The Sources & How We Know

The silence is partly accident and partly system. The Persians conducted their empire in Aramaic written on leather and papyrus, media that rot; what survived did so almost by mistake, clay tablets baked hard in the destruction of Persepolis, inscriptions cut into living rock. But the deeper cause is that the Persians did not practise history in the Greek manner at all. Briant's survey of the evidence opens from exactly this point: unlike most conquering peoples, the Persians left no written testaments of their own history in the narrative sense of the word; unlike the Assyrian kings, the Great Kings kept no annals of their deeds.^[1] Ctesias boasted of consulting "royal records" in which the Persians set down their ancient affairs, but no other trace of such historical archives exists, and the archives that are attested, satrapal and royal alike, were administrative: registers of decisions, grants and taxes rather than chronicles.^[1] The Persian past was carried instead in oral tradition, in the songs and recitations through which the deeds of kings and heroes passed from generation to generation, and in the royal proclamations it was conflated with the genealogy of the dynasty. Briant's formulation has become the classic statement of the predicament:

Written in the immovable and infinite time of the King, the history of the Persians was never situated in the measured time of History by the Persians themselves.

Pierre Briant, From Cyrus to Alexander, p. 7^[1]

The consequence is stark: "one must reconstruct the narrative thread of Achaemenid history from the writings of their subjects and their enemies".^[1] Every surviving connected narrative of the Greco-Persian Wars, and of the Achaemenid court, is Greek. The "Persian view" can be recovered, but only obliquely: from the ideology of the royal inscriptions, from the priorities visible in the administrative tablets, from the archaeology, and from reading the Greek sources against their own grain. One caution comes first. It is tempting to reach for the Shahnameh of Ferdowsi (c. 1000 CE) or later Persian epic as a native memory of the Achaemenids, but no conduit from Achaemenid tradition to those texts can be demonstrated, and they cannot be used as evidence for events of the period.

The Persians did leave one kind of first-person testimony: the royal inscriptions, above all the great trilingual monument of Darius at Behistun. These are primary evidence of the highest value, but they are propaganda in the exact sense that they record what a king wished posterity to believe, in formulaic court language. Old Persian cuneiform was itself probably an artificial royal creation, devised under Darius to give the ruling people a script of their own; nearly everything written in it belongs to a closed ceremonial sphere, monumental inscriptions "intended to reflect the unchanging majesty of Persian imperial power", which record no political events, no changes, no administration, with the single great exception of Behistun.^[2] Nor does the Old Persian version carry special authority: the Elamite and Babylonian versions are equally the king's word, and at Behistun the Elamite text was probably cut first.^[2]

What the inscriptions do tell, with unmatched directness, is the Achaemenid royal ideology: that kingship is the gift of the Wise Lord, Ahura Mazdā; that the cosmos is ordered by arta, the Truth or right arrangement, and besieged by drauga, the Lie; and that the king's office is to uphold the one and destroy the other. Darius casts every rebel as a partisan of the Lie:

Thou who shalt be king hereafter, protect thyself vigorously from the Lie; the man who shall be a Lie-follower, him do thou punish well, if thus thou shalt think, "May my country be secure!"

Darius I, DB §55, trans. Kent^[3]

He even builds a claim of truthfulness into the monument itself, which is precisely the kind of claim a contested account needs to make:

Thou who shalt hereafter read this inscription, let that which has been done by me convince thee; do not thou think it a lie.

Darius I, DB §56, trans. Kent^[3]

That the monument protests its own honesty is a warning, not a reassurance. Behistun is Darius's justification for a violent and contested seizure of the throne: he had the man he killed declared an impostor, a magus named Gaumāta who "lied" in claiming to be Cyrus's son Bardiya. The inscription is the earliest and most eloquent statement of the arta-against-the-Lie ideology at the heart of Achaemenid kingship, and it is also, quite possibly, the empire's founding lie about itself. It is read as ideology first and as fact only where other evidence corroborates.

The richest narratives are Greek, and they are the hardest to use well. Herodotus is the indispensable spine: his Histories (c. 440s–420s BCE) are the fullest account of the empire down to Xerxes, and a surprising amount is corroborated by the Near Eastern record: the satrapy-and-tribute structure, the Royal Road, the daric, the Gaumāta usurpation against Behistun, the schooling of Persian boys. He was a genuine researcher who distinguished what he saw from what he was told, so far from hostile to Persia that Plutarch would later fling philobarbaros, "barbarian-lover", at him as an accusation.^[1] He preserves the Persians' own severe ethic:

Their sons are carefully instructed from their fifth to their twentieth year, in three things alone, — to ride, to draw the bow, and to speak the truth.

Herodotus, Histories 1.136, trans. Rawlinson^[4]

Yet the same tradition that made Herodotus the father of history also counted him a liar: Lucian's satire consigned him to the liars' punishments in the afterlife, with Ctesias for company.^[5] He wrote two to three generations after Darius, inside a Greek frame of freedom against servitude; as Kuhrt puts it, "his work is marked by a series of didactic motifs, which shape the way he structured his account": imperial expansion breeds decline, monarchy slides into despotism, the liberator becomes a conqueror.^[2] His army numbers are wildly inflated; his gold-digging ants are folklore; and his famous "Constitutional Debate", in which Persian grandees weigh democracy, oligarchy and monarchy, reads to most modern scholars as Greek political theory in Persian dress, for all that Herodotus himself insisted, twice and emphatically, that it really took place.^[6]

The other Greeks compound the difficulty. Ctesias of Cnidus, a physician at the court of Artaxerxes II, should have been gold and was dismissed by a century of scholarship as sensational and garbled, the ancestor of the historical romance; the modern reassessment reads his Persica less as failed history than as a record of the stories Persians around the court actually told, c. 400 BCE, about their own past. Xenophon is three different witnesses in one man: the eyewitness of the Anabasis (invaluable on the land, the satraps and the roads), the moral novelist of the Cyropaedia (an idealised Cyrus, never to be cited as fact), and the observer of Persian estate-management and the paradise garden. Earliest of all is Aeschylus, whose tragedy *The Persians* (472 BCE) staged the defeat at Salamis from inside the Persian court, eight years after the battle, by a veteran of the war: the messenger's cry that "the whole barbarian host is lost" opens the earliest surviving Greek portrait of Persia, and it is already an invention, Persian grief imagined for an Athenian audience.^[7]

Two structural limits run under the whole tradition. First, language: Greek writers worked without Persian, Elamite or Akkadian, and the Greek world made no effort to acquire them; "Greek remained the only language of civilization for every Greek-speaking man", as Momigliano observed of the whole classical engagement with the East.^[8] Everything Greek authors report of Persian records and traditions came to them through interpreters, informants and hearsay. Second, selection: what interested a Greek audience was the court, the harem, the King's table and the fabulous wealth; the tax registers and province lists drew far less interest, so the tradition preserves anecdote where a historian wants structure. The balance of the modern verdict is nonetheless not dismissal. However distorted the accounts, they contain a wealth of real information; what the reader must reckon with, in Kuhrt's phrase, is "not invention, but the 'spin' that has been put on events".^[2]

The deeper problem with the Greek tradition is not error but perspective, and naming it was the founding act of the modern discipline. Heleen Sancisi-Weerdenburg's essay "The Fifth Oriental Monarchy and Hellenocentrism" (1987) took its title from Henry Rawlinson's Victorian history of Persia and turned Edward Said's Orientalism onto the ancient evidence: the Orient of sweet perfumes, intrigue and harem-ladies is a European construction, for Europe "one of its deepest and most recurring images of the Other", and its pedigree runs back through the historiography of Persia to the Greek fourth century itself.^[9] The stereotype outlives every particular fact: "It is not so much the Orient that has eternally similar characteristics, it is rather the Western vision of it that remains the same".^[9] Her conclusion made the methodological point that still governs the field: "It is not so much the facts that distort historical reality as the outlines into which they are fitted", and the remedy is to analyse the literary and intellectual mould into which the Greek data were poured, a task that goes beyond checking Greek claims against thin Iranian evidence.^[9]

Janett Morgan's Greek Perspectives on the Achaemenid Empire: Persia Through the Looking Glass (2016) carries the same critique to its furthest point. Morgan's governing image is the looking glass: the Greek sources are a mirror that reflects Greek concerns as much as any Persian reality, and, worse, a mirror whose distortions the modern reader has often inherited unawares. She opens with Lord Curzon at Persepolis, who in a single visit found the ruins insignificant, then magnificent, then monotonous, three views of a site that never moved, because "it was Curzon who shaped Persepolis to fit the needs of his tale".^[10] Her sharpest argument is that the Greco-Persian Wars have been made into "the" explanation for every Greek image of Persia (the cruel despot, the decadent luxury, the dominating wives, the ever-present eunuch), and that this "Great Event" is a modern fairy tale of ancestry, in which Greek victory becomes the fountainhead of European civilisation; she quotes Brunt: "In 480–79 the Greeks saved themselves and the future of European civilisation from Oriental conquest."^[10] Against this she sets a longer view: the tropes of otherness and despotism grow out of internal Greek dialogues of status and identity that predate the Achaemenids and outlast them, driven "more by political change and elite competition than by victory in the Graeco-Persian Wars", so that a Greek image of Persia is often evidence for rivalries in Athens or Sparta, not for anything east of the Aegean.^[10] The rule that follows is exact: the Greek sources are best used as evidence for how Greeks constructed "Persia", and trusted for Achaemenid fact only where the contemporary record agrees.

Around the Greek narratives stands a ring of contemporary Near Eastern texts, one for nearly every people the empire absorbed, and each carries its own interest. From Babylon come the chronicles and the Cyrus Cylinder, the barrel of clay on which Cyrus, entering Babylon in 539 BCE, presents himself as the chosen of Marduk, restorer of neglected gods and returner of exiles. The Cylinder is the cautionary case of the whole field: celebrated in modern times as a charter of human rights, it is in fact standard Mesopotamian royal boilerplate, its liberating phrases the conventional language priests used to bless any acceptable conqueror, and Kuhrt's 1983 analysis, which dismantles the modern human-rights legend, is decisive here.^[11] The chronicle tradition that narrates the fall of Babylon is itself the work of Marduk's priesthood, hostile to the king Cyrus displaced, Nabonidus; the Babylonian evidence for the conquest is a dossier of partisan texts agreeing against a common enemy; it is no neutral record. From the Hebrew Bible come Ezra and Nehemiah, genuinely valuable for the Across-the-River province and the policy of restoration, and preserving Aramaic documents whose authenticity is still debated, and the moment in Deutero-Isaiah where a Jewish prophet hails a Persian king as the Lord's anointed:

Thus saith the LORD to his anointed, to Cyrus, whose right hand I have holden, to subdue nations before him.

Isaiah 45:1, KJV^[12]

Esther and Daniel, by contrast, are later court-novellas: convincing in their Persian colour, unreliable as event-history. From Egypt comes the only native voice of a conquest anywhere in the empire, the statue-inscription of the collaborator Udjahorresnet, courtier of Cambyses and Darius, along with Darius's multilingual canal stelae and the burial records of the sacred Apis bulls. Every one of these is a witness with a purpose: priestly grievance, dynastic legitimation, a subject people's hope, a courtier's self-portrait. None of that disqualifies them; it defines how they are read. The method, here as everywhere, is to read the purpose along with the text.

The most candid evidence is the evidence no one meant posterity to read. Buried under the terrace at Persepolis lay two administrative archives in Elamite: the Fortification tablets (a corpus running to tens of thousands of pieces, dating 509–493 BCE, the heart of Darius's reign) and the smaller Treasury tablets (c. 492–457 BCE). They are ration-dockets, travel-authorisations and payroll: grain, wine and livestock issued to workers, officials, priests and the royal household, silver wages that mark the spread of coined money. They were not composed to impress anyone, and so they correct the Greek picture at its most vulnerable points. Against the Greek stereotype of secluded, passive Persian women, the tablets show workgroups headed by a female chief whose ration far outstrips her workers', extra rations issued to new mothers, and a great lady, Irdabama, sealing her own orders and commanding a workforce of several hundred.^[13] Against the image of an enforced royal monotheism, the same ledgers record offerings for Ahura Mazdā and for Elamite and other gods side by side, a genuinely plural cult in the court's own country. The archive answers few of the questions a narrative would; but where it touches the Greek account, the archive wins.

Persepolis is the greatest of the archives but not the only one. From Nippur in Babylonia come the Akkadian records of the Murašû family, a business firm managing military land-grants and noble estates in the later fifth century: the fiscal machinery of the empire seen from below. From Egypt come the Aramaic papyri of the Judaean garrison at Elephantine, with their family archives, their temple to Yahweh and their correspondence with the Persian authorities, and the leather letters of the satrap Arsames, running his Egyptian estates from Babylon. And the pattern keeps repeating: a customs account of duties levied on Phoenician and Ionian ships, found reused under a literary text, is so far unique evidence for the taxation of trade; a cache of fourth-century Aramaic documents from Bactria, published only in the 2000s, revealed a fully articulated satrapal administration in the far north-east, exactly where archaeologists had long argued Persian government was a fiction.^[2] Kuhrt draws the warning of that discovery for the whole discipline:

In studying the Achaemenid empire, we must always be prepared to have cherished views totally overturned by such chance discoveries.

Amélie Kuhrt, The Persian Empire, p. 14^[2]

Beyond the texts lies everything the spade and the eye recover: the platform and reliefs of Persepolis, the tombs at Naqsh-e Rostam, the palaces of Susa and the garden-plan of Pasargadae, the seals, the metalwork, the coinage. This evidence is mute (it tells no story and names few names), but it is contemporary and it is not rhetoric, and it does two things the texts cannot. It confirms and dates: Babylonian astronomical diaries, cross-referenced to regnal years, anchor the chronology to the day, and excavation confirms Greek craftsmen at work in the Persian palaces exactly as the Susa foundation charter claims. And it speaks the empire's own visual language of kingship and order (the king above the winged symbol, the ranked tribute-bearers of the Apadana), imagery that carries Achaemenid ideology without a Greek intermediary. Its limits are real. Interpretation is contested, beginning with the identity of the figure in the winged ring; and the empire is strangely hard to see in the ground outside its capitals. Distinctively Persian finds in the provinces are thin, a paradox Sancisi-Weerdenburg fixed in a phrase, "when one decides to look from the bottom, it is often hard to see the empire", and archaeologists long divided over whether that thinness meant superficial rule or merely a government that did not travel with its own pottery.^[14] The Bactrian documents answered the question for one region, in favour of the second reading: an administration invisible to archaeology had been there on parchment all along.^[2]

The method that holds all this together is triangulation: no single source is trusted alone; the Greek narrative supplies the story, the inscriptions the ideology, the tablets the daily reality, the archaeology the anchor, and a claim is credited in proportion to how many independent kinds of evidence converge on it. The evidence does not assemble into a jigsaw with one picture on the box; in Kuhrt's phrase, "we gain a series of overlapping and differing perspectives", which sometimes reinforce a classical or biblical passage and sometimes flatly contradict it.^[2] Where a Greek source and the contemporary record conflict, the contemporary record wins; where the Greek source stands alone, the reader follows Briant's maxim and tries to "distinguish the Greek interpretive coating from the Achaemenid nugget of information".^[1] And because "historians cannot choose their sources", the Greek tradition is never simply discarded: for long stretches of political history it is all there is, and the discipline consists in using it without being used by it.^[1] Three worked cases show the method running, each treated at full length in its own entry.

On the surface, the accession of Darius is the best-corroborated event in Achaemenid history. Darius's own Behistun narrative and Herodotus's account, composed two generations apart in different languages for different audiences, agree on the essentials: a magus impersonated Cyrus's son Bardiya, reigned briefly, and was killed by seven Persian conspirators, of whom Darius became king. Babylonian business tablets, dated day by day to the reigns of the successive claimants, then fix the chronology of the crisis with a precision unmatched anywhere else in the ancient world.^[2] It looks like triangulation at its best. The trap is that the convergence is not independence. Darius circulated his official account through the empire, on stone at Behistun, in Aramaic on papyrus as far as Elephantine, and every later narrative, Herodotus's included, descends from the version the winner published. What the tablets independently confirm is that a king recognised as "Bardiya" really reigned, and for longer than a brief imposture suggests; what nothing outside Darius's own testimony confirms is that this Bardiya was an impostor at all. Hence the sceptical reading of Bardiya / Gaumāta that has dominated scholarship since the 1980s: that "the person murdered by Darius and his friends was none other than Cyrus' son and Cambyses' brother, Bardiya", and that "it is, therefore, Darius himself who is the true rebel".^[2] Agreement between sources proves nothing until their descent is traced: two witnesses reciting the same press release are one witness.

The second case runs the other way: a contemporary document set against a Greek narrative. Herodotus's Cambyses is a madman whose crowning sacrilege was to stab the sacred Apis bull of Memphis, which died of the wound; the story anchors his whole moral portrait of the king, and through him a Greek theorem about despotism ending in impiety and madness. The Serapeum of Memphis, where the embalmed Apis bulls were interred, was excavated in the 1850s, and the epitaph of an Apis buried during Cambyses' occupation, conventionally dated to 524 BCE, was found in place. It shows Cambyses, in Egyptian royal dress, kneeling in devotion: the bull died in the king's sixth year, was embalmed through the full rites, and was buried in a granite sarcophagus that Cambyses dedicated as pharaoh, the epitaph closing, "It was done according to everything his Majesty had said".^[1] In the strong form of the twentieth-century rehabilitation this reads as a plain refutation: Briant's verdict on the collision is blunt, "Herodotus recorded spurious information", and far from killing the Apis, Cambyses buried one according to the rules, as a conqueror taking his place in Egyptian kingship.^[1] What the monuments prove, though, is narrower: they disprove the story of blanket sacrilege, since a king who provided in proper form and in his own name for an Apis burial did not massacre the cult, but they cannot exclude every version of an outrage. Nor is the epitaph a neutral check on Herodotus. It is an official product of the very accommodation it attests, the conquest domesticated in stone by the priesthood and the collaborator who served the new pharaoh, a document with an argument of its own to make. That is the harder and better lesson: the correcting source is interested too, and using it means reading its purpose as closely as the narrative's. The dated Apis stelae have since been re-examined and the mid-century certainty has softened; the reopened file is carried under the conquest of Egypt by Cambyses II.^[15]

The third case shows criticism working even where no document contradicts. Herodotus counts Xerxes' invasion force at 1,700,000 infantry, and with crews, cavalry and camp followers arrives at over five million souls at Thermopylae. No source refutes the figure; arithmetic and logistics do. "For a long time, Herodotus's statistics were judged unacceptable, if only for purely logistical reasons", Briant notes: the fleet total of 1,207 is a literary number with a Homeric pedigree, feeding the horde would have taken millions of litres of flour a day, and the one land battle fought to a finish, Plataea, implies Persian forces of the same order as the Greek, perhaps some tens of thousands under Mardonius rather than millions.^[1] The inflation itself is evidence, of what the invasion became in Greek memory. A companion case concerns the same king's reputation as a destroyer of temples. For over a century, scholarship repeated that Xerxes avenged the Babylonian revolts of 484 BCE by demolishing Esagila, Marduk's great sanctuary, and carrying off the god's statue, and hung a general theory of Achaemenid decline into religious intolerance on the event. Kuhrt and Sherwin-White's re-examination (1987) found the story resting on a chain of Greek anecdote and modern conjecture: the Babylonian evidence knows nothing of a destruction, the Ptolemaic Canon (which struck out sacrilegious Assyrian kings as "kingless") gives Xerxes his full regnal years, and a temple record under Artaxerxes I shows Esagila functioning normally, so that "this evidence should in fact be used to argue for continuity of cult at Babylon".^[16] A stereotype, the hubristic oriental despot of Aeschylus and Herodotus, had organised the scholarship on Xerxes I until the documents were re-read without it.

None of this method is timeless; most of it is barely two generations old. For two centuries after the decipherment of cuneiform, the empire was still read through the Greek mirror, as an oriental despotism of luxury and decline, and the new Persian evidence was, in Kuhrt's image, fitted into "a kind of template of Persia and its rulers" that the classical authors and the Old Testament had built, "fitted or forced to fit".^[2] The editors of the first Achaemenid History volumes described the old dispensation with dry precision: the Greeks were held to be almost infallible, first because they were Greeks and second because they were contemporaries, and if the monuments disagreed with Herodotus, so much the worse for the monuments.^[9]

The correction was organised, deliberate, and datable. From 1983 to 1990 Heleen Sancisi-Weerdenburg, joined by Amélie Kuhrt, convened the Achaemenid History Workshops in Groningen, London and Ann Arbor: some thirty scholars a year, Assyriologists, classicists, Iranologists and archaeologists deliberately mixed, publishing their proceedings in the Achaemenid History series. The principles were common property: "They took a skeptical approach to the classical literary sources", asked how the empire might be interpreted without the Greek narratives, and tried "to gather different kinds of evidence and look at it afresh without assuming that Greek and Latin literature provided the basic framework" which other evidence merely supplemented or corrected.^[17] The paradox the founders embraced was that escaping the Greek sources demanded more Greek philology, not less: "in order to move away from the dominant and oppressive Greek perspective, it is more than ever necessary to pay attention to Greek historiography and analyze it in a new way", as the introduction to the volume on the Greek sources put it, the only way, in Sancisi-Weerdenburg's phrase, "to dehellenise and decolonialise Persian history".^[9] The workshops' demolition work was concrete and particular: Xerxes' Babylonian sacrilege, the "decadence" model of the post-Xerxes empire, and Sancisi-Weerdenburg's own question "Was there ever a Median Empire?", which dissolved the Median empire of the textbooks into a projection of Herodotus's narrative onto silent Near Eastern evidence.^[17]

The synthesis followed. Pierre Briant's Histoire de l'Empire perse (1996; English 2002), the work of a regular participant, became "a sort of summa achaemenidologica", the centre of the field: where earlier surveys gave confident narrative, Briant asks on every page how the sources may be interpreted, and his research notes run to 174 pages.^[17] Kuhrt's sourcebook The Persian Empire (2007) then put the primary documents of every language and genre side by side in translation, so that any reader can begin from the evidence rather than from a narrative built on one class of it.^[2]

What the workshops began is now simply how the subject works. Josef Wiesehöfer's survey leads with the indigenous evidence and treats the Greek image of Persia as a topic in its own right; Maria Brosius reconstructed the royal women from the tablets against the Greek harem tales; Wouter Henkelman's studies of the Elamite Fortification archive rewrote the religion of the Persian heartland; Lloyd Llewellyn-Jones used the Persian evidence to put the court, the very subject the Greeks distorted most lovingly, on a documentary footing.^[11] The pendulum has also begun to swing back, in a controlled way. Some scholars now find the founding generation "too defensive of the Achaemenids and too cautious about comparisons with later periods", and a strand of work, Harrison's Writing Ancient Persia among its statements, questions whether the new orthodoxy sometimes replaced Greek bias with an idealised empire.^[17] The productive form of the reaction reads the Greek sources once more as bearers of genuine Persian material, oral traditions, court stories, institutional detail, recoverable when checked against the documents: Ctesias rehabilitated as a witness to Persian storytelling rather than condemned as a failed Herodotus is the emblematic case. And Morgan turns the source-critical lens on the modern historian: the scholar's own inherited images, Curzon's Persepolis, the "Great Event" of 480, still shape what the evidence is allowed to say.^[10] The critique of the sources has become, in the field's most recent turn, a critique of the readers.

The gaps are structural, not accidental. The evidence clusters brutally: in time on the late sixth and fifth centuries, in space on the empire's north-west; the fourth century, the era Alexander ended, is the worst documented of all, and the eastern half of the empire is known mainly from the moments Greek armies marched through it.^[2] Whole institutions hang on single finds: one customs account, one pouch of satrapal letters, one archive of one firm in one Babylonian city. The deepest uncertainties are of two kinds. Some are evidential accidents that new finds could mend, as the Bactrian parchments mended one, and as the tens of thousands of unread Persepolis fragments still might. Others are built into the record itself. Whether Darius told the truth about the man he killed cannot be settled by any likely discovery, because every narrative descends from Darius. Where Xerxes' daiva-shrine stood, whom the winged figure shows, how far the kings were Zoroastrians in any later sense, whether the Aramaic documents in Ezra are authentic: these stay open because the classes of evidence that would decide them do not touch.^[18] The discipline's maturity shows in leaving them open, uncertainty stated being knowledge of a kind, and in remembering that its firmest results were once cherished views. Morgan ends her book with Shipley's warning, and it serves for the whole field: "no historian should ever imagine that he or she has said the last word on a subject".^[10]