāna) that kept the human breath from defiling the flame.^[4] The furniture of the rite, the stepped fire-holder, the barsom, the mouth-veil, is continuous from these Achaemenid images through the much later scene Strabo describes among the Magi of Cappadocia, priests making their incantations before an ever-burning fire, rods in hand and felt drawn over their lips.^[9] Strabo's is a Hellenistic scene that cannot be read straight back into Darius's Persis, and the roofed sanctuary with its ever-burning fire belongs to a later stage of the religion than the open-air worship of the early empire (see the sacred fire). But the ritual grammar, the guarded flame, the rods, the veiled mouth, reaches back to what the reliefs and the tablets already attest.

When the Achaemenid empire fell the priesthood endured and spread, to Cappadocia where Strabo found them, to Central Asia, and on into the Parthian and Sasanian ages. At Nisa a Magus signs for great quantities of wine in the state stores; under the Sasanians the Magi became a powerful estate, exempt from the head-tax, their chief priest bearing the title mowbed, from magu-pati, "chief of the Magi". In the last quarter of the third century CE the Magus Kirdēr rose to that office and pressed a militant Zoroastrian orthodoxy as the state church, leaving four rock inscriptions of his own. With them the priesthood at last speaks in the first person, setting down its own deeds and its own creed, seven centuries after Gaumāta and the first appearance of the word on the Behistun cliff. The silence of the Achaemenid Magi is thus not permanent, only early: the institution that served Cyrus and Darius is the same that produced Kirdēr, carried down the centuries by descent and by the words its members kept in memory.