Herodotus, The Histories

Almost nothing is securely known about Herodotus, and what is usually reported comes not from his own book but from a single ancient biography in the tenth-century Byzantine lexicon known as the Suda, compiled some fourteen centuries after his birth. It makes him the son of Lyxes and Dryo, of a leading Halicarnassian family; a nephew or cousin of the epic poet Panyassis; a man who left home because of the tyrant Lygdamis, spent time on Samos, returned to help expel the tyrant, and, disliked at home, sailed west to the new colony of Thurii, where he died. David Asheri, whose commentary is the standard modern account of the life, treats these notices with caution: the Suda draws on Hellenistic scholarship rather than on Herodotus, and its facts are a mixture of inference, guesswork and later legend.^[1] The one firm rule, Asheri holds, is to look for the man in his own work, and there he is oddly present and oddly hidden at once. Greek historiography inherited from epic the convention that an author should not speak of himself unless the subject required it, and Herodotus mostly obeys it; yet his voice, questioning, weighing, digressing, is felt on every page, so that to read the Histories is like hearing him talk.^[1]

Halicarnassus, on the coast of Caria in south-western Asia Minor, was a Dorian foundation with a strong Carian element and by Herodotus' day a subject city of the Persian king, ruled by a dynasty of Carian tyrants, of whom Artemisia, who fought for Xerxes at Salamis, was the most famous. That a man born inside the empire and under a Persian client dynasty should have written its history from the Greek side is one of the shaping facts of the work. The names of his father and kin, attested in a Halicarnassian inscription, are themselves Carian rather than Greek, and Asheri infers from the family's lack of any claim to divine ancestry that Herodotus was probably not of the old aristocracy.^[1] He wrote not in the Doric of his city but in Ionic, the literary prose dialect of the Ionian enquiry he was continuing.

The famous dates are a scholarly construction, and both ends are soft. No source records Herodotus' birth; the figure of 484 BCE is an ancient back-reckoning. Hellenistic chronographers fixed a writer's akmē, his prime, by convention at his fortieth year, tied it to a datable public event, in this case the foundation of Thurii in 444/3, and counted back forty years to a notional birth. Asheri calls the figure "the source of the calculation of the fictional date of his birth", reached by counting forty years back from the foundation of Thurii.^[1] The other end is no firmer. The latest events the Histories mention belong to the first years of the Peloponnesian War, 431 and 430, so composition continued at least that late; the year of his death cannot be fixed at all, only bounded: he did not die before about 430.^[1] The conventional "c. 484 to c. 425" should be read with that looseness on both sides.

What the ancient biographers never mention is the one activity the work itself makes central: travel. Herodotus says, in the first person, that he went to Egypt and up the Nile to Elephantine, sailed to Tyre to check a claim about Heracles, saw the battlefields of the Nile Delta and the inscribed letters at Thebes in Boeotia.^[1] How far to believe these declarations is an old question, and the honest answer, Asheri argues, is to weigh them case by case rather than accept or reject them wholesale.^[1] He certainly knew Ionia, Caria and the Aegean at first hand; that he penetrated far into the Persian heartland is doubtful, and the shape of his Persian material suggests he did not.

The work opens with a single sentence that serves as its title and its programme:

What Herodotus the Halicarnassian has learnt by inquiry is here set forth: in order that so the memory of the past may not be blotted out from among men by time, and that great and marvellous deeds done by Greeks and foreigners and especially the reason why they warred against each other may not lack renown.

Herodotus, Histories 1.1, trans. Godley^[2]

Three purposes are folded into that sentence. The first is to rescue human achievement from oblivion, a debt to the epic idea of kleos, undying fame. The second is stated with a deliberate care that shapes everything after: the deeds worth saving are those of Greeks and barbarians alike, a pledge of even-handedness that Asheri reads as programmatic, a declaration that in the face of human greatness the differences between peoples do not decide whose deeds are worth recording.^[1] The third is the causal question, and above all the reason why they fought, which makes the Histories the first Greek prose work to organise itself around an explicit problem of cause rather than a bare chronicle.

The plan follows the growth of Persian power. Herodotus begins with Croesus of Lydia, the first barbarian he holds to have wronged the Greeks, then turns to the Persians who overthrew him: Cyrus the founder, Cambyses II, the conqueror of Egypt, Darius, and Xerxes, whose invasion fills the last three books. The reigns of the Persian kings are the spine, and the ethnographic set-pieces hang off them at the point of conquest, so that Egypt is described when Cambyses attacks it and Scythia when Darius campaigns there. Michael Flower states the observation structurally: "the whole structure of the Histories is built upon the birth, growth, and checking of the Persian empire", the Persians "the driving force of the history", and it can hardly be an accident that the work's first narratives and its last alike are "focalised through the eyes of the Persians".^[3] The word Herodotus uses for these units is not "book" but logos, an account or section; the division of the whole into nine books, each later named for one of the nine Muses, is not his. It is the work of the Alexandrian librarians of the third and second centuries BCE, and the Muse-names are not attested as titles before the second century CE.^[1] The familiar reference system, Herodotus in nine books, is an editorial scaffold laid on the text long after the author, and his own cross-references cut across it.

That structure has always raised the question of unity. Half the work is digression, and a reader who came for the Persian Wars must pass through the customs of the Egyptians, the rivers of Scythia and the wonders of Arabia to reach them. Asheri reframes the difficulty as a false one: the ethnographic and the historical are two phases of a single enquiry, the description of a people shading into the narrative of its conquest, so that the logoi are not interruptions of the history but the history's necessary breadth.^[1] The work is not a war monograph with excursuses but an enquiry into the whole inhabited world, given shape by the war that drew that world together.

The Greek word Herodotus chose for what he was doing, historíē, meant enquiry, investigation, research, with no restriction to the human past; only in the fourth century did its sense narrow toward "history" in the modern sense, and, as far as the evidence goes, the term appears for the first time in Herodotus.^[1] He did not set out to write "history"; he published the results of an enquiry, and the modern word descends from his usage rather than the reverse. Nor was the enquiry a private study: Rosalind Thomas observes that the proem's other keyword, apodexis, the setting-forth of the results, "bears the connotations of display and demonstration"; the work belongs to a world of public performance and competitive proof.^[4]

The method, as he describes it, ranks two kinds of evidence. First-hand sight, opsis, he holds more reliable than hearsay, akoē, because it does not depend on the honesty or memory of another; but he is no naïve empiricist, and where sight fails he reasons from probability, to eikos, and states a considered opinion of his own.^[1] His characteristic move is to lay out competing versions and to report them all, even those he does not credit, on a principle he states once and asks the reader to carry through the whole work:

For myself, though it be my business to set down that which is told me, to believe it is none at all of my business; let that saying hold good for the whole of my history.

Herodotus, Histories 7.152.3, trans. Godley^[5]

That sentence is the hinge of his practice and of his reputation. It licenses him to preserve traditions he cannot verify, which is why so much survives in the Histories that would otherwise be lost; and it is also, in his critics' eyes, the alibi of a man who reports marvels and shifts the burden of belief onto the reader. Both readings have a point. Herodotus does sometimes reach a verdict, choosing the more plausible of rival accounts and supporting the choice with an argument or a proof; and he often declines to, setting the versions side by side and leaving the judgement to the reader or to posterity.^[1]

The informants he invokes need scrutiny of a kind he could not always give them. His "learned men", the logioi, are frequently the Egyptian temple staff and Babylonian priests whose word he takes with respect; Oswyn Murray notes that in each case Herodotus chooses them because they seem likely to hold an authoritative version of the past, and, more valuable still, that he keeps the traditions of different places separate rather than blending them into one smooth account.^[6] The weakness is that such priests were often minor functionaries of limited learning, and that Herodotus reached them through hired interpreters whose renderings, of a hieroglyphic inscription, say, he had no means to check.^[1] The model of the method, in Murray's phrase, is impeccable however faulty the execution.^[6] And the execution commands the age's whole language of proof: Thomas shows Herodotus "quite familiar with the use of tekmeria", decisive evidence, with testimony and argued necessity besides, a fluency that "would also upset the conveniently neat contrast so often made between Thucydides and Herodotus", the rigorous analyst against the casual entertainer.^[4]

The honorific is Cicero's, and it came with a barb attached. In the mid first century BCE Cicero called Herodotus the "father of history", pater historiae, crediting him with founding the genre; but in the same breath he observed that the work was full of fabulous tales, innumerabiles fabulae, and elsewhere doubted the authenticity of the very oracles Herodotus reports.^[7] The title, in other words, was never a warranty of accuracy. It honoured Herodotus for inventing a kind of writing, not for getting the facts right, and the man who coined it thought the Histories stuffed with fable. Every later use of "father of history" carries that double edge, whether or not the user remembers it.

Asheri's summary catches the paradox of the ancient reception in one line: Herodotus "became the 'father of lies' long before becoming the 'father of history'", and no ancient historian was censured with such violence, yet none was less easy to ignore.^[1] The reason is partly that Herodotus does something new and hard to categorise. Aristotle could call him a mere teller of tales, a mythologos, in one work and treat him as the type-specimen historian in another. Momigliano, looking back over the whole tradition, fixed the quality by a contrast, writing of Polybius that he "is the prototype of the historian who never marvels, just as Herodotus is the prototype of the historian who always marvels", a description that names both the strength, an insatiable curiosity about the world, and the liability, a taste for the wonderful that a stricter age would distrust.^[8]

The counter-tradition is as old as the honour. It begins with Ctesias of Cnidus, a Greek physician at the Persian court a generation after Herodotus, who wrote his own Persica in avowed opposition to the Histories, branded his predecessor a liar and a teller of fables, and claimed better sources at first hand.^[9] The irony, as later scholarship notes, is that Ctesias staked his rival authority on the same eyewitness-versus-hearsay distinction Herodotus had made his own, and did so on far weaker ground; the man who called Herodotus a liar is reckoned by modern historians the wilder of the two.^[10] Simon Hornblower compresses the irony to a clause, that Ctesias "roundly called Herodotus a liar" but "his own versions often presuppose him"; and Momigliano's verdict is harsher still: "There is no evidence that Ctesias tried to understand the Persians of his time, as Herodotus had done."^[11][12] The charge, once made, stuck. Josephus records the settled ancient impression that "all agree on Herodotus' mendacity", and Lucian, in a comic underworld, lodged Herodotus among the damned:

And the greatest punishments of all were endured by those who had ever in their lives told lies and historians who had not told the truth, amongst whom were Ctesias of Cnidus, Herodotus, and many others.

Lucian, True Histories 2.31, trans. Llewellyn-Jones & Robson^[13]

The sharpest ancient attack is Plutarch's essay On the Malice of Herodotus. Its charge is not simple falsehood but bias, kakoētheia, malice: to a Boeotian patriot who thought history should praise the virtuous, Herodotus' even-handedness read as slander, a preference for the unflattering account, a readiness to credit the barbarian and the Athenian at the expense of Thebes and Corinth.^[14] The opening states the thesis, that the famously pleasing style is the very mask of the malice: it is "the highest malignity", Plutarch declares, to pretend "to simplicity and mildness and be in the mean time really most malicious", and since Herodotus aims that malice chiefly at the Boeotians and Corinthians, he thinks himself obliged to defend "our ancestors and the truth", a motive, Hornblower drily notes, declared in that order.^[14][11] What follows is arguably the first formal diagnostic of historiographic bias ever drawn up: the malicious narrator, Plutarch rules, uses "the most odious terms" where gentler ones would serve, drags in discreditable matter irrelevant to the story, omits the creditable, and prefers the worse of two versions.^[14] The concrete counts include the charge that stuck longest, that Herodotus is "so favorable to the barbarians" that, "acquitting Busiris of those human sacrifices and that slaughter of his guests for which he is accused", he shifts the wickedness onto Greeks.^[14] Even the essay's parting shot concedes the artistry it distrusts: "as in roses we must beware of the venomous flies called cantharides; so must we take heed of the calumnies and envy lying hid under smooth and well-couched phrases and expressions".^[14] Read against the grain, the pamphlet is backhanded evidence of the very quality modern readers prize: Herodotus was attacked in antiquity precisely because he was, by the standards of Greek patriotic writing, unusually fair to the enemy. Briant makes the point directly, that Herodotus shows no systematic hostility to the Persians, which is what earned him Plutarch's label of philobarbaros, barbarian-lover.^[15]

The modern revival of the "father of lies" came through the nineteenth century's worship of Thucydides. Ranke and his successors crowned the austere analyst of the Peloponnesian War the greatest of historians, and beside him Herodotus looked credulous; A. H. Sayce and others recast him as a fraud who had invented his travels and his sources, and consigned the Histories to what German scholarship called Schwindelliteratur, swindle-literature.^[1] The charge that Herodotus had lied about where he went and whom he spoke to had in fact a long pedigree, reaching back through the Renaissance, where the humanist Juan Luis Vives had already called him mendaciorum pater, the father of lies.^[1] Its most rigorous modern form is Detlev Fehling's study of 1971, translated into English in 1989, which argued that Herodotus' source-citations, his constant "the Persians say", "the priests told me", are in large part a literary fiction, a rhetorical device of authentication rather than a record of real informants.^[16] Fehling's own summary is more exact than the slogan: in the source-citations, he concluded, almost everything "is accounted for by three basic rules: the principle of citing the obvious source, the principle of maintaining credibility, and the principle of regard for party bias". His strongest single regularity does much of the arguing: the peoples of Herodotus' nearer zone are only ever cited at first hand, those of the farther zone only ever at second, with "not a single exception to the rule" either way, a patterning, Fehling holds, that no honest field-notes would produce.^[16] Fehling's book set off the fiercest source-critical quarrel of the late twentieth century, and its adherents, with François Hartog and Stephanie West, were dubbed the "Liar School of Herodotus".^[17]

The Liar School provoked an equally forceful reply. W. K. Pritchett's The Liar School of Herodotos of 1993 went through Fehling's demonstrations case by case; Albert de Jong, no easy believer, judged the rebuttal decisive, "a blow from which the work will not be able to recover".^[18] Pritchett's central argument is a charge of incoherence: "We find a fatal inconsistency in the arguments of Fehling, West, and other members of the liar school." The school's Herodotus lies to an audience that knows he is lying, yet invents his citations precisely to deceive that same audience into belief: a transparent entertainer and a systematic fraud at once, and he cannot be both. The rebuttal bites in the particular too: the most mocked of Herodotus' marvels, the gold-digging "ants" of India, dissolves under investigation, for "It clearly was a rodent, not an insect, and there is nothing 'fictive' about the creature"; the Himalayan marmot really does throw up gold-bearing soil, and the Greek "great ant" is a name no stranger than the "great sparrow" that named the ostrich.^[18] Others sought not to refute Fehling but to reframe the question. Stephanie West, herself once counted among the sceptics, drew the distinction that has largely held: even if Herodotus sometimes invented the citation of a source, that does not by itself mean he invented the information, which may still be sound however it was dressed.^[19] Nino Luraghi's 2001 volume gathered work that placed the Histories in the oral, social and intellectual culture of its time and asked how orality and literacy shaped knowledge of the past.^[20] Luraghi's own position amounts to a genuine third way between honest notebook and literary fiction. The collective citations, "the Persians say", "the Egyptians say", are on this reading "references to the social surface of such knowledge, that is, to the several groups that consider each single story or piece of information to be true and significant": neither interview minutes nor inventions, but markers of whose past a story is, the narrator "letting the human groups whose past he relates take the stage and tell the story".^[21] Maurizio Giangiulio's verdict on invention splits the difference with care: "Although Herodotus surely did not invent the very content of the information he gives in his text, he used what he had at his disposal to construct narratives that appear more coherent and detailed than any tradition" his local informants ever kept, a conclusion he observes is even compatible with Fehling's principle of the obvious source.^[22] Behind these lay Momigliano's caution that ancient readers did not always come to a history expecting, or wanting, literal truth, so that measuring Herodotus by a Rankean yardstick mistakes the genre.^[17]

The upshot is not a verdict but a spectrum. Herodotus' standing as a storyteller and literary artist is scarcely questioned; as a historian by modern standards assessments have placed him, in Matt Waters' phrase, "along the entire spectrum from keen observer to naïve dupe", from antiquity to the present.^[23] A. R. Burn's older verdict strikes the same balance: Herodotus is not strictly the father of history, Burn allows, but "the first great artist or architect of history, the scale of whose edifice dwarfed all that had been done before".^[24] Asheri, weighing the whole reception, sets against the "father of lies" a considered claim for the "father of history" in a stricter sense. Herodotus, he argues, taught later historians a discipline none had stated before him: that a problem must be researched, that research rests on evidence, that evidence must be weighed for reliability, that direct observation outranks hearsay, that sources deserve to be reported faithfully even when doubted so that they can later be tested, and that established facts must be convincingly explained. For all his debt to his Ionian predecessor Hecataeus, Asheri concludes, Herodotus is "the true 'first inventor' of the critical method".^[1] The modern reader who wants to use him need not settle the label; the point of the debate is that reliability in Herodotus is not a single quantity but varies sharply by the kind of matter in hand, which is what the rest of any assessment must show.

For the Achaemenid empire Herodotus is the earliest and fullest narrative source, and the shape of his knowledge is uneven in ways the reader must keep in view. Amélie Kuhrt sets out the frame: his aim was to commemorate the Greek victories of 490 and 480 to 479, so his fullest and most reliable information clusters in the earlier phases of the empire and along its north-western frontier, the zone where Persia met the Greek world.^[25] Beyond that, his account is organised by a set of Greek didactic themes that bend the material as they run through it: freedom against servitude, the hubris of imperial expansion, the decay of monarchy into despotism and so to disaster.^[25] These are not neutral reporting frames; they decide which episodes are told and how.

The correction Kuhrt urges is not to dismiss the distortion but to read through it. What the Greek sources give, she argues after Bosworth, is often not invention but the "spin" put on real events, so that a hostile or dramatised account can still preserve a substrate of information once the moralising is discounted.^[25] Herodotus was also no lone explorer with news of unknown worlds; he wrote for an audience already argued over by others who had written on Persia, Charon of Lampsacus, Dionysius of Miletus, Hellanicus of Lesbos, and his Persian material is deployed inside the intellectual debates of the sophistic age rather than offered as raw travelogue.^[25] Christopher Tuplin has distinguished four registers of Achaemenid material in the Histories, information on religious and social customs, administrative data, narratives of the growth of power, and discussions of the character and institutions of the Persian state, and each has to be judged on its own footing.^[26] The consequence for method is exact: every Persian logos has to be read as a Greek literary construction with a documentary core, and the core reached, wherever possible, by testing the text against the Persians' own record.

The test cuts both ways, and the method needs its failure case beside its confirmations. The canonical one is the death of the Apis bull. Herodotus reports that Cambyses, maddened, stabbed the sacred calf of Memphis, which "lay in the temple and died of the blow on the thigh", and that "the priests buried him without Cambyses' knowledge".^[27] A funerary stela and sarcophagus from Cambyses' reign instead record full royal honours for an Apis burial. In Flower's summary, "the epitaph on the grave stele and the inscription on the sarcophagus of this very Apis survive, and they record that Cambyses buried the bull with elaborate funeral rites". Flower follows the strong twentieth-century rehabilitation: the monuments undo the portrait of blanket sacrilege, and he suspects that Herodotus was "grossly misled by his Egyptian sources", priests with every reason to blacken the conqueror.^[3] The chronology of the bulls, however, no longer permits the stronger conclusion that the officially honoured Apis must have been Herodotus' alleged victim. A successor buried under Darius I was said to have been born before that Apis died, and a separate young calf has therefore been proposed; the possibility remains unproved. The contemporary monuments carry priority for the burial they attest, yet do not prove that no other outrage occurred.^[28] "The reader should thus be aware", Flower concludes, "that Herodotus' narrative cannot necessarily be taken at face value as a factual record of early Persian history."^[3] Even where the record confirms him, confirmation has a ceiling: institutions can be corroborated, the narrative built over them cannot. One could hardly call Dickens' A Tale of Two Cities a work of history merely because there was a French Revolution and a king named Louis XVI, and, as Flower puts it, "the kind of deep corroboration of Herodotus' narrative that would confirm its status as non-fiction is unattainable".^[3]

The single richest ethnographic set-piece for the historian of Persia is the short account of Persian religion and custom in the first book. Herodotus opens with what strikes a Greek most, a series of absences:

It is not their custom to make and set up statues and temples and altars, but those who make such they deem foolish, as I suppose, because they never believed the gods, as do the Greeks, to be in the likeness of men; but they call the whole circle of heaven Zeus, and to him they offer sacrifice on the highest peaks of the mountains.

Herodotus, Histories 1.131, trans. Godley^[29]

This is the earliest outside description of Iranian worship, and it is right in its main lines: an aniconic, open-air cult, without cult images or temple buildings in the early period, directed to the sky and to the elements. Where Herodotus adds a rite he is again a useful witness. No sacrifice, he reports, may be made without a priest to chant over it, and the worshipper may pray for no good thing for himself alone:

A Magus stands beside it and chants a history of the gods, which is how they refer to their invocation. Indeed, it is not their custom to perform sacrifices without having a Magus present.

Herodotus, Histories 1.132, trans. Purvis^[30]

The detail that the sacrificer must pray "for the welfare of the King and the entire population of Persians", not for himself, catches something real about a religion bound up with the person of the king, and matches the king-centred piety of the royal inscriptions.^[30][25] Herodotus is drawn, too, to the marks of difference that make a Greek reader look twice: the Persians deliberate on weighty matters twice over, once drunk and once sober, and adopt only what pleases them in both states; they hold their birthdays the greatest of feasts; and, in a phrase he plainly relished, no nation so readily takes up foreign customs, the Median dress and the Egyptian armour worn without shame.^[31]

Two of his data reach to the heart of Persian ideology. The making of a Persian he reduces to three things:

They educate their boys from five to twenty years old, and teach them three things only, riding and archery and truth-telling.

Herodotus, Histories 1.136, trans. Godley^[31]

And the deepest disgrace is falsehood: "They hold lying to be foulest of all, and next to that debt", the debtor being despised, he adds, because need drives him to lie.^[31] The horror of the lie is the ethical core the royal inscriptions carve as the contest of Truth against the Lie, and Herodotus, from outside and in Greek, has caught its outline. The convergence is the more striking because he could not have read the inscriptions that say the same. Two more of his observations find the same echo. The claim that Persians honour other peoples in proportion to their nearness, holding the farthest-dwelling the basest, "finds corroboration in those Persian royal inscriptions that list the countries and peoples over which the King held sway", lists that with a single exception begin with the peoples nearest Persia and move progressively outwards; and his picture of a kingship bound to continual conquest, each king adding to the empire, "is confirmed in a general way by Persian royal inscriptions".^[3]

Yet the chapter is a mixture, and the Persian record both confirms and corrects it. Kuhrt notes that the headline claim, no images and no altars, is contradicted in detail by the excavated fire-altars and by seal designs from the Persepolis Fortification archive that show worship before divine statues; the "no temple" point survives, but the flat aniconism does not.^[25] The god Herodotus calls "Zeus" is Ahura Mazdā, and here the interpretatio graeca, the habit of naming a foreign god by his nearest Greek equivalent, both helps and misleads. Albert de Jong, whose commentary weighs the chapter clause by clause, shows that Herodotus' inference is his own and is wrong: told that the Persians used no images, Herodotus concluded that they did not conceive their gods in human form, whereas Avestan text and Achaemenid art alike are anthropomorphic; the deduction imposes a Greek logic on evidence that says nothing of the kind about Persian belief.^[32] De Jong's overall verdict on the passage is measured and worth keeping in mind for the whole of Herodotus on Persia: it is important above all for its great antiquity rather than its informative depth, yet it preserves genuine observations, and the individual cults it names can be corroborated from the Iranian side even where its picture of the religion as a system is built on Greek assumptions about how cults spread.^[32]

The most quoted single slip sits in this same chapter, and it is instructive precisely because it is small. Reporting that the Persians had learned to sacrifice to a heavenly Aphrodite, Herodotus lists the goddess's names among the nations: she is "called by the Assyrians Mylitta, by the Arabians Alilat, by the Persians Mitra".^[29] But Mithra is no goddess of love. He is a major male Iranian god, of the covenant, the oath and the sun, and no Iranian source makes him female or identifies him with Anāhitā, the goddess whose functions better fit the passage. De Jong reports the near-consensus of specialists: Herodotus has introduced the name of Mithra and made a mistake in presenting him as a female divinity, whether by confusing him with Anāhitā or by conflating a closely associated divine pair.^[32] The error is exactly the kind an outsider makes, working at second hand and fitting an unfamiliar pantheon onto a familiar one; it is a fair measure of the limits of Herodotus' grasp of Persian religion, and a warning that his theology under its Greek names must always be handled with care.

Three bodies of Persian matter make Herodotus essential for Darius' reign, and the first is the accession. Herodotus tells how Cambyses died, how a Magus impersonating the dead prince Bardiya (his "Smerdis") seized the throne, how seven Persian nobles conspired and killed the usurper, and how Darius emerged as king. The killing became, he says, a Persian festival:

They celebrate a great festival on it, which they call the Massacre of the Magians; while the festival lasts no Magian may come abroad, but during this day they remain in their houses.

Herodotus, Histories 3.79, trans. Godley^[33]

The narrative can be set directly against Darius' own version, carved on the rock at Behistun, and the agreement is one of the strongest single points in Herodotus' favour. The essentials match: the usurper is a Magus posing as Cambyses' murdered brother, the murder was kept secret, a small band of nobles brought him down. More striking still is the prosopography. Herodotus' list of the conspirators corresponds, name for name with a single exception, to the helpers Darius names at Behistun; of the six nobles besides Darius, five, Otanes, Gobryas, Hydarnes, Intaphrenes and Megabyzus, are the same men, and only Herodotus' sixth, Aspathines, differs from Behistun's Ardumaniš, a substitution Stephanie West plausibly puts down to Aspathines' later prominence.^[1][19] Herodotus does markedly better here than Ctesias, whose list has drifted further from the record.^[19] Kuhrt confirms the match from the inscription's own roster of Darius' helpers.^[25]

The confirmation is real but must be read with a caution Briant states sharply. Agreement between Herodotus and Behistun does not prove either true, because Darius' inscription is itself a partisan document, the victor's account of a seizure of power, and both the Greek tradition and the monument may descend from the same official version put out by the winning faction.^[15] The stronger form of the doubt is now widely held: in Flower's summary, "it is commonly argued that the false Smerdis truly was the son of Cyrus and brother of Cambyses, and that Darius denied this in order to legitimate his own seizure of the throne", the case Briant himself develops at length, though Christopher Pelling has challenged the consensus.^[3][15] Nor do text and monument agree even at the level of event: Herodotus has Cambyses kill his brother after reaching Egypt where Behistun sets the murder before the expedition, and places the killing of the Magus at Susa where Darius puts it in Media.^[3] Where Herodotus departs from the terse inscription he is at his most Greek: the set-piece by which the impostor is unmasked, the concubine sent to feel in the dark whether the sleeping Magus has had his ears cropped, is, in Briant's words, more fairy tale than history, a folk-motif of the kind that clusters in Herodotus' Persian court-tales.^[15] Even the tale's anchoring detail turns out to be Greek: Momigliano, reporting a demonstration of Albert Demandt's, noted that an unmasking by mutilated ears "makes sense only within the Greek iconographical tradition", and confessed to finding the result disturbing, having himself always taken the scene for a typically Oriental story.^[34] The pattern will recur across the Persian material: the skeleton of events is often sound, the flesh of motive and scene often invented.

Attached to the accession is the most argued passage in all of Herodotus on Persia, the debate in which three of the conspirators weigh the merits of democracy, oligarchy and monarchy before the winners settle on kingship. Herodotus knew it would strain belief and said so:

When the tumult was abated... the rebels against the Magians held a council on the whole state of affairs, at which words were uttered which to some Greeks seem incredible; but there is no doubt that they were spoken.

Herodotus, Histories 3.80, trans. Godley^[35]

Otanes speaks for popular rule, and his praise of it turns on a word that is unmistakably late-archaic Greek political theory:

But the virtue of a multitude's rule lies first in its excellent name, which signifies equality before the law; and secondly, in that it does none of the things that a monarch does.

Herodotus, Histories 3.80, trans. Godley^[35]

Darius carries the day for monarchy with an argument as schematic as the rest:

I hold that monarchy is by far the most excellent. Nothing can be found better than the rule of the one best man.

Herodotus, Histories 3.82, trans. Godley^[35]

And the vote is counted like an Athenian show of hands: "Having to judge between these three opinions, four of the seven declared for the last."^[35]

The historicity of this debate is the crux of the whole "Herodotus on Persia" question, because the two readings pull opposite ways. On one side stands what most scholars conclude: the vocabulary is Greek, isonomia, equality before the law, being the slogan of free constitutions in fifth-century Greek thought and unimaginable in the mouths of Persian grandees in 522.^[1] Maria Brosius states the sceptical case at its bluntest: the constitutional debate is an expression of Greek political thought, there was never any choice of constitution or of who should be king, and Darius controlled events throughout; Otanes was never a candidate for the throne.^[36] Briant agrees that the tone and vocabulary are typically Greek, and that behind the constitutional set-piece lies at most a real dynastic question, who should rule in the absence of a direct heir, not a philosophy seminar.^[15] Flower's formulation of this pole is the crispest: that Herodotus twice insists, against his own readers' incredulity, on Otanes' motion to abolish monarchy in favour of democracy "shows how little he understood the mentalité of sixth-century Persia".^[3] On the other side, the debate is not simply a foreign body dropped into Persian history. Asheri's commentary reads it as an intentional integration, placed at the centre of the third book as an interpretive key to the great crisis of the succession, and notes that Herodotus himself insisted on its truth and pointed, in defence, to the Persian satrap Mardonius later installing democracies in the Ionian cities.^[1] Momigliano read the insistence the opposite way, as the very point: Herodotus "very pointedly declares the Persians capable of discussing the relative merits of democracy, oligarchy and monarchy like any well-trained Greek sophist" because his thinking is "basically committed to the mutual understanding of Greeks and Persians".^[37] Kuhrt registers the balance: some have argued that Herodotus preserves a historical reality in Greek dress, hung on the hypothesis that the false Bardiya had attacked noble privilege, but most reject strict historicity; what the debate does powerfully is dramatise the crisis the assassination opened.^[25] A grain of genuine Persian matter may survive even so, for the story ends with a real and verifiable datum, the standing privilege of the house of Otanes as the one Persian family free to refuse obedience to the king.^[25]

The second great body of Persian matter is administrative: Herodotus' account of how Darius divided the empire and fixed its tribute. He credits Darius with the invention of a systematic fiscal order, and the Persians, he says, remembered the reforming king with a shopkeeper's nickname:

It is by reason of this fixing of tribute, and other like ordinances, that the Persians called Darius the huckster, Cambyses the master, and Cyrus the father; for Darius made petty profit out of everything.

Herodotus, Histories 3.89, trans. Godley^[38]

He then lists the provinces, which he calls satrapies, with their assessments, and closes with a single grand total: "Darius collected a yearly tribute of fourteen thousand five hundred and sixty talents; I take no account of figures less than ten."^[39] It is the fullest classical account of the empire's fiscal skeleton, and it rests on something real. The prevailing view, Kuhrt reports, is that the list derives from an official document that Herodotus has "hellenised", and the figures are precise enough to be built on genuine fiscal records; Briant, weighing the same list, judges the numbers broadly reliable and drawn, however indirectly, from real archives such as those of Sardis.^[25][15]

The distortion is in the framing, not the data. Herodotus' twenty "satrapies" do not map onto the administrative reality: the list does not harmonise with the lists of peoples in the Old Persian royal inscriptions, nor with the gift-bearing delegations carved at Persepolis, and it seems to describe tax-districts rather than the governed provinces the Persians themselves recognised.^[25] The clearest sign that the order is Greek and not Persian is where it begins. Briant notes that Herodotus opens with the Ionians, whereas every Achaemenid document begins its survey of the empire with Persia, the heart of royal power; the historian from Asia Minor has organised the empire from his own western edge inward.^[15] Stephanie West draws the same conclusion, that the list cannot straightforwardly reproduce a Persian original and that the inscriptional lists of peoples are emphatically not to be equated with a division into satrapies.^[19] The famous huckster jibe, likewise, is Herodotus' own moralising colour rather than a recorded Persian saying, part of the triptych, Cyrus the father, Cambyses the master, Darius the trader, by which he characterises three reigns.^[15] The reform itself, and the debate over what stands behind the list, are weighed in the entry on Darius I.

The third body of administrative matter is the road system. Herodotus describes the great highway from Sardis up to Susa, its stages and its guarded, hostelry-lined length:

All along it are the king's stages and exceeding good hostelries, and the whole of it passes through country that is inhabited and safe.

Herodotus, Histories 5.52, trans. Godley^[40]

And the mounted courier relay that ran along it drew from him the most quoted sentence in the whole Persian ethnography:

There is nothing mortal that accomplishes a course more swiftly than do these messengers... these are stayed neither by snow nor rain nor heat nor darkness from accomplishing their appointed course with all speed... This riding-post is called in Persia, angareion.

Herodotus, Histories 8.98, trans. Godley^[41]

Here the Persian evidence confirms Herodotus and then dwarfs him. The Persepolis Fortification tablets record the sealed travel-authorisation system, the ration-issuing stations, the officials moving under royal seal, exactly the apparatus the Royal Road implies, and they document it running from Sardis to Susa and on across the eastern empire Herodotus never describes, from India to Susa, from Bactria to Susa, from Susa to Kandahar.^[25] The reality was an imperial network; Herodotus knew one road of it. Briant makes the limit explicit: in this passage Herodotus' eastern horizon extends no further than Susa and Babylonia, and he appears to know nothing of the roads across the Iranian plateau and Central Asia.^[15] A further caution comes from the older commentary of How and Wells, who observed that the Sardis-to-Susa road was far older than the empire, built on Anatolian routes the Persians inherited rather than created.^[42] The description is sound as far as it reaches; its reach is the measure of what lay outside a Greek enquirer's view.

Herodotus' Persia is also, and inescapably, a Greek self-portrait. The insight is Hartog's, whose study of the Scythian books set out to uncover "the rhetoric of otherness at work in the text, to determine some of the figures it employs, and to dissect some of its procedures; in short, to note the operative rules in the fabrication of the 'other'".^[43][44] Chief among those figures is inversion, the "handy figure" by which a traveller transcribes otherness as anti-sameness, the barbarian as the Greek turned upside down; though Hartog himself warns that the text "does not rest on a systematic, mechanical inversion", and sets the figure beside comparison, analogy, the measure of the marvellous and the labour of translation, a caution his summarisers have not always kept.^[43] The reading ties the Scythian mirror to Persia directly: the Persian Wars, placed later in the work, "play the role of matrix to the narrative of Darius's war" in Scythia, so that Darius pursuing an ungraspable enemy prefigures Xerxes among the Greeks, the Scythians standing where the Greeks will later stand.^[43] Applied to Persia the reading is older and more consequential, because the Greek image of Persia became the Western image of "the East". Heleen Sancisi-Weerdenburg put it flatly: in the Greek sources Persia is portrayed as the negative mirror-image of Greece, a well-constructed example of what a later age would call Orientalism.^[45] The most durable of these constructions is the theme of despotism and decadence, the notion that the once-hardy Persians grew soft under the luxury of empire and monarchy hardened into tyranny. Murray traces its prototype to the Lydian narrative of the first book, where Herodotus first works up the model of an eastern society and of oriental despotism before fastening it onto Persia; Brosius says Herodotus, together with Aeschylus, helped shape the freedom-versus-despotism and Europe-versus-Asia antitheses inherited by later European thought.^[6][46] Kuhrt and Sherwin-White read the despotism and decadence of Herodotus' Xerxes as a deliberately developed narrative motif rather than neutral reportage.^[47]

The decadence half of the theme, though, may belong more to Herodotus' readers than to Herodotus, and the counter-position deserves its statement. Flower argues that "it is too simplistic to say, as many moderns have done, that Herodotus attributes their defeat to softness caused by luxury." In the battle narratives the Persians, unlike their subject allies, fight bravely and to the death at Marathon, Plataea and Mycale, and Herodotus' own explanation of the defeat at Plataea is technological rather than moral: "In spirit and strength the Persians were not inferior [to the Spartans], but being without protective armour and in addition being inexperienced, they were not similar to their opponents in skill."^[3]

The mirror has an Athenian face as well: Sara Forsdyke argues that Herodotus' imperial Persia doubles as a warning about the aggressor empire of his own day, Xerxes' speech announcing the invasion running on the same logic of ancestral achievement and endless expansion by which Periclean Athens justified its rule. "Through his portrait of Xerxes, Herodotus draws a tacit parallel between Athens under Pericles and Persia under Xerxes that suggests the injustice and dangers of Athenian imperialism".^[48] On that reading the Persian logoi are at once a self-portrait by inversion and an admonition in the present tense.

Whether this makes Herodotus unfair to Persia is a question the sources answer with care. The crude reading, that the Histories set European freedom against Asian slavery as a clash of principles, is itself a modern schematisation. Asheri points out that Herodotus knew despotism to be a Greek phenomenon too and love of freedom to be found among barbarians, and that Medes, Babylonians, Egyptians and Scythians all fight for their independence in his pages; the antithesis is later ideology read back into a subtler text.^[1] Herodotus' closing chapters do the dissolving themselves: after Plataea the Spartan regent Pausanias refuses to impale Mardonius' head, such deeds being "more fitting for barbarians to do than for Greeks", yet within a few pages the Athenian Xanthippus, father of Pericles, has the Persian Artayctes crucified while his son is stoned to death before his eyes, and, as Flower observes, "the distinction between Greek and barbarian has become difficult indeed to discern".^[3] And the charge of philobarbaros cuts the other way: what Plutarch condemned as malice was Herodotus' willingness to grant the Persians their dignity and their tragedy.^[15] The last words of the whole work make the point. The Histories end not on a note of Greek triumph but on the mouth of Cyrus, warning the Persians that empire will unmake them:

Soft lands breed soft men; wondrous fruits of the earth and valiant warriors grow not from the same soil.

Herodotus, Histories 9.122, trans. Godley^[49]

The Persians, Herodotus adds, took the point and chose "rather to be rulers on a barren mountain side than slaves dwelling in tilled valleys".^[49] Flower presses the ending's second register: the Persians did take Cyrus' advice and were not unmade by the soft plains, so the close cannot encode an explanation of their failure in Greece; rather "The Histories thus end with an image of the Persians as a tough people under a wise king. They could not and should not be underestimated in the future", a live warning written while the empire stood intact and powerful.^[3] That the empire's story should close on a Persian king's tragic wisdom, rather than on the victor's boast, is the very quality the modern historian most needs, and the reason a book so often wrong remains so nearly irreplaceable.

Nowhere is Herodotus less to be believed than on numbers, and the great case is the size of Xerxes' army. Totting up his infantry, cavalry, marines and camp followers, Herodotus reaches an exact and impossible sum:

The number of those whom Xerxes son of Darius led as far as the Sepiad headland and Thermopylae was five millions, two hundred and eighty-three thousand, two hundred and twenty.

Herodotus, Histories 7.186, trans. Godley^[50]

No such host could have been fed, watered or moved. Matt Waters observes that even reduced by a factor of ten the numbers remain too high for effective logistics, and that the swollen figures serve a purpose: the greater the barbarian horde, the greater the Greek victory over it, which is why the motif of the army drinking rivers dry runs through the narrative.^[51] Llewellyn-Jones reaches the same verdict, that the catalogue is not to be read literally but as a portrait of the immensity of power Herodotus imagined the Great King to command.^[52] George Cawkwell puts the sceptical position without softening: the "countless, voiceless, hopeless" hordes must go, and with them the comic apparatus of an army numbered by being herded through a walled enclosure and of rivers drunk dry.^[53] Modern estimates for the land army run from perhaps fifty thousand to two hundred thousand, most toward the lower end.^[51]

The numbers are shaped as much as inflated, and the shaping can sometimes be caught. Herodotus' fleet of 1,207 triremes, one of the few figures many scholars once accepted, rests, How and Wells noted, on a single line of Aeschylus rather than any muster-roll; Sean Manning shows that the figure echoes both Aeschylus and the roughly twelve hundred Achaean ships of the Iliad, a literary inheritance masquerading as a count.^[42][54] Manning notes, too, the suspicious cleanness of the Marathon casualty figures, 192 Greek dead to 6,400 Persian, a ratio of exactly thirty-three and a third to one, far higher than any figure Thucydides or Xenophon reports for battles between Greeks, and reads it as a designed rather than a counted number.^[54] The clinching detail is that Herodotus treats his own impossible total as real: he computes the daily ration of grain the whole host would consume, and, How and Wells observed, even fumbles the division, dividing 5,283,220 by forty-eight and setting down a remainder for the quotient.^[42] The fantastic figure is not his alone; Ctesias independently gives the same 1,700,000 for the foot, and Dandamaev sets out the received totals as they stand in the tradition, so the numberless Asiatic army is a shared Greek habit rather than a personal quirk.^[10][55]

The set speeches raise the same problem as the numbers, in a subtler form. Herodotus fills his narrative with debates, war-councils and single addresses, none of which he could have heard, and the modern judgement is that they are compositions rather than records. Cawkwell states it plainly for the invasion: the discussions between Xerxes and his advisers, Mardonius, Artabanus, Artemisia, are almost certainly entirely fictitious and must be excluded from any reconstruction of Persian aims and strategy.^[53] The Constitutional Debate is the same phenomenon at the level of political theory, an argument in Greek terms placed in Persian mouths, defended by Herodotus against his own readers' disbelief.^[35]

To treat the speeches as lies, though, is to mistake the convention. Asheri argues that Herodotus, like Thucydides after him and by the shared practice of the new genre, composed speeches to stage opposed moral and political points of view, and that no ancient historian can be charged with dishonesty for a debate freely written up rather than transcribed; even Thucydides admitted doing as much.^[1] The speeches also carry the tragic architecture of the work. Murray shows that the figure of the wise adviser ignored, Artabanus warning Xerxes against the invasion, Demaratus explaining the Spartans, is the device by which Herodotus builds foreboding into a story whose end the reader already knows; the god who strikes the tallest, in Artabanus' warning, is the theme of the whole.^[6] The speeches are not evidence for what Persians said; they are evidence for how Herodotus thought, and for the moral pattern he found in the rise and fall of the greatest power the Greeks had faced.

To use Herodotus at all is to reckon with the kind of source he was, an enquirer working overwhelmingly from oral tradition, and the most searching modern treatment is Murray's essay on Herodotus and oral history. Even scholars who doubt it grant, Murray notes, that Herodotus at least presents his work as an oral account, maintaining the fiction of spoken testimony even where he draws on written sources.^[6] What comparative study of living oral traditions shows is a horizon. Reliable memory, unattached to a written record, reaches back about three generations, or on a generous estimate 150 to 200 years, and the pattern fits Herodotus exactly: his information is in reasonable order from his own day back to the mid-seventh century, and the period before about 650 is a realm of isolated stories and conjecture.^[6] Nearness in time does not by itself improve the record; Murray observes that Herodotus' account of Polycrates of Samos, whom he might have reconstructed from men who had known him, is no more intrinsically historical than his tale of the Corinthian tyrants at the edge of his knowledge.^[6] And accuracy of transmission is not the same as truth: a group memory transmits a fiction as faithfully as a fact.^[6]

For Persia specifically, Murray defines what channel of memory Herodotus did and did not have. He had no access to an official royal chronicle or to a priestly tradition, and his thin grasp of Persian religion shows it; what he had were the variant traditions current in the high Persian aristocracy, easier for a Greek to reach than the imperial bureaucracy, families such as that of Zopyrus whose tales lie behind his account of Darius' accession and the siege of Babylon.^[6] The character of that material is precise. The Persian logoi are not continuous history but episodic court-tales, novellas of palace plot and cruel vengeance, of harem intrigue and the wise counsellor, where women decide events and folk-motifs supply the joins.^[6] The apparent exceptions, the satrapy list, the Royal Road, the army catalogue, are lists cast in documentary form rather than genuine documents, built on the model of an archive but not lifted from one.^[6] Thomas, whose Herodotus in Context set the enquirer among the oral performers and the natural scientists of the late fifth century, supplies the wider frame: Herodotus belonged to that same world of enquiry and shared its methods, his concrete material couched in terms that could be carried directly into the more abstract debates of the sophists.^[4] What all this yields is not a licence to trust or to reject but a rule of use: a Herodotean tradition is worth what its channel of transmission is worth, and must be weighed as memory rather than read as a document.

What that rule looks like on the page is best seen in the first book, where the rise and death of Cyrus are told almost wholly in the register of legend. The founder reaches Herodotus already storied: the dreams of Astyages, the royal infant handed to a herdsman to be exposed, the boy revealed by his kingly bearing in a children's game, the war brought on by the revenge of the wronged Harpagus; then Croesus saved on the pyre; and at the last the queen Tomyris, who plunges the dead king's head into a skin of blood to avenge her son. Herodotus knows the register he is working in and says so at both ends of the reign. He opens by choosing among versions, following those Persians who "desire not to make a fine tale of the story of Cyrus" though there were "no less than three other accounts of Cyrus" he could give; and he closes it with a shrug: "Many stories are related of Cyrus' death; this, that I have told, is the worthiest of credence."^[56] Murray names the form these tales take: the rise of Persia and the usurpation of Darius are both "retold within the framework of the Palastgeschichte", the court-novella, so that "the account of Cyrus' divinely ordained rise to power is transformed by being subordinated to a story of revenge and the faithless vizier".^[6] The stories woven around Croesus are, on the same reading, "quite different in character and clearly Greek in origin", moral tales of the god who strikes down the prosperous.^[6] The worked cases, the birth-legend, the pyre, the irreconcilable deaths, are examined in the entry on Cyrus; what belongs here is the rule they teach: in Book 1 the folk-tale is not an intrusion into the history but the medium of it, and nothing told in that register, however circumstantial, can be pressed for fact.

The measure against which Herodotus was judged, in antiquity and after, was Thucydides. Thucydides began his history where Herodotus ended his, at the Greek capture of Sestos, and so inaugurated the "cycle" by which each Greek historian took up the story where his chosen predecessor left off; but he began it as a rebuke.^[1] Without naming Herodotus, Thucydides dismissed those who write for the pleasure of the moment rather than for use, and claimed for his own austere work the title of a possession for all time, a ktēma es aiei; the barb is generally read as aimed at Herodotus' public recitations.^[57][1] The nineteenth century took Thucydides' side and made the contrast damaging to Herodotus once more.

The ancient record, though, looks more hostile than ancient practice was, for a structural reason Hornblower catches in a maxim of Jacoby's: "polemic usually names its object, borrowings are anonymous". The attacks survive under Herodotus' name; the borrowings mostly do not, and their scale could be vast: Pausanias the periegete, by one calculation, took material from Herodotus without citation on at least 82 occasions. By the imperial period the quarrel had burnt out: "Herodotus was a classic who could be argued with on his own terms, pillaged, and adapted freely", and independence no longer had to be proved by calling him a fool or a liar. Plutarch's broadside came late for the same reason in reverse: under Rome, with the Parthians cast as the new Persians, the Persian-Wars tradition was still politically alive and Herodotus' version of it still worth contesting.^[11]

The modern rehabilitation answers the rebuke on its own ground. Kurt Raaflaub argues that Herodotus does implicitly what Thucydides makes explicit: he too studies the patterns of human behaviour that make history a guide to the future, holding up his ethnographic and historical material as a mirror in which his audience might understand itself, an educational purpose close to the Thucydidean one.^[58] The twentieth century recovered, too, the Herodotus the nineteenth had lost: the father of history, and no less of ethnography and Kulturgeschichte, the founder of the comparative study of peoples, a standing marked by the naming of a modern geopolitical journal Hérodote after him.^[1] The most recent turn, the narratological reading associated with Carolyn Dewald, attends to the authorial voice that interrupts the narrative hundreds of times to weigh, doubt and qualify, and reads the Histories not as an unmediated transcript of events but as a set of authorial choices held up for inspection, the sophisticated modern form of taking Herodotus' own scepticism seriously.^[59][17]

The Histories survive through a medieval Greek manuscript tradition whose oldest important witness dates to the tenth century, supported by a substantial body of papyrus fragments recovered from Egypt, some of them centuries older than any manuscript. The papyri are earlier but not automatically better, Asheri cautions; they reflect the same recension, and the language of even the best manuscripts is a mixed Ionic and Attic that no editor can restore to a certain original.^[1]

The division into nine books named for the Muses is a librarians' scaffold of the Hellenistic age, and no part of the author's plan. The Greek text re-entered western Europe through Lorenzo Valla's Latin translation, made for the papacy and printed at Venice in 1474, before the first printed Greek editions of the sixteenth century.^[1] The modern critical text long rested on C. Hude's Oxford Classical Text of 1908, whose book-chapter-section numeration remains the citation standard; the current standard critical edition is N. G. Wilson's Oxford Classical Text of 2015.^[17] Among modern English translations, A. D. Godley's Loeb of 1920 to 1925 is in the public domain and prints the Greek on the facing page; the leading current renderings are Aubrey de Sélincourt's Penguin, Robin Waterfield's Oxford World's Classics with an introduction by Dewald, and Andrea Purvis' translation in Robert Strassler's heavily annotated Landmark Herodotus. Herodotus is cited, by universal convention, by book, chapter and section: Histories 3.80 is book three, chapter eighty.