The Battle of Salamis (480 BCE)

On a day near the end of September in 480 BCE, in the narrow water between the coast of Attica and the island of Salamis, the fleet Xerxes had gathered from every maritime people of his empire was beaten by a smaller Greek force, while the King watched from a seat set up on the shore. The engagement is the best-remembered of all the battles of the Achaemenid wars in the west, and it has carried, ever since, a weight of meaning the event itself may not bear: the salvation of Greece, the triumph of free men over a despot's slaves, the turning-point of European history. That reading is Greek, and it is old, older than Herodotus. This entry sets it beside the harder question that governs the whole subject. Everything we know of Salamis comes from the winners and their descendants; there is no Persian account of the battle, not one line; and the two sources that matter most, the tragedian Herodotus read at a century's distance and, earlier still, an Athenian play performed for an Athenian audience that had fought in the battle, are both works of Greek self-celebration. To recover the Persian side is therefore not to find a lost Persian source but to read the Greek ones against their grain: to ask what the fleet actually was, why the King chose to fight in that place, what the defeat did and did not decide, and how a court that had no word for surrender would have understood a day that went wrong under the King's own eyes.

The two witnesses, and the silence behind them

The source problem at Salamis is unusually stark even for this empire. The earliest account is a tragedy. Aeschylus' Persae ("The Persians") was staged at Athens in 472 BCE, eight years after the battle, by a poet who by tradition had himself fought against the Persians (at Marathon, and quite possibly at Salamis), before an audience many of whom had pulled an oar in the straits. It is the oldest surviving European drama, and the only Greek tragedy on a contemporary historical subject; it is also, and this is the pointed fact, a play that puts the lament for the disaster into Persian mouths, staged in the theatre of the victors. Its long messenger-speech, reporting the battle to the widowed queen and the chorus of Persian elders at Susa, is the first narrative we possess of any event in Achaemenid history told by a contemporary. It is a magnificent poem and a piece of civic triumph at once, and the two cannot be separated. Its Xerxes is the type the Greek tradition fixed, the young and imprudent king misled by bad advisors, whose overconfidence in his gigantic forces and whose sacrilege bring on disaster, so that at Salamis "the freedom of Hellas was at stake."^[1] When the Persae has its messenger cry that "so great a multitude of men never perished in a single day,"^[2] we are hearing not a casualty return but Athens exulting, in the voice of the enemy, at the enemy's ruin.

Herodotus, writing his Histories two generations later, gives the fullest account, more than forty chapters of it (8.56-96), dense with named captains, quarrelling admirals, oracles and marvels. He drew on Greek informants, family traditions and the poets, and possibly on Greeks who had served in the King's fleet; his narrative is incomparably rich and, in the modern judgement, incomparably tendentious, shaped by his theme of free Hellas against Asian despotism. Later writers add variants (Diodorus, drawing on the fourth-century Ephorus of Cyme, gives a clearer because more succinct version), but they descend from the same Greek stream. There is no counter-current. As the standard corpus of the sources puts it, for the whole political and military history of this period "we are dependent on Graeco-Roman writers, with Herodotus' masterly narrative dominating the picture."^[3] The consequence for Salamis is that every motive and gesture attributed to the King comes to us already interpreted by people for whom his defeat was the best thing that ever happened.

The fleet that lost was not Persian

The first correction the sources force is the plainest and the most often missed: the fleet defeated at Salamis was not a Persian fleet in any ethnic sense. The Persians of the Iranian plateau were horsemen and foot-soldiers, not sailors; the empire's sea-power was its subject coastlands, levied and led by their own kings and captains. Herodotus' own account makes this unmistakable. When Xerxes goes down to the ships at Phaleron before the battle, the commanders who come before him and are seated "as the king had assigned rank to each one" are "first the king of Sidon, then he of Tyre, and after them the rest" (8.67).^[4] The line of battle was drawn up by nations: "Opposite the Athenians had been ranged the Phenicians, for these occupied the wing towards Eleusis and the West, and opposite the Lacedemonians were the Ionians, who occupied the wing which extended to the East" (8.85),^[4] with the Cypriots, the southern Anatolians and probably some Egyptians in the centre. The best modern reconstruction, working from Herodotus and Aeschylus' detail that the fleet was ordered in three files, places the Phoenicians on the right, the Ionian Greeks on the left, and the other Mediterranean subjects between.^[5]

The irony runs deep and the Greek sources feel it. On the losing side were Greeks in large numbers, the Ionians and islanders of the empire's Aegean coast, fighting for the King; on the winning side were the mainland Greeks who called them traitors. One of the King's most conspicuous captains was herself a Greek-speaking dynast, Artemisia of Halicarnassus, queen of a Carian city, whom Herodotus, her own fellow-townsman, singles out for admiration (below). The word "Persian" for the men who drowned in the straits is thus a category the Greek narrative supplies for its own purposes. What sank at Salamis was the naval manpower of Phoenicia, Cyprus, Egypt, Cilicia and Ionia, subjects of the King drawn from precisely the coastlands the empire had most recently and most lightly absorbed.

Themistocles' stratagem and the choice of the narrows

The Greek fleet had fallen back on Salamis after the drawn engagements off Artemisium and the loss of the pass at Thermopylae; behind it, Athens had been evacuated, taken, and its Acropolis burned. The allied captains were divided and frightened, many of the Peloponnesians wanting to sail south and defend the Isthmus of Corinth rather than fight for a land already lost. The Athenian commander Themistocles, according to the tradition, forced the issue by a ruse, sending a servant, Sikinnos, secretly across to the King's fleet with a false message: that the Greeks were about to flee, that they were quarrelling among themselves, and that Xerxes had only to bar the exits to destroy them and even to win over the medizers among them. Herodotus gives the words:

"The commander of the Athenians sent me privately without the knowledge of the other Hellenes... to inform you that the Hellenes are planning to take flight, having been struck with dismay; and now it is possible for you to execute a most noble work, if ye do not permit them to flee away." (Herodotus 8.75, trans. Macaulay)^[4]

The King, on this account, took the bait: in the night his fleet put to sea, occupied the passages, and landed troops on the islet of Psyttaleia in the channel mouth, so that by dawn the Greeks were surrounded and had no choice but to fight in the narrows. The stratagem is the hinge of the Greek story, the cunning of the free man turning the despot's own numbers against him, and its historicity is debated. Aeschylus already has the deceiving Greek (8.75 in Herodotus; verses 355-360 in the play), so the motif is old; but whether the message decided anything is another matter.

Here the Persian-side reading parts company with the Greek most sharply, not by denying the ruse but by refusing to let it explain the King's decision. Herodotus himself lets slip that Xerxes had resolved to fight before the message arrived, during the fleet's own preliminary night deployment.^[5] The choice to give battle in the constricted water, throwing away the fleet's advantages in numbers and open-sea speed, requires an explanation the trickery does not supply. Part of it was practical: a blockade would have been slow, the sailing season was closing, and the King had a long march home to think of. But the deeper logic, argued at length in the most recent study of the campaign from the Persian side, was ideological. For an Achaemenid king, battle was the appointed proof of divine favour, and the reluctant enemy penned on his island fitted an ancient royal image, inherited from the Assyrian and Babylonian past, of the heroic sovereign who crosses the waters to trap the fugitive foe "like fish in a net."^[5] The very difficulty of the narrows was part of the attraction, enhancing the display of royal fearlessness and trust in the god. On this reading the decision to fight at Salamis was not the blunder of a man gulled by a slave's message but the considered act of a king performing kingship, who could no more decline the challenge than deny his own charisma. That it went wrong is a separate fact from why it was made.

The King as spectator

Every Greek account agrees that Xerxes watched the battle, seated on high ground above the strait. Herodotus places him "sitting just under the mountain opposite Salamis, which is called Aigaleos" (8.90), with scribes at his side recording the names of captains who distinguished themselves so that they might be rewarded, and their fathers and cities with them.^[4] The reference-standard treatment of Greco-Persian relations sums the scene: "the Persians were beaten under the very eyes of Xerxes, sitting enthroned on the beach."^[6] The Greeks make the watching King the emotional centre of the battle, the audience of his own catastrophe, and Aeschylus draws the picture the whole tradition kept:

"Xerxes groaned aloud when he beheld the extent of the disaster, for he occupied a seat commanding a clear view of the entire army, a lofty headland by the open sea. Tearing his robes and uttering a loud cry, he straightaway gave orders to his force on land and dismissed them in disorderly flight." (Aeschylus, Persae 465-471, trans. Smyth)^[2]

The watching is not a Greek invention, and it should not be explained away as mere literary staging, as some have tried. Royal observation of battle was a genuine Achaemenid practice, the transposition to the field of the iconography of the royal audience, the King enthroned to receive his subjects and dispense justice.^[5] He showed himself watching because the watching was itself an act of rule: it bound the outcome to his person and his god, and his presence was meant to make his subjects fight the harder. Herodotus half-grasps this when he writes that at Salamis the King's men "surpassed themselves," "every one being eager and fearing Xerxes, and each man thinking that the king was looking especially at him" (8.86).^[4]

What the Greeks did with the spectacle was to invert it. The King who staged his presence as the guarantee of victory became, in their telling, the impotent witness of defeat, and the throne that was to have overawed the enemy now framed his humiliation. This is the crux of the Persian-side reading: the visibility of the fleet, and the proximity of the King and his court to the water, meant that when the display went wrong it could not be hidden. Staged before the King and his household, the defeat was not merely material but an unintended spectacle of failure, and it fell on the one point where Achaemenid ideology was most exposed, the claim that the King ruled the sea by the favour of the god.

The battle

The fighting was long, confused and murderous, and no ancient writer can trace the movement of particular squadrons. The Greeks, drawn up in the narrow water with their flanks protected by the island and the shoals, met the King's fleet as it pressed up the channel. Aeschylus' messenger, the earliest witness, describes the opening and the ruin:

"It was a ship of Hellas that began the charge and chopped off in its entirety the curved stern of a Phoenician boat. Each captain drove his ship straight against some other ship. At first the stream of the Persian army held its own. When, however, the mass of our ships had been crowded in the narrows, and none could render another aid, and each crashed its bronze prow against each of its own line, they splintered their whole bank of oars." (Aeschylus, Persae 409-416, trans. Smyth)^[2]

The decisive factor, on this account and on the modern reconstruction alike, was crowding. The King's larger fleet, funnelled into the constricted channel and formed in successive lines, lost the room to manoeuvre; as the front line took losses and tried to withdraw, the ships behind pressed forward for their own chance to fight under the King's eye, and the two collided. Herodotus makes the same point: "those who were stationed behind, while endeavouring to pass with their ships to the front in order that they also might display some deed of valour for the king to see, ran into the ships of their own side as they fled" (8.89).^[4] The Greeks, fewer and in order, held their line and rammed; the King's fleet, its formation broken, was battered piecemeal, the sea so thick with wreckage and bodies that the messenger recalls the Greeks in small craft striking the swimmers "as if our men were tuna or some haul of fish" (Persae 424-426).^[2]

Yet even here the Persian side is not simply the story of a rout. The King's own Greek and Phoenician ships rammed too and took prizes, and several Greek vessels were boarded and captured before the day turned; two Samian captains in the King's service, Theomestor and Phylakos, were rewarded afterwards, one made tyrant of Samos, the other enrolled among the orosangai, the "benefactors of the king," and given land (8.85).^[4] The fleet inflicted real damage before it broke, a fact the Greek insistence on their own light losses tends to bury. Among the notable dead on the King's side was Ariabignes, a son of Darius and a half-brother of Xerxes, who commanded the Ionians on the left; his death, and that of the Cilician king Syennesis and the Calyndian king Damasithymos, meant the loss was felt at once in the King's own family and among his client rulers (8.89).^[4]

Artemisia of Halicarnassus

One figure dominates Herodotus' account out of all proportion to her force of five ships, and she is the sharpest single instance of the Greek sources' complexity: Artemisia, queen of Halicarnassus and its neighbouring Carian towns, a woman fighting for the King, and Herodotus' own compatriot. Before the battle she alone of the commanders advises against fighting at sea at all, and her speech, as Herodotus stages it, is a clear-eyed reading of the fleet's real weakness, its dependence on unreliable subject navies:

"Spare thy ships and do not make a sea-fight; for the men are as much stronger than thy men by sea, as men are stronger than women... thou, who art of all men the best, hast bad servants, namely those who are reckoned as allies, Egyptians and Cyprians and Kilikians and Pamphylians, in whom there is no profit." (Herodotus 8.68, trans. Macaulay)^[4]

The King honours the advice but follows the majority, thinking that his men had held back off Euboea only because he was not there to see, and "now he had made himself ready to look on while they fought a sea-battle" (8.69).^[4] In the battle itself Artemisia, pursued by an Athenian ship and unable to escape, rams and sinks a ship of her own side, a Calyndian vessel carrying the king Damasithymos; the Athenian captain, supposing her a Greek or a deserter, turns away, and no one of the sunk crew survives to accuse her. Xerxes, watching and told that the ship she sank was an enemy's, is said to have exclaimed the line the tradition preserved above all others:

"My men have become women, and my women men." (Herodotus 8.88, trans. Macaulay)^[4]

The Athenians, Herodotus adds, were so galled that a woman should sail against their city that they set a prize of ten thousand drachmas on her capture (8.93).^[4] The Artemisia material is a caution against reading the sources flat: it is at once a genuine memory of a Carian queen who commanded at Salamis, a piece of Halicarnassian local pride in Herodotus' own hand, and a device by which the historian scores his heaviest points against Persian masculinity, the King's men shamed by a woman, the King himself gulled into praising a friendly ship's destruction as a feat of arms. The queen fighting for the King is thus both the most vivid actor in the Persian fleet and a figure shaped, at every turn, to the Greek narrative's purposes.

What Salamis did and did not decide

The popular memory of Salamis is that it saved Greece and finished the invasion; neither is quite true, and the Persian side of the account depends on seeing why. The war did not end. Xerxes withdrew to Asia with much of the army, but he left his kinsman Mardonius in Greece with a picked force to winter in Thessaly and renew the campaign in the spring, and it was the following year, 479, that the issue was actually settled, on land at Plataea, where Mardonius was defeated and killed, and on the Ionian coast at Mycale. Salamis was a naval defeat in a war decided by armies a year later; it is the later, monumental memory that compressed the whole into one glorious day in the straits.

Nor was the fleet annihilated. The scale of loss is genuinely uncertain, because the Greek sources, distant from the events and celebratory in tone, had no access to Persian records and did not even try to count the King's dead ships, preferring anecdote. Aeschylus proclaims total destruction; Herodotus implies a loss of perhaps half, yet reports the fleet still fielding some three hundred ships the next season; Diodorus gives more than two hundred lost. Comparison with the loss-rates of later Greek naval battles suggests the true figure lay somewhere in this range, serious but not catastrophic, and that the King's fleet, though badly hurt, inflicted significant losses of its own and remained a force the Greeks still feared.^[5] The synthesis of the reign puts the military verdict bluntly: "despite the resounding defeat, the military outcome was not catastrophic. The Persian army was practically intact... As for the navy, it was certainly not completely destroyed: the Greeks still feared it."^[7] The picture of a shattered armament fleeing home is a Greek dramatisation, which Herodotus half-undoes himself by preserving rival stories of the return and recording that Thessaly, Paeonia and Macedon stayed loyal along the King's route.^[3]

What Salamis did decide was harder to weigh and, from the Persian side, more real than any ship-count. It broke the momentum of the invasion at the political level, and it broke something in the royal image. The King had assembled the greatest fleet the world had seen, in a display staged as proof that he ruled earth and sea by the god's consent, and then, in his own sight, he lost. In the terms of Achaemenid ideology this was close to unthinkable, for the standard explanation of a military defeat across the whole Near East was the abandonment of the ruler by his god. As the modern analysis frames it, the formula of the Behistun inscription now played out in reverse: the god did not help, the King's fleet did not defeat the enemy, the enemy did not flee, there was no pursuit and no capture of the rebel leaders.^[5] The defeat also strained the empire's cherished picture of its subjects brought together in common purpose by the King: Herodotus' anecdotes of the battle's aftermath, the Phoenicians denouncing the Ionians as traitors and being beheaded for it (8.90), Artemisia sinking a fellow-vassal's ship, the contingents blaming one another, hint, if they reflect real post-battle rumour, at a corrosive effect on the ideal of imperial cohesion.^[5]

The Greek picture of the King, and how to read it

The most seductive and least trustworthy element in the sources is their portrait of Xerxes' own conduct: the panicked despot, raging and weeping, executing scapegoats, fleeing in terror lest the Greeks break the bridges over the Hellespont and trap him in Europe (8.97). Herodotus carries the news of the sea-fight to Susa in a second messenger, at which "they all tore their garments and gave themselves up to crying and lamentation without stint" (8.99);^[4] Aeschylus supplies the robe-tearing and the disordered flight. This is, in the modern judgement, "sheer fiction, running contrary to every Near Eastern precedent for reactions to defeat and management of the royal image."^[5] The first rule of a military setback for an Assyrian, Babylonian or Persian king was to not acknowledge it: Darius at Behistun records nineteen victories and never a single casualty. A real Achaemenid response would have run to propitiation and purification, to the transfer of blame onto subordinates balanced against rewards for the loyal, and above all to the direction of attention toward the campaign's genuine successes and the victory still expected from Mardonius; of the rituals the Greek sources say nothing, preferring stories of royal cowardice and cruelty that better suited the image of hubris brought low.^[5]

The retreat itself was neither flight nor rout. Xerxes had always meant to return to Asia after the taking of Athens; the march back to Sardis took months and was followed by a long residence there, from which the King continued to direct the war.^[7] Delegating a war's completion to a general while the King began his homeward progress had ample precedent, and would, had Mardonius won, have added to the King's own credit as the victories of Darius' lieutenants had at Behistun.^[5] Herodotus stages the post-battle council as a debate in which Mardonius offers to stay and subdue Greece with three hundred thousand men, and Artemisia, consulted again, advises the King to withdraw and leave Mardonius to the risk:

"If he subdue those whom he says that he desires to subdue... the deed will after all be thine, master, seeing that thy slaves achieved it: and on the other hand if the opposite shall come to pass... it will be no great misfortune, seeing that thou wilt thyself remain safe." (Herodotus 8.102, trans. Macaulay)^[4]

The speech is Herodotus' composition, but its logic is sound Achaemenid strategy, and the course it recommends is the one actually taken. That the historian could put a clear and correct royal calculation into Artemisia's mouth, in the same pages that mock the King as a weeping fool, is a fair measure of how much the Greek sources contain and how carefully they must be read: the facts are often reliable, the frame around them almost never neutral.

Why the reading matters

Salamis concentrates, in a single day and a single silence, the difficulty of studying this empire from the ground its enemies won. The battle was real, its outcome was real, and the damage to the King, material and ideological, was real; but almost every word we have about it was written to celebrate the men who inflicted that damage, the earliest of them in a play staged in the theatre of the city Xerxes had burned, with Persians made to mourn before an Athenian audience. To read Salamis from the Persian side is not to deny the defeat but to strip away the freight the victory later carried: to see the fleet as the subject navies it was, to grant the King a considered reason for fighting where he did, to measure the loss against the war's real length and the fleet's real survival, and to locate the true wound not in the arithmetic of sunken ships but in a display of divinely warranted power that failed under the King's own eyes, and had then to be repaired by the further victory that never came.