Anāhitā

The full Avestan name is a string of three words that later worked as a title: arədvī sūrā anāhitā, rendered by the philologists as 'the moist, the strong, the undefiled', or 'the high, powerful, undefiled'.^[1] Only the last of the three, anāhitā, survived as her ordinary name, passing into the hellenised Anaïtis and the Middle and New Persian Anāhīd/Nāhīd, which remains the ordinary Persian word for the planet Venus. The word is a negation, an- prefixed to a participle āhita-, 'stained, defiled', so that the name means 'the undefiled' or, in Kent's blunt gloss, 'the Spotless'.^[2] Albert de Jong sets the three epithets out and notes how they 'continued to function as her name', the goddess being known in Pahlavi as either Ardwisur or Nāhīd.^[3]

Behind the Avestan title lies a reconstructed Old Iranian divine name, *Anāhiti, a short-i feminine that is not quite the Avestan Anāhitā and whence the Greek Anaïtis directly comes. William Malandra reads the very mismatch of the two forms as a philological trace of the fusion the rest of the evidence shows: either two distinct goddesses have been run together in the Avesta, or the goddess entered the Zoroastrian canon late enough to leave the joining still visible in her declension.^[2] Behind that, on Mary Boyce's account, stands an older figure still. She traced the water-goddess back to the Indo-Iranian personification of the world-river, cognate with the Indian Sarasvatī and with a reconstructed Iranian *Harahvatī, 'the great mythical river which flows down from high Harā into the sea Vouru-Kaša', whose proper name was gradually eclipsed by her cult-epithets until Arədvī and Sūrā served in its place.^[4] On this reading the water-and-fertility Anāhitā is genuinely ancient, a goddess the Iranians carried with them out of their shared prehistory with the Indians.

The description is of water, and specifically of one great mythical water. She is 'the large river, known afar', as broad as all the waters that flow upon the earth, that runs down with force from the height of the mountain Hukairya into the world-encircling sea Vouru-Kaša. From that one river, the hymn says, all the waters of the earth flow out. Because she is the water of life, she is the source of fertility, and the hymn makes the connection with an anatomical directness that later ages found startling. Ahura Mazdā himself, in the opening stanza, bids Zarathuštra sacrifice to her as the goddess:

Who makes the seed of all males pure, who makes the womb of all females pure for bringing forth, who makes all females bring forth in safety, who puts milk into the breasts of all females in the right measure and the right quality.

Ābān Yašt (Yašt 5.2), trans. Darmesteter 1883^[1]

This is the core of the old concept: a river-goddess who is by that very nature a goddess of procreation, of birth and of the nourishment of the young. The waters she governs are not a metaphor but the physical streams and springs of the earth, and the reverence paid to them, sacrifice made beside running water, was among the oldest and best-attested of Iranian observances.

Her great document is the fifth Yašt, the Ābān Yašt or 'Yašt of the Waters', the longest of the hymns after the Mihr Yašt, and one of the most valuable, because interwoven with its praise is a roll-call of the legendary heroes of Iran who sacrificed to the goddess and asked her help. Ahura Mazdā leads the list, sacrificing to her before the creation; then come the primordial kings and champions of the national legend, Haošyangha, Yima, Θraētaona, Kərəsāspa, the Turanian Frangrasyan (the Afrāsiyāb of later epic), Kavi Usan and Kavi Haosravah, down through Zarathuštra himself and his patron Kavi Vīštāspa. Each is granted his boon, or, in the case of the empire's enemies, refused it. The hymn is thus both a liturgy and a compendium of Iranian myth, and its heroes are the same figures who fill the later Persian tradition.

The goddess of the Yašt is no still idol. She is a driver of chariots, mistress of the storm-waters:

Who drives forwards on her chariot, holding the reins of the chariot. She goes, driving, on this chariot, longing for men and thinking thus in her heart.

Ābān Yašt 5.11, trans. Darmesteter 1883^[1]

The chariot is drawn by four white horses of a single colour and a single blood, which the hymn glosses as the wind, the rain, the cloud and the sleet, so that the goddess rides the weather itself. She is 'strong and bright, tall and beautiful of form'. This is a divinity of enormous natural power, elemental and swift, and the martial energy that later attached to her, the warrior-queen who grants victory in battle, sits easily enough on a being who commands the flood and the storm.

Yet the hymn is not of a piece, and its unevenness matters, because it carries much of the argument about who Anāhitā really was. Two things sit oddly in it. The first is that a good part of the text is not native to a water-goddess at all: de Jong observes that 'the main part of Yt. 5 ... is a compilation of texts from the hymn to Aši', that is, from Yašt 17, the hymn to Aši, the ancient goddess of Reward and Fortune, whose fertility-blessings and martial associations were transferred wholesale to Anāhitā as her western cult overtook Aši's eastern one.^[3] The Avestan hymn we have is, on this view, a made thing, an old goddess of the river dressed in the borrowed verses of a goddess of luck and battle.

The second is even more revealing. Near the end of the hymn, after all the elemental imagery, come verses that describe the goddess not in motion but standing still, in Boyce's rendering 'ever observed',^[4] arrayed like a great lady in a golden embroidered mantle, with golden earrings, a golden necklace, a golden crown of a hundred stars, and garments of the finest beaver-skins. She appears in these late stanzas:

Ardvi Sura Anahita, who stands carried forth in the shape of a maid, fair of body, most strong, tall-formed, high-girded, pure.

Ābān Yašt 5.126, trans. Darmesteter 1883^[1]

The detail of the costume is so circumstantial, and so unlike the wind-driven river of the opening, that scholars from Friedrich Windischmann in 1858 onwards have concluded that the writer was describing a statue, a consecrated cult-image standing robed and jewelled in a shrine.^[4] If that is right, then these verses were composed to accompany the very image-cult that the classical sources credit to Artaxerxes II, and the hymn preserves within itself the moment when an aniconic Iranian goddess acquired a body in gold and beaver-fur. Boyce read the contrast exactly so: in time, she wrote, 'Anahita' became prominently a water divinity,

while Aredvi Sura gained the cult-title of 'the Lady', a characteristic Mesopotamian usage

Boyce, A History of Zoroastrianism II, 203^[4]

and developed the martial traits of a warrior-queen that a mere goddess of water and fertility need never have had. The single Avestan hymn, on this analysis, is a palimpsest of at least three figures: the Indo-Iranian river *Harahvatī, the fortune-goddess Aši whose verses were borrowed, and a western planetary-and-warrior goddess whose image and pomp were grafted on last.

Here the real difficulty begins. For all the antiquity that Boyce claimed for the water-goddess, Anāhitā is nowhere named in the royal inscriptions of the first Achaemenids. Darius and Xerxes invoke Ahura Mazdā over and over, and beyond him only 'the other gods who are', whom they never single out. Anāhitā is not among the named; and the daiva inscription of Xerxes, the one royal text that speaks directly of cult, orders the worship of Ahura Mazdā and says nothing of any goddess. For the first century and a half of the empire, the water-goddess is invisible in the record of the crown.

That silence has always seemed the single hardest fact about her, and it has usually been stated flatly: she is absent from the inscriptions, and absent too, on the published texts, from the Persepolis Fortification archive, the administrative record that does register offerings to a range of local and Iranian divinities.^[5] The absence looks total, and it is the fixed point against which every reconstruction of her cult has to work.

The archive silence, however, is less clean than it looks, and the qualification is a substantial one. The Fortification tablets do record offerings poured out at rivers and waters, several of them, alongside the sacrifices to named gods. What they do not do is say to whom the river offerings were made. Wouter Henkelman, whose study of the archive's Elamite-Iranian religion is the fullest, puts the problem carefully:

The question as to who is worshipped in the river offerings is not explicitly answered by either Strabo or the six Fortification texts just mentioned.

Henkelman, The Other Gods Who Are, 378^[6]

A literal offering 'to water', he notes, could honour a deified river, the Iranian tradition knew several, such as the Oxus, or the water-spirits protecting wells and streams, or a named goddess invoked beside them. The one river offering in the archive that does name its recipient names not her but the Elamite god Humban; Anāhitā appears only as a modern candidate for the anonymous remainder, and Henkelman flags even that with a question mark of his own.^[6] Yet she is precisely the kind of divinity the later Zoroastrian tradition addresses at running water, so she may well stand behind some of these offerings without ever being written down. The evidence does not name her; it also does not exclude her, and the emphatic Herodotean picture of a goddess wholly imported later has to be held against an administrative practice of water-worship already in place. The archive turns the confident 'she is absent' into a more exact and more useful 'she is not named'.

Then, under Artaxerxes II (r. 405/4–359/8 BCE), she surfaces, and surfaces suddenly, in a fixed formula. In several of his inscriptions from Susa the king sets Anāhitā and Mithra beside Ahura Mazdā, so that for the first time the anonymous 'other gods' are given names. The clearest instance is the trilingual A²Sa, cut on the column-bases of the rebuilt apadana at Susa. There the king records that the hall his forefather Darius had built was burnt in his grandfather's time, that by the favour of Auramazda, Anāhitā and Mithra he has raised it again, and prays for their protection:

May Auramazda, Anahita and Mithra protect me from all evil, and that which I have built, sorcery is not to destroy …

Artaxerxes II, A²Sa, at Susa; trans. Kuhrt 2007^[7]

Roland Kent's standard edition renders the same triad 'Ahuramazda, Anaitis, and Mithras', in the hellenised forms the classical world knew.^[7] The same three gods, invoked together, recur in A²Sd and A²Ha. The abruptness of this is the central fact of the subject, and the interpretations of it divide the field. On one reading the change is a genuine religious revolution: Artaxerxes II broke with the reticence of his ancestors and promoted two gods, one of them a novel image-cult, to the summit of royal religion. On the other reading the gods were old and popular all along, and what changed was only the register of the inscriptions, which now followed the naming habit of the Yašts rather than the older manner that addressed the great Ahura and left his companions unnamed. The inscriptions themselves do not settle it, and they are not even internally consistent: Ahura Mazdā keeps the first place, the formulas vary from text to text, and the sole surviving inscription of the next king, Artaxerxes III at Persepolis (A³Pa), invokes Ahura Mazdā and Mithra but drops Anāhitā altogether.^[7] Why she should vanish again so quickly, no one can say.

More recent scholarship has tended to read the debut with Briant's caution rather than Boyce's confidence. Maria Brosius states the minimal position plainly: the first appearance of the two gods' names in Artaxerxes II's texts 'allows no conclusions about the introduction of his cult, or that of Anahita, in his reign, but only the suggestion of a possible elevation of these two deities'.^[8] An epigraphic first mention is evidence for a change in what the crown chose to inscribe; whether what Persians worshipped had itself changed is another question.

Henkelman presses the scepticism further, and narrows the evidential base as he does so. There is, he holds, no conclusive evidence for a general cult of Anāhitā anywhere, the Achaemenid heartland included; and the royal epigraphy on which the whole reconstruction leans is thinner than it is usually made to look. She is securely named in only two texts, both from Susa (A²Sa and A²Sd); the Hamadan provenance of the third, A²Ha, is unproven and perhaps doubtful.^[9] The named goddess of the inscriptions, in other words, rests on two building-texts of one king at one city, which is a slender platform for an empire-wide innovation.

The one text that seems to explain the innovation is a fragment of the Babylonian priest Berossus (early third century BCE), preserved at second hand by the Christian writer Clement of Alexandria. After noting that the Persians, Medes and magi had once used no cult-statues at all, Berossus reports that Artaxerxes, son of Darius, was the first to set up an image of 'Aphrodite Anaïtis' in Babylon, and required her worship:

to require such worship from the Susians, Ecbatanians, Persians and Bactrians and from Damascus and Sardis.

Berossus, FGrH 680 F11, via Clement, Protrepticus 5; trans. Kuhrt 2007^[10]

Here is a royal edict, from an apparently official source, prescribing a statue-cult of the goddess across the great centres of the empire from the Babylonian heartland to the Aegean coast. It fits the appearance of Anāhitā in the inscriptions, it fits the statue-verses of the Yašt, and it has traditionally been taken as the charter of her imperial cult. Boyce built an elaborate reconstruction on it: that the impulse came from the Babylonian mothers and wives of the late kings, above all Parysatis, familiar from infancy with the statues of Ishtar, and that the assimilation of the western planetary goddess to the Avestan river-yazata was carried through by learned Persian magi so that orthodox Zoroastrians could worship her image through the Avestan invocation of Arədvī Sūrā Anāhitā.^[4] A shrewd corollary follows from her own analysis: in a heartland already full of Mesopotamian goddesses, the 'setting up' of images at Babylon and Susa may in places have been the appropriation of existing statues of Nanaia or Ishtar, and the introduction of magi into their temples, rather than the carving of anything new.^[4]

Later notices give the cult flesh. Plutarch has Artaxerxes II enthroned at Pasargadae in the temple of 'a warlike goddess' whom the Greeks took for Athena, and consigning a rival's mistress to be priestess of 'Anaïtis' at Ecbatana; Strabo describes her sanctuaries and their sacred prostitution in Anatolia, though that last detail is doubted, the Susa manumission-inscriptions once read as binding freed women as temple hierodules of Nanaia now being explained wholly by ordinary Greek law;^[4] and Aelian preserves the strange detail of the tame lions kept at her shrine in Elymais, 'which greet and wag their tails at those entering'.^[10] The lion is the beast of Ishtar, and its presence at Anāhitā's shrine is one of the surest signs of how far the two goddesses had merged.

Whatever its scale, the innovation credited to Artaxerxes II was a breach in one of the best-marked features of old Persian worship: its refusal of images (see the sacred fire). To plant a robed and jewelled cult-statue of a goddess in shrines from Babylon to Sardis was to import a Mesopotamian habit into a religion that had done without it, and the Avestan tradition registered the strain. The statue-verses of the fifth Yašt (5.126–129) are the trace of it inside the holy text: an image described so exactly that it can only have been composed for a real one, folded into the hymn to an elemental river that needs no body at all.

Boyce drew the consequence out into a wider argument about the shape of the religion. The orthodox answer to the new image-cult, on her reading, was to found temples of their own that housed no sculpted idol but only a consecrated fire, the one icon a follower of Zarathuštra could accept, so that the rise of the enclosed fire-temple is itself a reaction to Anāhitā's statues. The figure floating in the winged ring over the royal tombs, long read as Ahura Mazdā and now contested, belongs to a different and older visual language; the goddess's statue is something new, and its novelty is the point. What can be said is that the image and the elemental river stand together, uneasily, in her cult from the fourth century on.

Against the confident picture of a datable, empire-wide statue-cult Pierre Briant has entered a thorough scepticism, and his caution is the necessary corrective to the older synthesis. The trouble, he argues, is that the Berossus edict cannot be checked against anything. There is the royal formula, and there is the report of the edict, and the two are made to prove each other; but there is no independent, datable evidence that statue-shrines of Anāhitā were actually founded in the towns Berossus lists at the time of Artaxerxes II. The one document that has been pressed into service, the Greek inscription of the Lydian hyparch Droaphernes from Sardis, records the dedication of a statue but calls it an andrias, the statue of a human, not of a god; its date wavers between the reigns of Artaxerxes I and Artaxerxes II; and the copy that survives was recut centuries later under Rome. It cannot bear the weight put on it. Briant's verdict is unsparing:

we are utterly unable to find confirmation or illustration anywhere else of the change suggested by the royal inscriptions and made explicit in the Berossus quotation.

Briant, From Cyrus to Alexander, 677^[11]

Worse, the sources jar against each other. Herodotus and Strabo both flatly deny that the Persians used images or temples at all; another fourth-century writer, Dinon, says they worshipped two gods in statue-form, but names them as Fire and Water, not Anāhitā and Mithra. Plutarch's temple at Pasargadae implies the goddess's shrine there already existed and may have been founded earlier. The result, Briant insists, is that we cannot even say how new Artaxerxes II's 'novelty' really was, whether he founded a cult, regularised a diffuse one that had grown up piecemeal in the west, or merely brought into the royal formulary two gods long worshipped by his subjects. The famous edict may be genuine and may be transformative; but it hangs, unsupported, over a documentary void.

Behind the cult-question lies the deeper one of origins, and here the poles are drawn most sharply between Boyce and de Jong, though they agree on more than the contrast suggests. Both hold that the Anāhitā of the empire is a composite; they differ on the weight of the two halves. Boyce's instinct is to begin from the ancient Indo-Iranian river-goddess, the genuinely old *Harahvatī, and to treat the Ishtar-material as a later western overlay assimilated to a pre-existing Avestan figure. De Jong reverses the emphasis. Reviewing the evidence, he notes that Anāhitā is absent from the Old Avestan texts, that even in the Younger Avesta her presence outside her own hymn is 'surprisingly limited', and that her overwhelming popularity is a western and Anatolian and Armenian phenomenon, disproportionate to any footing in the east. His conclusion runs the other way:

All this makes it very likely that Anahita was originally a Western Iranian goddess.

de Jong, Traditions of the Magi, 105^[3]

Her 'overpowering role as a warrior queen and as a goddess of love and healing', he adds, cannot be squared with the modest river-spirit of the Avestan name, and points instead to the great goddesses of the Near East, to Babylonian Ishtar, to Elamite Nanaia, to the shared association of these figures with the planet Venus. On this account the cult that Artaxerxes II promoted was western Iranian to the core, a goddess of the planet Venus learnt from Mesopotamian neighbours, who was fitted into the Zoroastrian canon only when her popularity forced the priests to find her an Avestan name, the water-yazata Arədvī Sūrā, whose epithet anāhita she could conveniently share. The Elamite Nanaia, worshipped for centuries at Susa and long fused with Ishtar, is the likeliest single model: it is at Susa and Babylon, Nanaia's own ground, that the image-cult is best attested, and the tame lions of the Elymaean shrine are her beast as much as Ishtar's.

The difference between the two positions is real but narrow, and it is worth locating precisely. Neither scholar denies the Indo-Iranian river-goddess; neither denies the Mesopotamian overlay; the question is only which came first in the west and which is the graft. What both grant is that the goddess of the empire is a composite, that her image and much of her power came from Ishtar and Nanaia, and that the water-yazata of the ancient name and the western queen of the cult were never quite the same being.

The Greeks are both an indispensable witness to the cult and a warning about how it reaches us. They fitted Anāhitā to their own goddesses, calling her now Aphrodite (in her guise as a goddess of love), now Artemis, now Athena (as a warrior), so that the same divinity wears three Greek masks in the sources and the modern reader must keep undoing the translation. The classic instance of the confusion is the earliest. Herodotus, describing Persian religion, begins with a statement that is broadly right and that Anāhitā's own history bears out:

They have no images of the gods, no temples nor altars, and consider the use of them a sign of folly.

Herodotus 1.131, trans. Rawlinson 1862^[12]

The old Persian religion he describes is aniconic, its gods the sun and moon, earth, fire, water and the winds:

They likewise offer to the sun and moon, to the earth, to fire, to water, and to the winds. These are the only gods whose worship has come down to them from ancient times.

Herodotus 1.131, trans. Rawlinson 1862^[12]

Then, Herodotus goes on, at a later period the Persians took up the worship of a heavenly goddess, Aphrodite Ourania, which he thought they had learnt from the Assyrians and the Arabians, a goddess 'whom the Arabians call Alitta, and the Persians Mitra'.^[12] That last name is a famous error. The celestial goddess the Greeks equated with Aphrodite Ourania is Anāhitā, not Mithra, who is a male god of the covenant; working at second hand, Herodotus has attached the wrong Persian name to the right cult. De Jong, who examines the passage at length, finds that most specialists read it exactly so, as a confusion in which the name of the male god Mithra has been given to the female Celestial goddess, the two being the most popular divinities of western Iran and easily muddled.^[3] What is striking, and what de Jong stresses, is that beneath the error Herodotus is substantially right about Anāhitā herself: he correctly makes her a latecomer, a goddess borrowed from the Near Eastern neighbours and not part of the oldest Persian worship, which is precisely the western-graft account of her origins. The blunder over the name and the accuracy about the cult sit side by side, a measure of how much and how little the best Greek source can tell us.

It was in the west that Anāhitā grew into a great power. The evidence for her cult, thin in the Iranian heartland and silent in the east, thickens the closer one comes to the Aegean and the Caucasus. Boyce noted that Strabo derived the great Persian temple at Zela in Pontus from a sanctuary founded in thanksgiving during Cyrus's Anatolian campaigns, an artificial walled mound rather than a building, and that from the sixth century onward there was a continuous Zoroastrian presence in Asia Minor from which much of what is recorded of the faith's observances comes.^[4] By the Hellenistic and Roman periods her temples stood across Lydia and Cappadocia and Pontus, at Hierocaesarea and Hypaepa and Zela, served by hereditary priests and, in the Greek accounts, by sacred slaves. At Zela she shared her altar with two other Persian gods, Omanos (behind whom stands Vohu Manah) and Anadatos, and her image-shrines kept up the recognisably Zoroastrian ritual, the priestly garb and the barsom and the chanting, that Strabo alone records for such places.^[3] Her statue was struck on the coins of several Anatolian cities.

Armenia went furthest of all. There, under her Armenian name Anahit, she was the chief goddess of the country, honoured with gold statues and an annual festival, so central that a whole province, Anahtakan, was named for her. De Jong points to precisely this western and Armenian weight as the argument for her western origin: a goddess so dominant in Anatolia and Armenia, and so faint in the Iranian east, is unlikely to have travelled westward from an eastern homeland.^[3] The Anāhitā the classical and later sources describe is overwhelmingly this western goddess, warlike and amorous and richly imaged, a long way from the elemental river of the hymn's opening stanzas.

Whatever her beginnings, Anāhitā outgrew them all. In the Sasanian period she is a goddess of the first rank in the Iranian heartland itself: her great shrine at Estakhr in Fārs, near Persepolis, was tended by the family that founded the dynasty, and kings sent the skulls of defeated enemies there as trophies, the martial Ishtar showing through the water-goddess to the end.^[4] The Sasanian founder-legend made Pābag, the ancestor of Ardashir, the fire-priest of this temple of 'Anāhīd' at Staxr, a goddess said to have been worshipped in Persis already under Artaxerxes II.^[9] That legend is the one continuous thread the later tradition drew from the fourth-century innovation to the dynasty that would make her a goddess of state. Yet the image-cult that Artaxerxes II had begun did not itself survive. By the Sasanian centuries the statue-shrines had been given up: de Jong observes that in that period 'statue cult was entirely abandoned and all image shrines disappeared'.^[3] The great Anāhitā sanctuary at Estakhr became, like the others, a temple of fire rather than of an image.^[9] The goddess kept her rank; her borrowed body was quietly set down, and the aniconism that the fourth-century innovation had breached closed over the cult again.

Her other traces endured. In the Zoroastrian calendar she keeps the tenth month and the day Ābān, the waters; her name lives on in Armenian Anahit and in Persian Nāhīd, still the ordinary word for the planet Venus. The silver-gilt vessels of late antiquity that show a robed woman holding a flower or pouring water, conventionally called 'Anāhitā', give a sense of the golden, jewelled lady of the statue-verses, even where the identification is a modern guess rather than an ancient label. She remained what she was made into over the long Achaemenid and post-Achaemenid centuries: the water of life given a face, an old river-spirit and a borrowed Near Eastern queen fused into the one goddess whom Persians, Greeks and Armenians alike could not do without.