In the last weeks of 1098, on the edge of a marsh in northern Syria, a priest watched men wade into cold standing water to drag out corpses. He knew exactly whose corpses they were. Two weeks earlier, his own army had thrown them in. These were the dead of a town called Maarat al-Numan, killed when the town fell, hauled outside the walls and dumped into the swamp so the streets could be cleared.

Now, in the middle of December, with the food gone and the countryside stripped bare, the same army was pulling those bodies back out of the water. The flesh had been rotting for a fortnight. The priest wrote down what the men did with it anyway. His name was Raymond of Aguilers.

He was chaplain to one of the most powerful noblemen in Europe. And the sentence he set down about that marsh is one of the most disturbing things ever written by a participant in a medieval war. He said the people ate those fetid bodies. And he said they ate them greedily.

He was not the only one who wrote it down. A knight’s chaplain, an anonymous soldier, a priest from Chartres, a monk in Normandy interviewing veterans, an abbot in France reworking the story a decade later, and most remarkably, the commanders of the army themselves in an official letter to the pope. Nearly every man who sat down to write an account of the First Crusade in the twenty years after it happened had to deal one way or another with the fact that at a small Syrian market town in the winter of 1098, Christian holy warriors ate the dead. Some of them place it during the siege.

Some place it after. Some say it was done in shameful secret. One insists it was done in the open on purpose in front of horrified onlookers. One claims he heard about it from the mouths of the men who did it.

What he heard involved cooking pots and spits. They do not agree on much. They agree that it happened. That town, Maarat al-Numan, sits in what is now Idlib province in Syria, about twenty miles south of Idlib City, on the old road between Aleppo and Homs.

People have lived there for a very long time. Its name folds together an old Aramaic word for cave, after the caverns that honeycomb the ground beneath it, and the name of an early Muslim governor, a companion of the Prophet. Before the crusaders arrived, it was a prosperous, ordinary place ringed with orchards of figs, olives, and pistachios, with a wall, a ditch, markets, mosques, and somewhere under ten thousand people who had every reason to believe the wall would hold. The men wading into that water had been three years earlier farmhands and small-town knights, most of whom never expected to leave the villages they grew up in.

The road that ended in the swamps of Maara began on the 27th of November 1095 outside the town of Clermont in the Auvergne, where Pope Urban II stood up in front of a crowd of clergy and laymen and asked the knighthood of Europe to march east. We do not actually know exactly what he said. Four or five versions of the speech survive, all written down years later by men who already knew how the story ended, and each version conveniently emphasizes whatever its author thought the crusade had turned out to be about. But the core of it is recoverable.

The Byzantine Emperor Alexios Komnenos had sent envoys asking for Western soldiers to help push back the Seljuk Turks who had swallowed most of Anatolia. Urban took that modest request for mercenaries and turned it into something enormous: an armed pilgrimage to liberate the eastern churches and Jerusalem itself, with a promise attached that would echo for centuries. March east, fight, and the penance for your sins is remitted. Die on the road, and you die absolved.

You have to sit for a second with what that offer meant to an eleventh-century mind. These were people who lived in genuine daily terror of hell, knights especially, men whose entire profession consisted of exactly the violence the church kept telling them was damning their souls. Urban was offering them a way to do the one thing they were good at—killing—and have it count for their salvation rather than against it. Chroniclers described the crowd at Clermont shouting “Deus vult, God wills it,” and men tearing strips of cloth into crosses to sew onto their shoulders right there in the field.

Whatever Urban expected, what he got was beyond anyone’s planning. Modern estimates for the armed pilgrimage that eventually gathered run somewhere between sixty and eighty thousand people, with the actual knights, the trained heavy cavalry, numbering perhaps seven to ten thousand. The rest were foot soldiers, servants, grooms, priests, monks, wives, children, the elderly, the poor. Entire families sold everything they owned to buy a cart and go.

That detail matters more than almost anything else. A huge fraction of the people who marched toward Jerusalem were not fighters at all. They were the poor. And the poor, three years later, are the ones history would blame for what happened at Maara.

The first wave to move did not even wait for the official departure date. In the spring of 1096, a ragged mass of tens of thousands set off behind a small, filthy, magnetic preacher called Peter the Hermit, a man who rode a donkey and whose followers treated the hairs of that donkey as relics. This so-called People’s Crusade lurched east through the Rhineland, and on the way parts of it committed the first atrocities of the whole enterprise, turning on the Jewish communities of Worms, Mainz, and Cologne. Thousands of Jews were murdered in their homes and synagogues by men wearing crosses, some given the choice of baptism or death, some not given any choice at all.

Bishops tried and mostly failed to protect them. So the crusade carried a capacity for massacre inside it from the very first weeks, long before anyone was starving. Hunger will explain a great deal of what happens at Maara. It will not explain all of it.

Peter’s horde reached Constantinople in August 1096, looted its way through the suburbs, and was ferried across the Bosporus by an emperor who wanted them gone. In Anatolia, they behaved less like pilgrims than like a plague, torturing and killing local Christians as readily as Turks. In October, the Seljuk Sultan Kilij Arslan dealt with them. At Civetot, his horsemen caught the main body in a valley and destroyed it in a morning, killing everyone where they stood and marching the young and pretty captives off to the slave markets.

Out of perhaps twenty thousand people, a few thousand survivors huddled in a fortress by the sea until Byzantine ships took them off. Peter the Hermit survived, having conveniently been back in Constantinople begging supplies when the blow fell. The real armies came on behind through the autumn and winter of 1096 into 97. Four or five great contingents, each behind some of the most formidable noblemen in Europe.

Godfrey of Bouillon, Duke of Lower Lorraine, pious and hard, richer in reputation than in land. His brother Baldwin, colder and cleverer, who would end up a king. Raymond of Saint-Gilles, Count of Toulouse, the oldest and wealthiest of them all, a man in his fifties who had fought Muslims in Spain, who brought his wife and his chaplain, our marsh-side witness Raymond of Aguilers, and who let it be known he intended to die in the Holy Land. Robert of Normandy, son of William the Conqueror, amiable and chronically broke, who had pawned his entire duchy to his brother to fund the trip.

Robert of Flanders. And then the one to watch: Bohemond of Taranto, a Norman from southern Italy, physically enormous by the standards of the age, charming, ruthless, and very publicly short of a principality. Bohemond had spent years fighting against the Byzantine emperor he was now marching to assist. And nobody, least of all the emperor, believed for one second that he was in this for the remission of his sins.

Bohemond and Raymond of Toulouse are going to spend this entire story circling each other like two dogs over one bone. And the town of Maara, along with everyone in it, is going to end up between their teeth. With them traveled the papal legate, Adhemar, Bishop of Le Puy, the one man with the moral authority to hold this collection of egos together. And in the long train behind the knights, that mass of the poor: tens of thousands of them on foot, pushing barrows, carrying children.

Every mile of the next two years is going to grind that column down. When the food runs short, the poor starve first. When the horses die, the knights walk. But the poor were already walking.

By the time this army reaches the walls of Maara, it will have been hungry off and on for two years, and some of its people will have crossed lines that men in the chronicles could barely bring themselves to describe. The armies converged on Constantinople through the winter of 1096 and into the spring of 97. Alexios handled them the way you handle armed guests you do not trust: fed them, flattered them, extracted oaths from their leaders that any former imperial territory they captured would be returned to him, and shipped them across the water as fast as decency allowed. Bohemond swore the oath with a straight face, and everyone did the math on what that was worth.

A portrait of Bohemond survives from those Constantinople weeks, drawn from life, and the man it describes is going to preside over the worst broken promise in this story. Anna Komnene, the emperor’s daughter, was a teenager when Bohemond passed through her father’s court. Decades later, writing her history, she produced the most vivid physical description of any figure of the age, a passage in which fascination and loathing wrestle openly. He was, she says, a full head taller than the tallest men, white-skinned and flushed, his yellow hair cropped short against the fashion, clean-shaven, wide in the chest, and something about him, she insists, was hard and savage even in his charm, his very laugh sounding to her like a threat.

Her father took the measure of him precisely, showing him a room filled floor to ceiling with treasure and then giving it all to him, a gift calculated to expose what the Norman was made of. Their first target was Nicaea, the old imperial city the Turks had made their capital barely a hundred miles from Constantinople. The siege gave the crusaders their first taste of what this war was going to be. When a Turkish relief force was beaten back, the crusaders cut the heads from the enemy dead and lobbed them over the walls into the city with catapults.

And then, just as they prepared their grand final assault, they looked up one June morning and saw the emperor’s banners flying from the towers. Alexios had negotiated the city’s surrender behind their backs, specifically to deny them the sack they were counting on. To a hungry army that had bled at the walls and expected the customary reward of plunder, it felt like theft. The resentment against the Greeks that took root at Nicaea never healed, and it fed directly into the disaster of discipline and supply that put men in a position eighteen months later to eat the dead.

A week after Nicaea fell, the army set out across Anatolia in two columns. On the 1st of July, at Dorylaeum, Kilij Arslan tried to erase them. His horse archers caught Bohemond’s column alone at dawn and enveloped it, wave after wave of arrows pouring into a mass of knights, pilgrims, women, and baggage animals compressed against a marsh. For five or six hours, Bohemond’s people simply endured it.

Then Raymond’s and Godfrey’s column came over the hills into the Turkish flank, and the Sultan’s army broke and ran. The crusaders looted the Turkish camp and found it full of treasure. What they could not loot was food. It was after Dorylaeum, in the burned and evacuated highlands of central Anatolia in August, that the great hunger of this story really begins.

Kilij Arslan could not beat them in the field, so he did the older, crueler thing: he emptied the land in front of them. Wells filled in or fouled, crops burned, villages abandoned. The army marched for weeks through country that gave them nothing, in heat the northern Europeans had never experienced. And it started to die, not dramatically in battle, but by the roadside of thirst and exhaustion.

On the worst stretch, chroniclers describe men and women collapsing with blackened mouths, pregnant women going into labor beside the track, and mothers so maddened by thirst that infants were abandoned where they were born. The horses died, killing the entire basis of these men’s way of fighting. Knights ended up riding oxen. Baggage was strapped to goats, to pigs, to dogs.

An army that had left Europe with several thousand of the finest horses in Christendom would arrive at its next great test with a few hundred. They came down out of the mountains in October 1097 and arrived, diminished and ragged, before one of the great cities of the ancient world: Antioch. And here the story slows right down. Maara makes no sense without Antioch.

The men who ate the dead in December 1098 had spent the previous fourteen months learning at Antioch what the absolute floor of human endurance looked like. Almost everything that happens at Maara happens first, in embryo, at Antioch: the famine, the eating of things no one should eat, the mutilation of corpses, the massacre in the streets, even in some accounts the first quiet consumption of human flesh. Antioch was the rehearsal. The city sat on the Orontes River, and its fortifications were like nothing any of these Westerners had ever seen: walls built under the emperor Justinian in the 500s, running along the river, climbing the slopes, and going straight up the sides of Mount Silpius to a citadel a thousand feet above the streets, with some four hundred towers.

The Turkish governor, Yaghi-Siyan, had prepared for their coming with practical ruthlessness, expelling or imprisoning much of the city’s Christian population and turning the great cathedral of St. Peter into a stable. The crusaders arrived on the 20th of October 1097, looked at those walls, and decided to blockade the gates and starve the city out. The siege of Antioch turned into a contest over which side would starve first, and the besiegers lost.

For the first few weeks, the Arantes Valley in late October was heavy with the harvest, and the army gorged itself with no thought for the months ahead. When the food within reach was gone, right around Christmas of 1097, the country had nothing more to give. Winter in northern Syria is not the postcard desert people expect. It rained for weeks on end.

Tents rotted and collapsed. Mail rusted on men’s backs. The camp turned into a sucking field of mud. An earthquake rattled the region that winter.

And through all of it, the hunger tightened. The price list survives. A single small loaf of bread sold for a bezant, a gold coin. The tongue of a camel: four solidi.

A goat: sixty. An egg: two. A walnut: one denarius. These are famine prices, the prices of a market in which money has almost stopped meaning anything.

And most of the army had no money at all. So they ate the horses, the precious, irreplaceable warhorses, and when those were too few, the donkeys and the mules. They added dogs and rats to the menu. Men picked through animal dung for undigested seeds of grain and ate those.

They boiled the leaves of fig trees, vines, and thistles into a green sludge. And they took the dried hides of horses, camels, donkeys, oxen, and buffalo, skins that had been sitting stiff for five and six years, and stewed them until something could be chewed off. One contemporary estimate has one in seven of the army dying of starvation through that winter. Among the poor, who could not buy a camel’s tongue at any price, it must have been far worse than that.

Inside the city, Yaghi-Siyan had kept a portion of Antioch’s large Christian population as insurance, and as the siege wore on, he squeezed them. The Greek patriarch of the city, an old man named John, was imprisoned, and the accounts describe him being hung over the battlements in an iron cage, suspended above the crusader lines. Christian churches inside the city were stripped. Out in the camp, the leaders tried everything the age knew to try.

In late December, Bohemond and Robert of Flanders took a large force away south on a grand foraging expedition up the Orontes Valley, toward the district of Maara itself—the first time the army’s hunger physically pointed at our town—and they blundered into a Muslim relief army coming north from Damascus. The crusaders won the fight at Albara on the last day of the year and lost the point of the expedition. The region was alerted. The wagons came back close to empty.

January was worse than December. The clergy diagnosed the famine the way eleventh-century clergy diagnosed everything—as a message about sin—and the leadership responded with a moral purge of the camp. Fasts were proclaimed, gambling banned, and in one of the stranger episodes of the whole expedition, the women were driven out. Adhemar and the princes ordered the camp cleared of what the chroniclers call the occasions of sin, and married men were separated from their own wives.

The camp economy, meanwhile, had reshaped itself around the shortage, and its shape carried straight through to who ate what at Maara a year later. In a famine market, the army’s own hierarchy turned into a food chain. The princes and the greater knights had treasure, and treasure bought camel tongue at four solidi and kept a core of fighting men alive. Below them, the ordinary knights sold their arms and their honor piecemeal, entering the paid service of richer lords.

And below them, the mass of the poor had nothing to sell at all except their labor and their prayers, and they starved first and worst and invisibly. A year of that, then the second famine inside the walls in June, harder and faster. Then the epidemic all summer. Then the autumn hunger.

The people who finally marched on Maara at the end of November 1098 had been living inside that machine for over a year and had watched it eat their families. And it had taught them, in the only school that matters, exactly what a human life weighed against a sack of grain. What saved the crusaders at Antioch was treachery arranged by the man best suited to arranging it. Bohemond had opened a secret channel to a man inside the city remembered as Firuz, an Armenian convert commanding a stretch of wall including a tower called the Two Sisters.

He agreed to let Bohemond’s men over his wall. Bohemond then went to the other leaders and, without naming his contact, made them an offer: whoever gets us inside keeps the city. Raymond of Toulouse refused on principle; the city had been imperial territory, and their oaths said it went back to Alexios. Then word arrived of just how close a colossal Muslim relief army under Kerbogha of Mosul was, and principle got thinner.

The leaders agreed. If the city fell to Bohemond’s scheme, Bohemond held it. On the night of the 2nd of June, 1098, the army marched away from the city as if departing, doubled back in the dark, and at dawn on the 3rd, sixty men went up a leather ladder onto Firuz’s wall. A gate was opened from inside, possibly by the city’s resentful Christians, and the army poured in.

What followed was the sack of a great city by men who had starved outside it for eight months. The crusaders killed everyone they met in the panic of that morning, Turks and the city’s own Christians alike, because in the dark and the frenzy, nobody was checking. The Gesta’s author says, “All the open places of the city were so choked with corpses that no one could endure the stench, and a man could not walk the streets except over the dead. ” Yaghi-Siyan fled through the hills, fell from his horse, and was left behind by his own men.

Local villagers found him lying there, recognized him, cut off his head, and carried it to Bohemond. Four days later, on the 7th of June, Kerbogha’s enormous army arrived and wrapped itself around the walls. The besiegers became the besieged. The second famine of Antioch, June of 1098, was worse than the winter one, and faster, and it is the famine in which, according to more than one account, some crusaders first tasted human flesh.

Horse blood was drawn and drunk or cooked. The last horses and donkeys were slaughtered. Leather was boiled—shoes and belts this time, not just old hides. Men paid fortunes for a dead dog.

Knights who had led hundreds now begged. Deserters went over the walls on ropes in the dark in such numbers that the army coined a nickname for them: the rope danglers. And it was exactly then, in the deepest pit of the siege, that the visions started. A poor Provençal servant named Peter Bartholomew came to Raymond of Toulouse and the bishop of Le Puy with a story.

Saint Andrew, he said, had been appearing to him for months and had shown him in a vision the burial place of the Holy Lance, the actual spear that pierced the side of Christ, hidden beneath the floor of the Cathedral of St. Peter right there in Antioch. Adhemar, the papal legate, listened and did not believe a word of it. Raymond of Toulouse believed, or chose to.

On the 14th of June, a dozen men, among them our chaplain Raymond of Aguilers, dug a pit in the cathedral floor all day and found nothing. Toward evening, Peter Bartholomew himself jumped down into the hole in his shirt, barefoot, told everyone to pray, and came up with a piece of iron. Draw your own conclusions about the archaeology of that moment. The chroniclers of the time already split on it along suspiciously political lines: the Provençals treasuring the Lance, and Bohemond’s Normans rolling their eyes.

What no one disputes is the effect. Word of the Lance went through that dying army like a current. Men who had been waiting to die decided God had personally signaled that they would not. On the 28th of June, 1098, the crusaders opened the gates of Antioch and walked out to fight Kerbogha’s much larger army on foot, with the Lance carried in the front line by Raymond of Aguilers himself.

It should not have worked. But Kerbogha’s coalition was a temporary arrangement of rivals who half wanted each other dead, and when the crusader line came on and would not stop, the coalition’s wings started peeling away in a chain of defections and panic. One of the most powerful armies assembled in that generation dissolved without a real battle. The crusaders had taken Antioch.

They had also, it turned out, taken a prize that would tear them apart. Bohemond wanted it and pointed to the bargain struck before the betrayal of Firuz. Raymond of Toulouse pointed to the oath sworn to Alexios. Neither man would move.

And through that same summer, an epidemic swept the region, possibly typhus, and it killed, among thousands of others, the one man who might have arbitrated the whole mess: the legate Adhemar of Le Puy, on the 1st of August, 1098. With Adhemar gone, the crusade had no referee, no adult in the room, only rival princes, hungry masses, and an increasingly loud contingent of visionaries claiming to relay instructions directly from dead saints. The army of tens of thousands had been camped in the same corner of northern Syria for a year. The land was eaten out.

Prices in Antioch climbed back toward famine levels. The rank and file watched the princes squabble over the title to a city while everyone below them slid back toward starvation, and they began to say louder and louder the obvious thing: we did not sell our farms and bury our children beside the road in Anatolia so that Bohemond could have a city. Jerusalem. When do we march on Jerusalem?

By November there was open talk among the commons of choosing their own commander and going on without the princes altogether. At a council on the 1st of November, the princes patched together a compromise that satisfied no one. And Raymond of Toulouse, feeling his grip on the army’s loyalty slipping, made a decision that was equal parts piety, politics, and pantry. He would take his forces south, resume the pilgrimage road, and along the way capture territory of his own.

The pilgrims got their movement toward Jerusalem. Raymond got his land grab. And the first substantial town on the road south, rich, walled, and confident, was Maarat al-Numan. Raymond had in fact already taken his first bite of the south.

In late September 1098, the count had marched a force up onto the limestone plateau southeast of Antioch and stormed a town called Albara. The assault succeeded quickly, and what followed set the template. Raymond of Aguilers, who was there, records that the count killed many of the inhabitants and sold many others into slavery at Antioch. Then Raymond had the town’s great mosque consecrated as a church, and he installed a priest from his own following, Peter of Narbonne, as its bishop—the first Latin bishop ever created in the East.

Albara mattered to Raymond as proof of concept. It matters to our story as a rehearsal: the commander, the season, the storming, the sorting of a population into corpses and merchandise. When the leaders of the crusade later sat down to write their report to the pope, they folded the two towns into a single clause, “at Barra and Mara. ” One was the sketch, the other the finished work.

The town of Maara had already beaten the crusaders once, and that victory five months earlier killed more of its citizens in December than any weakness in its wall. Back in July of 1098, a Provençal knight named Raymond Pilet had gathered a private army of foot soldiers and adventurers and struck out south to carve something for himself. Swollen with confidence, he marched on Maara. The townsmen, stiffened by Turkish troops from the countryside and from Aleppo, met his ragtag force outside the walls on the 27th of July and broke it.

Many died of thirst itself in the rout. Maara, in other words, had faced the Franks in open battle and had watched them run. When the real army appeared before their walls at the end of November, the people of Maara were not inclined to be terrified. The town declined every suggestion of surrender, and its defenders stood on the ramparts and mocked the besiegers below.

One tradition adds a nastier detail: that they insulted the crosses the crusaders carried, in full view of the army. The army that appeared before Maara’s walls on the 27th and 28th of November 1098 was a fraction of the great host but a hardened fraction. Raymond of Toulouse and Robert of Flanders brought their forces down from the north, and Bohemond, who did not trust Raymond out of his sight, followed within a day. Realistic estimates for the whole force run to a few thousand fighting men and an uncounted mass of pilgrim camp followers.

Facing them on a wall with a deep ditch in front of it stood a garrison that was mostly the town itself: a citizen militia of merchants, artisans, and farmers, reinforced by whatever Turkish and Arab fighters had come in from Aleppo and the nearby castles. No relief army was coming, and everyone probably knew it. After Kerbogha’s humiliation, no amir was eager to try his luck against the Franks in the field. Maara would stand or fall alone.

The first thing the siege killed was the orchards. An army in winter needs firewood beyond counting, and an army building engines needs timber. The groves that ringed Maara, the olive and pistachio groves that had made the town’s name and its living for generations, went down under crusader axes in the opening days. Trees older than anyone inside the walls were dismembered for cook fires and for the frame of the machine that would break the town.

The wells outside the walls went next, seized or spoiled. And by the end of the first week, Maara sat inside a widening ring of its own destroyed livelihood. The first assault came almost immediately, on the 28th of November. Raymond and Robert of Flanders threw their men at the wall the very morning after arrival.

It did not work. Ladders went up short of numbers, the ditch fouled the attack, and the townsmen threw them back. Bohemond arrived that evening, and the leaders tried again, another direct storm, and failed again. The defenders of Maara, the same townsmen who had chased off Raymond Pilet in July, had now beaten the main crusader army twice in two days.

So the siege settled into a blockade. And here the shape of this whole story tightens like a knot: since the besiegers had almost no food, the attackers were starving while they attacked. December closed in, the second winter. Foraging parties ranged farther and farther and came back with less and less, or did not come back.

Raymond of Aguilers writes about the poor of the army during those two weeks scattering across the countryside just to find anything at all, digging for roots and grain caches in abandoned villages, chewing on stalks and seeds picked out of the dirt and the dung. The Antioch diet again, one year on. Meanwhile, the work at the wall went on under a rain of everything the town could throw. The crusaders labored to fill sections of the great ditch with brush, rubble, and the stones of nearby buildings.

The defenders concentrated on exactly those men, dropping cut stone, pouring down quicklime that burned into eyes and wet skin, and, in a detail every account seems to relish, hurling down hives of furious bees. Parleys happened, or something like them, in glimpses the sources barely bother to record. The town had beaten these people twice in July and again on the 28th of November, and did not believe the wall would fall. The crusaders, for their part, no longer had the supplies to accept anything short of everything.

By the second week of December, the leaders knew, and said aloud in council, that the army could not sit there many more days without dissolving; that the choice had narrowed to taking the town at once or abandoning the campaign in humiliation. The tower was Raymond’s answer. Unable to feed a long blockade, the leaders went for engineering. Raymond of Toulouse paid for the construction of a siege tower, a rolling wooden castle built higher than Maara’s wall on four wheels, with knights stationed on the upper platform and, according to the accounts, priests and clergy assigned to accompany the machine, praying aloud as it moved.

Hunger built the machine that broke Maara. On the 11th of December, the crusaders rolled the tower up against the wall. Knights on the top hurled stones down onto the defenders on the rampart. Javelins and arrows swept the wall-walk.

Men below worked with iron hooks to drag stones out of the battlements. And under all that cover, the real work happened at ground level, where sappers got into the ditch and began undermining the wall itself. The defenders fought all day. At some point in the afternoon, with the garrison’s attention nailed to the sector where the tower loomed, crusaders with ladders went up a stretch of wall on the far side of town that had been stripped of men.

A knight named Gouffier de Lastours entered history as the first man over the wall of Maara, a distinction that made his family’s reputation for generations. By evening, the crusaders held a long stretch of the ramparts, and the defense of Maara came apart. The Turkish and Arab fighters and the town’s leading men pulled back off the walls in the dusk. The town’s people withdrew into their houses and into the caves.

Night fell with the town breached and not yet taken, in that suspended, dreadful interval that students of siege warfare will recognize since it recurs across the centuries: the hours between the wall falling and the sack beginning, when everyone inside knows what the morning is. Valuables went under floors and into wells. Coins were sewn into clothing or swallowed against the morning. The passage down into the caves filled through the night with families and the old, feeling their way into the cold rock their ancestors had cut.

Some of the remaining fighters and notables negotiated through the dark for a promise Bohemond would break by noon. Others simply waited behind their doors with whatever weapon a merchant or a weaver owns. The sack of Maara did not even wait for morning. The famished mass of the army poured through the breaches and over the walls in the dark, ahead of any order and outside anyone’s control, and began to plunder.

In the sources, you can watch the two sieges of Maara happen at once: the princes’ siege, a matter of towers and terms; and the poor’s siege, a matter of empty stomachs and two years of dying by the road. On the morning of the 12th of December, Bohemond offered terms to what remained of the town’s leadership. Those who gathered in a designated palace near the gate and ceased fighting would have their lives. The sources on the Muslim side and several on the Christian side agree on what those terms proved to be worth.

The people who surrendered were stripped of their goods. Then the killing began anyway, and once it began, it went through the whole town. Ibn al-Athir, the great Arab historian writing a century later, says the Franks put the population to the sword for three days, killing in his figure more than a hundred thousand people. Ibn al-Qalanisi, much closer to the events, describes it with the specific bitterness of broken faith: the town’s people took refuge in their houses under a promise of protection, and the Franks broke every assurance, plundered everything they found, and raised crosses over the town.

The Latin chroniclers hardly dispute the substance. The Gesta’s author says flatly that his own side killed everyone they met there, man and woman, so that a person could scarcely walk the streets except by stepping on the dead. And he preserves, without any visible discomfort, the detail that damns his own lord most completely. Bohemond, he says, sent word through interpreters to the leaders of the town that they should gather with their wives and children and goods into a palace above the gate, and he would save them from death.

The town’s notables did exactly that. What Bohemond’s men then did was take everything they had, gold, silver, and their ornaments. And as for the people, some they killed, and others they ordered taken to Antioch to be sold. The palace of protection emptied into slave coffles.

The pledged word of the most famous soldier in Christendom converted overnight into inventory. The numbers need honest handling. Western tradition settled on a figure of around twenty thousand dead at Maara. Ibn al-Athir’s hundred thousand is vastly larger.

Modern historians point out that Maara’s entire population before the siege was probably under ten thousand people, swollen somewhat by villagers who had fled inside the walls for safety. So every one of the traditional figures outruns the number of people who could physically have been there. But numbers in medieval chronicles work less like statistics and more like volume knobs. A hundred thousand is the sound of a scream carried across a century.

What the figures agree on from every direction is totality. Maara was not raided or subdued. It was emptied. The men were killed more or less systematically in the days after the fall, and the women and children who survived the first frenzy were gathered up as property.

Slave caravans went north from Maara that winter. Families were broken up on the block a month after they had been standing in their own kitchens. Some of the town’s people did what people in that region had done in every catastrophe for a thousand years and went underground, into the caves and tunnels that gave Maara its name. It did not save them.

The crusaders knew, or quickly learned, that the survivors had taken their families and whatever gold they had down into the rock. Fire and smoke were pushed into the cave mouths to force people out, choking whole families up into the light where the men were killed and the rest taken. Those suspected of hiding valuables were tortured for them. The belief had spread through the army that Saracens hid their gold by swallowing it, so the dead were opened.

Bellies were slit and searched. At Maara, at Antioch before it, and most infamously at Jerusalem seven months later, dead men and women were treated as containers, and the search for coins in human viscera became, for some of the army’s poor, a form of prospecting. The chroniclers report this in the tone of men describing weather. A few days after the fall, then, this was Maarat al-Numan: a town of the dead in December, held by the starving.

Corpses filled the streets faster than they could be dragged away and were pitched into the ditch, into wells and cisterns, fouling the town’s own water, and into the marshy ground outside the walls. The victors tore the place apart looking for the wealth a town this prosperous was supposed to contain and found much less than they had dreamed—granaries lighter than expected after a season of regional war, portable wealth hidden, buried, swallowed, or already carried off—and crucially, not remotely enough food. The army had taken Maara to solve its hunger, and Maara did not solve it. Within days of the massacre, the conquerors of the town were starving inside it, wintering among the thousands they had killed.

And the princes, with a timing that defies parody, resumed their real war, the one against each other. Bohemond held the walls and towers. Raymond held the interior and installed his own man, the newly minted bishop of Albara. And the two greatest lords of the crusade settled in to bicker over the title to a charnel house.

And in that place, in those weeks, the thing happened that this story is named for. Take the sources one at a time, in roughly the order they were written. Start with the earliest, the anonymous Gesta Francorum, written around 1100 by a knight who followed Bohemond and was there. His account is the shortest and the plainest, one sentence dropped into the narrative of the aftermath: “While the army lingered at Maara, some of our people, driven by terrible necessity, cut flesh from the dead bodies, sliced it into strips, and cooked and ate it.

” One clause, and he moves on. His close cousin in the sources, a Puy-en-Velay priest named Peter Tudebode, specifies that it was the bodies of Saracens that were eaten, only the enemy dead. An army starving indiscriminately had by December buried a great many of its own people. If pure famine was the whole engine here, the freshest and nearest dead would have served whoever they had been in life.

The insistence, repeated across several accounts, that the eaten were Muslims suggests that even at the absolute floor of human extremity, a line was being maintained. And a line maintained is a choice being made. Next, Fulcher of Chartres. Fulcher was a priest on the crusade, though by this point he was hundreds of miles away at Edessa serving Baldwin, so his account rests on the testimony of men he knew and marched with.

He is generally a sober, careful writer, not given to horror for its own sake. His sentence visibly costs him something. “I shudder to tell it,” he says, and then he tells it. “Many of our people, harried by the madness of starvation, cut pieces of flesh from the buttocks of the Saracens lying there dead, and cooked them.

And when the pieces were barely warmed through, hunger took over, and they devoured the flesh half raw. ” Every clause of that is doing work. The buttocks: the gluteal muscle, the largest cut of meat on a human body, a detail with the awful specificity of testimony rather than invention. The cooking attempted and abandoned: mouths that could not wait for the fire to finish.

And Fulcher’s timing places this during the siege, in the hungry camp before the wall fell, and he adds that in this the besiegers were more damaged than the besieged. Now the eyewitness, Raymond of Aguilers, chaplain in the Provençal camp at Maara, carrier of the Holy Lance at the battle of Antioch, a man whose chronicle runs hot with visions, portents, and absolute certainty that God’s hand steered every hour of this expedition. He places the eating after the fall of the town, during the starving weeks of the occupation. His version opens with the image this whole story started with: “Famine gripped the army so hard that the people ate the bodies of the Saracens, already stinking, which had been thrown into the swamps beside the city two weeks and more before, and they ate them greedily.

” Two weeks. The bodies pulled from that cold marsh water in mid-December had been dead since the fall of the town, softening in the wet the whole time. And the chaplain’s word for how his fellow Christians consumed them is “greedily. ” He is the only writer who gives us appetite.

Then Raymond does something none of the other early writers do. He tells us who was watching. “This spectacle,” he says, “disgusted many of our own people and of strangers. And word of it went out.

” And he puts a speech in the mouths of the Muslims of the region, asking one another who could stand against a people so obstinate and so cruel that a year of siege could not turn them from Antioch, and who now lived on human flesh. God, Raymond concludes, had put the terror of the Franks into every race. The man was there. He describes public eating, an audience, revulsion spreading outward, and then he banks the revulsion as a strategic asset: terror as a gift from God.

Whether anyone at Maara planned it that way in the moment, the army’s own chaplain, writing within a couple of years, understood perfectly well what the cannibalism had bought them, and priced it without blinking. The accounts written back in Europe over the following decade, by men who had not been there but who interviewed those who had, did not soften the story. They escalated. Ralph of Caen, a Norman clergyman who served Tancred and wrote his deeds sometime after 1112, delivers the single most horrifying sentence in the entire dossier, and he stakes his credibility on his sources for it.

He says he had it from the very perpetrators of the shame. At Maara, Ralph writes, “Lack of food drove our people to make meals of human flesh, and the adults were put into the cooking pot, and the children were fixed on spits and roasted, and both were devoured. ” Boiled adults. Roasted children.

Scholars have gone both ways on whether that resembles a literary image migrating between texts, a stock picture of ultimate barbarism, or two independent witnesses describing similar horrors. The Muslim chronicles of Maara dwell at length on the massacre and the broken promises and say almost nothing about eating. The Arabic tradition that does mention the eating attributes it to hunger, to the consumption of the dead. So Ralph claims the confession of participants, describing something beyond every other account, and the surrounding record supports the eating of corpses and goes silent exactly where Ralph goes furthest.

Whichever way you resolve it, you are left somewhere terrible. Either men told Ralph the truth about pots and spits, or the crusade’s own veterans found the documented reality so unbearable that the retellings mutated into monstrosity within a generation. Albert of Aachen, working in the Rhineland from returned crusaders’ reports, adds one genuinely strange grace note. The Christians at Maara, he writes, did not shrink from eating the bodies of the Saracens and Turks they had killed, cooked in fire.

“And even the dogs they caught. ” The construction places dog-eating as the further extremity, the thing his audience would find harder. And maybe for an eleventh-century European audience that was true. That little sentence shows you the hierarchy of horror operating in the minds these reports landed in, and how differently that hierarchy was built from ours.

Then comes Guibert of Nogent, writing around 1108, a brilliant and endlessly quotable snob who never went east. Guibert takes the cannibalism and performs surgery on it. It happened, he concedes, but rarely and in secret, and the men who did it did it stealthily, knowing it was abominable. So the army as a whole stands acquitted.

And it was not the knights, God forbid, nor the respectable poor. He has a specific population to hand it to. And with him, for the first time in our story, we meet them under their name: the Tafurs. The Tafurs are one of the strangest presences in medieval literature, and half the difficulty of talking about them is deciding what genre they belong to.

In the sober chronicles they barely exist, in Guibert, and then massively in the vernacular epics composed over the following century they swell into a whole barefoot nation within the army. The name itself seems to come from a word for beggar or vagabond. Guibert describes them almost anthropologically: a horde of the utterly destitute who marched without weapons, without money, without shoes, living on roots and the refuse of the camp, filthy, covered in sores, and terrifying to everyone, including their own side. He gives them a king.

King Tafur, the story runs, had been a Norman knight who lost his horse and with it his rank, and instead of clawing his way back up, he went the other way entirely: gave away his arms and armor, dressed in sackcloth, and took command of the poor. The princes, Guibert says, valued the Tafurs. They carried loads, filled ditches, worked the siege machines, and swarmed over walls. And the Muslims feared them beyond all the rest, believing the Franks kept a tribe of flesh-eating devils on purpose.

The Chanson d’Antioche, the great crusade epic composed in something like its surviving form in the 1180s, three generations after the events, takes the cannibalism and hands it entirely and enthusiastically to King Tafur’s people. And it moves the scene from Maara back to the siege of Antioch, and it plays it, God help us, almost as comedy. The starving Tafurs in the poem come to Peter the Hermit, of all people, and ask what they should do. And Peter looks at the field of Turkish dead and tells them that the bodies lying there, properly cooked and salted, will eat well enough.

So the Tafurs flay the corpses, boil some and roast others over fires in full view of the city walls, and sit there before all of Antioch, eating the enemy dead and pronouncing the meat, in one line that has scandalized readers for eight centuries, “tastier than spiced peacock. ” The poem gives us the Turkish garrison watching this, crying out to their gods, fathers watching what happens to the bodies of their sons. And the city’s commander sends to Bohemond in outrage. And Bohemond in the poem spreads his hands.

“Nothing to do with me,” he says in effect. “These people are not ours to command. Even we cannot control them. They frighten us too.

” And having disclaimed them, the poem’s crusader princes quietly bank the fear, exactly as Raymond of Aguilers had banked it in sober prose eighty years earlier. The Chanson d’Antioche in its surviving form is credited to a poet named Graindor de Douai, reworking older material for audiences in the France of the 1180s. And those audiences were, to a striking degree, the actual descendants of the crusaders: noble houses whose founding glory was an ancestor at Antioch or Jerusalem, listening after dinner to a professional singer perform the family’s finest hour. Crusade epic was ancestral advertising, and it obeyed advertising’s rules.

The ancestors must be magnificent and the miracle real. The cannibalism could not be cut—it was too famous—so the poem does the next best thing and gives it entirely to people no listener in the hall was descended from. The earliest sources describe the eaters as “some of our people,” “our army,” “the Christian people. ” No Tafurs.

The word and the blame arrive together a decade later with a writer safe in a French abbey, and then flower in poems written for the grandchildren of the crusaders. Somebody had to carry Maara. The princes could not, the knights would not, and so it settled on people with no surnames, no chroniclers of their own, and no descendants endowing monasteries to argue their case. One more document, and then we can leave the eating, though it does not really leave the story.

That letter written by the crusade’s own leaders from Laodicea in September 1099, to the pope, the official report of the whole expedition over the names of an archbishop, of Godfrey, and of Raymond of Toulouse himself. About our town, the drafting hand does not flinch. “While we delayed at Barra and Mara,” the leaders wrote, “famine grew so great in the army that the Christian people ate the putrid bodies of the Saracens. ” The commanders of the First Crusade, in an official dispatch to the head of their church, reported the consumption of rotten human flesh by their own people in one flat clause, wedged between logistics.

With no secrecy claimed and no monsters invented, the sentence simply signed, sealed, and carried to Rome. Whatever else remains uncertain about Maara, the event itself was entered into the record by the men responsible for the army that did it, within nine months, addressed to God’s own vicar. The crusaders’ own writers set Maara against the longer history of siege and famine. Survival cannibalism under siege was not unknown to these men.

Every educated churchman in the army knew Josephus, whose account of the Roman siege of Jerusalem in the year 70 contains the most famous atrocity story in ancient literature: the noble woman Mary of Bethuba, who, starving in the encircled city, killed, roasted, and ate her own infant son. Nor was it only ancient. Within living memory of the crusaders’ own grandparents, the great European famine of the 1030s had produced reports, recorded by the monk Ralph Glaber, of human flesh dug up, of travelers murdered for food, of a man at the market of Tournus caught selling cooked human meat, for which he was burned. The eleventh-century mind, in other words, had a filed category for this.

Famine cannibalism was a known horror at the absolute edge of the human map, something that happened to the besieged and the godforsaken. What had no precedent at all was what the chronicles of the First Crusade had to record: an army of the cross, on the most sacred journey Christendom had ever organized, doing the edge-of-the-map act in enemy country, to enemy dead, some of it possibly in the open, and then marching on to triumph. The besieged eat the dead in the old stories. At Maara, the besiegers did, and won.

A strong current of modern scholarship argues that many people on the First Crusade believed sincerely and increasingly that they were living inside the last days of the world; that the march to Jerusalem was the opening act of the apocalypse; and that the army itself was the instrument of prophecies now visibly coming true. Two years of impossible survivals had fed that belief. Dorylaeum, the Lance, the defeat of Kerbogha by starving men on foot. Each one filed as a miracle, each one raising the temperature.

And in the apocalyptic frame, the ordinary rules were already suspended. The poor were exalted, since the last days belong to the meek, and the enemy dead were already damned, their bodies so much matter. A man inside that frame, starving at Maara, watching the visionaries relay orders from dead saints, did not necessarily experience the marsh as a moral abyss. He may have experienced it as one more impossible thing in a year of impossible things.

Life in the captured town through late December was its own quiet horror even without the marsh. Winter rain on a townscape of the unburied. The wells fouled with the dead, so that water itself carried the town’s fate in it. Soldiers billeted in houses whose owners they had killed, sleeping under blankets pulled from beds whose families were in the ditch.

And in the command tents, the quarrel over possession of this place actually intensified as its worthlessness became clear, since by now Maara had stopped being a prize and become a symbol, a token in the game between Raymond and Bohemond over Antioch. Bohemond at one point around the turn of the year simply pulled his men out of the towers he held and went back north to Antioch. The two greatest soldiers of the expedition settled into a cold war, with a dead town and the road to Jerusalem as hostages. The deadlock broke from below, and the way it broke is one of the most extraordinary scenes of the whole crusade.

Raymond, trying to purchase the loyalty of the other princes and reassemble the march under his own banner, convened them at a castle up the road in mid-January and offered money: hard figures, thousands of gold solidi to Godfrey and Robert of Normandy, smaller sums scaled down the pecking order. The recorded offers survive. Some took the money. Godfrey declined.

And while the princes haggled, the poor at Maara, the survivors of Anatolia and two Antioch famines and the December we have just lived through, decided they had had enough of being wintered in a slaughterhouse while rich men negotiated. And they began, without orders, with their bare hands and whatever tools they could find, to demolish the town. They pulled Maara’s fortifications apart, stone by stone, the sick and the feeble joining in. Raymond of Aguilers describes even those too weak to lift working stones off the walls, levering blocks loose and rolling them into the ditch day after day, with the logic of people who had correctly identified the actual obstacle between them and Jerusalem: walls.

As long as Maara had walls, it was worth holding. And as long as it was worth holding, princes would sit and hold it. So the poor deleted the town as a military object. The bishop of Albara reportedly tried to stop them and was ignored.

Raymond of Toulouse, returning from his haggling to find his prize being disassembled by its own occupiers, drew the only conclusion available. On the 5th of January, he ordered the demolition completed, and the remaining fortifications of Maarat al-Numan came down. And what fire could finish, fire finished. The demolition was slow, dangerous work done by people at the end of their strength.

Falling masonry killed and maimed some of the wreckers. The chronicle preserves the astonishing image of the sick, unable to stand, working stones loose from the lower courses and rolling them down into the ditch from their knees. It was also, in its way, the most eloquent statement anyone made in the whole affair. No sermon or speech survives from the mass of the army.

What survives is the act itself: a starving crowd dismantling a fortress to force their lords back onto the road to Jerusalem. And it worked. Within three weeks, when the column finally moved, the town it left behind had been unmade twice over: its people in December, its stones in January. Then, on the 13th of January 1099, the count of Toulouse, the richest crusader of them all, the man who had spent a month wrangling over the ownership of ruins, walked out of the gate of Maara barefoot, dressed as a simple pilgrim, and took the road south.

His army followed, and Robert of Normandy and Tancred with their contingents fell in over the following days. The barefoot exit was a formal act of penance, the posture of a man approaching a shrine. The chroniclers present it, and Raymond surely intended it, as devotion recovered, the crusade remembering itself. But stand where we have been standing for the last hour, and it is hard not to see what the gesture was penance for, and what the man’s bare feet were walking away from.

The road to Jerusalem from that point is one of the eeriest stretches of the whole story, precisely on account of what did not happen. This shattered, half-starved column, a fraction of the force that left Europe, now marched for months through hostile country and was barely fought. Amirs whose walls the crusaders could never have taken sent envoys out to meet them with bread, with horses, with guides, with money, buying the Franks’ passage past their towns. The lords of Shaizar, the fortress on the Orontes just down the road from Maara, made terms almost immediately.

Word of Antioch had already made the Franks uncanny, an army that should have died twice and had not. Word of Maara made them something worse. Raymond of Aguilers told us straight out what his side believed the cannibalism had purchased: the terror of the Franks in every race. And the behavior of the towns on the road south is that sentence acted out on the map.

Fight the Franks and your city becomes Maara: your sons in the ditch, your daughters on the block at Antioch, your dead in their cooking fires. Or open your gates, pay, feed them, and watch them walk on toward Jerusalem. Town after town did the sums. The tribute itself made the point physically.

Convoys came out to the marching column from cities the crusaders never entered. For the first sustained stretch since Europe, the army that had boiled thistles outside Antioch and done worse at Maara walked the coast road with full saddlebags, provisioned by the very people their reputation terrified. The crusade’s worst crime paved the road to its greatest victory. One piece of unfinished business from Antioch met its end on that road.

The Lance and Peter Bartholomew had kept ascending after the great battle, Peter relaying ever more detailed instructions from his rotating cast of saints. During the long, futile siege of Arqa in the spring of 1099, the muttering finally came to a head, and Peter magnificently overplayed his hand. He demanded the ordeal, judgment by fire, to settle the Lance forever. On the 8th of April, Good Friday, the army built two walls of blazing olive branches with a gap of about a foot between them, and Peter Bartholomew, in his tunic, carrying the relic he had pulled from that pit, walked into the fire while thousands watched.

He came out the other end, which by the rules should have ended the argument, and then he collapsed. He died twelve days later of his burns and injuries. And the Lance’s authority largely died with him, at least outside the Provençal faithful. Set his ending beside the marsh: fire and human flesh, twice in a single winter and spring, at Maara with the dead, at Arqa with a living man, both times in front of a crowd, both times as proof of something.

These people had come to believe that ultimate questions could be settled in meat. What followed once the column was truly moving has a dreamlike quality in the sources. The Amir of Homs sent horses. The Amir of Tripoli sent gold, mules, and provisions and opened his markets.

The country changed around them, the Lebanese coast in early spring, green and heavily stocked, and men who had spent two winters chewing hide walked through orchards, eating until they were ill. On the 7th of June 1099, the army came over the ridge northwest of Jerusalem and saw it. Chroniclers described the whole column weeping, men falling to their knees, three years and uncounted tens of thousands of lives from Clermont. And among the people crying on that ridge at the sight of the holy city were necessarily whoever had waded into the marsh at Maara six months before.

Five weeks later the city fell, and the men of the marsh went over the wall with everyone else. The sack that followed became a byword within its own generation. The chroniclers on the winning side, and this bears repeating, the winning side, described the streets and the temple platform running with blood. Raymond of Aguilers reaching for the Book of Revelation, exalting that in the courts of the temple men rode in blood up to the knees, up to the horses’ bridles.

The worst of it concentrated on the temple platform. Crowds of the city’s people fled in the final hours to the roof of the al-Aqsa mosque, and Tancred, taking their surrender—and not incidentally a colossal ransom—gave them his banner to display as a sign of his protection. It protected them for one night. At dawn, other crusaders climbed to the roof and killed everyone under the banner, men and women, several hundred people by the accounts.

The city’s Jews fared no better. And days after the fall, with the frenzy supposedly spent, Albert of Aachen records a decision taken in cold blood: the leaders, fearing the captives still held in the city, ordered the remaining prisoners killed. A second, administrative massacre swept through people who had survived the first. One exception glows strangely in all this darkness.

The Fatimid governor of Jerusalem, Iftikhar al-Dawla, held out in the Tower of David with his bodyguard and negotiated terms with Raymond of Toulouse: treasure for safe passage. And Raymond, the lord of Maara’s ruins, kept the bargain to the letter, escorting the governor and his men out of the dying city and seeing them safely on the road to Ascalon. The one great safe conduct of the campaign that was honored in full. And at Jerusalem, the swallowed-gold conviction we watched at Antioch and Maara reached its industrial form.

Fulcher of Chartres describes it with that terrible matter-of-factness of his. It was wonderful, he says, meaning astonishing, to see the squires and the poorer footmen, once they learned of the Saracens’ trick, split open the bellies of the dead to pull bezants out of the entrails. And because opening one body at a time was slow, the dead were heaped into piles and burned, so that the coins could be raked more easily out of the ashes. No eating is recorded at Jerusalem.

Food had reached the army by then. Across the whole crusade, for all its massacres, the eating of the dead stays fixed to that one winter, at that one town, at the absolute bottom of the famine. Wherever this army was fed, it did many dreadful things, and it did not do that one. And then we cross the line one more time, to look at how the other side remembered our town, since Maara’s second life, its life in Muslim memory, runs longer than the crusader states did and is arguably still going.

The nearest chroniclers in time, like Ibn al-Qalanisi in Damascus, fixed the elements that mattered to them: the wooden tower, the fall, and above everything the broken guarantee, safety pledged and then the killing, the plunder, the crosses raised over a Muslim town. In their pages, the town does not fall in December of 1098, but in the month of Muharram of the year 492 after the Hijra. Two calendars, two worlds, each fully convinced it stood at the center, meeting at the wall of a pistachio town, with the calendar the town used losing everything, and the memory of that loss keeping its own count ever since. In August of 1099, weeks after Jerusalem fell, the chief qadi of Damascus, Abu Sa’d al-Harawi, arrived in Baghdad at the head of a party of refugees from Syria and marched them into the court of the caliph to demand a response to what the Franks had done.

It was Ramadan, and al-Harawi and his companions, travel-worn and deliberately scandalous, broke the fast publicly in the caliph’s own audience, eating in daylight before the court, precisely to force the assembly to feel a fraction of the outrage it was declining to feel. Men wept in that audience hall, the accounts say, as the refugees described what they had fled. The caliph appointed a commission of inquiry. The commission, in the way of commissions, produced nothing.

Baghdad had its own wars. The practical answer of the Islamic East to the First Crusade in 1099 was silence. The silence lasted a generation, and the counter-crusade when it finally came came slowly: Zengi taking Edessa in 1144, Nur al-Din unifying Syria, Saladin taking Jerusalem back in 1187. And the men who built that movement built it in part out of memory, out of the grievance file that al-Harawi had carried to Baghdad and Baghdad had shelved.

And Maara stayed in the file the whole time. The modern book that carried this memory to the widest audience, Amin Maalouf’s The Crusades Through Arab Eyes, published in 1983, gives its chapter on this campaign the title “The Cannibals of Maara,” and argues that the town’s fate, carried by poets and oral memory, fixed the image of the Franks for centuries. Some historians have pushed back, noting the curious fact we have verified for ourselves along the way: that the detailed, gloating, ashamed, contradictory testimony about the eating survives overwhelmingly in Christian ink, the perpetrators’ own literature, while the Arabic sources mostly recorded the massacre and moved on. Which means the strangest custodians of this atrocity’s memory have always been the heirs of the men who committed it, writing it down generation after generation, unable to leave it alone.

And the town itself endured, as towns do. Maara was retaken for Islam within a generation, rebuilt and rewalled. It went on trading and growing pistachios through Zengids, Ayyubids, Mamluks, and Ottomans, keeping at its heart the tomb and the memory of its blind poet. Anyone who visited the town in the last century before the war walked, mostly unknowing, through the layers of this story.

Maarat al-Numan grew into a busy provincial town of tens of thousands, with olive presses and pistachio brokers on the Aleppo highway, an old great mosque at its heart, and a museum housed in the Khan Murad Pasha, a grand Ottoman caravanserai, holding one of the finest collections of Roman and Byzantine mosaics in the Middle East. Under the streets, the caves. Above them, ordinary life. History does not mark the places where it did its worst.

It just moves on and lets people live there. For nine hundred years in Maara, people did. And here, the poet’s afterlife tracked the town’s. Abu al-Ala al-Ma’arri, born in the town in 973, blinded by smallpox as a small child, and by the time of his death in 1057 one of the most celebrated poets in the Arabic world.

A skeptic who doubted every creed with magnificent even-handedness, a recluse who called himself the hostage of his prisons—blindness, his house, and later he added, the body itself—and of all things, a strict vegetarian, a man who for decades refused to eat flesh or eggs or honey, or to wear leather, on the principle that no creature should suffer for his comfort. His masterpiece in prose, the Epistle of Forgiveness, sends a contemporary on a guided tour of the afterlife, interviewing dead poets about their sins and their grammar. His great verse collection, the Luzumiyyat, he composed under a self-imposed double rhyme scheme, thousands of lines of pessimism polished like knives, doubting the sects, and in some verses the resurrection itself. Conservative critics called him a heretic for a thousand years while poets kept memorizing him.

When he went to Baghdad around the year 1000, the capital’s literary world received him as a phenomenon and offered him the patronage every poet lived on, and he refused it, on the principle that praise for pay was a form of lying. After eighteen months, he came home to Maara, to a small inherited income, a diet of lentils and figs, and a house he rarely left. He never married, on principle. He never traveled again.

He asked that his grave carry one line only: that this wrong was done to me by my father, and I did it to none. Meaning that being born was the only injury in his ledger, given to him, passed on to nobody. The people of Maara kept his tomb. In the twentieth century, modern Syria claimed him as a national glory, and in the 1940s, the sculptor Fathi Muhammad cast a bronze bust of the old man, blind eyes, heavy beard, and it stood in the center of Maarat al-Numan, the town’s most famous face watching over its square.

Then came the Syrian civil war, and Maara found itself once again on the road armies use, fought over, shelled. Its museum looted and blasted, its people scattered into the same fields their ancestors had fled through. And on the 12th of February 2013, fighters of the al-Nusra Front, the local franchise of al-Qaeda, came for the poet. They took a stone-cutting tool to the statue of al-Ma’arri and cut the head off.

Photographs went around the world within hours: the bearded bronze head on the ground beside its plinth. Explanations varied. Some said the militants condemned him as a heretic for the poetry. Some said it was clan politics, or the rumor of treasure inside the bronze.

And perhaps it hardly matters which. A thousand years nearly after the man’s death, and 915 years after crusaders ate the dead in his hometown’s marshes, men with absolute certainty in their eyes stood in the square of Maara and beheaded the one citizen it ever produced who had warned, in meter, against absolute certainty. His line about the world’s two sorts of people had been quoted for centuries: those with brains and no religion, and those with religion and no brains. The statue stood headless in the square.

And the town emptied around it. And somewhere under all of it, under the square and the houses and the ruined museum, the old caves are still there, holding whatever the winter of 1098 left in them.