Around 5000 BC, the first farming culture of Central Europe, the Linearbandkeramik, left behind a small group of deposits that do not look like burials: pits and ditches filled with people who died violently, or whose bodies were broken up long after death. Five of them, Talheim, Asparn/Schletz, Schöneck-Kilianstädten, Halberstadt and Herxheim, are enough to show how different the events were from one another, and why "the end of a culture" is a harder phrase to defend than it sounds.
\nThe Linearbandkeramik, or Linear Pottery culture (LBK), is named for the incised bands on its pots. Its farmers spread from Hungary to the Rhine and beyond over several centuries, and their ordinary dead were laid on one side with knees drawn up, often with a pot or a stone tool beside them. That is why the five deposits matter. Each one departs from the normal pattern, and each departs from it in a different way.
\nFor a long time such departures were explained away. Many prehistorians treated the early Neolithic as a world of plentiful land and open villages, where organised violence belonged to a later age of metal and fortifications. The mass graves are the evidence that forced a rethink, and the argument over what they mean is still running.
\n| Site | \nCountry | \nMinimum number of individuals | \nWhat the evidence shows | \n
|---|---|---|---|
| Talheim | \nGermany | \n34 | \nMen, women and children killed by axe and adze blows and thrown into a pit | \n
| Asparn/Schletz | \nAustria | \nat least 67 | \nVictims left unburied in the ditch of an enclosure | \n
| Schöneck-Kilianstädten | \nGermany | \nat least 26 | \nBlunt-force and arrow injuries, with deliberate breaking of the lower legs | \n
| Halberstadt | \nGermany | \n9 | \nAlmost all adult males, with clustered blows to the back of the head | \n
| Herxheim | \nGermany | \nhundreds | \nBones processed and deposited in pits; the cause is disputed | \n
The counts for Talheim, Asparn/Schletz and Kilianstädten are set side by side in a 2023 comparative table. Our own record of prehistoric collective violence lists the same sites with the trauma reported and a certainty rating.
\nAt Talheim in Baden-Württemberg, 34 people lay in a single pit, with injuries of the kind that stone axes and adzes make. The wounds are concentrated at the back of the skull, which is hard to produce in an even fight. The full account is in our article on the Talheim pit.
\nAt Asparn/Schletz in Lower Austria the pattern is different in one respect: the dead were never buried at all. They lay in the ditch of a defended settlement and were gnawed by scavengers before soil covered them. The excavators recovered at least 67 individuals from only a part of the ditch circuit, and they have suggested that more remain. That larger figure is an extrapolation, not a count. See Asparn/Schletz for the detail.
\nBoth sites are usually read as the destruction of an entire community. At Talheim, the missing group would be the attackers, who left no remains.
\nThe Schöneck-Kilianstädten grave, in Hesse, was published much later, in 2015. The excavators reported at least 26 people killed by blunt force and arrow injuries and placed in a commingled grave. What set it apart was the deliberate breaking of the lower limbs, a pattern the authors said had not been documented at such sites before. They read it as torture or mutilation, although the bones cannot show whether the victims were still alive when the legs were broken.
\nOne more observation recurs across these sites. Young women appear to be missing from the dead at Kilianstädten, and the excavators suggested that this may point to abduction. That is an inference from an absence, so it should be held loosely.
\nHalberstadt, in Saxony-Anhalt, looks like a different kind of event. The 2018 study counted nine individuals, almost all adult males, and found that the blows to the head clustered at the back of the skull in a regular way. The authors read this as execution, followed by careless disposal of the bodies. They also reported that the victims were largely not local people.
\nThat reading points to a controlled killing of captives, not the sudden destruction of a village. Set beside Talheim, where the wounds are spread across the skull in a way that suggests chaos, Halberstadt shows that the culture produced at least two distinct forms of lethal violence. The press coverage at the time made the same comparison.
\nHerxheim, in the Palatinate, is the exception that stops the group from being a tidy list of massacres. Here a settlement with a pit enclosure contained the remains of a very large number of people, whose bones were cut, broken and deposited in pits.
\nTwo readings exist, and the excavation team itself is split. One group argued that the processing marks show mass cannibalism. Another asked whether the site was instead a ritual centre, with remains brought from elsewhere and treated in a long mortuary process, and it questioned whether the evidence supports a war reading at all. No weapon trauma of the Talheim kind has been demonstrated there. We have set out the argument in Herxheim: cannibalism or death ritual, and our dataset rates the row low for exactly this reason.
\n| Period | \nWhat the record shows | \n
|---|---|
| c. 5600 BC | \nLBK farming spreads across Central Europe | \n
| c. 5000 BC | \nTalheim, Asparn/Schletz and Kilianstädten are dated to roughly this period | \n
| c. 5000 BC and after | \nHerxheim and Halberstadt belong to the same cultural horizon | \n
| Later centuries | \nThe LBK gives way to successor cultures | \n
The dates here are deliberately approximate, because the sites are not dated to the year, and the ordering within the final phase is not settled.
\nIt is tempting to turn these deposits into a story: a farming society strained by pressure on land or resources, tipping into raiding, and finally collapsing. Some researchers have argued for warfare in the LBK on the strength of ditches and mass deaths, and that view helped to overturn the older picture of a peaceful Neolithic. But five sites spread across a culture that lasted for centuries cannot prove a single cause. The evidence does not show that the events were connected, that they were simultaneous, or that violence ended the culture. Another reason for caution is that the sites survive because they are unusual, so the picture may be skewed toward the dramatic.
\nThe word that sits underneath the whole debate is perimortem, an injury made around the time of death. Telling it from later damage is hard, and we explain why in Perimortem: the hardest word in osteology. The same difficulty affects the older case of Jebel Sahaba, where a restudy changed how a famous massacre is read.
\nWhat the five graves show is variety. One shows a community destroyed, another bodies left to the scavengers, a third deliberate mutilation, a fourth a probable execution and the fifth a deposit that its own excavators cannot agree on. Together they establish that lethal group violence existed among Europe's first farmers. They do not establish a single cause, and they do not show that the culture ended because of it. A cautious reading holds both facts at once.
\nDid the Linear Pottery culture end in war? The evidence does not say so. Several sites show lethal group violence around 5000 BC, but five deposits cannot show that violence ended a culture that lasted for centuries.
\nWhich of the five sites is the most disputed? Herxheim. Its excavators have argued both for cannibalism and for a ritual treatment of the dead, and no clear weapon trauma has been shown there.
\nHow do we know the deaths were violent and not accidental? Archaeologists look for perimortem injuries, such as blows to the skull and arrow wounds, which show no healing. The method has limits, because bone damage after burial can imitate injury.
\nWhy are young women thought to be missing? At Kilianstädten the excavators noted that young women were absent from the dead and suggested abduction as one explanation. This is an inference from absence and not a demonstrated fact.
\nIn the 1970s the West German Federal Criminal Police Office, the BKA, turned computers on a problem that had defeated its detectives: where do people who live under false names hide? Its answer, the Rasterfahndung or "grid search", compared the records of millions of ordinary citizens against a profile of how a fugitive might behave. The method was born in the hunt for the Red Army Faction, returned after the attacks of 11 September 2001, and was twice reined in by the country's highest court.
\nIn 1971 Horst Herold became president of the BKA, a post he held until 1981. He believed that a modern police force should work like a research institute, and he built up computerised files on persons, objects and events. The Red Army Faction gave him a reason to use them. Its members lived underground, rented flats under invented names and moved between safe houses, so the old methods of watching known contacts brought few results. The group's own statements, which we discuss in our piece on how the RAF wrote its own record, told the police what the group claimed. They did not tell the police where its members slept.
\nThe idea of the grid search was to reverse the question. Instead of searching for a name, the police defined a pattern and let a machine find who fitted it, or, more usefully, who did not fit the pattern of an ordinary resident.
\nThe logic can be shown with the best-known case, the Frankfurt search of 1979. Investigators reasoned that a person hiding under a false identity would still need electricity, and might pay the bill in cash to avoid a bank trail. Police therefore collected the records of tenants who paid their electricity in cash, a pool that sources put at between sixteen and eighteen thousand people. They then compared these names against the registration offices and other public lists. Most tenants were real residents, and fell away. Those who could not be found in any directory were the interesting ones, since a flat rented under a false name leaves no trace elsewhere.
\nAccording to the accounts we have, only two people were left at the end of this process. One of them was a wanted member of the group. Whether the grid search truly produced the arrest is disputed: a critical analysis in the journal of the civil-liberties group CILIP argues that the arrest came from a targeted manhunt, and that how many people went into such searches, how many remained and how many hideouts they uncovered has never been made fully public. The honest summary is that the method has a famous success story and very little audited evidence of effectiveness.
\n| Date | \nEvent | \n
|---|---|
| 1970 | \nHesse passes what is generally counted as the first data protection law in the world | \n
| 1971 | \nHorst Herold becomes president of the BKA | \n
| 1977 | \nThe federal data protection law is passed; the German Autumn raises the pressure on the police | \n
| 1979 | \nThe Frankfurt grid search for flats rented under false names | \n
| 15 December 1983 | \nThe Constitutional Court rules on the census and recognises informational self-determination | \n
| Autumn 2001 | \nAfter 11 September, the states order grid searches for "sleepers" | \n
| 4 April 2006 | \nThe Constitutional Court limits preventive grid searches to cases of concrete danger | \n
The years of the RAF were also the years in which German data protection law took shape, and the two developments were connected. The state of Hesse passed its data protection act in 1970, and the federal law followed in 1977. Both were a reaction to the arrival of computing in public administration, a worry we also describe in our account of the European Commission's contest over computers. Some of the most persistent critics of Herold's methods came from within this new field of law. They asked what it meant for a free state to compile files on citizens who were suspected of nothing.
\nThat question was taken up by a different dispute. In the early 1980s Germans protested against a planned national census, and in December 1983 the Federal Constitutional Court ruled on it. The court held that, under modern conditions of data processing, everyone has a right to decide in principle who may use their personal data, the right of informational self-determination. The census ruling was not about terrorism, but it supplied the vocabulary with which every later grid search was judged. The same decade also saw anxiety about intelligence gathering elsewhere, as in the American debate over NSA surveillance in the Nixon years.
\nAfter the attacks of 11 September 2001, in which several of the pilots had lived and studied in Hamburg, the German states ordered a new round of grid searches. This time the aim was prevention. The police were not looking for a known fugitive but for so-called sleepers, people who might be preparing an attack while living quietly. The profile was drawn by features such as age, sex, a student status and religion and country of origin, and universities, registration offices and other bodies were asked to hand over records. A civil-liberties review put the number of people whose data were collected and classified nationwide at around 8.3 million.
\nThe results were thin. According to an answer given by the interior minister of Hesse, ten states found no person who could be classed as a sleeper, and in five others the people identified with an Islamist background were already known from earlier investigations. Only in Hamburg were contacts and suspected supporters of the attackers found. A study by the Max Planck Institute for criminal law treated the post-2001 search as the only nationwide use of the preventive grid search of its kind.
\nOne of the people affected, a student in North Rhine-Westphalia, took his case to Karlsruhe. In its order of 4 April 2006 the Federal Constitutional Court held that a preventive grid search interferes seriously with informational self-determination. It is compatible with the constitution, the court ruled, only if there is a concrete danger to high-ranking legal interests such as the life or freedom of persons or the existence of the state. A general threat situation, such as the one after the attacks of 2001, or tensions in foreign policy, is not enough.
\nThe ruling did not ban the method. It made it a tool for emergencies with specific evidence rather than a routine tool of suspicion, and it showed that the principles worked out in 1983 still bound the police a generation after the RAF's peak.
\nThe grid search is an episode of the Cold War era that has outlived it. It began when a democratic state under attack, at the time of the events described in the German Autumn of 1977 and the Stammheim trial, looked for an answer in technology. Our record of these operations shows how many of the group's attacks went unsolved, which helps to explain why the police were drawn to an instrument that promised to find people nobody could name. The later years, in which the third generation produced almost no convictions, only strengthened that pull.
\nThe grid search left two legacies. It gave the German police a model of data-driven investigation that other countries later copied under other names. And it gave the German constitutional order a body of case law that treats personal data as something citizens hold against the state, even in the shadow of terrorism. The evidence that it caught terrorists is limited and disputed; the evidence that it forced the law to define its limits is not.
\nWhat does Rasterfahndung mean? The word means a "grid search" or dragnet investigation. Police compare large sets of personal data, such as utility or registration records, against a profile to filter out the few people who match it.
\nDid the Rasterfahndung catch members of the Red Army Faction? The best-known case is the Frankfurt search of 1979, which is said to have left two people, one of them a wanted member of the group. Critics dispute that the arrest was a result of the grid search, and the full results of such searches were never published.
\nWhy did the Constitutional Court restrict it in 2006? The court found that a preventive grid search interferes seriously with the right to informational self-determination. It allowed it only where a concrete danger to important legal interests exists, not merely a general threat after 11 September 2001.
\nHow many people were screened after 2001? A civil-liberties review put the figure at around 8.3 million people nationwide. According to the Hessian interior ministry, most states found no sleepers, and the study rests on official replies and press reports.
\nOn 8 November 1658 a Dutch fleet under Jacob van Wassenaer Obdam forced its way through the Sound and broke the Swedish blockade of Copenhagen. The Republic had no love for Denmark and little for its king. It fought because the narrow water between Zealand and Scania was the gate to the Baltic, and the Baltic was where Amsterdam earned its living.
\nThe Sound, the strait at the eastern end of the Danish islands, was the busiest chokepoint in northern Europe. Almost every ship sailing between the North Sea and the Baltic passed it, and Denmark charged a toll on the cargo at Helsingør. Dutch ships carried most of that traffic: rye and wheat from Poland and the Baltic coast, timber, tar, hemp and iron going west, and salt, herring, cloth and wine going east.
\nThat trade is the reason the Republic could feed its towns while its farmers raised cattle and flax. Historians such as Jonathan Israel have long treated the Baltic "mother trade" as the foundation of Amsterdam's wider commerce, which is why a threat to the Sound was treated as a threat to the city itself. The older medieval order that once governed these waters is described in our piece on the Hanseatic League; by the 1650s Dutch skippers had largely taken its place.
\nThe Dutch did not want Denmark to be strong. They wanted nobody to be strong enough to close the strait. A weak, toll-taking Denmark that could be bargained with suited them better than a Sweden holding both shores.
\nSweden's king Charles X Gustav had spent the 1650s fighting in Poland, then turned on Denmark in 1657. In the winter of 1657-58 his army crossed the frozen Belts to Zealand, a march that astonished contemporaries. Denmark sued for peace, and the Treaty of Roskilde of February 1658 forced it to give up Scania, Halland, Blekinge and Bornholm, together with Norwegian territory around Bohuslän and Trondheim.
\nFor the Dutch, the alarming part was geography. Once Scania was Swedish, Sweden held the eastern shore of the Sound and the king still hankered after the western. In the summer of 1658 he resumed the war, landed on Zealand and besieged Copenhagen. If the city fell, Denmark would cease to be a power, and the Sound would belong to Stockholm.
\nThe States General sent a fleet to the Sound under Wassenaer Obdam, with Witte de With as vice-admiral. Our account of Witte de With covers the man; here the point is the machinery behind him. The ships came from several of the Republic's colleges of admiralty, the semi-independent boards described in our guide to the five Admiralties, and the expedition was driven in no small part by the merchants of Amsterdam, whose admiralty paid for much of the effort.
\nThe fleet sailed in the middle of October and reached the approaches to the Sound early in November, with soldiers aboard for the besieged city.
\nThe Swedish fleet under Carl Gustaf Wrangel lay off the strait. Sources give the Dutch about 41 ships and the Swedes about 45, though counts of ships and guns vary between accounts, and the numbers should be treated as approximate. The Dutch sailed in three squadrons and the Swedes in four.
\nThe date also differs by source. The battle is usually given as 8 November in the New Style calendar used on the Continent, which was 29 October in England and Sweden, and both dates appear in the literature.
\nA small Danish squadron was present but, according to the accounts we have read, was held back by contrary wind and could only watch. The fight was therefore Dutch against Swedish. The Dutch pressed through the narrows and engaged the Swedes in close, hard action. Witte de With's flagship Brederode ran aground and was lost along with its admiral; the Rijksmuseum's large painting of the battle shows the Eendracht engaged with two Swedish warships. Casualties are commonly put at roughly 1,400 for the Dutch and 1,200 for the Swedes, with one Dutch ship lost and several Swedish ships captured. These are the figures of the standard summaries, not of a careful modern count.
\nBy nightfall the Dutch held the water in front of Copenhagen. Reports say that around two thousand Dutch soldiers and a quantity of supplies were landed in the city, and the Swedes pulled back from their forward positions.
\n| Date | \nEvent | \n
|---|---|
| Winter 1657-58 | \nSwedish army crosses the frozen Belts to Zealand | \n
| 26 February 1658 (Old Style) | \nTreaty of Roskilde; Denmark cedes Scania and more | \n
| Summer 1658 | \nCharles X Gustav resumes the war and besieges Copenhagen | \n
| Mid-October 1658 | \nDutch relief fleet sails north | \n
| 8 November 1658 (New Style) | \nBattle of the Sound | \n
| 11 February 1659 | \nSwedish assault on Copenhagen fails | \n
| 1659 | \nConcert of The Hague: England, France and the Republic press for a peace | \n
| 27 May 1660 | \nTreaty of Copenhagen ends the war | \n
What the Republic did next is as telling as the battle. It did not seize Danish territory, take over the toll, or set up a base on the Sound. Its aim was a balance. Together with England and France it backed the Concert of The Hague in 1659, which urged a peace built on Roskilde but with the Sound kept open to shipping, and the Treaty of Copenhagen in 1660 largely followed that line.
\nThe result was a settlement that served Amsterdam without making it a conqueror. Sweden kept the Scanian provinces and so a coast on the Sound, but it did not take Copenhagen; Denmark kept its capital and its toll. A century of Dutch policy had aimed at this outcome, and 1658 was the one occasion on which the fleet had to be used to secure it.
\nThe episode also shows the limits of Dutch sea power. The relief worked because the Republic could mobilise a large fleet quickly for a clear purpose. Fourteen years later, in the crisis told in the disaster year of 1672, the same state was fighting for survival on land and sea at once. A comparison with the Company's wars in Asia, such as Cape Rachado in 1606, shows how differently the Republic fought when its core interests, not its Company's, were at stake. The commercial empire behind both is explained in our history of the Dutch East India Company.
\nOur record of Dutch Golden Age sea battles lists the Sound among the actions of the century, and it is one of the few in which the Republic fought openly to prevent a rival from controlling trade rather than to win it for itself.
\nThe Battle of the Sound was a short, bloody action with a durable result. It did not make Denmark a friend or Sweden an enemy of the Republic; it kept a strait open. Amsterdam's merchants had learned that the cheapest way to protect a trade empire built on bulk cargo was to prevent any single power from owning its chokepoints, and in November 1658 they paid for a fleet to prove it.
\nWhy did the Dutch Republic intervene on Denmark's side? The Republic's Baltic trade passed through the Sound, and Sweden controlling both shores would have let it tax or close that passage. Denmark was not so much an ally as the lesser danger.
\nWho won the Battle of the Sound? The Dutch, in the sense that matters: they reached Copenhagen and ended the sea blockade. Losses on both sides were heavy, and the Dutch lost Witte de With.
\nDid the Dutch take anything from the war? No territory. They secured a peace in which the Sound stayed open, which was the whole purpose of the expedition.
\nWhy are the dates given differently? The Continent used the Gregorian calendar while England and Sweden still used the Julian one, so the battle falls on 8 November in one and 29 October in the other.
\nHubert Walter was the one man who stood at both ends of Richard the Lionheart's reign: a bishop who helped to end the Third Crusade by bargaining with Saladin's side, and then the minister who kept England solvent, quiet and governed while its king was absent, captive or at war. Walter is remembered far less than the king he served. Yet when Richard's reputation as a crusader and soldier is weighed, the English state that paid for it was largely Walter's work.
\nWalter came from a Norfolk family and was brought up in the household of Ranulf de Glanvill, Henry II's chief justiciar and a relative of his, who trained a generation of royal clerks. The Exchequer was his schooling, and its habits of written record and audit stayed with him for life. He rose in the church, becoming Dean of York and then, in 1189, Bishop of Salisbury, the year Richard became king.
\nThat background matters. Walter was not a warrior-prelate by temperament. He was a lawyer and accountant who found himself on a crusade.
\nWalter sailed east with the English host in 1190. During the siege of Acre the army lost its leader of the church when Archbishop Baldwin of Canterbury died, and the Bishop of Salisbury became the most senior English churchman in the camp. Modern accounts describe him as acting as a link between the king and the wider host; the claim that he effectively led the English clergy there is a fair summary, but the chronicles do not hand him an official title for it.
\nHis conduct in the war was practical rather than heroic. The king's campaign in the east is recorded in our accounts of Richard's conquest of Cyprus, the victory at Arsuf and the fighting outside Jaffa, and Walter is not a central figure in any of them. Our record of Richard's operations shows how few of them were set-piece sieges. His moment came afterwards, in the long and tiring diplomacy that followed Richard's two retreats from the road to Jerusalem.
\nWhen the truce described in our account of the Treaty of Jaffa opened the Holy Sepulchre to unarmed pilgrims, parties of crusaders went up to Jerusalem. Walter led one of them. Both Ambroise and Bahā' ad-Dīn describe the visit, and Bahā' ad-Dīn records that Saladin received the bishop with interest, asked him about the king of England and gave the English clergy permission to maintain priests at the shrines. The details of the conversation are reported by an admirer of Saladin, so the warmth of the encounter should be read with that in mind.
\nRichard himself did not go. The bishop therefore saw what the king never saw: the city that the whole expedition had been fought for.
\nRichard sailed in October 1192 and was seized in Austria and handed to the emperor. Walter did not go straight home. He is recorded as reaching the king in captivity in Germany, after which Richard sent him back to England to help to raise the ransom and to take charge of the kingdom. Richard's mother, Eleanor of Aquitaine, was active in the same effort, as we describe in our article on Eleanor and the patronage of noblewomen.
\nIn 1193 Walter was chosen Archbishop of Canterbury and by the end of that year was also chief justiciar, the king's deputy in England. For the next five years, with Richard almost wholly absent after 1194, he was in practice the ruler of the kingdom. The ransom had to be found, then a war against Philip II of France had to be paid for, and this was the first great test of an administration driven by money.
\nWalter's lasting reputation rests on a series of procedural inventions, not on a single great act. The best known follow.
\n| Year | \nMeasure | \nWhat it did | \n
|---|---|---|
| 1194 | \nGeneral eyre and Articles of Eyre | \nRoyal justices toured the counties; the articles also provided for the election of coroners in each county | \n
| 1194 | \nRegulation of Jewish debts | \nRecords of loans to Jews were to be kept in official chests, giving the crown a claim on the registers | \n
| 1195 | \nKeepers of the peace | \nKnights were sworn to arrest offenders, an ancestor of the later justices of the peace | \n
| 1195 onward | \nFeet of fines | \nCopies of land agreements were kept by the court, creating a lasting archive of title | \n
Each of these should be read with the usual caution. Historians disagree about how far each was Walter's personal initiative and how far the work of the royal court around him, and the coroners and the Jewish records in particular served the crown's finances as much as any wish for fair dealing. The crown wanted information it could tax and enforce, and written records provided it.
\nThe effect, nevertheless, was cumulative. Walter's years left England with more written records, a more regular visit of the king's justices and a deeper layer of local officials drawn from the knightly class. This was the kind of society that, within a generation, could demand and then police the charter described in our article on Magna Carta. It also formed part of the dynastic empire explored in our piece on the Plantagenet marriage network.
\nIn 1198 Walter resigned as justiciar, under pressure from Pope Innocent III, who objected to a high churchman holding secular office. When Richard died in April 1199, Walter helped to secure the succession of the king's brother John and crowned him at Westminster in May. John then made him Chancellor, and in that office Walter's clerks began to keep the regular rolls of royal letters that survive from the reign.
\nWalter died on 13 July 1205 and was buried at Canterbury. The chronicle tradition reported that John was relieved to be rid of such a capable and independent minister, but that story comes from later writers and is better treated as legend than as evidence. What is certain is that the reign changed character without him, and the opposition to John that ended at Runnymede in 1215 grew in a kingdom whose methods Walter had done much to shape.
\n| Date | \nEvent | \n
|---|---|
| c. 1160 | \nBorn in Norfolk | \n
| 1189 | \nBecomes Bishop of Salisbury | \n
| 1190 | \nSails east with Richard's crusade | \n
| September 1192 | \nLeads a pilgrim party to Jerusalem after the truce | \n
| 1193 | \nArchbishop of Canterbury and then chief justiciar | \n
| 1194 | \nEyre and coroners established | \n
| 1198 | \nResigns as justiciar | \n
| 1199 | \nCrowns John; becomes Chancellor | \n
| 13 July 1205 | \nDies | \n
Richard's fame was won on the coast of Palestine. The money, the men and the order behind it came from England, and Walter was the person who made the kingdom deliver. His career shows that a medieval crusade was a project of government as much as of battle, and that the long-term benefit of Richard's reign for England was administrative rather than military. The crusade left no lasting conquests. The records, local officers and courts of Walter's years remained.
\nDid Hubert Walter fight in the Third Crusade? He went on the crusade as a bishop and is best known for his role as a negotiator and organiser, not as a commander in battle. The main chroniclers place him in the camp and in the diplomacy after the truce, not in the front rank at the great battles.
\nWhat did Hubert Walter do for Richard I at home? As chief justiciar from 1193 he governed England while Richard was absent, raising the ransom and then the money for the war in France. He also oversaw the 1194 eyre, which established the office of coroner.
\nWhy did Hubert Walter resign in 1198? Pope Innocent III objected to an archbishop holding the highest secular office in the kingdom, and Walter stepped down as justiciar. He returned to royal service the following year as John's chancellor.
\nDid he really meet Saladin? Bahā' ad-Dīn records that Saladin received the Bishop of Salisbury at Jerusalem in 1192. Because the account comes from Saladin's own circle, the exact words attributed to the conversation are uncertain even though the meeting itself is well attested.
\nIn 1101, two of the most prominent men who had abandoned the First Crusade turned around and went east again, and neither came home. Hugh of Vermandois, brother of the king of France, died of his wounds at Tarsus in the autumn of 1101. Stephen of Blois, one of the richest counts in Europe, was killed at Ramla the following spring. Their deaths belong to a forgotten campaign, the Crusade of 1101, that is usually dismissed as an epilogue to the capture of Jerusalem. This article follows the two counts through it, and says plainly where the chronicles stop talking.
\nBoth men had taken the cross in 1096, and both had left before the goal was reached. Hugh led the French royal contingent and had travelled ahead to Constantinople as an envoy. He was later sent from Antioch to Alexios I Komnenos and did not return to the army. Stephen was the more notorious case. During the long siege of Antioch in 1098 he withdrew, officially because he was ill, and on the road home he met the emperor Alexios and reported that the crusaders were doomed, a report that helped persuade Alexios to turn his relief force back. Antioch then fell to the crusaders and held, and Stephen's reputation never recovered.
\nA man who returned home with his vow unfulfilled was in an awkward position. Church teaching treated the crusade as a vow, and in 1100 Pope Paschal II was urging those who had turned back to go again. According to Orderic Vitalis, Stephen's wife, Adela of Normandy, daughter of William the Conqueror, shamed him into going. That story is one Norman monk's version and no other chronicle confirms it; it may reflect how contemporaries judged him rather than what was said in the bedchamber. What is not in doubt is that both counts took the cross a second time. Some historians read this as a matter of honour and pressure, and the evidence does not allow a more private motive to be recovered.
\nThe expedition was not one army but several, gathering in 1100 and setting out in the spring and summer of 1101. A large Lombard force under the archbishop of Milan went first, joined at Constantinople by Raymond of Toulouse and a Byzantine escort. A Nivernois contingent under William II of Nevers followed, and a third force of Aquitanians, Bavarians and others moved behind them under William IX of Aquitaine and Welf IV of Bavaria. Stephen of Blois travelled with Stephen of Burgundy, and Hugh of Vermandois marched with the French element.
\nThe plan was to cross Anatolia by the route the First Crusade had fought along, then press south to Antioch and Jerusalem. The First Crusade had succeeded because it moved as one body and was fed by a Byzantine and Armenian network of supply. The second wave did neither. Against it stood Kilij Arslan I of Rum, who had learned from the first passage, and the Danishmend emirs of the north.
\n| Date | \nEvent | \n
|---|---|
| 1096 | \nHugh and Stephen take the cross; Hugh reaches Constantinople as an envoy | \n
| June 1098 | \nStephen leaves the siege of Antioch | \n
| 1100 | \nPaschal II presses those who turned back to return | \n
| Spring 1101 | \nLombards and others reach Constantinople | \n
| Summer 1101 | \nThe Lombard-led army is defeated in northern Anatolia | \n
| Summer and autumn 1101 | \nThe Nivernois and then the Aquitanian columns are destroyed near Heraclea | \n
| Autumn 1101 | \nHugh of Vermandois dies at Tarsus | \n
| Easter 1102 | \nStephen of Blois reaches Jerusalem | \n
| May 1102 | \nStephen of Blois dies at the battle of Ramla | \n
The sources disagree on some of the months of the Anatolian battles, as is common for this campaign, and the table gives the broad sequence rather than days.
\nBy the late summer the picture was grim. The columns, marching separately, were picked off one after another in the dry interior, short of water and under constant mounted attack. Hugh's death came out of the Aquitanian disaster at Heraclea. Albert of Aachen's account, which the dataset cites for him, records him dying of wounds at Tarsus in Cilicia, to which he had struggled after the fight. In the dataset record this is listed as "Died of wounds", and the place of death, Tarsus, is the one point on which the sources are not in conflict.
\nBeyond that the chronicle is thin. It does not say who tended him, and it gives no deathbed scene that can be trusted. Later writers tell of his burial at Tarsus; the detail rests on the same few sources. For a count who had stood close to the French crown, the record of his end is short, and the text should not be padded to hide that.
\nStephen's story is better attested because Fulcher of Chartres was in Jerusalem. Having been separated from the main losses, Stephen reached Antioch and then Jerusalem, where he spent Easter 1102. He then took part in the defence of the new kingdom. Baldwin I, king of Jerusalem, rode out to meet an Egyptian army near Ramla in May 1102 with a small force and, by Fulcher's telling, underestimated the enemy. The crusader cavalry was overwhelmed. Stephen of Blois was killed in the fighting. The dataset lists him as "Killed in action" and cites Fulcher II.19, and this record follows that. Fulcher is sympathetic to Baldwin and wrote with hindsight, so how far the king bore the blame for the defeat is something modern historians weigh, not something the chronicle settles.
\nStephen thus did what he had once failed to do: he saw Jerusalem and died fighting for it. Our record of these operations places both counts under the heading "Expedition of 1101", even though Stephen's death falls in 1102, because the dataset sorts men by the crusade they had joined rather than by calendar year.
\nThe 1101 expeditions are often called the crusade of the faint-hearted, a label that fits the participants' reputation more than their record: tens of thousands set out, and few arrived. The defeat showed that the road through Anatolia could not be crossed by a mob any more than by an army, a lesson already paid for at Civetot in 1096. It also showed that the success at Antioch and Jerusalem had been fragile. Future crusades preferred sea routes, and the Genoese and Venetian fleets became part of how Latin settlement was supplied, a development that eventually carried a crusade as far as Constantinople in 1204.
\nThe sufferers were not only the counts. Pilgrim columns included women and non-combatants, whose fate the chroniclers rarely trouble to name, a gap explored in the role of women in the crusades. For the settlers who held the coast, the arrival of reinforcements was always uncertain, and the military orders such as the Knights Templar would in time fill the gap that these failed relief efforts exposed. Disease also carried off leaders on earlier campaigns, as the death of Adhemar of Le Puy at Antioch shows, and the same ledger of lords who died away from home runs through the later siege of Acre in 1291.
\nHugh of Vermandois and Stephen of Blois are remembered, when they are remembered, for what they gave up. The second campaign is the more honest measure of them. Both went back, one died on the road and one at the gates of the kingdom he had helped to win, and the chroniclers who recorded the deaths left several questions open. Where the sources are silent, whether on Hugh's last days or on the exact day of Stephen's, they should be left open.
\nWhy did Hugh of Vermandois and Stephen of Blois join the Crusade of 1101? Both had left the First Crusade before Jerusalem and were under pressure to fulfil their vows. Orderic Vitalis says Adela of Normandy pushed Stephen to go, though that account is not independently confirmed.
\nHow did Hugh of Vermandois die? Albert of Aachen, the source the dataset cites, records that he died of wounds at Tarsus after the defeat of the Aquitanian column in Anatolia. The chronicle gives few further details.
\nHow did Stephen of Blois die? He was killed at the battle of Ramla in May 1102, fighting alongside Baldwin I against an Egyptian army. Fulcher of Chartres, who was in Baldwin's household, is the main source.
\nWhy did the Crusade of 1101 fail? The armies marched separately through Anatolia, without the unity or supply of the First Crusade, and were defeated one by one by Turkish cavalry. The details of each battle are disputed.
\nTitus Labienus was the senior legate of Caesar's Gallic army, the officer who held the northern front against the Treveri, won a battle on the Seine in 52 BC while Caesar was failing at Gergovia, and then, in January 49 BC, abandoned his commander for Pompey. He is remembered chiefly for that last act. But the operational record is more interesting than the betrayal. Our record of Caesar's campaigns in Gaul and Britannia credits Labienus as the Roman commander of three separate operations, and every claim about them rests on one witness: Caesar himself, with a final book added by his staff officer Aulus Hirtius.
\nLabienus served as Caesar's legatus, a senatorial deputy who held delegated command rather than an independent province. He was already at Caesar's side in 58 BC, when Caesar left him in charge of the fortifications at Geneva against the Helvetii (BG I.10), and he led a flanking column in the opening manoeuvres of that war (BG I.21-22). From then on he is the officer Caesar most often entrusts with a detached command. That trust was itself a political fact: a legate with several legions and a free hand could win glory, and Caesar needed such men because he could not be everywhere.
\nCaesar's narrative style matters here. The Commentaries are in the third person and spare with praise. When Labienus is named as the victor, the reader should remember that Caesar chose to name him, and that Caesar wrote with his own reputation at Rome in mind. The sources leave out what Labienus thought, wanted or said in private.
\nThe Treveri lived in the Moselle valley around modern Trier, on the border between the Gallic world and the Rhine. They could call on Germanic mercenaries and fielded some of the best cavalry in Gaul, which is why Caesar posted Labienus there. In the winter of 54 BC the Treveri leader Indutiomarus, who had been a rival of the pro-Roman Cingetorix, tried to turn the revolt of the Eburones into a general rising. After the destruction of Sabinus and Cotta (the story told in our article on the ambush at Atuatuca), survivors reached Labienus's camp, and Indutiomarus came to besiege it (BG V.37, 53).
\nLabienus's answer, as Caesar tells it, was patient deception. He stayed behind his ramparts and let the Treveri grow contemptuous, then ordered his cavalry to concentrate on one target: Indutiomarus himself. When the Gauls approached in disorder, the Romans broke out, killed Indutiomarus as he tried to cross a river, and the siege collapsed (BG V.57-58). Caesar presents it as a lesson in using patience to defeat a bolder enemy.
\nIn 53 BC the Treveri were still unsettled. Labienus, with one legion and a cavalry force, used a feigned withdrawal to draw the Treveri across a river on unfavourable ground, and defeated them. Caesar then reports that the Treveri submitted and that Cingetorix was given authority over them (BG VI.7-8).
\nThe crisis of 52 BC is where Labienus's independence shows. While Caesar fought Vercingetorix in central Gaul, he sent Labienus north with four legions to deal with the Senones and Parisii (BG VII.34, 57). Labienus marched along the Seine towards Lutetia, the Parisian settlement on what is now the Île de la Cité in Paris, but met a Gallic force under the elderly Camulogenus, whom Caesar calls an Aulercan, holding a marsh that blocked the way (BG VII.57-58).
\nRebuffed, Labienus withdrew, seized boats at Metiosedum and crossed the river. He then marched down the opposite bank towards Lutetia, and the Gauls burned the town and its bridges and faced him across the Seine (BG VII.58). Caesar adds that Labienus had meanwhile learned of the Roman reverse at Gergovia and the revolt of the Aedui, and that he now needed to rejoin the main army rather than be cut off (BG VII.59). His solution was a night operation: he sent part of the force upriver with noise and boats to hold Gallic attention, while the main body crossed lower down. The next morning the Gauls were forced to fight on his terms, and Camulogenus was killed in the battle (BG VII.60-62).
\nThe exact site of the fight is not certain; Caesar's description does not permit a confident identification, and scholars have placed it on different stretches of the river. It is the operation most often cited when historians argue that Labienus was a capable independent commander (Gilliver; Goldsworthy).
\nCaesar's own narrative stops after Alesia. Book VIII is by Hirtius, and it must be read as the work of a loyal staff officer. Hirtius reports that in 51 BC Labienus again defeated the Treveri and their German allies in a cavalry action (BG VIII.45), and that in 50 BC Caesar placed him in charge of Cisalpine Gaul (BG VIII.52). In the same passage Hirtius mentions a rumour that Caesar's enemies were courting Labienus, and says Caesar did not believe it. Whether the rumour was already true is something we cannot establish from the text.
\n| Year | \nOperation | \nReference | \n
|---|---|---|
| 58 BC | \nFortifications at Geneva; flanking move against the Helvetii | \nBG I.10, 21-22 | \n
| 54 BC | \nCampaign against Indutiomarus and the Treveri | \nBG V.53-58 | \n
| 53 BC | \nDefeat of the Treveri | \nBG VI.7-8 | \n
| 52 BC | \nBattle of Lutetia | \nBG VII.57-62 | \n
| 51 BC | \nCavalry victory over the Treveri (Hirtius) | \nBG VIII.45 | \n
| 50 BC | \nPosted to Cisalpine Gaul (Hirtius) | \nBG VIII.52 | \n
| 49 BC | \nLeaves Caesar for Pompey | \nBC I.15 | \n
When civil war began in January 49 BC, Labienus left Caesar's camp and went over to Pompey. Caesar mentions it in the Civil War, and again reports Labienus as a prominent Pompeian in 48 BC (BC I.15; III.13). He fought against his former commander for the rest of his life, in the Pompeian cause in Greece and Africa, and according to later tradition died at Munda in Spain in 45 BC. Caesar's text offers no explanation; the later ancient writers, who wrote under the Empire, offer several, and none of them is contemporary evidence.
\nHistorians therefore argue. Some have emphasised that Labienus came from Picenum, a region where Pompey's family had great influence, so that his old loyalty to Pompey reasserted itself when Caesar crossed the Rubicon. Others suggest that Labienus saw that Caesar's march on Rome was an illegal act and chose the constitutional side, or that he resented being treated as an instrument rather than a partner after eight years of victories. The evidence does not settle the question, and each reading says as much about the historian as about the legate.
\nLabienus's Gallic career is best read as a case study in delegated command. Caesar's victories depended on officers who could win without him, and Labienus's three credited operations show what that meant: patience against Indutiomarus, a deception against the Treveri, a bold river crossing at Lutetia. They also show the limit of the source. We know Labienus mainly through the man he later fought, and that man chose how to present him. Alongside the lost legion at Atuatuca, the naval victory in Decimus Brutus's Veneti campaign and the conquest described in Publius Crassus in Aquitania, his commands illustrate how Caesar ran a war across a whole country. The Treveri frontier also belongs to the longer story of Roman operations on the Rhine, told in Drusus and Germanicus on the Rhine. The wider political world of this period, including the letters in which contemporaries watched the crisis of 50-49 BC unfold, appears in Cicero's private correspondence.
\nWho was Titus Labienus? He was Caesar's senior legate in Gaul from 58 BC, a deputy with delegated command who held the northern front and won independent victories over the Treveri and at Lutetia. He joined Pompey against Caesar in 49 BC.
\nWhich Gallic operations did Labienus command? Our dataset credits him as Roman commander of the 54 BC campaign against Indutiomarus, the 53 BC defeat of the Treveri and the 52 BC Battle of Lutetia (BG V.53-58; VI.7-8; VII.57-62). Hirtius adds a further Treveri victory in 51 BC (BG VIII.45).
\nWhy did Labienus leave Caesar? The sources do not say. Caesar gives no reason, and the explanations offered by historians, including a Picene link to Pompey and a constitutional objection to the march on Rome, are interpretations rather than evidence.
\nIs the account of Labienus reliable? Only partly. Caesar and Hirtius were both writing for Caesar's benefit, and Hirtius wrote knowing what Labienus later did, so the praise and the silences both need to be read with that in mind.
\n"Perimortem" means "around the time of death", and it is the word on which almost every claim about a prehistoric massacre rests. It is also a word that bone cannot fully back: a skeleton records the condition of the bone when it broke, not the hour, and those two things are not the same.
\nOsteologists sort injuries into three groups. Antemortem injuries happened while the person lived and have begun to heal. Bone that has remodelled, with rounded edges and new growth around a break, is the signature of someone who survived. Postmortem damage happened to dry bone, long after death, through the weight of soil, the plough, the trowel or the weather. Perimortem sits between them: an injury with no sign of healing, made while the bone still behaved like living bone.
\nThe first two are easy to describe. The third is a boundary that moves. Fresh bone is not hard like a stone. It holds collagen and moisture, and when it breaks it tends to break the way a green branch does. Dry bone breaks the way old timber does. The forensic literature lists the differences: fresh bone tends to give curved or oblique fracture lines with smooth surfaces and a fracture face of the same colour as the outer bone, while dry bone tends to give right-angled, rougher breaks with a paler fracture face (Ubelaker and Adams set out such indicators in 1995). Those are tendencies, not switches.
\nThe practical problem is that "fresh" is a property of the bone, not of the calendar. A body left in a warm, dry place and a body in a cool, wet pit lose their moisture and collagen at very different rates. A fracture made days or weeks after death can therefore look perimortem. Anyone who sets out to hack apart a corpse that is still fresh produces what looks exactly like trauma around death, whether or not that trauma killed the person.
\nAn experiment published in 2008 shows how thin the margin is. Bruce Wheatley broke 76 deer femora at different intervals after death: a "fresh" group less than four days dead, and a "dry" group dead 44 days or a year. The fracture features did separate the groups statistically. But with one exception, all seven features of so-called wet-bone fractures also turned up on bones that were a year old, and he concluded that a perimortem call on an individual bone should be made with caution. It was an experiment on deer, not on people, so it is a warning and not a measurement. It is still a useful one.
\nSo the word has a built-in limit. It says that an injury was inflicted while the bone was fresh. It does not say who did it, why, or whether the person was alive.
\nDespite that, the term does most of the work in the archaeology of violence. Our record of collective violence before 3000 BCE admits a site only where the excavators identify perimortem injury attributable to violence, and it states in the same breath that healed trauma shows a violent life, not a violent death. The word is a gatekeeper because nothing else is available: there are no witnesses and no written records, only bone.
\nWhere the evidence is strong, the reading is not seriously disputed. At Talheim in Germany, around 5000 BC, the pit of thirty-four people carries injuries concentrated on the head, of the kind that axes and adzes make, in bodies that were dumped rather than laid out. The pattern, the unhealed breaks and the manner of deposition point the same way. Nothing in that case depends on a single ambiguous fracture.
\nWhen bone is read against its own burial. At Nataruk in Kenya, the excavators reported in Nature in 2016 that ten of twelve articulated skeletons showed evidence of violent death at a lagoon edge roughly ten thousand years ago. In the same journal, Christopher Stojanowski and colleagues argued that some of the cranial lesions are more likely to be post-depositional, and specifically that three crania show damage from soil compression. The excavators replied. The perimortem label is the whole argument here, and the two sides disagree about whether the damage was done before or after the bone entered the ground. Our register records Nataruk at low certainty for that reason, and its inclusion is not an endorsement of either reading.
\nWhen the bone was broken on purpose, after death. At Herxheim in the Rhineland, hundreds of people's bones were smashed and cut around 5000 BC. Part of the excavation team, in a 2009 paper in Antiquity, read this as cannibalism. Other members of the same team read it as an elaborate secondary rite for remains brought from elsewhere. The Herxheim dispute is a case where "perimortem" is accurate, because the bone was fresh when worked, yet tells nothing about how anyone died. Processing marks show what happened to a body, not what killed it.
\nWhen the same skeletons carry both kinds of injury. The 2021 restudy of Jebel Sahaba on the Nile found healed and unhealed injuries in the same cemetery, and a quarter of the skeletons with lesions carried both. The healed breaks are antemortem, the unhealed ones are perimortem, and between them they changed the reading from one event to a long pattern of violence. Here the distinction between the two words is exactly what turned the interpretation.
\n| Question | \nAntemortem | \nPerimortem | \nPostmortem | \n
|---|---|---|---|
| Bone condition | \nAlive | \nFresh, still elastic | \nDry, brittle | \n
| Healing | \nPresent (remodelling) | \nAbsent | \nAbsent | \n
| Typical fracture | \nHealed callus | \nCurved, smooth, colour matches | \nAngular, rough, paler face | \n
| What it can support | \nA violent life | \nAn injury near death, or on a fresh body | \nBurial, plough or excavation damage | \n
| Main risk | \nTaken as cause of death | \nTaken as cause of death | \nMistaken for earlier trauma | \n
None of this means that prehistoric massacres are fiction. It means that the strong cases are strong because several lines agree: the injury type, the weapon the wounds suggest, the absence of healing, the body positions and the way the dead were deposited. The enclosure at Asparn-Schletz belongs in that company.
\nThe weak cases are the ones in which perimortem appearance stands alone. A historian writing about a single fractured skull should say what the bone shows: that it broke while still fresh. Whether that happened in a fight, in a ritual, in an accident with a grave or under a bulldozer is a separate argument, and it needs evidence from the context rather than from the break.
\n"Perimortem" is the most useful word in the osteology of violence and the most overworked. It names a condition of bone, not a moment, and it cannot separate a killing from the treatment of a body that was still fresh. The best prehistoric cases rest on much more than the word, and the most contested ones, from Nataruk to Herxheim, are contested precisely where it carries the whole weight. A careful reader asks, of every claim, what the word is standing in for.
\nWhat does perimortem mean? It describes an injury inflicted at or around the time of death, while the bone was still fresh and had not healed. It is a judgement about the condition of the bone, not an exact time of death.
\nHow do archaeologists tell perimortem from postmortem damage? They look at the fracture outline, the surface and colour of the break, and the absence of healing, then at the burial context. Experiments on animal bone show that these features overlap, so the call is made with caution and with other evidence.
\nDoes a perimortem injury prove that someone was killed? No. It shows that the bone was fresh when it was damaged, which can also be true of a body cut apart or moved shortly after death. Establishing violence needs further evidence, such as weapon-type injuries, embedded points or the pattern across many bodies.
\nWhy does healed trauma not count as evidence of a massacre? A healed injury means the person survived it, which indicates a violent life but not a violent death. At Jebel Sahaba, healed and unhealed injuries together are what suggest repeated violence over time.
\nOn 24 April 1975 six members of the Rote Armee Fraktion walked into the West German embassy in Stockholm, took twelve people hostage and demanded the release of 26 prisoners in West Germany. Twelve hours later the building was burning, two hostages and one attacker were dead, and a second attacker was dying. The demand had been refused, and Sweden's police had been shown to be unready for an event of this kind.
\nThe word "Stockholm" in hostage history usually points to the 1973 bank robbery at Norrmalmstorg, which gave "Stockholm syndrome" its name. This is a different event, two years later and a few kilometres away: a political seizure by a West German armed group, on foreign soil, aimed at a government that was not Sweden's. That Sweden had no part in the dispute is the first fact about it.
\nThe group called itself Kommando Holger Meins, after a prisoner who had died on hunger strike in Wittlich prison on 9 November 1974. Naming an operation after a dead comrade was the group's settled habit, discussed in our piece on the Kommando statement. The six were Siegfried Hausner, who led the group, Karl-Heinz Dellwo, Hanna Krabbe, Bernhard Rössner, Lutz Taufer and Ulrich Wessel.
\nTheir demand was the release of 26 imprisoned members and associates, who were to be flown out of West Germany. It was addressed to the Bonn government, led by Chancellor Helmut Schmidt, and it came at a moment when the group's leading figures were facing trial, a trial that eventually needed a purpose-built courtroom building at Stammheim. The attackers said they carried 15 kilograms of TNT and would blow up the building if the police intervened.
\n| Time (24 April 1975) | \nEvent | \n
|---|---|
| About noon | \nSix armed people enter the embassy on Skarpögatan, in the Gärdet district, and take the staff hostage; Ambassador Dietrich Stoecker is among the twelve held | \n
| Afternoon–evening | \nThe demand is passed to Bonn; the German government refuses to release the prisoners | \n
| During the siege | \nMilitary attaché Colonel Andreas von Mirbach and economic attaché Heinz Hillegaart are killed by the occupiers | \n
| Shortly before midnight | \nThe explosives detonate and the building catches fire; Ulrich Wessel is killed, Siegfried Hausner is gravely burned | \n
Two points about this record need care. The sources agree that the two attachés were killed by the occupiers, but the order of events inside the building, and what exactly was said at each deadline, rests largely on later testimony and on the court record, not on any neutral observer. And the sources describe the explosion as apparently accidental; the occupiers' own intentions at that moment are not something an outside historian can establish with certainty.
\nThe central political fact is a refusal. The government in Bonn did not release the prisoners, and it held to that. Whatever one makes of the decision, it set a pattern that was repeated in the autumn of 1977, which we trace in the German autumn, day by day: the state would not trade prisoners for hostages, and the cost of that policy would be borne by the people held.
\nThe Swedish side is the least known part of the story. The scholar Dan Hansén, in his study of the episode, concludes that for everyone involved it was obvious that the Swedish police did not have the capacity to act professionally in a terrorist situation. The group wanted something no Swedish official could grant, which left the police with little to do but wait.
\nThe more striking point is what followed. Hansén's argument is that the decade after the siege was marked by stasis in Swedish counter-terrorism policy making. Other states moved faster. West Germany had created its GSG 9 unit after the 1972 Munich Olympics massacre, and Norway installed a comparable force in 1975, in the wake of what had happened in Stockholm. Sweden's own response, on this account, was unusually limited for a crisis so plainly instructive.
\nThat is an interpretation by one research team, and it is worth keeping apart from the narrative facts above. But it is a useful correction to the assumption that a dramatic failure automatically produces reform.
\nHausner was flown to West Germany, where he died of his injuries on 4 May 1975, at Stammheim. The four surviving attackers, Krabbe, Dellwo, Taufer and Rössner, were tried by the Higher Regional Court in Düsseldorf and on 20 July 1977 each received two life sentences for two murders, with hostage-taking as an aggravating circumstance.
\nFor the people killed, the record is plain and should be said plainly. Andreas von Mirbach and Heinz Hillegaart were embassy officials, not parties to the dispute the attackers were pursuing.
\nThe siege is the clearest example of the RAF's second phase, as the German federal agency for civic education describes it: an organisation that now acted to free its imprisoned members, not merely to attack targets. Four features stand out.
\nThis is why the siege is a separate row in our record of these operations. The later decades, covered in the third generation, look very different, because convictions became rare.
\nStockholm 1975 is remembered, when it is remembered, as one more RAF operation. Read closely it is two stories. One is German: a refusal to negotiate that the state held to at the cost of two lives. The other is Swedish: a police force that was publicly shown to be unready and a political system that, on one careful reading, learned remarkably little from it.
\nIs this the event that gave "Stockholm syndrome" its name? No. That term comes from the 1973 bank robbery at Norrmalmstorg. The 1975 embassy siege was a political seizure by a West German group and is a separate event.
\nWho was killed at the embassy? Military attaché Colonel Andreas von Mirbach and economic attaché Heinz Hillegaart were killed by the occupiers. Among the attackers, Ulrich Wessel died in the explosion and Siegfried Hausner died of his injuries ten days later.
\nDid the German government give in? No. The Bonn government refused to release the 26 prisoners whose release was demanded, and the siege ended with the explosion shortly before midnight.
\nWhat did Sweden change afterwards? According to Dan Hansén's study, very little for about a decade, even though the police had shown they lacked the capacity to deal with such an event. Norway, by contrast, set up a comparable unit in 1975.
\nWitte Corneliszoon de With (1599-1658) circled the globe, helped capture a Spanish silver fleet, led the Dutch at Kentish Knock and died with his flag still flying in the Sound, yet his countrymen never warmed to him. He was feared by his crews, shunned by his peers and openly insubordinate to his superiors, and the Dutch Republic used him anyway, because it had few commanders who could do what he did.
\nDe With was born at Brielle on 28 March 1599, in the town that the Sea Beggars had taken in 1572 and that became a cradle of Dutch naval men. According to the standard biographical accounts he was a childhood rival of Maarten Tromp, who came from the same town, and the rivalry never really ended. He went to sea in 1616 on trading voyages to the East Indies and India, and in 1620 took part in an expedition against the Barbary corsairs.
\nIn 1626 he was vice-admiral of a small spice fleet that, by the account most biographies give, came home by way of the Pacific and so completed a circumnavigation. That alone made him unusual among Dutch commanders, who were mostly bred on the cold, short routes of the North Sea. He had also, by then, seen the Spanish possessions on the American Pacific coast from the sea. The Republic's war with Spain was a war for trade as much as for territory, as we describe in how the Dutch East India Company transformed global trade networks.
\nIn 1628 he served in the fleet of Piet Hein when it took the Spanish silver fleet in the Bay of Matanzas, the capture we examine in Piet Hein and the Silver Fleet. In 1637 he returned to the navy proper and, in 1639, fought beside Tromp at the Battle of the Downs, where a Spanish armada was destroyed in English waters. By then he was a Holland officer of the first rank.
\nThe Dutch navy of those years was not a single fleet. It was run by separate admiralties, each with its own money, ships and local loyalties, a structure laid out in the five admiralties of the Dutch Republic. That matters for De With: he was a Hollander in a force where Zeeland, Friesland and Holland officers watched each other closely, and his temper did not help.
\nIn 1647 the States-General sent him to Dutch Brazil with a fleet that was poorly equipped, to help the colony against the Portuguese. What followed was a long quarrel with the colonial council. His ships were not supplied, the fleet decayed, and in November 1649 he sailed home against orders with the two ships still seaworthy.
\nAt The Hague he went to complain about the colonial government and was arrested instead. He was charged with insubordination and desertion on 259 counts and came close to a death sentence. According to the accounts we follow, the States of Holland intervened, pointing out that only they could condemn one of their admirals to death. In February 1651 he was acquitted of most charges and lost only the pay for the period involved. By September 1651 he was back on convoy duty.
\nWhether he was a deserter or a man left to rot by a bankrupt colonial administration is a question historians have answered differently. The West India Company was short of money and quarrelled with everyone, and De With was not a man who bore this quietly.
\nThe Dutch naval historian J. C. M. Warnsinck, as the standard summaries have it, judged that De With was feared and hated by his inferiors, avoided by his equals and always full of defiance toward his superiors. Crews on more than one occasion are said to have refused him as a flag officer on their ship. These judgements come from later historians and from Orangist and Holland party feeling, so they deserve some caution. But the pattern is consistent with his Brazil record.
\nHis talent was never in doubt. He was bold in the attack, a capable seaman and unusually well travelled. The Republic's difficulty was that boldness and obedience rarely arrived in the same officer. A navy built on committee compromise had little room for a man like him, and had little choice but to use him anyway.
\n| Date | \nEvent | \n
|---|---|
| 28 March 1599 | \nBorn at Brielle | \n
| 1616 | \nGoes to sea; voyages to the East Indies | \n
| 1626 | \nVice-admiral of a spice fleet; returns round the world | \n
| 1628 | \nServes under Piet Hein at Matanzas | \n
| 1639 | \nFights beside Tromp at the Battle of the Downs | \n
| 1647-1649 | \nDutch Brazil; returns home against orders | \n
| February 1651 | \nAcquitted of most charges | \n
| 8 October 1652 | \nCommands the fleet at Kentish Knock | \n
| 8 November 1658 | \nKilled at the Battle of the Sound | \n
When the First Anglo-Dutch War began, Tromp was the Republic's best-known commander. After he failed to bring the English to battle off the Shetlands in August 1652, with the Dutch fleet scattered by weather, the States-General suspended him. The command passed to De With, then the Holland vice-admiral. The appointment was political from the start, and it put him over the man he had rivalled since boyhood.
\nThe battle was fought on 8 October 1652 by the Dutch calendar (28 September in England) on the shoals of the Kentish Knock, off the Thames estuary. De With had roughly sixty ships against an English fleet under Robert Blake of somewhat more. Many of his captains were of doubtful quality, and the Dutch fleet became disordered; a gale and the shoals did the rest. The Dutch lost two ships and withdrew. Our own record of these operations in the dataset of Dutch Golden Age naval battles lists the battle as an English victory and names him as the Dutch commander.
\nIt was a defeat, not a disaster, and he was not alone to blame: captains who would not follow signals plagued the Dutch navy throughout the war. But it was enough. Tromp was reinstated and won at Dungeness on 30 November. The Republic's political and commercial pressures were already heavy, as the disaster year 1672 shows later on, and no committee in The Hague was inclined to forgive a loser.
\nThe last act came in the Baltic. In the Second Northern War, Sweden besieged Copenhagen, and the Dutch, who lived on Baltic grain and timber, could not allow the Sound to fall under one power. Lieutenant-Admiral Jacob van Wassenaer Obdam led a relief fleet, with De With as vice-admiral commanding the van.
\nOn 8 November 1658 the two fleets met at the Sound north of Copenhagen, and the Dutch won, relieving the city. The price was De With's life. His flagship Brederode, a ship of some 54 guns, was held by the current and grounded under the guns of several Swedish ships, among them the Draken and the Wismar. Fire broke out, and De With was shot dead in the fighting. The Swedes took the flag. By tradition he fought until he fell; the Dutch sources that tell this were, of course, written by his own side.
\nThe Swedes embalmed his body and displayed it at Elsinore before sending it to Copenhagen in January 1659. He was later buried in Rotterdam. The Rijksmuseum and the National Maritime Museum in Greenwich both hold paintings of the battle, which show how much the Sound was a battle of crowded ships in a narrow channel.
\nHis reputation after death was mixed. In Rotterdam his name was given to a street and, later, to a well-known arts centre, which dropped the name in the 2020s. The museum Gevangenpoort in The Hague has written about him as a naval hero with a tainted name. We would add that the label cuts both ways: the same man who was court-martialled was also one of the few Dutch admirals to die in the line of battle.
\nDe With was not a likeable man and perhaps not an easy officer to serve, but the Republic never found a replacement for his skill. His career reads as a record of a navy that depended on talent it could not control. For what Dutch wealth made possible at home, see scientific societies and knowledge networks in the Dutch Golden Age, and for the earlier Spanish naval threat that shaped the Republic's origins, the Spanish Armada of 1588.
\nWhy was Witte de With so unpopular? Contemporaries and later historians describe him as harsh to his crews, quarrelsome with his equals and defiant of superiors. Some of that picture comes from later writers with their own party loyalties, so it should be read with care. His Brazil conduct and the 259-point charge sheet did nothing to improve it.
\nDid De With lose the Battle of the Kentish Knock? Yes, the Dutch were beaten there on 8 October 1652, losing two ships. Blame is usually spread across a disordered fleet, doubtful captains and bad weather, as well as the admiral, who had taken over after Tromp was suspended.
\nHow did Witte de With die? He was killed at the Battle of the Sound on 8 November 1658, when his flagship Brederode was grounded and boarded by Swedish ships. His body was embalmed by the Swedes and later sent to Copenhagen, then to Rotterdam.
\nWhat was his quarrel with Maarten Tromp? Both were from Brielle and the biographies describe a lasting rivalry. It mattered politically in 1652, when De With took the command from which Tromp had been suspended.
\nTwice in six months, Richard the Lionheart brought the army of the Third Crusade to Beit Nuba, about twelve miles from Jerusalem, and twice he turned it around without fighting a battle. The retreats have been read as cowardice, as piety turned to politics, or as the plain realism of a commander who understood a siege better than his critics did. The sources agree on the sequence and disagree, sharply, on whom to blame.
\nAfter the victory at Arsuf in September 1191 and the refortification of Jaffa, the crusaders held a strip of coast. Jerusalem lay inland, in the hills, and had been in Muslim hands since Saladin took it in October 1187. For the crusaders it was the stated purpose of the whole expedition, and for Saladin it was the proof of his legitimacy as a champion of Islam. Both sides understood what was at stake.
\nRichard, however, had reasons to move slowly. His army was a coalition: Angevin and English troops under his own command, the French contingent left behind by Philip II under Hugh III of Burgundy, the Templars and Hospitallers, and the barons of the surviving Latin kingdom, whose interests did not always match those of visitors from the West. His conquest of Cyprus earlier that year had given him a base, but not a united army.
\nRichard moved inland from Jaffa in the autumn and the army spent Christmas near Latrun. In early January 1192 it reached Beit Nuba, which the Latin sources call Betenoble. Saladin, for his part, had dismissed much of his field army to winter quarters in December, a decision that the Muslim sources present as a sensible response to the season and the Latin sources later describe as an opportunity that was missed.
\nWhat happened next is the part of the story that depends on the chroniclers. The weather turned savage. The Itinerarium describes torrential rain, hail and cold that rusted armour, spoiled food and rotted tents, and the army, already worn by months on campaign, was losing men to sickness and desertion. A council was held. According to the Latin accounts the Templars and Hospitallers, together with the barons of the Latin kingdom, advised against a siege: even if the city fell, the army would melt away once the pilgrims had fulfilled their vows, leaving too few men to hold it against Saladin's return. Richard accepted the advice. The army fell back to Ramla and then to Ascalon, which Richard spent the spring rebuilding.
\nBetween the two advances much changed. In late April 1192 Conrad of Montferrat, elected king of Jerusalem only days before, was murdered in Tyre. Henry of Champagne, Richard's nephew, married Conrad's widow and took over the title. Richard took Darum in May and then, in June, led the army inland again. Bahā' ad-Dīn reports that Saladin was warned on 6 June that the Franks were moving in force, and the Muslim command, which had expected the winter campaign to be the last, began to prepare for a real siege. Richard's army captured a large Egyptian supply caravan at Tell al-Khuwaylifa in the same month, which hurt Saladin's logistics and raised crusader spirits.
\nThe army again reached Beit Nuba and again, the sources say, debated. This time the argument was about method as much as principle. Richard, by the account most favourable to him, pointed to the heat and the shortage of water: Saladin's men had been fouling or blocking the wells around the city, and an army besieging Jerusalem in July would be thirsty and exposed. Hugh of Burgundy and the French wanted to attack at once. Richard's alternative, reported in some sources, was a strike on Egypt, the economic base of Saladin's power. The army split into two factions, neither strong enough to carry out its plan. Richard is said to have offered to go to Jerusalem as a simple soldier but not to lead the army. Without a single commander, the host drew back to the coast in early July.
\nHistorians weigh the same few explanations.
\nSupply and water. Jerusalem sits on a limestone ridge, far from the coast and its fleet. Every sack of flour and every barrel of water had to come up from Jaffa through country where Saladin's cavalry could cut the road. This was the argument Richard himself is reported to have made.
\nHolding the city. A siege might succeed, but the army that took Jerusalem would need to hold it for years. The military orders, who had learned this at Hattin, said that the crusaders did not have the men to do so.
\nPolitics. Richard's brother John was intriguing with Philip II, and the king knew that he could not stay in the East indefinitely. Gillingham argues that this pressure explains his preference for a negotiated settlement over a gamble. Our record of these operations lists both advances as campaigns withdrawn without battle, which is a useful reminder that the Third Crusade's most celebrated moments were also its least decisive.
\nCommand. The coalition had no single will. When the French and the Latin barons wanted one thing and Richard another, the army could neither take Jerusalem nor agree to do otherwise.
\nA later legend has Richard climb a hill and see the city in the distance, then cover his face with his shield, saying that he was not worthy to look on a city he could not free. Nothing in the surviving contemporary chronicles proves it, and it should be read as a later story that expresses what people felt about the retreat rather than what Richard did.
\n| Date | \nEvent | \n
|---|---|
| October 1187 | \nSaladin takes Jerusalem | \n
| September 1191 | \nBattle of Arsuf; Jaffa is refortified | \n
| December 1191 | \nSaladin dismisses much of his army; crusaders spend Christmas near Latrun | \n
| Early January 1192 | \nFirst advance reaches Beit Nuba | \n
| January 1192 | \nCouncil; the army withdraws to Ramla and Ascalon | \n
| April 1192 | \nConrad of Montferrat is murdered; Henry of Champagne becomes ruler of the kingdom | \n
| May 1192 | \nRichard takes Darum | \n
| June 1192 | \nSecond advance reaches Beit Nuba; Egyptian caravan is captured | \n
| Early July 1192 | \nArmy withdraws to the coast | \n
| August 1192 | \nThe battle of Jaffa | \n
| September 1192 | \nThe truce signed at Jaffa | \n
The two marches show a commander who could bring an army within sight of its objective and who judged, twice, that reaching it was not the same as keeping it. The judgement was probably sound on military grounds, and the kingdom that survived on the coast outlasted Richard by a century, until the fall of Acre in 1291. But it was also a judgement that left the crusade without its central prize, and later crusades, from the Fourth Crusade onward, show how little the movement could do without a leader whom all its factions would follow. The Templars, who advised caution at Beit Nuba, would spend the next hundred years defending the same narrow coast.
\nHow close did Richard get to Jerusalem? The crusader army reached Beit Nuba, about twelve miles from the city, on both advances: in January 1192 and again in June 1192. Some later accounts say the walls could be seen from nearby heights, but the contemporary sources do not agree on this.
\nDid Richard ever see Jerusalem? A well-known story says he refused to look at it. The surviving contemporary chronicles do not record that, and historians treat it as legend. Richard left the Holy Land in October 1192 without having entered the city.
\nWhy did the military orders advise against a siege? According to the Latin chroniclers, they feared that the army could not hold Jerusalem once the pilgrims went home. They also knew that Saladin could cut the road from the coast, which made supply and water a constant problem.
\nWas Richard's decision a mistake? Historians disagree. Gillingham defends it as sound military judgement, while others point to Saladin's weakened army and see a missed opportunity. The evidence, mostly written by partisans on one side or the other, cannot settle the question.
\nIn the spring of 1099 the First Crusade spent three months outside a small fortress in what is now northern Lebanon, lost two of its lords there, and left on 13 May having captured nothing. Arqa was not Jerusalem and not Antioch. It stood in the way of no one's pilgrimage vow, and the siege was the work of one ambitious count. The deaths of Anselm of Ribemont and Pons of Balazun at its walls are among the plainest evidence of what a stalled siege cost the men who fought it.
\nBy the winter of 1098–99 the crusade had come apart. Adhemar of Le Puy, the papal legate, had died at Antioch, Bohemond of Taranto had kept the city for himself, and the army that was supposed to march on Jerusalem sat idle for months while its princes argued. The rank and file eventually forced the issue. Raymond of Toulouse, the wealthiest of the leaders, moved south along the coast road and, on 14 February 1099, began to besiege Arqa.
\nHis motives are debated. A common reading is that Raymond wanted a lordship of his own on the Lebanese coast, centred on Tripoli, to offset Bohemond's gains at Antioch, and that Arqa was the first step toward it. Runciman gives the siege that political shape. Raymond's chaplain Raymond of Aguilers, our main narrative source, presents the whole affair in the language of the pilgrimage instead, and that framing should be read as the count's advocate speaking.
\nThe besiegers were not a single army. Tancred and Robert of Normandy had entered Raymond's service, since he could pay them. Godfrey of Bouillon refused that arrangement and spent part of the period elsewhere on the coast. The result was a siege run by a coalition whose members disagreed about what it was for.
\nNeither Anselm nor Pons has a chronicle of their own. They survive in a few lines of the sources and in two rows of our record of these operations, which counts the leaders of the crusading armies who died on campaign.
\nAnselm of Ribemont held the lordship of Ribemont and Ostrevant, in the north of France. He is better documented than most men of his rank for one reason: he wrote home. His letters to Archbishop Manasses II of Reims describe the capture of Nicaea and the siege of Antioch, and they are among the few eyewitness accounts of the crusade written by a lay lord during the campaign rather than recollected afterward. They end in 1098. Anselm died on 25 February 1099, eleven days after the siege began, as the standard modern reference lists have it.
\nPons of Balazun held the lordship of Balazun in southern France and served in the Provençal army of Raymond of Toulouse. His name is attached to one of the crusade's principal narratives. Raymond of Aguilers named him as the colleague with whom he wrote the Historia Francorum. When Pons was killed at Arqa, Raymond carried the work on alone, which is why the chronicle's later chapters read as one man's voice.
\nThe dataset records both men as killed in action at Arqa, and it cites the chronicle for each death: for Anselm, Raymond of Aguilers, Historia Francorum, chapter XVI (Recueil des historiens des croisades, Historiens occidentaux III, p. 277); for Pons, the same author, chapter XV.
\nThe wording matters. Modern summaries commonly say that both men were struck by stones thrown from or around the defences, and that Pons died of the blow of a stone missile. This article repeats that only as a summary of the secondary literature. The dataset row for each man says "killed in action" and gives a chapter reference; it does not give a mechanism, and I have not confirmed the mechanism against the Latin text. Where the chronicle falls silent on how a man died, the honest statement is that the source names the death and not its cause.
\nWhat the sources do show is the kind of siege it was. A fortress on a hill, defended by a garrison that sallied out and threw missiles down, could kill lords as easily as foot soldiers. The leaders' rank did not keep them back from the walls. A siege that had no clear way to end exposed them for as long as it lasted.
\nThe siege is also remembered for a quarrel that had nothing to do with its walls. Peter Bartholomew, the Provençal who had reported the discovery of the Holy Lance at Antioch, was still with the army, and doubts about the relic had not gone away. In April, Arnulf of Chocques, a clergyman in Robert of Normandy's household, challenged him to prove the lance genuine by ordeal of fire. Peter went through the fire and died of his injuries some days later.
\nRaymond of Toulouse had staked a great deal on the lance, and Raymond of Aguilers had defended it in his chronicle. Peter's death damaged both. Historians read the episode as a blow to the count's standing as the crusade's leader, and some argue it hurt him more than the stalled siege. Others note that the sources are partisan on both sides, so the exact effect is difficult to measure.
\nMeanwhile the siege produced nothing. The garrison held, and the army outside it was losing time, men and cohesion. Arqa was lifted on 13 May 1099. The crusaders left the fortress untaken and marched south, passing along the coast and then inland toward Jerusalem, which they reached on 7 June and took on 15 July. Tripoli, the larger prize behind Arqa, did not fall to them until 1109.
\nThe fortress had cost the crusade two lords and three months. It had also shown that the princes could not be held to a single plan without the legate who had once held them together. When the army finally moved, it moved because the poorer pilgrims and the lesser knights demanded it, as it had before at Antioch.
\n| Date | \nEvent | \n
|---|---|
| 1098 | \nAnselm of Ribemont writes his letters to Archbishop Manasses of Reims | \n
| 1 Aug 1098 | \nAdhemar of Le Puy dies at Antioch | \n
| 14 Feb 1099 | \nRaymond of Toulouse begins the siege of Arqa | \n
| 25 Feb 1099 | \nAnselm of Ribemont is killed at Arqa | \n
| 1099 | \nPons of Balazun is killed at Arqa (the sources give no more exact date) | \n
| April 1099 | \nPeter Bartholomew undergoes the ordeal by fire and dies of his injuries | \n
| 13 May 1099 | \nThe siege is lifted without result | \n
| 7 June 1099 | \nThe crusaders reach Jerusalem | \n
| 15 July 1099 | \nJerusalem is taken | \n
Arqa is easy to skip when the First Crusade is told as a story of Antioch and Jerusalem. It was a side-quarrel fought for a count's ambitions, and it ended with no victory to remember. For that reason it shows what the great battles do not: how the ordinary arithmetic of siege warfare removed lords from the army, men like Anselm and Pons, who left no deeds for a poet and a few lines for a chronicler. The same arithmetic runs through the whole history of the crusader states, from the destruction of the first wave at Civetot to the fall of Acre in 1291. Behind each siege stood the pilgrims and the women who travelled with the camp, and the military orders that later took over the burden of holding such fortresses.
\nWhere was Arqa and why did the crusaders besiege it? Arqa lay in what is now northern Lebanon, inland from the coast near Tripoli. Raymond of Toulouse besieged it from 14 February 1099, and many historians think he hoped to carve out a lordship around Tripoli to rival Bohemond's Antioch, though his chaplain's chronicle does not say so.
\nWho were Anselm of Ribemont and Pons of Balazun? Anselm was lord of Ribemont and Ostrevant and the author of surviving letters to Archbishop Manasses of Reims. Pons was lord of Balazun in southern France and a collaborator of Raymond of Aguilers on his chronicle. Both died at Arqa.
\nHow did they die? The dataset records both as killed in action at Arqa, citing Raymond of Aguilers, chapters XV and XVI. Modern summaries often mention stones thrown at the besiegers, but the chronicle reference alone does not let this article state the cause with confidence.
\nWhy did the siege fail? The sources point to a divided command, a garrison that held out, and the damage done to Raymond's authority when Peter Bartholomew died after the ordeal by fire. The crusaders lifted the siege on 13 May 1099 and marched for Jerusalem.
\nIn the summer of 56 BC, a legate of about thirty, with twelve cohorts and a large cavalry force, marched into Aquitania, the south-western third of Gaul, and defeated the Sotiates, the Vocates and the Tarusates in a single season. His name was Publius Licinius Crassus, and three years later he would die at Carrhae beside his father, the triumvir Marcus Crassus. Because Carrhae is the part of his life everyone remembers, his Gallic service has almost vanished behind it. Yet it is the only stretch of his career for which we have a detailed account of him giving orders, and Caesar's third book devotes most of eight chapters to it (BG III.20-27). The campaign appears in our record of Caesar's operations in Gaul and Britannia as the conquest of Aquitania, commanded by Crassus rather than by Caesar.
\nPublius Crassus had come to Gaul as one of Caesar's legates, an appointment that tied Caesar to the richest man in Rome. In 58 BC, at the battle against Ariovistus, he was in command of the cavalry and saw that the Roman third line was needed; he ordered it forward, and Caesar credits that decision with restoring the day (BG I.52). In 57 BC Caesar sent him with a single legion to the Atlantic coast, and the tribes of Armorica, including the Veneti, offered submission (BG II.34). That submission was brittle: the next winter it broke over requisitioned grain, and the naval war against the Veneti followed.
\nFor 56 BC Caesar had several fronts. He himself dealt with the Veneti and the coastal tribes, Titurius Sabinus went against the Venelli and their neighbours in Normandy, and Crassus was sent south. Caesar's stated reason is that Aquitania was large and populous, and that he wished to prevent its tribes from sending reinforcements to the rest of Gaul (BG III.11). Aquitania differed from the lands to its north: Caesar opens his work by noting that its people differed from the Celts and Belgae in language, customs and laws (BG I.1). Modern linguists generally connect the personal names on later Aquitanian inscriptions with ancestral Basque, which agrees with him.
\nCrassus began with recruitment. He drew reinforcements from Tolosa (Toulouse), Carcaso (Carcassonne) and Narbo (Narbonne), the communities of the Roman province nearest to Aquitania, and brought them into Sotiates territory (BG III.20). The Sotiates met him with a large cavalry force and attacked his marching column. After a cavalry fight in which the Romans eventually prevailed, the Sotiate infantry, who had been hidden in a valley, came out and the battle was renewed. It ended in a Roman victory, and Crassus moved on their main town (BG III.20-21).
\nThe siege showed something about the Aquitani. When Crassus advanced siege-sheds and towers against the walls, the defenders made sorties and then tried to undermine his works with tunnels, which Caesar says they were expert at because of their copper mines (BG III.21). Crassus's diligence defeated the tunnelling, according to Caesar, and the Sotiates, seeing no chance of success, asked for terms. He accepted their surrender and ordered the weapons handed over. While this was going on, the king Adiatunnus attempted a sortie from another side of the town with 600 soldurii, bound followers who, in Caesar's description, shared in all the good things of life with the man to whom they had devoted themselves and died with him if he died (BG III.22). Caesar says the sortie was driven back, and that Adiatunnus was nevertheless given the same terms as the rest.
\nCrassus then moved against the Vocates and Tarusates. Their reaction is the more interesting part of the campaign. They sent for help across the Pyrenees, and received men from Hispania Citerior, among them leaders who, in Caesar's words, had served under Quintus Sertorius and were thought to be experienced commanders (BG III.23). Sertorius had waged a long guerrilla war against Rome in Spain until his murder in 72 BC, and Caesar presents these veterans as men who intended to fight the Roman way, with fortified camps and cut supply lines.
\nThat put Crassus in a difficult position. His force was smaller than the enemy's, his supply lines were exposed, and he held a council about whether to fight at once or wait (BG III.24). Caesar tells us that he judged delay would strengthen the enemy and that his own men were eager. The next day he drew up his army but found the enemy staying behind their fortifications. An assault on a well-held camp was risky. According to Caesar, Crassus therefore sent a force round the back of the enemy camp, where the fortifications were weaker, while the main body attacked the front and kept their attention (BG III.25-26). The result was a rout. Caesar says that of an enemy force of more than 50,000 drawn from the Aquitani and the Cantabri, barely a quarter survived (BG III.26). Treat that figure with caution: it is Caesar's, and ancient battle counts regularly grow with the glory of the victor.
\n| Date | \nEvent | \nSource | \n
|---|---|---|
| 58 BC | \nCrassus commands the cavalry at the battle against Ariovistus | \nBG I.52 | \n
| 57 BC | \nCrassus, with one legion, receives the submission of the Armorican tribes | \nBG II.34 | \n
| 56 BC | \nSent into Aquitania with twelve cohorts and cavalry | \nBG III.11 | \n
| 56 BC | \nBattle and siege against the Sotiates | \nBG III.20-22 | \n
| 56 BC | \nDecisive battle against the Vocates, Tarusates and Spanish allies | \nBG III.23-27 | \n
| 55 BC | \nMarcus Crassus holds his second consulship with Pompey | \n\n |
| 54-53 BC | \nPublius leaves Gaul to join his father's Parthian expedition | \nPlutarch, Crassus | \n
| 53 BC | \nDeath at Carrhae | \nPlutarch, Crassus | \n
Caesar says that nearly all of Aquitania then sent hostages and submitted, while some remote tribes neglected to do so because the season was far advanced (BG III.27). That qualification matters. The conquest was formal rather than thorough,. Caesar does not tell us how many men Crassus lost, a silence common in his accounts of victories, so the campaign's cost cannot be measured.
\nWhy does it matter, then? First, it shows what Caesar's legates could do alone. In the same summer Sabinus and Decimus Brutus also won independent operations, and the Aquitanian campaign is one of the clearest cases of a legate with his own army fighting a Roman-style battle against opponents who used Roman methods. Second, it shows how the Gallic War was in part a training ground. Crassus applied a probing approach to a tough fortified position, a flanking move and the use of his reserve, while the disaster at Atuatuca in 54 BC shows how differently a legate's decisions could end. The wider machinery of Roman command in the north is explored in our account of Rome's generals on the Rhine.
\nPublius left Gaul around 54 BC, when his father went east. Plutarch reports that Caesar sent him with a force of Gallic cavalry and that he was part of the army that Marcus Crassus led into Mesopotamia (Plutarch, Crassus 17-25). When the Parthians drew the Roman army out into open ground at Carrhae, Plutarch has Publius pursue what looked like a retreating enemy with his cavalry, archers and infantry, only to be cut off and surrounded by Parthian cataphracts and horse archers. Wounded, he is said to have had his shield-bearer kill him, and his head was later shown to his father (Plutarch, Crassus 25-26). Plutarch's story is a literary construction and it is hard to say how much of it is accurate. But it makes a point: the aggressive tactical skill that won in Aquitania was not enough against a mounted enemy that Roman commanders had not faced before. Marcus Crassus, whose catastrophic expedition has its own account here, did not long survive his son.
\nPublius Crassus has one of the oddest records in Roman history. He was a competent subordinate who won two operations in two summers, and he is remembered for the third one he lost. The Aquitanian campaign gives a more useful picture of him: the legate who sensibly asked for a council, who recruited locally, and who took the risk of splitting his force at the right moment. It also demonstrates how thinly the English-language literature covers the less famous years of the Gallic War. His Aquitanian victory deserves to be read alongside the better-known Gallic sieges, as an example of what Rome's junior commanders could do when left alone.
\nWho was Publius Crassus? Publius Licinius Crassus was the younger son of Marcus Crassus, the wealthy politician who formed the First Triumvirate with Caesar and Pompey. He served as one of Caesar's legates in Gaul from 58 BC, and he died at Carrhae in 53 BC.
\nWhere was Aquitania in 56 BC? Caesar's Aquitania lay between the Garonne, the Pyrenees and the Atlantic, a smaller area than the later Roman province of the same name. Its tribes, including the Sotiates, Vocates and Tarusates, were in the region of modern Gascony.
\nWhy did Spanish troops fight in Aquitania? The Aquitanian tribes sent for help across the Pyrenees, and according to Caesar they received troops from Hispania Citerior with officers who had served under Sertorius. Caesar's account is the only source, so the detail of this alliance cannot be checked.
\nIs Book III of the Gallic War reliable on this campaign? It is the only detailed narrative, but Caesar was not present and he wrote for a Roman audience. Numbers in particular, such as the 50,000 men said to have assembled, should be treated as claims rather than as measurements.
\nFor forty-four days in the autumn of 1977, from the kidnapping of an employers' president in Cologne on 5 September to the discovery of his body in France on 19 October, West Germany was governed by a crisis staff, a news blackout and a single question: how far may a democracy go to save the lives of its hostages? The weeks have been known ever since as the "German Autumn", and what the state decided in them is still argued over today.
\nThe autumn did not begin in September. The Red Army Faction's second generation had spent the spring and summer of 1977 on what the group itself called an offensive, and the federal agency for civic education treats these months as the peak of its violence. On 7 April the Federal Prosecutor General, Siegfried Buback, was shot dead in his car in Karlsruhe, together with his driver Wolfgang Göbel and the judicial officer Georg Wurster. On 30 July Jürgen Ponto, chairman of Dresdner Bank, was shot at his home in Oberursel during a kidnapping attempt that went wrong.
\nBehind both attacks stood the same demand. The leading figures of the group had been sentenced to life at Stammheim on 28 April, and the people outside wanted them out. Our article on the courtroom built for that trial describes the verdict; the autumn was, in effect, the group's answer to it.
\n| Date | \nEvent | \n
|---|---|
| 5 September | \nHanns Martin Schleyer is kidnapped in Cologne; his driver and three police officers are killed | \n
| From 6 September | \nChancellor Helmut Schmidt's crisis staff meets in Bonn | \n
| 30 September | \nA law is rushed through that allows contact between prisoners and the outside world to be cut off | \n
| 13 October | \nThe Lufthansa aircraft Landshut is hijacked on a flight from Mallorca to Frankfurt | \n
| 16 October | \nThe pilot, Jürgen Schumann, is shot dead in Aden | \n
| 17–18 October | \nThe aircraft lands in Mogadishu; the GSG 9 storms it shortly after midnight | \n
| 18 October | \nThree imprisoned leaders of the group are found dead or dying at Stammheim | \n
| 19 October | \nSchleyer's body is found in the boot of a car in Mulhouse | \n
On the evening of 5 September, Schleyer's car was stopped on a street in Cologne by a car blocking the road and a pram standing in the carriageway. His driver Heinz Marcisz and the police officers Reinhold Brändle, Roland Pieler and Helmut Ulmer were shot dead. Schleyer was president of both the employers' association and the federation of German industry, which made him one of the best-known businessmen in the country. He was also a former SS officer, a past the kidnappers used and that his family and many historians have since had to confront.
\nThe kidnappers signed their demand with a Kommando name, as the group did for almost every operation. The mechanism is explained in our piece on how the RAF wrote its own record. The demand was the release of imprisoned members of the group, who were to be flown to a country of their choosing.
\nThe reply of the Schmidt government was a mixture of delay and law. A crisis staff, with leaders of the opposition included, met almost daily. Officials negotiated through an intermediary in order to gain time while the police searched for the hostage, and the government's own documentation records that the state's aim was to keep contact open without conceding the central demand. At the same time the prisoners at Stammheim were cut off. A hurriedly passed law allowed all contact between certain prisoners and the outside world, including their lawyers, to be broken. It was a sharp departure from the ordinary practice of a state governed by law, and legal scholars have debated it ever since.
\nPress and broadcasters were asked to restrain their reporting. Many complied. Looking back, the cooperation of the newsrooms is one of the more striking features of the weeks, and one that had no real parallel later in the republic's history.
\nOn 13 October four Palestinian hijackers took over a Lufthansa Boeing 737, the Landshut, shortly after it left Mallorca for Frankfurt. They said they were acting in solidarity with the imprisoned RAF members and added demands of their own. The aircraft, with more than eighty passengers and a crew of five, then flew a long route, landing at Rome, Larnaca, Bahrain, Dubai and Aden.
\nIn Aden, on 16 October, the hijackers shot the captain, Jürgen Schumann, who had left the aircraft to inspect it. The co-pilot, Jürgen Vietor, then flew the aircraft on to Mogadishu in Somalia. The government had in the meantime sent the minister of state Hans-Jürgen Wischnewski to negotiate, and had secured the agreement of the Somali authorities to an operation by the GSG 9, the federal border police's new counter-terrorist unit, which had been formed after the Munich Olympics of 1972.
\nShortly after midnight on 18 October, the unit stormed the Landshut. Three of the four hijackers were killed, one woman was wounded and arrested, and all the passengers survived. It was the clearest success the state achieved in the whole of the autumn.
\nThe news reached Stammheim the same night. By the next morning three of the leading imprisoned members of the group, two men and one woman, were dead in their cells, and a fourth prisoner had been wounded but survived. The official investigations concluded that the three had killed themselves, and that is the account endorsed by the courts and by the government.
\nThe circumstances have been challenged since. How weapons and ammunition had reached prisoners held in one of the most heavily guarded buildings in the country, and whether the cells had been bugged in the way the earlier scandal suggested, are questions that supporters of the group and a number of independent observers kept alive for decades. We do not choose between these readings here. The official finding is suicide; the questions about the security failure that made it possible were never fully laid to rest.
\nWith the Landshut freed, the kidnappers' leverage was gone. On 19 October a message was passed on that Schleyer had been killed, and his body was found that day in the boot of a car in a street in Mulhouse, in French Alsace. He had been held for forty-four days. Exactly who fired the shots has not been settled in a way that satisfies every account, so this article names no one.
\nThe immediate result was a hardening. The group's second generation had failed in its central aim, and the prisoners it had tried to free were dead. The group continued, and its later years, which produced far fewer convictions than attacks, are covered in the third generation. Our record of these operations places the 1977 attacks in the sequence of the whole.
\nFor the Federal Republic the autumn was a test of whether the state could defend its citizens without becoming what its critics said it was. Contact bans, press restraint and a crisis staff acting outside the normal channels were accepted by most of the public at the time, and have been questioned since. The tension continued into the early 1980s, when NATO's response to the SS-20 missiles brought a different generation of protesters into the streets, and it was not finally resolved until the fall of the Berlin Wall removed the world in which the group had operated.
\nThe German Autumn is best understood as two crises that ran together: a kidnapping by a German group and a hijacking by Palestinian militants acting on its behalf. The state met both with determination and with measures that sit awkwardly beside its own constitutional principles. It saved the passengers in Mogadishu and lost the hostage in Mulhouse, and it remains disputed what the deaths at Stammheim say about the prison system. Where the evidence runs out, as it does for several of the killings of 1977, the honest account is to say so.
\nWhy is it called the German Autumn? The name was taken up in Germany for the weeks between the kidnapping of Hanns Martin Schleyer on 5 September and the events of 18 and 19 October 1977. It covers the kidnapping, the hijacking of the Landshut and the deaths at Stammheim as one connected crisis.
\nWho carried out the hijacking of the Landshut? Four Palestinian militants took the aircraft on 13 October 1977, claiming to act in solidarity with the imprisoned RAF members. Three were killed when the GSG 9 stormed the aircraft in Mogadishu, and one woman was wounded and arrested.
\nDid the Stammheim prisoners kill themselves? The official investigations concluded that three prisoners took their own lives on the night of 17 to 18 October 1977. Doubts about how the weapons reached the cells have persisted, and the question has never been settled to the satisfaction of all observers.
\nWho killed Hanns Martin Schleyer? His body was found in Mulhouse on 19 October 1977, and the group that kidnapped him claimed responsibility in its own statements. Which individual fired the shots has not been established in a way all accounts accept, so this article does not name anyone.
\nOn 8 September 1628, in the Bay of Matanzas on the north coast of Cuba, the Dutch admiral Piet Hein took the greater part of a Spanish treasure fleet with scarcely a shot exchanged. The haul was valued at about 11.5 million guilders, the largest single prize the Dutch Republic ever took at sea, and what happened to that money says more about the Republic than the capture itself.
\nSpain's American silver moved in convoy. Each year the fleet of New Spain gathered cargo from Mexico and the Caribbean, sailed for Havana, and crossed the Atlantic under escort. The system protected the treasure well against pirates, and less well against a state that was at war with Spain and had ships to spare.
\nThe Dutch Republic was that state. The Twelve Years' Truce had ended in 1621, the same year the States-General chartered the West India Company (WIC). The Company was a trading firm on paper and a war fleet in practice: its purpose was to hurt the Iberian monarchy where it was richest. The Battle of Abrolhos off Brazil was one result; the silver fleet was the ambition behind many others. Earlier Dutch squadrons had tried for it and failed. Spanish silver was also the engine of the empire's credit, the theme of our piece on Potosí and Spanish imperial finance.
\nHein sailed in 1628 as the Company's admiral, with Witte de With as his flag captain. The Spanish fleet was commanded by Juan de Benavides. The fleet was sighted off Cuba, and rather than reach Havana it ran for Matanzas Bay, where the ships were trapped against the shore. The Spanish crews and their officers abandoned or surrendered several vessels, and the Dutch took the treasure without the fight that had been expected.
\nThe details of the fighting differ from account to account, and the Dutch narrative, the only detailed contemporary one, was composed by the Company's own director. Johannes de Laet's Iaerlijck Verhael is the source most later Dutch histories lean on, and he wrote for a Company that wanted the capture read as skill and God's favour. That is worth remembering when reading the usual phrase that the fleet was taken "without bloodshed". It tells us the Dutch lost few men, and tells us little about Spanish casualties or about how much the Spanish commanders could have done.
\nThe figure usually quoted is 11,509,524 guilders, the sum the cargo fetched when it was brought home and sold. The inventory commonly given has about 177,000 pounds of silver, 66 pounds of gold, pearls, thousands of hides, chests of sugar, and thousands of bags of indigo and cochineal. Those particulars come from Dutch summaries of the sale rather than from an archival ledger I could check, so treat the exact counts as a reported tally rather than an audited one.
\nIt is worth noticing what the total contains. Silver was the headline, but a large share of the value was in dyes, hides and sugar, goods the Company could sell in Amsterdam like any merchant. The prize was as much a cargo to be marketed as a treasure to be counted, and that is one reason it was converted into money so quickly.
\nNor was 11.5 million guilders profit. It is the gross value of the prize. The Company had to pay for the fleet that took it, and prize money was divided by the rules of the admiralties and the Company's charter, with shares for the crews, the admiral, the Company and the state. Commemorative medals struck in the Republic, one of them now in the collection at Greenwich, celebrated Hein and the capture, so the event was marketed as a national triumph from the start.
\n| Use | \nWhat is reported | \n
|---|---|
| Company shareholders | \nA dividend reported at 50 per cent of share value for that year | \n
| The land war | \nMoney for the field army; Dutch histories tie the prize to the campaign that took 's-Hertogenbosch in 1629 | \n
| The Company itself | \nShips and resources for the Brazilian venture, with Olinda and Recife taken in 1630 | \n
| The state | \nA share for the stadholder and the States | \n
Each line needs qualifying. The 50 per cent dividend is repeated in Dutch popular histories and museum texts, and it is plausible for the scale of the prize, but readers should treat it as a reported figure. That the silver paid for the siege of 's-Hertogenbosch is likewise a common Dutch claim; the siege was a large operation, and the prize made cash available for the army at a time when the Republic's revenue was under strain. That is a link between windfall and war, not an accounting proof that specific coins paid specific soldiers.
\nThe longer view is less flattering. Some historians argue that the windfall flattered the Company's finances instead of mending them. A single lucky year set an expectation of dividends that ordinary trade could not meet, and the Company's later fortunes depended on the war in Brazil, which was costly. The more stable Dutch wealth of the age came from the VOC and its trade networks, not from prizes.
\nFor Spain the loss came at a bad moment. Its crown lived on advances against future silver, and the fleet's cargo was largely spoken for before it left the Americas. The bankruptcies of the Habsburg crown, of which Philip II's default of 1557 is the classic earlier example, show how dependent it was on the timely arrival of ships. We should be careful not to make one capture the cause of Spain's decline: the crown survived, sent its fleets again, and the historical debate about Spanish decline points to many causes. What Matanzas did was to embarrass Madrid and demonstrate that its sea lanes could be broken.
\nHein did not enjoy his fame for long. Appointed lieutenant-admiral of Holland in 1629, he was killed on 17 June of that year in an action against the Dunkirk privateers off Ostend, reportedly by a cannon ball. He was buried in Delft in a monument that turned the silver fleet into a national memory. The ballad about the silver fleet, still sung in the Netherlands, is a later inheritance, and its image of the Dutch as winners has outlasted the Company itself.
\nHis success was not a template. The five admiralties of the Republic could not easily repeat a Company operation of this kind, and no Dutch admiral took the silver fleet again. Matanzas is entered in our record of Dutch Golden Age sea battles as a Dutch victory, and it is one of the few rows in which the outcome column is unambiguous. It also belongs with the Spanish Armada of 1588 in the long story of how the Dutch and the English chipped at Iberian sea power, while the Iberian Union meant that Portuguese and Spanish possessions could be attacked as a single target.
\n| Date | \nEvent | \n
|---|---|
| 1621 | \nTruce ends; WIC chartered | \n
| 8 Sept 1628 | \nHein captures the fleet in Matanzas Bay | \n
| 1629 | \nSiege of 's-Hertogenbosch; Hein killed off Ostend on 17 June | \n
| 1630 | \nDutch hold Olinda and Recife | \n
| 1631 | \nAbrolhos | \n
Matanzas was not a battle so much as a transfer of wealth. It shows how a chartered company with a war fleet could turn Spanish silver into Dutch dividends, army pay and Brazilian ambitions within months. It also shows the limits of a windfall: it did not create lasting Dutch strength, and it did not by itself end Spain's. The 11.5 million guilders bought a great deal, but chiefly time and confidence.
\nHow much was the silver fleet worth? The figure usually given is 11,509,524 guilders, the sale value of the whole cargo, which included dyes, hides and sugar as well as silver. It is a gross figure, not the profit, since the costs of the expedition and the shares of the crews and state came out of it.
\nWas there a battle at Matanzas? Not in the sense of a fleet action. The Spanish fleet was trapped in the bay and the treasure taken with very little fighting. The Dutch account, written by a director of the Company, is our main contemporary source, so the Spanish side of the story is thinly documented.
\nWhat did the money pay for? Dutch histories connect it with a large dividend for Company shareholders, funds for the army that took 's-Hertogenbosch in 1629, and the Company's Brazilian ventures. These are reported uses, not a ledger-by-ledger trace.
\nWhat happened to Piet Hein? He was made lieutenant-admiral of Holland and was killed in June 1629 fighting Dunkirk privateers off Ostend. He was buried in Delft.
\n: De Laet's narrative is cited from the 1644 Leiden edition.\n: The medal is described in the Greenwich collections catalogue.\n: Het Scheepvaartmuseum's account of Hein gives his career and death.
\n" }, { "id": "https://writesidesofhistory.com/battle-of-jaffa-1192-richard-on-foot/", "url": "https://writesidesofhistory.com/battle-of-jaffa-1192-richard-on-foot/", "title": "The Battle of Jaffa, 1192: Richard's Last Fight on the Sand", "summary": "In August 1192 Richard the Lionheart relieved Jaffa with a handful of knights and crossbowmen, then held a beach against Saladin's army on foot.", "date_published": "2026-09-29T00:00:00.000Z", "image": "https://writesidesofhistory.com/images/generated/battle-of-jaffa-1192-richard-on-foot.webp", "tags": ["medieval"], "content_html": "In early August 1192, Richard the Lionheart sailed south to a Jaffa that Saladin had just taken, went ashore with a few dozen knights and a couple of thousand crossbowmen, and beat off an Ayyubid army many times his size on the beach outside the town. It was the last fight of the Third Crusade, and it was fought by a king who had, days earlier, been preparing to go home.
\nJaffa was the crusaders' prize from the previous autumn. Richard had refortified it after the victory at Arsuf, and it was the port through which any advance on Jerusalem would be supplied. In the summer of 1192 he had been negotiating with Saladin and bickering with the French and with the Duke of Burgundy's men, and he had withdrawn to Acre.
\nSaladin, seeing the coast weakly held, marched on Jaffa and attacked it in the last days of July. The sources disagree on exact days, and the modern narratives differ by a day or two, but the sequence is clear. His miners brought down part of the wall, his troops broke into the town, and the garrison withdrew into the citadel. According to Bahā' ad-Dīn, a truce was agreed under which the defenders would surrender if no help came by a set date. Meanwhile the town itself was plundered, and the Ayyubid emirs lost control of their men.
\nThe lesson of that week is one the Latin chroniclers and the Muslim chronicler agree on: the loss of Jaffa was a matter of days, not of a long siege. If the citadel fell, the coastal position that had been so hard to win would go with it.
\nRichard's response is the part that made the later legend. Ambroise and the Itinerarium both describe a king who was ill, short of men and short of money, and who nonetheless took what he had. A relief force went by land along the coast, while Richard himself went by sea in a small fleet of galleys. Modern accounts give him around fifty-five knights, some hundreds of men-at-arms and about two thousand Genoese and Pisan crossbowmen, but the exact numbers come from chroniclers who wrote to glorify him and should be treated as approximate.
\nWhen the galleys reached Jaffa, the shore was lined with Ayyubid troops and the banners of the town flew over the harbour. The chroniclers tell of a priest who got out of the citadel and swam or was smuggled to the ships with the news that the citadel still held. Richard, in the well-known version of the story, did not wait for the galleys to be brought in: he jumped into the surf and led his men ashore. The scene is best read as the Itinerarium and Ambroise present it, as a set-piece, and the historical core is simpler: a small landing force, led by the king in person, drove the Ayyubid troops out of the harbour area and relieved the citadel.
\nThe second half of the story is less often told. Saladin's men were in possession of a town they had looted, and they were not in formation. Richard retook Jaffa quickly, and the garrison of the citadel, which had been on the point of capitulating, was saved.
\nRichard did not shelter behind the walls. He pitched a small camp outside the town, along the shore, and the chroniclers explain why: the town was full of the dead and the stench of the fighting, and its walls had just been breached. Some historians argue it was also a matter of pride, a king who could not appear to be besieged.
\nSaladin gathered his army and, on a morning a few days later, moved to attack the camp. The commonly cited figure, from the Latin sources, is that Richard had only about ten horses, and Gillingham treats the numbers as unreliable but the disparity as real. The Muslim army was mostly cavalry, and it was several times larger than the force on the sand.
\nThe Ayyubid attack came at dawn, and the crusaders were nearly caught asleep. What followed is the famous image of the battle. Richard had his knights kneel in a line with the butts of their spears and lances planted in the sand, their shields overlapping, and between the knights he put his crossbowmen, one to load and one to shoot. The horsemen who charged the line met a hedge of points that horses would not go into; the crossbowmen shot them from behind it.
\nRichard mounted when he could, and the chroniclers describe a king riding along the line, then charging out and back again, and personally striking down an emir. The claims are heroic and the Itinerarium is the source for most of them. Bahā' ad-Dīn, from the other side, does not dispute that the attack failed, and he admits that the emirs were unwilling to press the charge at the price required. He writes of Saladin's anger at the reluctance of his commanders, and of a day on which the sultan's men held back.
\nBahā' ad-Dīn also preserves a detail that fits the picture of mutual respect the later legend loved. When Richard's horse was killed, Saladin's brother al-Adil, or Saladin himself in some tellings, sent him horses as a gift. It is a story that both traditions like, and the evidence that it happened as told is thin; it is best treated as a tale about chivalry in the sources rather than as a fact about the battle.
\nThe crusaders' losses were small and the Ayyubid army withdrew. Richard, however, could not follow up. He had almost no cavalry, he was seriously ill, and the men who had held the line were too few to hold Jaffa and go anywhere else at the same time.
\nThat is why the fight matters less for its tactics than for what it did to the bargaining. Neither side could finish the other. Saladin could not take the coast, Richard could not take Jerusalem, and both wanted to go home. The talks that had run on and off for a year were concluded within weeks, and the treaty sworn at Jaffa left the coastal strip in crusader hands and Jerusalem in Saladin's. Our record of these operations lists the relief of Jaffa as the last battle in Richard's campaigns, and it is one that he fought and won in person.
\nIt also belongs to a longer story. The coastal kingdom that Jaffa protected lasted until the fall of Acre in 1291, and the men who defended it, among them the Templars, were the heirs of exactly the kind of small, disciplined force Richard used on the sand. The crusade that had started with the taking of Cyprus from Isaac Komnenos ended, in a sense, with the same king making the most of the fewest men.
\n| Date | \nEvent | \n
|---|---|
| September 1191 | \nRichard reaches Jaffa after Arsuf and begins refortifying it | \n
| Summer 1192 | \nTalks with Saladin continue; Richard withdraws to Acre | \n
| Late July 1192 | \nSaladin storms Jaffa; the garrison holds the citadel | \n
| 1 August 1192 | \nRichard lands from the sea and retakes the town | \n
| Early August 1192 | \nSaladin's dawn attack on Richard's camp is beaten off | \n
| 2 September 1192 | \nThe truce is sworn at Jaffa | \n
Jaffa is the fight in which Richard's reputation as a commander is easiest to defend and hardest to measure. The chroniclers who made it a legend were his partisans, and the numbers they give are the numbers of a story. But the outline holds across the Latin and the Muslim accounts: a town lost, a rapid counter-stroke, a disciplined line against a larger army, and an enemy that would not press the attack. It bought no city, and it did not need to. It bought the negotiating position from which the crusade could end on terms both sides could accept.
\nWhen was the Battle of Jaffa fought? Richard landed on about 1 August 1192, and Saladin's dawn attack on his camp followed within days. Modern accounts give slightly different dates for the fight itself, and the chronicles themselves are not precise about it.
\nHow many men did Richard have? The usual modern figure is about fifty-five knights, some hundreds of men-at-arms and around two thousand crossbowmen, against a much larger Ayyubid force. These numbers come from partisan chroniclers, so they are best treated as an order of magnitude rather than a count.
\nDid Richard really fight on foot? The chroniclers describe him fighting for part of the day without a mount after his horse was killed, and Bahā' ad-Dīn's account of horses being sent to him is often repeated. Both are good stories, and the details are less secure than the outline of the fight.
\nDid the battle end the Third Crusade? In effect it did. Neither side could win, and the truce followed within weeks. It left Jerusalem in Muslim hands and kept a coastal strip for the crusaders.
\nAdhemar of Le Puy fought at the front of the crusader line at Dorylaeum and survived an eight-month siege at Antioch, only to die a month after the city was finally secured — not from a Turkish blade but from an epidemic that swept the captured streets in the summer heat of 1098. As papal legate, the bishop of Le Puy had been the one figure in the crusader army whose authority no baron seriously disputed. When he died on 1 August 1098, the expedition lost the closest thing it had to a unifying head, and the record of what killed him is thinner than the record of what he did while alive.
\nAdhemar was born around 1045 into a noble family connected to the counts of Valentinois, and he was elected bishop of Le Puy around 1080, a diocese in the volcanic hills of the Auvergne with a long pilgrimage tradition. He had probably already travelled to Jerusalem once before 1087, giving him a rare, direct familiarity with the route his own crusaders would later take. When Pope Urban II preached the crusade at the Council of Clermont on 27 November 1095, Adhemar was in the crowd, and, according to the chroniclers who later shaped the story, he was the first man present to take the cross. Urban had approached him beforehand and asked him to travel east as his personal representative, so the gesture at Clermont formalized an arrangement already settled.
\nThat appointment made Adhemar something no other figure in the crusade possessed: standing that came from Rome rather than land, lineage or military reputation. He traveled east attached to the army of Raymond of Toulouse, the richest of the crusading nobles, but his authority as legate ran across the whole expedition. In an army assembled from independent princes who answered to no single commander, that made him, in practice, the crusade's only permanent chairman.
\nThe role was not a ceremonial one. At the Battle of Dorylaeum on 1 July 1097, Bohemond of Taranto's advance guard was nearly overwhelmed by a Seljuk force under Kilij Arslan I before the rest of the army, marching some distance behind, could arrive. Adhemar is credited with leading part of the relief force that struck the Turkish flank and turned the battle, fighting in person — a habit chroniclers noted approvingly in a bishop whose office normally kept him out of combat.
\nThe crusade reached Antioch in October 1097 and settled into a siege that dragged into the following June. Conditions inside the camp deteriorated badly over the winter of 1097–98: supplies ran short, foraging parties had to range further as the countryside was stripped bare, and desertions became common enough that named lords, not just anonymous soldiers, slipped away. Throughout, Adhemar's legatine authority gave him a role no purely military commander could fill — mediating between princes whose rivalries threatened to break the siege apart before the walls did.
\nAntioch finally fell on 3 June 1098, taken through a betrayal arranged with a Turkish officer inside the walls rather than by assault. The victory brought no relief: a large Muslim army under the atabeg Kerbogha of Mosul arrived almost immediately afterward and besieged the crusaders inside the very city they had just captured, cutting them off from the food they had hoped Antioch would finally supply. It was during this second siege that a Provençal pilgrim named Peter Bartholomew announced visions directing him to a relic buried beneath the floor of the city's Church of St Peter — the lance that had pierced Christ's side at the crucifixion.
\nAdhemar's reaction was skeptical rather than credulous. He is reported to have known that a Holy Lance was already venerated in Constantinople, which made a second, freshly discovered one difficult to credit, and Raymond of Aguilers, who supported Peter Bartholomew throughout the episode, records the bishop's doubt candidly. Whatever Adhemar privately believed, he judged the discovery's effect on a starving, encircled army more useful than its authenticity was doubtful, and he did not move to suppress it. When the crusaders broke out of the city on 28 June 1098 and routed Kerbogha's larger force in the open field, the lance went into the fighting as a battle standard, by some accounts in Adhemar's own party, regardless of what its custodian privately thought it was.
\nVictory did not end the suffering. Antioch in July 1098 was crowded with soldiers and refugees, its water and sanitation strained past what a siege-battered town could support, in the full heat of a Syrian summer. An epidemic broke out among the crusaders in the weeks after Kerbogha's defeat, and on 1 August 1098 it killed Adhemar. Fulcher of Chartres, writing his chronicle years later, records the bishop's death among that summer's losses at Antioch, though Fulcher had already left the main army with Baldwin of Boulogne for Edessa and was not present to witness it. Peter Tudebode's chronicle likewise places Adhemar's death at Antioch that August, among an unspecified number the same sickness took. Neither source names the disease with confidence; modern accounts most often suggest typhoid, but the chroniclers closest to the event did not commit to a diagnosis, and this record does not invent one where they were silent.
\nHe was buried in the cathedral of St Peter at Antioch, the church beneath which the Holy Lance had reportedly been found weeks earlier. His death did not end his role in the story. Raymond of Aguilers reports that Peter Bartholomew later described visions of Adhemar appearing after death, in the company of the Virgin Mary, confirming what the bishop had doubted in life — that the lance found at Antioch was genuine. Historians reading Raymond's account today treat these posthumous appearances less as evidence of a changed opinion than as a device by which the relic's defenders enlisted the crusade's most credible dead voice on their side.
\nAdhemar's authority had kept the crusade's rival princes from breaking apart over Antioch's fate even before the city fell; his death removed the one figure both Bohemond of Taranto and Raymond of Toulouse were prepared to defer to. Bohemond, who had led the party that forced Antioch's gates open, claimed the city for himself, and without a legate to arbitrate, Raymond's competing claim had no forum left to be heard in. The dispute stalled the march to Jerusalem for months, until ordinary soldiers and poorer pilgrims, according to the chroniclers, forced the leadership's hand by threatening to go on without them. Our record of these operations lists Adhemar's death from disease alongside dozens of others whose ends came far from any battlefield, a pattern the march to Jerusalem would repeat many times over.
\nBefore Adhemar reached Antioch, the crusade's first wave had already been destroyed at Civetot, and the earliest crusading bands' passage through Europe had triggered the Rhineland massacres that Urban II's call never intended. The camp Adhemar held together at Antioch depended, as at every crusader siege, on the unnamed women who traveled with the army and rarely appear in the chronicles his fellow legates left behind. The authority vacuum his death created was eventually filled by institutions such as the military orders that came to police and finance the crusader states, states that survived, in shrinking form, until the fall of Acre in 1291 — passing through crises as severe as the sack of Constantinople in 1204 along the way.
\n| Date | \nEvent | \n
|---|---|
| c. 1080 | \nAdhemar elected bishop of Le Puy | \n
| 27 Nov 1095 | \nCouncil of Clermont; Urban II names Adhemar papal legate | \n
| 1 July 1097 | \nBattle of Dorylaeum; Adhemar's relief force helps turn the battle | \n
| Oct 1097 | \nSiege of Antioch begins | \n
| 3 June 1098 | \nAntioch falls to the crusaders | \n
| 5–28 June 1098 | \nKerbogha besieges the crusaders inside Antioch | \n
| 28 June 1098 | \nCrusaders break out and rout Kerbogha's army | \n
| 1 Aug 1098 | \nAdhemar dies of disease in Antioch | \n
Adhemar of Le Puy is remembered, when he is remembered at all, for a relic he doubted and a battle he did not live to see resolved into a march on Jerusalem. That framing undersells a man whose real achievement was administrative: for three years he kept an army of rival princes from destroying itself from within, using an authority that came from Rome and nowhere else. The siege that made Antioch famous killed thousands through hunger and the sword; it took the bishop himself only after the fighting was won, in an epidemic the chronicles record but do not fully explain.
\nWas Adhemar of Le Puy a soldier or a churchman? Both, in practice. As bishop of Le Puy he held no military rank, but chroniclers describe him fighting in person at Dorylaeum and leading troops during the siege of Antioch, alongside his formal role as Pope Urban II's legate and the crusade's chief mediator between rival princes.
\nDid Adhemar believe the Holy Lance found at Antioch was genuine? The evidence suggests he doubted it, partly because a Holy Lance was already venerated in Constantinople. Raymond of Aguilers, an eyewitness who supported the relic's discoverer, records Adhemar's skepticism directly, even as the bishop allowed the discovery to stand because of its effect on his starving army's morale.
\nWhat exactly killed Adhemar of Le Puy? Fulcher of Chartres and Peter Tudebode both record his death from disease at Antioch on 1 August 1098, amid an epidemic that followed the city's second siege. Neither chronicler names the specific illness; modern historians most often suggest typhoid, but that remains an inference rather than a detail the surviving sources confirm.
\nHow did Adhemar's death affect the crusade's leadership? It removed the one figure whose authority both Bohemond of Taranto and Raymond of Toulouse were willing to accept over Antioch's disputed control. Without a legate to mediate, the rivalry stalled the march to Jerusalem for months until pressure from the ordinary soldiers forced the princes to move on.
\nIn the autumn of 54 BC, a Roman column of some 6,000 to 9,000 men walked out of a fortified camp in what is now Belgian Limburg on a promise of safe passage, and into a ravine from which almost none of them walked out again. The camp was called Atuatuca, held for the winter by a legion and five cohorts under the joint command of Quintus Titurius Sabinus and Lucius Aurunculeius Cotta, in territory belonging to the Eburones (BG V.24, 26-37). By most counts it was the only complete legion Rome lost in the eight years of the Gallic War. The man who engineered it, Ambiorix, king of the Eburones, is far less remembered than the wars he helped provoke — and the ambush belongs to the operational history our record of Caesar's campaigns in Gaul and Britannia tracks across eight fighting seasons, as Rome's single costliest afternoon of the war.
\nThe disaster had its roots in logistics rather than tactics. Caesar had just returned from a second, larger expedition across the Channel into Britain, and that year's harvest across Gaul was poor (BG V.24). Rather than concentrate his eight legions in one or two large winter camps as usual, he dispersed them more widely than ever before, so each garrison could forage a smaller area. One of those garrisons — a legion recently raised in Cisalpine Gaul, reinforced with five independent cohorts — was posted among the Eburones, a tribe whose king, Ambiorix, had until then presented himself as a cooperative client of Rome (BG V.24, 27).
\nThat decision to scatter the army thinly, made by Caesar himself, was precisely what a hostile local leader needed: the Eburones' territory lay near the Rhine frontier, within reach of Germanic tribes such as the Sugambri, and farther from the support that other garrisons — including Titus Labienus's camp among the Treveri and Quintus Cicero's camp among the Nervii — could offer if trouble broke out.
\nAbout two weeks after the legion settled into camp, Ambiorix and his co-king Catuvolcus attacked Roman foraging parties, then broke off and asked for a parley (BG V.26-27). What Ambiorix told Sabinus there was a carefully built lie: a general uprising, coordinated across Gaul, had already struck every Roman winter camp at once, and a large force of Germans was crossing the Rhine to join it. He presented himself as a reluctant client acting under pressure from his own people, and offered the Romans free passage to the nearest friendly camp if they left at once, before the Germans arrived (BG V.27-29).
\nCotta argued strenuously against moving: no order from Caesar authorized abandoning a fortified position, and a well-stocked camp could withstand a siege far more safely than a column could survive an ambush in open country (BG V.28). Sabinus overruled him, and after a night of debate the legion marched out at dawn, its baggage train strung out behind it, trusting an enemy king's word over its own ramparts (BG V.28-31) — a decision historians have puzzled over ever since, given that Ambiorix had ambushed the same Romans' foraging parties only days earlier.
\nAbout two miles from camp, the road descended into a long defile bordered by wooded high ground, where Ambiorix had positioned his warriors along both slopes. Once the column's center had committed to the low ground, the Eburones and their allies struck from both sides at once, pinning vanguard and rear together so the strung-out Romans could neither advance, retreat, nor form a coherent line (BG V.32-33). Caesar describes a fight lasting from dawn well into the day, the Romans forming a hollow square around their baggage and beating back repeated charges while slowly losing cohesion.
\nCotta was struck in the face by a sling bullet early on and fought until he was killed later in the battle. Sabinus, seeing the position collapse, sent an officer to ask Ambiorix for a second parley; Ambiorix accepted, then had Sabinus and his companions surrounded and killed once they had laid down their weapons (BG V.35-37). The remaining Romans fought on until nightfall, when the survivors took their own lives rather than be captured; Caesar states that only a handful escaped the ravine, reaching Labienus's camp with news of the disaster (BG V.37).
\n| Date (54 BC) | \nEvent | \n
|---|---|
| Autumn | \nCaesar disperses eight legions into widely separated winter camps after a poor harvest; a legion and five cohorts are posted among the Eburones (BG V.24) | \n
| Within ~2 weeks | \nAmbiorix and Catuvolcus attack Roman foragers, then request a parley (BG V.26-27) | \n
| Parley | \nAmbiorix tells Sabinus that every Roman camp is under attack and offers safe passage; Cotta objects (BG V.27-28) | \n
| Following dawn | \nThe legion abandons its camp and marches out with its baggage train (BG V.31) | \n
| Same day | \nAmbiorix's forces ambush the column in a ravine; Cotta is killed fighting, Sabinus is killed after a second parley (BG V.32-37) | \n
| Nightfall | \nSurvivors take their own lives rather than surrender; a small remnant reaches Labienus (BG V.37) | \n
| Following days | \nEmboldened by the victory, the Nervii and allies besiege Quintus Cicero's camp, which holds until Caesar force-marches to relieve it (BG V.38-52) | \n
| 53 BC | \nCaesar returns with a larger army to campaign against the Eburones; Catuvolcus poisons himself, Ambiorix escapes into the Ardennes (BG VI.29-44) | \n
Word of the Eburones' victory spread fast, and within days the Nervii besieged Quintus Cicero's nearby camp — brother of the orator whose private letters later dwelt anxiously on his fate — until Caesar force-marched a relief column to break the siege (BG V.38-52). The name Atuatuca resurfaces the following year in a way that complicates locating the original camp: in 53 BC Caesar built a supply depot, also called Atuatuca, whose garrison was itself besieged by Sugambri raiders from across the Rhine (BG VI.32-41). Whether this was the same site, a nearby one, or simply a shared name is still debated; a 2024 reassessment by archaeologists Nico Roymans and Marc Martens argues against identifying either camp with the later Roman town of Atuatuca Tungrorum, modern Tongeren, and the actual ground of Sabinus and Cotta's disaster remains unlocated.
\nCaesar treated the loss as a debt owed publicly. In 53 BC he led a large force back into Eburones territory with orders, in his own words, to erase the tribe's name from among the nations of Gaul, burning settlements, seizing livestock, and inviting neighboring peoples to plunder Eburones land rather than face Roman retaliation themselves (BG VI.29-34, 43). Catuvolcus, too old to face flight or further war, cursed Ambiorix for the scheme and poisoned himself with yew, a tree Caesar notes grows commonly across Gaul and Germany (BG VI.31). Ambiorix was never captured; when a Roman cavalry detachment finally located him, he slipped away into the Ardennes forest, and Caesar attributes the escape to luck rather than to any Roman failure (BG VI.29-30, 43).
\nHow completely that campaign destroyed the Eburones as a people is disputed. Some read Caesar's language of annihilation literally; others argue the tribe largely scattered into terrain Roman forces could not search, and that later Roman writers erased its name as a form of political damnation rather than describing an accomplished extermination. What is not disputed is that the Eburones vanish from the sources afterward, dispersed among neighboring peoples along the same Rhine frontier Roman writers would later mythologize as home to Germanic tribes of untamed virtue.
\nThe ambush at Atuatuca rarely gets the attention paid to Caesar's set-piece sieges, but in strict terms of Roman losses it was the war's single worst reversal — closer in scale to catastrophes like the destruction of Varus's legions at Teutoburg or Crassus's disaster at Carrhae than to any other episode in Caesar's own campaigns. It was won not by superior numbers or equipment but by a lie told to a commander predisposed to believe it, and it briefly threatened to unravel Rome's position in Belgic Gaul before the relief of Cicero's camp restored the balance. Ambiorix, unlike the naval commander who broke the Veneti fleet that same war, left almost no personal trace beyond this one triumph and the long, inconclusive war of reprisal it provoked.
\nWho was Ambiorix? Ambiorix was joint king of the Eburones, a tribe settled between the Meuse and the Rhine in what is now Belgian Limburg, who presented himself as a client of Rome before orchestrating the destruction of a Roman legion in 54 BC through deliberate deception rather than open battle (BG V.24-27).
\nHow many Roman soldiers died at Atuatuca? Caesar's account describes the destruction of an entire legion together with five additional cohorts, a force modern historians estimate at roughly 6,000 to 9,000 men, with only a small number of survivors reaching Titus Labienus's nearby camp (BG V.24, 37).
\nWhy did Sabinus agree to abandon a fortified camp? Ambiorix convinced Sabinus that every other Roman winter camp was already under attack and that Germanic tribes were crossing the Rhine, framing the offer of safe passage as the only way to avoid being overwhelmed; his co-commander Cotta argued against the move but was overruled (BG V.27-29).
\nWhat happened to Ambiorix afterward? Ambiorix was never captured. Caesar returned in 53 BC with a larger army bent on destroying the Eburones, and while his co-king Catuvolcus took his own life rather than flee, Ambiorix escaped a Roman cavalry pursuit into the Ardennes forest and disappears from the historical record after that campaign (BG VI.29-31, 43).
\nBetween 1996 and 2008, excavators working ahead of a construction project at Herxheim, in the German region of the Palatinate, emptied more than eighty pits ringing a Linear Pottery culture settlement and found them packed with broken human bone, broken pottery, and the bones of cattle and pigs, all mixed together. The people were farmers of the LBK, the first agricultural culture of Central Europe, and the pits date to the very end of that culture's existence, somewhere between 5300 and 4950 BC. What the excavators disagree about, and have disagreed about since the first full analysis appeared, is what the broken bones mean: the remains of a mass cannibal feast, or the debris of a death ritual that dismembered bodies without eating them.
\nHerxheim was found by surface survey in the 1980s and excavated in two campaigns: a rescue dig from 1996 to 1999 ahead of an industrial estate, directed by Annemarie Häußer, and a research excavation from 2005 to 2008 led by Andrea Zeeb-Lanz of the Rhineland-Palatinate state heritage service. What emerged was not a simple ditch but two systems of long, shallow, overlapping pits running around the settlement, together holding roughly eighty separate deposits of bone, pottery, tools and ornaments.
\nThe pottery in those pits is one of the strangest details. Some of it was made locally, but a substantial share came from LBK communities hundreds of kilometres away, and strontium isotope analysis of teeth from the human remains points the same way: a share of the people buried at Herxheim did not grow up nearby. Whatever happened here drew in objects, and on this reading people, from well beyond one settlement's territory, which is why the excavators call it a "central place" rather than an ordinary village cemetery.
\nThe scale of the human remains grew as work progressed. The 2009 study by Bruno Boulestin and colleagues, the first full analysis of the bone assemblage, put the minimum number of individuals at around 500 and suggested the true figure could approach 1,000. A fuller monograph on the remains, published in 2015, raised that minimum past 1,000 once further pits and re-examined material were included.
\nThe 2009 Antiquity paper argued that the human bone at Herxheim had been processed exactly as the animal bone in the same pits had been processed: skulls broken open in a pattern consistent with removing the brain, long bones snapped and the fracture surfaces scraped in the pattern left by marrow extraction, and cut marks on the bone surface matching the marks left by filleting meat rather than by simple dismemberment for burial. Skullcaps in particular were removed in a controlled way that the authors compared to the preparation of a container or cup. Taken together, the team argued, these were not the marks a community leaves on the bodies of its own honoured dead; they were the marks left by butchery for consumption, on a scale — hundreds of individuals — that pointed to a mass event rather than an isolated act.
\nOther members of the same research world read the identical bone assemblage differently. Jörg Orschiedt and Miriam Haidle argued, in a paper published two years before the cannibalism study, that fractures and cut marks of this kind are also exactly what a thorough secondary burial rite produces: bodies allowed to decompose or deliberately defleshed, then broken apart, sorted, and redeposited once the flesh was gone, as part of a mortuary practice rather than a meal. They pointed to a detail that the cannibalism reading has to explain away: jaws and the base of the skull are underrepresented in the Herxheim pits compared with what a random sample of broken skeletons should contain, which is easier to account for if certain bones were deliberately kept back or handled apart during a ritual than if the goal of the processing was simply to eat what was available. On this reading, Herxheim was a "central place" for handling the dead of a wide region — a Neolithic ossuary rebuilt through defleshing and redeposition — and no one at the site necessarily ate anyone.
\n| Date | \nDevelopment | \n
|---|---|
| c. 5300–4950 BC | \nHerxheim enclosure in use at the end of the Linear Pottery culture | \n
| 1980s | \nSurface finds of human bone lead to the site's discovery | \n
| 1996–1999 | \nFirst excavation campaign, directed by Annemarie Häußer, ahead of construction | \n
| 2005–2008 | \nSecond excavation campaign, led by Andrea Zeeb-Lanz | \n
| 2007 | \nOrschiedt and Haidle publish the secondary-burial reading | \n
| 2009 | \nBoulestin, Zeeb-Lanz and colleagues publish the cannibalism reading in Antiquity | \n
| 2012 | \nIsotope study finds a share of individuals were not local to Herxheim | \n
| 2015 | \nBoulestin and Coupey's monograph raises the minimum count past 1,000 | \n
| 2019 | \nZeeb-Lanz's own synthesis volume treats consumption as still unproven | \n
What makes Herxheim unusual is not just that outside critics doubt the cannibalism claim; the doubt runs through the excavation project itself. The 2009 field report describing the pits and their contents was co-authored by the whole excavation team, including both Zeeb-Lanz and Orschiedt — the same two researchers who, in different papers either side of that report, argued for opposite readings of the same bones. Zeeb-Lanz herself co-authored the Antiquity cannibalism paper that same year. Ten years later, in her own edited synthesis of the excavation, she took a more cautious position: the bodies were deliberately, violently taken apart, she argued, but proof the flesh was actually eaten was still missing. Boulestin did not retreat; his 2015 monograph with Anne-Sophie Coupey restated and expanded the cannibalism case with a larger sample of bone. The site's own director now sits closer to the sceptics than to the paper that first made her name at Herxheim, while her original co-author has not changed his conclusion.
\nThat split matters, because it is not a dispute between people who know the bones and people who do not. Everyone involved agrees the bodies were violently dismembered, the skulls opened, and the remains mixed with animal bone and imported pottery in the same pits. What they disagree about is narrower and harder: whether "processed like meat" also means "eaten like meat," a distinction the bone surface alone may not settle either way.
\nHerxheim does not identify who died there or why, only how their remains were handled afterward. Some researchers link the site to a period of stress at the end of the LBK, when population growth and resource pressure are thought to have strained relations between communities across Central Europe — a period that also produced outright massacres such as Talheim, where thirty-four people were killed and dumped in a single pit with none of Herxheim's careful bone processing. Whether Herxheim's dead died locally or arrived already dead from elsewhere, and whether the practice ran for years or was compressed into one episode, remain open questions; the comparison with better-resolved cases, such as the repeated raiding documented at Jebel Sahaba, only underlines how much interpretation Herxheim's bones still require. Other Neolithic gathering places built for reasons unrelated to violence, like the ceremonial enclosures at Göbekli Tepe or the early proto-urban centres discussed alongside the first cities before writing, are a reminder that drawing people and goods from far away does not, on its own, prove anything about violence. This site's own record of collective violence before 3000 BCE lists Herxheim with a low certainty rating precisely because the two readings remain equally defensible on current evidence.
\nHerxheim has produced one of the largest, strangest bone assemblages in European prehistory, and more than fifteen years of specialist study have not resolved what it means. The people who dug the pits are not united behind either the cannibalism reading that first made the site famous or the death-ritual reading that has always run alongside it; the project's own director has moved toward caution while her original collaborator has not. Until new evidence breaks the tie, the honest account of Herxheim is not a single answer but a documented disagreement among the people best placed to judge it.
\nWhat is Herxheim? Herxheim is a Linear Pottery culture (LBK) enclosure in the German Palatinate region, excavated between 1996 and 2008, where more than eighty pits contained the broken and mixed remains of at least 500, and by later estimates over 1,000, people alongside animal bone and pottery imported from hundreds of kilometres away.
\nWas Herxheim really a site of cannibalism? That is disputed. A 2009 study argued the human bone was processed exactly like the animal bone found with it, evidence the authors read as mass cannibalism, while other researchers on the same excavation project argue the same fractures and cut marks are consistent with a secondary burial ritual that defleshed and redeposited bodies without anyone eating the remains.
\nWhy does it matter that the excavation team disagrees? Because the disagreement is not between outside sceptics and the people who excavated the site — it runs through the project itself. The site director, who co-authored the original cannibalism paper, later took a more cautious position in her own synthesis of the excavation, while her co-author maintained the cannibalism reading in a further monograph.
\nHow does Herxheim compare to other Neolithic mass-fatality sites? Unlike Talheim, where thirty-four people were killed and dumped in one pit with no further processing, the Herxheim bones show extensive, deliberate breakage and cutting spread across many deposits over what may have been years, which is exactly why its cause is harder to pin down than a straightforward massacre.
\nAt Asparn an der Zaya in Lower Austria, a farming settlement dug two encircling ditches for defence around 5000 BC — and then the bottom of those ditches filled with its own dead. At least 67 people, men, women, children, and one adult whose sex could not be determined, were killed by blows to the head and left exactly where they fell. Nobody returned to bury them. Carnivores gnawed the exposed bone for months or years before soil finally covered what remained. The find is known today by the modern village above it: Asparn/Schletz.
\nThe enclosure at Schletz belonged to the Linearbandkeramik, or Linear Pottery culture, the first full farming society of Central Europe, named for the incised bands decorating its pottery. Its settlements were normally open, undefended, and their dead were normally buried with care: laid on one side, knees drawn up, a pot or a stone tool placed nearby. Schletz breaks that pattern twice over. The settlement was ringed by two ditches, evidently built for defence, and the people found in them were not interred at all.
\nA partial excavation of those ditches recovered the remains of at least 67 individuals: 27 children, 26 men, 13 women, and one adult of undetermined sex. All but one of the skulls examined show blunt-force trauma consistent with the stone-edged axes and adzes that Linearbandkeramik farmers used as tools; a single victim was killed by an arrow. The bodies were not laid out or covered. Many are disarticulated, with limbs scattered and bones bearing carnivore tooth marks — the signature of corpses left in the open long enough for scavengers to reach them.
\nBecause only a fraction of the enclosure's ditch circuit has been dug, the excavators have suggested that a full excavation could recover several times that number, on the order of a few hundred. That figure is an extrapolation from the density of remains in the excavated sections, not a count of bodies anyone has actually found, and it should be read as such.
\nFieldwalking around Asparn an der Zaya turned up concentrations of human bone in the early 1980s, and Helmut Windl began systematic excavation of the enclosure in 1983. Work continued intermittently until 2005, uncovering parts of the settlement together with sections of both ditches. Windl's own first account appeared in 1996, in a catalogue for a regional museum exhibition rather than an academic paper — a field reading of what he had just dug up, ahead of the specialist studies that followed. The detailed trauma analysis by Maria Teschler-Nicola and colleagues, published the same year, examined the injuries bone by bone and argued that the deaths resulted from violent assault rather than accident or disease.
\nThe most direct comparison is Talheim, in Baden-Württemberg, where 34 people of the same culture were killed with the same weapons and dumped in a single pit. Together with a third Linearbandkeramik site, Schöneck-Kilianstädten, these finds are why the peaceful image of Europe's first farmers no longer holds. Organised killing in prehistory reaches back further still, to the repeated raiding documented at Jebel Sahaba in the Nile valley, thousands of years before farming reached Central Europe at all.
\nWhat sets Schletz apart from Talheim is not the killing but what happened afterward. At Talheim the bodies were at least gathered into one pit. At Schletz they were left in the open along the base of the ditch, exposed to weather and animals for an extended period before sediment finally buried them. The taphonomic evidence — the marks left by scavenging, the scattering of bones that would otherwise stay articulated, the absence of any of the grave goods that mark ordinary Linearbandkeramik burials — points to a community that was not just killed but left with no one able, or willing, to perform the rites that would normally have followed a death.
\nThat absence is also demographic evidence. At Talheim and at Kilianstädten, young adult women are noticeably under-represented among the dead, which most researchers read as a sign that women of childbearing age were taken captive rather than killed. At Schletz they are not missing: women, including younger women, appear among the 67 in roughly the proportion a living community would have had. If the pattern holds as more of the ditch is examined, it suggests that whoever attacked Schletz did not spare anyone to take away — which is part of why some popular accounts of the site have reached for the modern word "genocide" rather than the more cautious "massacre" that the academic literature generally prefers.
\n| Date | \nDevelopment | \n
|---|---|
| c. 5210–5000 BC | \nLinearbandkeramik settlement at Schletz reaches its final phase, enclosed by two ditches | \n
| c. 5000 BC | \nAt least 67 people are killed and left unburied in the ditch fill | \n
| 1983–2005 | \nHelmut Windl carries out excavation of the settlement and ditch sections | \n
| 1996 | \nWindl publishes the first account of the find; Teschler-Nicola and colleagues publish the detailed trauma analysis | \n
| 2004 | \nWild and colleagues radiocarbon-date bone from Schletz alongside Talheim and Herxheim | \n
| 2007 | \nGolitko and Keeley cite Schletz as evidence for sustained warfare at the end of the Linearbandkeramik | \n
Whether Schletz is an isolated local catastrophe or one episode in a broader collapse at the end of the Linearbandkeramik is genuinely disputed. In 2004, Eva Maria Wild and colleagues radiocarbon-dated bone from Schletz alongside Talheim and Herxheim in Germany. The Schletz and Talheim deaths could be contemporaneous, despite lying more than 300 kilometres apart, while Herxheim's deposit appeared somewhat earlier. Read one way, a shared date for two distant sites suggests a wider regional crisis. Read more cautiously, statistical compatibility is not proof of a shared cause, and Herxheim's earlier date shows that not every Linearbandkeramik mass-fatality site belongs to the same moment.
\nMark Golitko and Lawrence Keeley later argued in Antiquity that Schletz, Talheim, and other fortified sites reflect a genuine escalation of warfare across the Linearbandkeramik world as good land grew scarce, at the same time as the culture's proto-urban neighbours further east were experimenting with very different, much larger forms of settlement. Other archaeologists are more cautious about extending one site's story into a continent-wide narrative, noting that each enclosure has its own excavation history and its own uncertainties about how the dead came to be there. Both readings agree on what happened at Schletz; they disagree about how far its lesson travels.
\nEven the strongest reading of the Schletz ditch cannot identify who carried out the killing, or whether it was a single raid or the end point of a longer conflict between neighbouring communities. No weapons or material culture recovered from the site can be traced to an outside group, which is one reason some researchers have floated internal conflict, between rival Linearbandkeramik communities, as at least as plausible as an attack by outsiders. The count of 67 is also a floor rather than a ceiling: it describes what has been excavated, not the full extent of what the ditch may still contain. Other prehistoric puzzles carry limits of a different kind: the drowned world of Doggerland has to be reconstructed entirely from trawled fragments and seabed cores, with no excavation of a living site possible at all.
\nThis site's own record of collective violence before 3000 BCE lists Asparn/Schletz alongside Talheim, Kilianstädten, and more than a dozen other prehistoric sites, each rated for how firmly the evidence supports a violent reading rather than an accident of preservation.
\nAsparn/Schletz does not resolve the question of whether the European Neolithic was peaceful or violent, because no single site can carry that weight. What it does show, with unusual clarity, is what a Linearbandkeramik community's ending could look like when it went wrong: an enclosure built for defence that failed to defend anyone, a death toll that fell on children and adults alike, and a burial that never happened. The ditch preserved what a grave would have hidden, and that is the reason a farming village on a low rise in Lower Austria is still being argued about more than seven thousand years later.
\nWhat is Asparn/Schletz? Asparn/Schletz is an Early Neolithic Linearbandkeramik enclosure near Asparn an der Zaya in Lower Austria, where excavations beginning in 1983 recovered the unburied remains of at least 67 people killed around 5000 BC, most by blows from stone axes or adzes.
\nHow does it differ from the Talheim death pit? Talheim's 34 victims were gathered into a single pit after death, while at Schletz the dead were left exposed along the base of the settlement's defensive ditches, where carnivore scavenging and disarticulation show that no burial rite followed the killing.
\nIs "genocide" the right word for what happened at Schletz? That is disputed. Some popular accounts use the term because women and children appear among the dead in roughly normal proportions, unlike at Talheim and Kilianstädten, but most archaeological literature prefers "massacre," partly because genocide is a modern legal category that risks reading a specific intent into evidence that cannot support it.
\nDoes Schletz prove that Neolithic Europe was generally violent? Not on its own. Researchers such as Golitko and Keeley have argued that Schletz fits a wider pattern of Linearbandkeramik warfare, while others caution that each mass-fatality site has its own history and that generalising from a handful of enclosures to the whole Neolithic goes beyond what the evidence can support.
\nIn 1975, West Germany did not have a courtroom secure enough to try the imprisoned leadership of the Red Army Faction, so the state of Baden-Württemberg built one from nothing, on the grounds of Stuttgart's Stammheim prison, in time for the trial to open that May. The windowless, reinforced building was never meant to be a public monument. It became one anyway, and outlived its own verdict by more than four decades before it was finally demolished.
\nStuttgart already had a functioning Higher Regional Court. What it did not have was a way to move four defendants from a high-security prison wing to that courthouse, several times a week, for a trial expected to run for years, without creating a standing opportunity for a rescue attempt. The indictment rested on the RAF's 1972 May offensive against American military and West German police targets; our record of these operations documents the offensive in full.
\nThe solution was architectural rather than procedural: put the courtroom next to the cells, so that no journey through the city was ever necessary. The federal agency's dossier on the trial states this directly — a new building beside the prison replaced the transport risk with a fixed, defensible perimeter. It was, in effect, an admission that the ordinary machinery of West German justice could not accommodate a trial the state considered existential.
\nConstruction moved fast by the standards of German public building. A single-storey, windowless hall of reinforced concrete and steel went up on the prison grounds through 1974 and into early 1975, with cabling strung over the roof to obstruct any attempt to approach or land near the building from the air. Reporting at the time put the cost of the building alone at around twelve million Deutschmarks, more than half of an overall trial budget estimated at roughly twenty million — extraordinary sums for what was, functionally, a single room built for a single case.
\nInside, the courtroom held space for the panel of judges, prosecutors, a large defense team, journalists and the public, all separated from the four defendants by security glass and armed guards. There were no windows to the outside world — only artificial light and poured concrete, in a building designed to keep the outside out rather than to let anything in.
\n| Date | \nEvent | \n
|---|---|
| 1974–early 1975 | \nWindowless courtroom built beside Stuttgart-Stammheim prison, reported at around 12 million Deutschmarks | \n
| 21 May 1975 | \nTrial opens against four defendants from the RAF's founding leadership | \n
| 1975–1976 | \nListening devices are installed in defendants' cells; discovery of the bugging becomes public in 1977 | \n
| 20 January 1977 | \nPresiding judge Theodor Prinzing withdraws after a defense bias motion succeeds; Eberhard Foth takes over | \n
| 28 April 1977 | \nVerdict delivered after roughly 192 trial days: life sentences for the surviving defendants | \n
| 2015–2019 | \nNew high-security courthouse built nearby for about €29 million, designed by Müller Reimann Architekten | \n
| April 2019 | \nNew building opens; the 1975 courtroom stands empty | \n
| 2020s | \nStuttgart's monument-protection authority grants permission to demolish the 1975 building | \n
| 2025–2026 | \nThe new courthouse hosts an unrelated high-security trial, reviving comparisons to 1975 | \n
A courtroom is supposed to be neutral ground on which guilt is decided, not assumed. The Stammheim building argued something different before a single witness was called: that the men and woman inside it were already, by the state's own engineering, treated as a threat requiring a fortress. Security glass separated the dock from counsel and the public gallery. Cameras and armed personnel were a fixture rather than an exception. The defense argued repeatedly, across the trial's 192 sessions, that the physical setting itself compromised the presumption of innocence — a complaint the digital edition of the trial protocol preserves in the defense's own recorded objections, not as later commentary but as a running feature of the proceedings.
\nThat tension — a building whose every wall makes a claim about danger, in a system that is supposed to presume none — did not end with this trial. It is close to the argument a legal scholar made in 2026 about the courthouse's successor on the same ground, describing the site as a stage on which the state performs its own confidence as much as it adjudicates a case.
\nThe building's isolation was meant to keep threats out. It also, for a period in 1975 and 1976, kept a secret in: investigators had placed listening devices in defendants' cells to monitor conversations with their lawyers, a fact that only became public in 1977. When it emerged, several of the defendants' own chosen lawyers — including Otto Schily, later a federal minister — walked out of the proceedings and refused to take further part until they had assurances that client conversations were no longer monitored. The affair briefly threatened the legitimacy of a trial the state had already spent enormously to secure, and it remains the clearest evidence that the building's defenses were not only aimed outward. A parallel, quieter version of the same instinct — a Cold War state widening its own surveillance reach under pressure — runs through the contemporaneous expansion of NSA wiretapping in the United States and through the intelligence-sharing habits NATO built up across the same decades.
\nThe trial ended on 28 April 1977 with life sentences for the surviving defendants, closing a case that had also cost one defendant's life — she died in her cell in May 1976, before the verdict was reached, in a death that remained disputed. The courtroom itself did not close with the case. Baden-Württemberg's judiciary kept using the same high-security hall for other proceedings that required its particular defenses for decades afterward, through the years in which Germany remained a country divided by the wider Cold War the RAF had positioned itself against, and beyond. A building raised for one case became, by default, the state's standing answer to any case it judged comparably dangerous.
\nBy the 2010s the 1975 building was outdated by the standards it had itself set: modern high-security courtrooms needed different technology, different accessibility, and better conditions for juries, judges and the public than a windowless concrete hall from the height of the Cold War could offer. Baden-Württemberg built a new courthouse immediately next to it, opened in April 2019 at a reported cost of around €29 million and designed by the Berlin firm Müller Reimann Architekten under the state's own property and construction office. The old building stood empty afterward. Stuttgart's monument-protection authority weighed whether the courtroom where a defining trial of the Federal Republic had unfolded deserved formal preservation, and ultimately granted permission for its demolition rather than requiring it be kept as a memorial — a decision that treated the building as infrastructure that had done its job, not as a site owed permanence.
\nThe new courthouse did not retire the symbolism that came with its location. In 2025 and 2026, the state security chambers sitting on the same ground in Stammheim heard an unrelated high-security case, and press coverage reached instinctively for the comparison to 1975 — the same address, the same architecture of glass separation and restricted access, applied to a different generation of defendants and a different kind of alleged offense. As with the earlier proceedings that once filled our account of the third generation's unsolved cases, the honest position is that some of what happens inside a courtroom like this stays genuinely contested, and a record built to explain the building should say so rather than adjudicate it. What the echo confirms is narrower and more durable: a state that builds a fortress for one trial tends to keep the fortress.
\nThe Stammheim courtroom was never meant to outlast its case, yet it stood for more than forty years, hosted trials its builders never anticipated, and was replaced only by another high-security courthouse built on the same logic next door. Its real subject was never only the four defendants tried inside it — it was how a democracy builds space for judging people it has already decided are dangerous, a question the Kommando statement's claims of responsibility never had to answer and the state could not avoid.
\nWhy did West Germany build an entirely new courtroom instead of using Stuttgart's existing court? Moving the defendants repeatedly across the city to the Higher Regional Court's normal building was judged too great a security risk, so the state built a windowless courtroom directly beside the prison holding them, eliminating the transport route rather than trying to secure it.
\nWhat was discovered about the defendants' cells during the trial? In 1977 it emerged that investigators had installed listening devices in defendants' cells during 1975 and 1976 to monitor privileged conversations with their lawyers. Several defense lawyers, including Otto Schily, walked out of the proceedings in protest until they received assurances the surveillance had ended.
\nWhat happened to the 1975 building after the trial ended in 1977? It remained in continuous use by Baden-Württemberg's judiciary for other high-security cases for decades, was finally replaced by a new courthouse that opened in April 2019, and was subsequently demolished after the city's monument-protection authority granted permission.
\nIs the Stammheim site still used for high-security trials today? Yes. The new courthouse that replaced the 1975 building in 2019 sits on the same ground beside the prison and has continued to host state security cases, including an unrelated trial in 2025–2026 that drew renewed comparisons to the original 1975 proceedings.
\nOn 7 September 1191, the rearguard of Richard the Lionheart's army broke formation on the coastal plain north of Jaffa without waiting for an order, and the general charge that followed it became the only pitched battle Richard fought and won outright in the Holy Land. The Battle of Arsuf is usually told as a duel between two famous names, but the men who actually held the column together that day — a Hospitaller Grand Master, a Flemish nobleman, a Burgundian duke — did as much to decide it as either king did.
\nRichard left Acre in the last week of August 1191, six weeks after taking the city, with roughly 20,000 men bound for Jaffa. The route ran down the coast so that the fleet could shadow the column offshore, carrying supplies and giving the army somewhere to retreat if it was broken. That single choice shaped everything about the march: the army moved in a tight rectangular formation, infantry and crossbowmen screening the landward flank, knights and baggage on the seaward side, never allowing Saladin's cavalry a gap to exploit — a stark contrast to the undisciplined column that was annihilated at Civetot almost a century earlier.
\nSaladin shadowed the column from the hills for more than two weeks, probing constantly and waiting for the discipline to crack. The Templars held the vanguard; behind them came the Breton and Angevin contingents, then the Poitevins, then the Anglo-Normans, with the Knights Hospitaller under their Grand Master, Garnier de Nablus, forming the rearguard that bore the worst of the harassment. On 7 September, as the column cleared the forest of Arsuf onto open ground near the ruined town of the same name, Saladin committed his full army to stop it there.
\nRichard is the name attached to Arsuf, but the battle was fought by a column of commanders whose names rarely make it past a footnote. Garnier de Nablus had led the Hospitallers since 1190 and had spent the march absorbing arrow fire and watching his order's horses die under him one at a time. Hugh III, Duke of Burgundy, commanded the French royal contingent left behind when Philip II sailed home after the fall of Acre — a substantial force that answered to no one but Richard by agreement, not by feudal obligation. Robert de Sablé led the Templar knights in the vanguard. And on the exposed inland wing of the rearguard rode James of Avesnes, a nobleman from Hainaut who had brought Flemish and Frisian crusaders east and had already fought through the siege of Acre.
\nNone of these men is a stranger to specialists in the Third Crusade — the same is true of Isaac Komnenos on Cyprus, taken by this army only months earlier — but none of them has the name recognition that keeps an article off the "thin stub" list either. Arsuf is really their battle: Richard's contribution was timing, not presence at the point of contact.
\nRichard's plan, as the Latin chroniclers describe it, was to absorb Saladin's attacks without breaking formation until the whole Ayyubid army was committed at close range, then answer with a single coordinated charge. It was a plan that asked an extraordinary amount of the rearguard, which took the heaviest fire for hours while the order to charge did not come.
\nGarnier de Nablus asked Richard more than once for permission to counter-attack and was refused each time. The Hospitallers held their ground losing horses to Turkish and Bedouin archery at close range until, according to the Itinerarium, the Grand Master and a knight named Baldwin de Carron broke first, spurring out of the line without a signal. The rest of the Hospitaller line went with them rather than watch two of their own ride into the enemy alone.
\nRichard's response was the decision that made the battle his. A premature, unsupported charge from one wing of a defensive column could have been destroyed in isolation. Instead he ordered every division forward at once, turning what should have been an act of indiscipline into a general assault delivered at exactly the moment Saladin's whole force was in the open. The Ayyubid line broke under the weight of it, and Richard is credited by the chroniclers with the harder discipline of the day: pulling his knights back into formation for a second and third charge rather than letting them scatter in pursuit.
\nNot every part of the line reformed cleanly. Fighting on the exposed inland wing, James of Avesnes was thrown from his horse and surrounded. Ambroise and the Itinerarium both turn his death into the set-piece of the battle: dismounted and fighting on foot, he is said to have killed several attackers before he fell, and later tradition has him crying out for Richard to avenge him as he went down — a line worth reading as chronicle embellishment rather than a transcript, since no one who heard it is likely to have survived to report it accurately. What is better attested is the aftermath: his body was recovered from the field the next day and buried at Arsuf, and the Latin sources agree that Richard himself was present at the burial, alongside Guy of Lusignan, the displaced king of Jerusalem who had joined the crusade. His death was treated at the time as the single costliest loss of an otherwise one-sided victory.
\nThe Itinerarium claims some 7,000 Ayyubid dead and thirty-two emirs killed against a few hundred crusader losses; Bahā' ad-Dīn, writing from inside Saladin's own camp, does not dispute a heavy defeat but gives no figure that confirms the Latin total, and modern historians treat both sides' numbers as rhetorical rather than a body count anyone actually took. What is not in dispute is the shape of the outcome: Saladin's field army was mauled and its morale damaged, but it was not destroyed, and it withdrew rather than surrendered.
\nThat distinction mattered more than the casualty count. Our record of these operations lists Arsuf alongside the treaty Richard eventually signed at Jaffa for exactly this reason: a battle that breaks an army's nerve without breaking the army itself buys time and ground, not the war — the coastal states it preserved would still be fighting for their existence a century later, at the fall of Acre in 1291. Richard reached Jaffa on 10 September and spent weeks refortifying it, and Saladin went back to the harassment and avoidance tactics that had worked before Arsuf, rather than risking a second pitched battle. Arsuf opened the coast road; it did not open the road to Jerusalem.
\n| Date | \nEvent | \n
|---|---|
| 12 July 1191 | \nAcre falls to the crusaders | \n
| Late August 1191 | \nRichard's army marches south from Acre along the coast | \n
| Early September 1191 | \nSaladin's forces shadow and harass the column from the hills | \n
| 7 September 1191 | \nBattle of Arsuf; the Hospitaller rearguard breaks and Richard commits a general charge | \n
| 7 September 1191 | \nJames of Avesnes is killed on the exposed inland wing | \n
| 8 September 1191 | \nJames of Avesnes's body is recovered and buried at Arsuf | \n
| 10 September 1191 | \nThe army reaches Jaffa and begins refortifying it | \n
Arsuf gave Richard the one thing his reputation is built on: a pitched battle, fought on his own terms, that he unambiguously won — a discipline the crusading movement would not always manage, as the Fourth Crusade's descent into the sack of Constantinople would later show. It did not give the crusade Jerusalem, and within a year Richard would be negotiating a truce rather than besieging the city. What it did was keep the coastal corridor open long enough for that negotiation to happen from a position of strength rather than collapse — and it did so because a Hospitaller Grand Master's nerve broke at the right moment, and a king who had planned for exactly that possibility was ready to turn it into a victory instead of a disaster.
\nDid Richard order the Hospitaller charge that started the battle? No. The Latin chronicles agree that Garnier de Nablus and a knight, Baldwin de Carron, broke ranks without waiting for Richard's signal after enduring hours of archery fire. Richard's achievement was reacting fast enough to turn an unauthorised charge into a coordinated general assault before it could be cut off and destroyed.
\nHow large were the two armies? Chroniclers give Richard's army as roughly 20,000 men and Saladin's force as larger, sometimes put as high as 25,000–35,000. Medieval battle numbers were rarely counted precisely and modern historians treat these figures as approximate at best.
\nWas Saladin's army destroyed at Arsuf? No. It suffered heavy losses and its morale was damaged, but it withdrew in reasonably good order and continued to harass Richard's operations for another year. Arsuf secured the coast road to Jaffa; it did not end the war.
\nWhy is James of Avesnes remembered from this battle specifically? He was the highest-ranking crusader killed at Arsuf, and the two main Latin chronicles both turned his last stand into the emotional centre of their accounts, complete with a dying appeal to Richard that later tradition treats as legend rather than a verified quotation.
\nThe Dutch Republic never built a navy. It built five. Between 1574 and 1597 five separate admiralty boards, one for Rotterdam, one for Amsterdam, one for Zeeland, one for the Noorderkwartier and one for Friesland, were fixed in law as the permanent, competing institutions that would fight the Republic's wars at sea for the next two centuries. There was no ministry of the navy, no single commander-in-chief who outranked the others by right, and no single treasury. There were five colleges, five sets of accounts, five sets of shipyards, and one States-General that had to negotiate with all of them before a fleet could put to sea.
\nThe reason was structural, not accidental. The Union of Utrecht of 1579 had bound seven sovereign provinces into a confederation, not a state, and each province guarded its own purse and its own militia jealously. When the Earl of Leicester tried, during his brief governorship in the 1580s, to fold naval administration into a single college answerable to him, Holland read it as a bid to override provincial control of trade and shipping. The reaction produced the opposite of centralization: Amsterdam formally organized its own admiralty in 1586, the Noorderkwartier followed with a board of its own in 1589, and Friesland completed the pattern in 1596. Rotterdam's board, administering what the Dutch called "de Maze," and Zeeland's, seated at Middelburg, had older roots still. On 13 August 1597 the States-General stopped fighting the fact and legislated it: an Instructie fixed five admiralty colleges as the Republic's permanent naval administration, each drawing delegates from its own province and from the States-General, each with its own admission of ships, its own dockyards and its own officers.
\nEach admiralty combined roles that a modern state would split among a navy, a coastguard, a customs service and an admiralty court. It fitted out warships, policed its stretch of coast, collected duties on shipping, and judged prize cases when a captured enemy vessel was brought into port.
\n| Admiralty | \nSeat | \nFormally established | \nProvince represented | \n
|---|---|---|---|
| Rotterdam (de Maze) | \nRotterdam | \n1574 | \nHolland (Maas district) | \n
| Zeeland | \nMiddelburg | \n1584 | \nZeeland | \n
| Amsterdam | \nAmsterdam | \n1586 | \nHolland (Amsterdam district) | \n
| Noorderkwartier | \nHoorn and Enkhuizen | \n1589 | \nHolland (northern district) | \n
| Friesland | \nDokkum, later Harlingen | \n1596 | \nFriesland | \n
Amsterdam, the richest of the five by a wide margin, generally carried the largest share of the fleet's cost and could usually outbuild and outman the others when it chose to. Rotterdam and Zeeland followed at a distance, each with its own dockyards and its own captains. Friesland and the Noorderkwartier were consistently the smallest and the least able to meet the shipbuilding quotas the States-General assigned them, a weakness that shows up repeatedly in the correspondence of the war years. None of the five was subordinate to another on paper, which is exactly why coordinating them in wartime took constant negotiation rather than a single order down a chain of command.
\n| Date | \nEvent | \n
|---|---|
| 1574 | \nThe Rotterdam admiralty, the oldest of the five, is organized as Spanish naval power is pushed back from Holland's coast | \n
| 1584 | \nZeeland's own admiralty college is seated at Middelburg | \n
| 1586 | \nAmsterdam formally organizes its own admiralty, resisting the Earl of Leicester's attempt at a single central board | \n
| 1589 | \nThe States-General authorizes a separate admiralty for the Noorderkwartier | \n
| 1596 | \nFriesland establishes its own board, completing the pattern of five | \n
| 1597 | \nThe States-General's Instructie of 13 August fixes all five admiralties in law as the Republic's permanent naval administration | \n
| 1652–1674 | \nThe three Anglo-Dutch Wars test the five-headed system against the single, centrally funded Royal Navy | \n
| 1795 | \nThe Batavian Revolution abolishes the admiralty system along with the Republic itself | \n
The admiralties were funded almost entirely by convooien en licenten, convoy dues and trade licenses levied on shipping passing through Dutch ports, which made them the closest thing the Republic had to a federal tax authority. In principle the revenue was earmarked for building and equipping warships; in practice the sums fell short of what the States-General's shipbuilding quotas demanded almost every year of a major war, and from 1638 the States-General took to leasing out the collection of convooien en licenten to tax farmers to squeeze more reliable revenue out of the same trade. The system tied the navy's budget directly to the health of Dutch overseas commerce, including the fortunes of companies such as the one chronicled in how the Dutch East India Company transformed global trade networks: when trade slowed, the fleet's funding slowed with it, at the exact moment an enemy blockade was most likely to be squeezing that trade in the first place.
\nThe human cost of the arrangement shows up in officers who rarely appear outside specialist naval histories. Job de Wildt served as secretary of the Amsterdam admiralty for decades in the second half of the seventeenth century, the administrator who actually kept ships, crews and paperwork moving through the college's bureaucracy and who helped organize the fleet that carried William III to England in 1688. Egbert Meeuwszoon Kortenaer rose through the Rotterdam admiralty's service to become Lieutenant-Admiral of de Maze in January 1665 and was killed commanding its squadron five months later at the Battle of Lowestoft, the opening fleet action of the Second Anglo-Dutch War. Neither man commanded the whole Dutch fleet, because under this system almost no one did in the way an English or French admiral could claim to. A joint expedition required each admiralty to agree to contribute ships in roughly its assigned proportion, appoint its own officers to lead them, and accept a single lieutenant-admiral, usually but not automatically drawn from Holland's boards, as commander of the combined fleet for that campaign only. Some historians argue this diffusion of authority slowed Dutch fleet concentration at critical moments in the Anglo-Dutch Wars; others point out that the same structure gave the Republic five independent shipbuilding and revenue bases that an enemy could not knock out with one strike.
\nThe admiralties also had to share the sea with two chartered trading companies that armed and fought their own ships under their own commissions rather than an admiralty's: the VOC in Asia and the WIC in the Atlantic and West Africa. Those company fleets fought battles the States admiralties never saw, from the Strait of Malacca to the coast of Brazil, actions catalogued in our record of these operations alongside the North Sea campaigns the admiralty boards ran directly.
\nThe five-admiralty system survived because no province would trade its own board, its own dockyards and its own share of convoy revenue for a single national one, even when the inefficiency of divided command was plain to contemporaries. It absorbed the strain of the Rampjaar of 1672, when the field army collapsed but the fleet, expensively maintained by all five boards through the stadholderless years, held the coast. It outlived the Golden Age prosperity that had funded it so easily through convoy and license revenue tied to booming trade, and it kept limping on for another century after that prosperity thinned. The system that five jealous provinces had built to avoid handing power to any one of them finally ended not by reform but by revolution: the Batavian Republic swept the five colleges away in 1795 and replaced them with a single national navy, the arrangement the States-General had specifically avoided in 1597.
\nWhy did the Dutch Republic have five admiralties instead of one? The Republic was a confederation of seven sovereign provinces under the Union of Utrecht, not a kingdom with a central government. Each of the maritime provinces insisted on controlling its own shipping revenue, dockyards and officer appointments, and an early attempt to fold naval administration into one college was read as an attack on that control. The compromise the States-General eventually wrote into law in 1597 was five permanent boards rather than one.
\nWhich admiralty was the most powerful? Amsterdam, by far the wealthiest of the five, generally funded and fielded the largest share of the fleet and could usually outbuild the others when it chose to. Rotterdam and Zeeland followed at a distance, each with its own dockyards, captains and revenue base. Friesland and the Noorderkwartier were consistently the smallest and struggled hardest to meet the shipbuilding quotas assigned to them.
\nDid the VOC and WIC have their own navies too? Yes, and this is a separate layer from the five admiralties. The Dutch East India Company and the Dutch West India Company armed and commanded their own fleets under their own charters, and fought company wars in Asia and the Atlantic that the States admiralties did not direct or pay for.
\nWhat replaced the admiralty system? The Batavian Revolution of 1795 abolished the Dutch Republic and its five admiralty colleges together. In their place the new Batavian state built a single centralized national navy, the very arrangement the provinces had specifically avoided when they legislated five boards in 1597.
\nOn 21 October 1096, a Seljuk force under Kilij Arslan I ambushed several thousand crusaders three miles from their camp at Civetot, on the Gulf of Nicomedia, and killed most of them, including the lord who had led the expedition's advance contingent across two continents. Walter Sansavoir, lord of Boissy-sans-Avoir, had set out from France seven months earlier at the head of one of the first armed bands to answer Pope Urban II's call at Clermont. He never saw the main crusading army, and neither did the two other lords the chronicles name among the dead at Civetot that autumn. Their deaths came almost a full year before Godfrey of Bouillon's forces reached the walls of Nicaea.
\nWalter Sansavoir's epithet was not a figure of speech. Boissy-sans-Avoir, his lordship west of Paris, was modest even by the standards of minor French nobility, and the name attached to him — "without possessions" — followed him east. When Peter the Hermit's preaching drew thousands of poor pilgrims across northern France and the Rhineland in the winter of 1095–96, Walter organized one of the first bands to actually leave, departing in March 1096, weeks ahead of Peter's own, much larger following.
\nThe logic was practical. A smaller, earlier group could move faster and forage more easily than the tens of thousands who would follow. Walter's contingent crossed Germany and Hungary without serious incident — King Coloman of Hungary, wary of the disorder these bands could cause, let them pass on condition of good behaviour — and it was only in Byzantine territory, near Belgrade, that supply disputes turned violent. A clash over foraging killed some of Walter's men before the Byzantine governor at Niš intervened, provisioned the column, and escorted it on to Constantinople, where it arrived around mid-July 1096.
\nEmperor Alexios I Komnenos had asked the West for soldiers, not a crowd of the poor, and he treated Walter's arrival as a problem to be managed rather than a relief force. He held the group outside the walls, supplied it, and waited for Peter the Hermit's larger host, which reached Constantinople in early August after a march that had turned genuinely bloody at Semlin and Niš. Once the two groups had combined — an uneasy mixture of French, German, Italian and Lombard bands with no agreed command — Alexios ferried the whole assembly across the Bosporus and settled it at a fortified supply base the Latin sources call Civetot, on the Gulf of Nicomedia. His instructions were explicit: wait for the organized baronial armies, and do not provoke Kilij Arslan I, the Seljuk sultan whose capital at Nicaea lay barely fifty miles away.
\nNot everyone in the camp died in battle. Albert of Aachen's chronicle, in the sixth chapter of his first book, records that Walter of Poissy, lord of Poissy and one of the expedition's own commanders, did not live to see the disaster that followed — he died of illness in the camp, one of an unknown number the sickness took before the Turks did. The chronicle gives no further detail; where a cause is not recorded beyond the fact of the illness, that silence is the whole of what survives.
\nCamp discipline did not hold. French bands raided toward Nicaea for plunder and food, and in September a mixed force of German and Italian pilgrims — estimated at several thousand — marched some four days from Civetot under Rainald of Broyes, lord of Broyes, and seized a small Turkish fort called Xerigordos, intending to use it as a forward base. Kilij Arslan responded quickly, sending a relief force that cut off the fort's water supply, which lay outside its walls. After eight days without water, the garrison capitulated on 29 September.
\nAlbert of Aachen's sixteenth chapter counts Rainald of Broyes among those killed when the fort fell, and that is the account this record follows; other tellings of the same siege describe Rainald surrendering and accepting conversion to save his life before the garrison was killed or enslaved regardless, and the sources do not resolve which came first. What is not disputed is the outcome for almost everyone else inside the walls: those who refused conversion were killed, and the women and children among them were taken into slavery.
\nTurkish agents reportedly slipped into the Civetot camp afterward and spread word that the raiding party had taken Nicaea itself and was dividing its riches — a rumor the surviving crusaders, hungry for the same plunder, chose to believe. Peter the Hermit was in Constantinople at the time, negotiating supplies with Alexios, and Walter Sansavoir is recorded arguing for patience until the sultan's real strength could be judged. He was overruled. On 21 October, some twenty thousand pilgrims, soldiers and camp followers marched out of Civetot toward Nicaea and into a wooded ravine where Kilij Arslan's army was waiting.
\nThe ambush was total. Albert of Aachen's twenty-first chapter describes Walter Sansavoir fighting to the end and dying pierced by seven arrows, his mail hauberk unable to stop them. Most of the column was killed where it stood; a remnant of perhaps three thousand fled back toward the coast and held out in a ruined fortification until a Byzantine relief fleet, alerted by Peter the Hermit, evacuated them across the strait. Our record of these operations lists both Walter Sansavoir and Rainald of Broyes among the war's earliest noble deaths — men who never reached the campaigns their epitaphs are usually filed under.
\n| Date | \nEvent | \n
|---|---|
| Nov 1095 | \nUrban II calls for crusade at Clermont | \n
| March 1096 | \nWalter Sansavoir's contingent departs France | \n
| Mid-July 1096 | \nWalter's group reaches Constantinople | \n
| Early Aug 1096 | \nPeter the Hermit's larger host arrives; the combined force is ferried to Civetot | \n
| c. mid-Sept 1096 | \nRainald of Broyes's raiding party captures Xerigordos | \n
| 21–29 Sept 1096 | \nSiege of Xerigordos; the garrison surrenders after its water is cut off | \n
| 21 Oct 1096 | \nAmbush at Civetot; Walter Sansavoir is killed and the camp is destroyed | \n
| Spring 1097 | \nThe main crusading armies arrive at Nicaea and pass the site | \n
When the main armies reached Nicaea the following spring, chroniclers described a hill of bleached bones near Civetot as the only monument to the men who had gone ahead of them. Peter the Hermit survived — his absence from the camp that October was the accident that spared him — and he rejoined the crusade in a much-reduced role, present at Antioch and Jerusalem but never again leading an independent force. The disaster at Civetot is usually treated as a prelude, a cautionary chapter cleared away before the "real" First Crusade begins with the massacres its passage triggered in the Rhineland and, later, the sieges that built the crusader states.
\nThat framing costs these men their own history. Walter Sansavoir, Walter of Poissy and Rainald of Broyes held real lordships in France before they mortgaged them to answer a papal summons, and each died in a way the chronicles recorded precisely because he was a lord, not one of the anonymous thousands beside him. Disease, siege and ambush killed them within seven months of leaving home, before Godfrey of Bouillon or the Knights Templar that followed the crusading movement had done anything to be remembered for. The crusader states these men never lived to see would stand for two centuries, through catastrophes as different as the sack of Constantinople in 1204 and the fall of Acre in 1291, and would depend at every stage on the labour of women who traveled with these same armies and are rarely named at all — a gap the crusading nobility's own record makes plain in its own way.
\nCivetot is remembered, when it is remembered at all, as a disaster that cleared the ground for the crusade proper — an undisciplined mob punished for ignoring an emperor's advice. That verdict is not wrong, but it flattens three men who held real titles and made a real journey into a single cautionary anecdote. Walter Sansavoir organized and led an army across two empires before he died fighting in a ravine near Nicaea. Rainald of Broyes held a captured fort for eight days without water before the sources lose track of exactly how he died. Walter of Poissy never reached a battlefield at all. None of them lived to see Jerusalem, and none of them is a name most accounts of the First Crusade bother to keep.
\nWhy did Walter Sansavoir leave France months before the main crusading armies? His band was one of several that set out in spring 1096, ahead of the baronial forces still mustering supplies and soldiers, on the reasoning that a smaller group could move and forage faster. That same speed left it without the numbers or discipline the later, organized armies had.
\nWhat happened to Peter the Hermit during the battle? He was in Constantinople negotiating supplies with Emperor Alexios I when the ambush occurred, so he was not present for it. He survived, later rejoined the main crusade when it reached Nicaea in 1097, and traveled with it to Antioch and Jerusalem, though he never again commanded an independent force.
\nDid anyone from the Civetot camp survive the ambush? A remnant of roughly three thousand escaped back toward the coast and held a ruined fortification until a Byzantine fleet, alerted by Peter the Hermit, rescued them and carried them back across the Bosporus. The great majority of the twenty thousand who marched out that morning did not.
\nIs Rainald of Broyes's death at Xerigordos disputed? Albert of Aachen counts him among those killed when the fort fell after an eight-day siege, which is the account this record follows. Other versions of the same episode describe him surrendering and converting to save his life first, and the surviving sources do not settle which is accurate.
\nIn the summer of 56 BC, a Roman fleet built that same year on the Loire met the war navy of the Veneti in the open water of the Gulf of Morbihan, and won not by ramming or archery but by cutting rope. Roman crews fitted long poles with sharpened hooks, normally used against siege walls, and used them to sever the halyards holding the Veneti's leather sails to their masts. It was an improvised answer to a real tactical problem: the Veneti had built their ships for the Atlantic, and those ships were, by Caesar's own account, better suited to the fight than anything Rome could put on the water that year (Caesar, The Conquest of Gaul, Book III.7-16). The engagement is recorded in the campaign table of our record of Caesar's operations in Gaul and Britannia as the naval campaign against the Veneti, commanded not by Caesar in person but by his young legate, Decimus Brutus.
\nThe trouble began the previous winter. In 57 BC the Veneti and other Armorican tribes of Brittany had submitted to Publius Crassus, son of the triumvir Marcus Crassus, handing over hostages as a guarantee of loyalty (BG II.34). Over the winter of 57-56 BC, Crassus's legion, quartered among the Andes near the Atlantic coast, ran short of grain, and he sent officers, including two named by Caesar as Quintus Velanius and Titus Silius, to requisition supplies from the coastal tribes (BG III.7-8). The Veneti seized them, and offered to trade the captured officers for the return of their own hostages. Several neighboring tribes — the Osismi, Lexovii, Namnetes, Ambiliati and others — joined the seizure and prepared for war rather than risk being singled out for punishment alone (BG III.9).
\nCaesar's own explanation is that the Veneti feared losing their hostages permanently and calculated, correctly for a season, that Rome's legions were poorly equipped for their terrain: tidal inlets, marsh and fortified promontories reachable only at low tide, ill-suited to infantry sieges (BG III.9, 12). Some modern historians note that Caesar had also just reconnoitered an expedition to Britain, and wonder whether the Veneti, who controlled the cross-Channel trade routes along this coast, had further reasons to resist a Roman fleet in their waters — a question the text itself does not answer.
\nCaesar's account gives an unusually detailed technical description of Veneti shipbuilding, and it is worth taking at face value even where his political framing should not be: the ships were built entirely of oak, with flat bottoms designed to sit safely on shoals and tidal mudflats, and bows and sterns raised high to survive Atlantic swells and storms that a Mediterranean galley could not have weathered (BG III.13). Their cross-beams were a foot wide and fastened with iron nails as thick as a thumb, their anchor cables were iron chains rather than rope, and their sails were made of raw or tanned hides rather than woven cloth or hemp — a design Caesar attributes either to a scarcity of flax or, more plausibly, to leather's greater resilience against Atlantic gales.
\nNone of this made the Veneti ships fast. What it gave them was strength and height: they rode higher out of the water than the Roman vessels, built hastily on the Loire that same spring and crewed largely with rowers drawn from the Roman province rather than veterans of Mediterranean naval warfare. Roman ships could not easily ram the heavier oak hulls, and missiles thrown upward at much higher decks did little damage. The Veneti fleet — Caesar puts the number in the low 200s of ships, gathered from across the Armorican coast and its allies — mustered at the mouth of the gulf to meet the Romans as they arrived (BG III.14).
\nCaesar states he watched the engagement from high ground on the shore with the bulk of his land army, unable to intervene directly (BG III.14). The Roman advantage lay only in oars and maneuverability, since a rowed vessel could turn inside a sailing ship becalmed or fighting the wind. Roman crews closed with individual Veneti ships one or two at a time rather than attempting a general melee, and it was here that Brutus's crews used the hooked poles — described in the sources as similar to the iron hooks Roman soldiers used against fortification walls — to grab and tear the ropes securing the yard to the mast (BG III.14). Once a Veneti ship's rigging was cut, its sail collapsed and the ship lost all way, since the Veneti had built no provision for rowing their oceangoing vessels under oar power alone (BG III.13-14).
\nTwo or three Roman ships would then surround each disabled vessel and board it, overwhelming the crew in close combat where Roman legionaries training as marines had the advantage. As more Veneti ships were disabled, others tried to flee under sail toward open water. At that point, Caesar records, the wind dropped to a dead calm, leaving the surviving Veneti ships stranded and unable to maneuver at all (BG III.15). Roman ships rowed down the survivors one by one through the rest of the afternoon; by evening, according to Caesar, only a handful of Veneti ships reached shore out of a fleet that had numbered in the hundreds.
\n| Date | \nEvent | \n
|---|---|
| 57 BC | \nVeneti and other Armorican tribes submit to Publius Crassus; hostages exchanged (BG II.34) | \n
| Winter 57-56 BC | \nCrassus's officers, sent to requisition grain, are seized by the Veneti (BG III.7-8) | \n
| Spring 56 BC | \nCaesar orders a fleet built on the Loire and recruits rowers from the Roman province | \n
| Summer 56 BC | \nThe Roman and Veneti fleets meet in the Gulf of Morbihan; Brutus commands the Roman ships (BG III.11, 14) | \n
| Summer 56 BC | \nHooked poles cut the Veneti sail-ropes; a sudden calm traps the surviving ships (BG III.14-15) | \n
| After the battle | \nThe Veneti senate is executed and the population sold into slavery (BG III.16) | \n
The defeat ended Armorican resistance for that campaigning season; with no fleet left and no army in the field, the coastal towns had nothing to defend them (BG III.16). Caesar's response was deliberately severe: he had the Veneti's ruling council put to death and sold the rest of the tribe into slavery, a punishment he explicitly ties to their violation of the customary immunity of envoys — the seized Roman officers — rather than to the military threat itself (BG III.16). It was among the harshest reprisals Caesar records against any Gallic community in the entire war, meant to deter revolt elsewhere — though other tribes further north and east tested Rome's frontier commanders again within a few years, including the legates who later held the line on the Rhine under Drusus and Germanicus.
\nThe battle is also one of the few securely dated naval engagements fought in open Atlantic water in antiquity, unlike the Mediterranean ship warfare familiar from Athens's naval buildup before Salamis or the triremes launched from Athenian shipyards. Those fleets fought with rams among light, fast galleys; Brutus's crews faced a wholly different naval architecture and solved it with an improvised weapon rather than a better ship.
\nThe naval campaign against the Veneti rarely gets the attention paid to the sieges of the later Gallic War, but it is one of the more revealing episodes of the conflict precisely because it was not won by Roman strength. Rome's ships were weaker, its crews less experienced with the ocean, and its commander a legate rather than Caesar himself. What decided the battle was a mechanical countermeasure to a specific enemy advantage, followed by weather the Romans could not have planned for. The brutal settlement that followed says as much about how Rome wanted to be seen by the rest of Gaul as it does about the Veneti themselves.
\nWho commanded the Roman fleet against the Veneti? Caesar assigned command to a legate he calls Brutus, generally identified as the young Decimus Junius Brutus Albinus, who would later serve under Caesar in the civil war and, in 44 BC, join the conspiracy against him. Caesar himself watched the battle from the shore with the land army rather than commanding at sea (BG III.11, 14).
\nWhy couldn't Roman ships simply ram or outfight the Veneti navy? The Veneti ships were built of solid oak with reinforced timbers, making them far sturdier than the lighter Roman vessels built hastily that spring, and their high sides put Roman missile troops at a serious disadvantage. Rome's only real advantage was that oared ships could maneuver better than sailing ships, especially in light or contrary wind (BG III.13-14).
\nWhat was the actual tactic that won the battle? Roman sailors used long poles fitted with sharpened hooks, adapted from equipment used in siege warfare, to catch and cut the ropes holding each Veneti ship's sail to its mast. Once disabled, a ship could not maneuver, and Roman crews would surround and board it while the crew was unable to escape (BG III.14).
\nWhat happened to the Veneti after their defeat? Caesar had the entire Veneti governing council executed and sold the rest of the population into slavery, a punishment he attributed specifically to their seizure of Roman officers rather than to ordinary battlefield defeat. It was one of the harshest reprisals recorded anywhere in the Gallic War (BG III.16).
\nIn August 1624, a forty-year-old bishop with a reputation for ruthless competence took his seat on the royal council of Louis XIII and did not leave real power again until his death eighteen years later. Armand Jean du Plessis, Cardinal de Richelieu, inherited a France still scarred by the religious civil wars of the previous century and threatened on every border by the Habsburg dynasty that ruled both Spain and the Holy Roman Empire. By the time he died in December 1642, French Protestants had lost their fortified towns but kept their faith, the crown's authority ran deeper into the provinces than it ever had, and France had committed itself to the European war that would eventually break Habsburg supremacy. None of that was inevitable. It was the product of one minister's patient, often brutal insistence that the interests of the state overrode every other loyalty — including, when necessary, his own church's confessional solidarity.
\nRichelieu was born into minor nobility in Paris in 1585 and was pushed toward the church almost by administrative accident: his family held the bishopric of Luçon as a hereditary benefice, and when his elder brother declined to take orders, the position fell to Armand instead. He was consecrated bishop of the small, poor Poitou diocese in 1607 at just twenty-one, using the post less as a devotional calling than as a training ground in provincial administration. His political break came through Marie de' Medici, regent for her young son Louis XIII, who secured his appointment as a secretary of state in 1616. A cardinal's hat followed in 1622, and in 1624, after his predecessor as chief minister was arrested on corruption charges, Richelieu took the position that would define the rest of his life.
\nHis early position was more precarious than the later legend of an all-powerful cardinal suggests. Louis XIII was a suspicious, reluctant patron, and Richelieu owed his survival less to royal affection than to his ability to make himself indispensable — and to eliminate rivals before they eliminated him.
\nThe Edict of Nantes, Henry IV's 1598 settlement with his former Huguenot co-religionists, had granted French Protestants not just freedom of worship but fortified towns and their own armed garrisons — a state within the state that Richelieu regarded as an intolerable threat to royal authority, regardless of his own office as a prince of the Catholic Church. When the Huguenot stronghold of La Rochelle rose in open rebellion with the promise of English naval support, Richelieu personally directed the siege that followed in 1627–1628. Unable to storm a well-fortified Atlantic port, he ordered a massive stone-and-timber sea wall built across the harbor mouth, reportedly inspired by accounts of Alexander the Great's causeway at Tyre, sealing the city off from resupply. The siege dragged on for over a year; when La Rochelle surrendered in October 1628, famine had killed a large share of a population that had numbered around 25,000 before the siege began.
\nThe victory was total but the peace that followed was calculated rather than vindictive. The 1629 Peace of Alais stripped Huguenots of their political and military privileges — their fortresses and private armies were gone for good — while leaving the religious toleration of the Edict of Nantes intact for another half-century, until Louis XIV revoked it in 1685. Richelieu had separated the religious question from the political one, itself a small revolution in how a Catholic minister could treat a Protestant minority.
\nRichelieu's boldest and most dangerous confrontation came not with Protestants but with the Catholic devout party at court, led by his former patron Marie de' Medici herself. By 1630 the queen mother had come to see Richelieu's foreign policy — an emerging willingness to ally with Protestant powers against the Catholic Habsburgs — as a betrayal of the faith, and during a severe illness of Louis XIII in September 1630 she pressed her son to dismiss the cardinal once he recovered. On 10–11 November 1630, in a scene later called the Day of the Dupes, Louis appeared to yield to his mother's demand. Richelieu's enemies at court celebrated prematurely; within a day the king reversed course, confirmed his minister in office, and sent his own mother into permanent exile, where she died in poverty in 1642, having spent her final years plotting fruitlessly against him.
\nThe episode crystallized the doctrine for which Richelieu became infamous across Europe: raison d'état, the idea that the state's security could justify actions — including a cardinal allying with Protestant princes — that a purely confessional politics would forbid. Critics accused him of cynicism; defenders argued he simply recognized, earlier than most contemporaries, that a multi-confessional Europe required a foreign policy based on power rather than creed.
\nRichelieu had been maneuvering against Habsburg power for years before France formally joined the fighting. As early as January 1631 he arranged the Treaty of Bärwalde, by which France subsidized the Swedish king Gustavus Adolphus's campaigns against the Holy Roman Emperor without French troops needing to fire a shot — a case of papal and cardinal diplomacy crossing confessional lines that scandalized more traditional Catholic opinion. That indirect strategy held until 1634, when a combined Habsburg army crushed the Swedish-led Protestant forces at the Battle of Nördlingen, threatening to knock France's remaining allies out of the war entirely. Richelieu concluded that only direct French intervention could prevent a total Habsburg victory, and on 19 May 1635 France declared war on Spain, followed by war against the Emperor the following year. The fighting outlived Richelieu by years, ending only with the Peace of Westphalia in 1648 for the Empire and the Peace of the Pyrenees in 1659 for Spain, but the strategic logic he set in 1635 held throughout: contain the Habsburgs by force rather than negotiation alone.
\nDomestically, Richelieu extended royal authority into the provinces through intendants — royal commissioners dispatched from Paris to oversee taxation and justice, bypassing local nobles and officeholders whose loyalty the crown could never fully guarantee. He also grasped, much as the crown's earlier use of the printing press to project state authority had shown, that controlling information mattered as much as controlling territory. In 1631 he lent his patronage to Théophraste Renaudot's Gazette, France's first officially sanctioned newspaper, which served in part as a vehicle for government-approved narratives.
\nHis most enduring cultural creation was the Académie française, founded through letters patent Louis XIII signed on 29 January 1635, with statutes formalized the following month. Its charge — to give French a fixed grammar and vocabulary capable of treating the arts and sciences with precision — reflected Richelieu's conviction that a unified, prestigious French language was itself an instrument of state power, not merely a matter of literary taste. The Académie has continued to operate, with interruptions, ever since, arguably outlasting even the political settlement later expressed in stone at Versailles.
\n| Year | \nEvent | \n
|---|---|
| 1585 | \nArmand Jean du Plessis born in Paris | \n
| 1607 | \nConsecrated bishop of Luçon at age twenty-one | \n
| 1616 | \nAppointed secretary of state under Marie de' Medici's regency | \n
| 1622 | \nNamed a cardinal of the Catholic Church | \n
| 1624 | \nBecomes Louis XIII's chief minister | \n
| 1627–1628 | \nDirects the siege of La Rochelle against the Huguenots | \n
| 1629 | \nPeace of Alais strips Huguenot military privileges | \n
| 1630 | \nThe Day of the Dupes; Marie de' Medici exiled | \n
| 1631 | \nTreaty of Bärwalde subsidizes Sweden against the Habsburgs | \n
| 1635 | \nAcadémie française founded; France declares war on Spain | \n
| 1642 | \nRichelieu dies in Paris, still in office | \n
Richelieu left France a paradox that later ministers inherited along with the state he had strengthened: a monarchy more centralized and more feared abroad than it had been in generations, purchased through decades of continuous war, heavy taxation, and the deliberate suppression of any institution — noble, Protestant, or parliamentary — that might check royal will. Louis XIV would inherit both the tools Richelieu built and the war debts they generated, and the absolutist court culture that reached its fullest expression at Versailles owed an unmistakable debt to the administrative machinery a cardinal had assembled a generation earlier. Whether Richelieu deserves credit as the architect of the modern French state or blame as the man who normalized reason of state over conscience remains, as it was in his own lifetime, a genuinely contested question.
\nWas Richelieu actually more powerful than King Louis XIII? Richelieu depended entirely on the king's continued favor and was recalled from the brink of dismissal at least once, most dramatically during the Day of the Dupes in 1630. Real authority still ran through Louis XIII, but Richelieu was uniquely skilled at making his continued ministry seem indispensable to the king's own interests.
\nDid a Catholic cardinal really ally with Protestant rulers? Yes — Richelieu subsidized the Swedish Protestant king Gustavus Adolphus against the Catholic Holy Roman Emperor beginning in 1631, and France later fought directly alongside Protestant powers after 1635. Some contemporaries and later Catholic critics saw this as scandalous; Richelieu justified it as necessary to prevent total Habsburg domination of Europe.
\nWhat happened to the Huguenots under Richelieu? After the fall of La Rochelle in 1628, the 1629 Peace of Alais removed Huguenot fortified towns and private armies but preserved the religious toleration granted by the Edict of Nantes. Full religious toleration would last until Louis XIV revoked the Edict in 1685, decades after Richelieu's death.
\nDid Richelieu write the Testament politique himself? The work is traditionally attributed to Richelieu and circulated as his political testament, but it was published posthumously and historians continue to debate how much of the text he personally dictated versus secretaries who assembled it from his papers.
\nIn the spring and early summer of 1096, months before the armies of the First Crusade reached the Holy Land, bands of armed crusaders and local townspeople turned on the Jewish communities of the Rhineland, killing an estimated several thousand people in Speyer, Worms, Mainz, Cologne, and smaller towns nearby. The violence had no sanction from Pope Urban II, who had called for the crusade at Clermont in November 1095 to fight Muslim powers in the East, not to attack Jews in Germany. It was instead the work of self-organized bands, most notoriously one led by a minor nobleman named Emicho, who argued that Christians should not march thousands of miles to fight unbelievers while leaving unconverted communities at home untouched. The result was the worst violence against Jewish communities in Christian Europe to that point, and a pattern of crusade-linked persecution that would recur for centuries.
\nUrban II's sermon at Clermont set off a wave of popular enthusiasm his own bishops and knights could not fully control. Alongside the feudal armies assembling through the winter of 1095–96, informal bands of pilgrims, minor knights, women, and other non-combatants gathered under preachers like Peter the Hermit to march east on their own timetable — a movement historians call the People's Crusade. Among the separate bands forming in the Rhineland was one led by Count Emicho of Flonheim, sometimes called Emicho of Leiningen, whose contingent had no connection to Peter's group and no episcopal approval. Emicho and his followers reasoned, according to the Hebrew chronicles that later described them, that a war for Christ should begin by forcing the conversion or destruction of Jewish communities living under Christian princes — an argument with no basis in Urban's own preaching. Several bishops moved quickly to shelter Jewish residents once the danger became clear, with very mixed success.
\nThe violence reached Speyer on 3 May 1096, when a mob attacked the city's Jewish quarter and killed around a dozen people who could not reach safety in time. Bishop Rüdiger Huzmann, who had personally invited a Jewish community to settle in Speyer in 1084 and walled off a protected quarter for them, moved immediately to shelter the survivors inside his own fortified residence. He went further than most fellow bishops would: chroniclers record that he had the hands cut off some of the attackers he could identify. Speyer's Jewish community survived 1096 largely intact because its bishop had both the will and the means to protect it — a contrast that events in Worms and Mainz would make painfully clear.
\nEmicho's band reached Worms on 18 May 1096. Some Jewish residents were killed in the town itself; most fled to the bishop's palace, but after roughly a week the crusaders broke in, and estimates drawn from the Hebrew chronicles put the dead at several hundred to around a thousand.
\nMainz suffered the worst single massacre of the campaign. Archbishop Ruthard, who held one of the most powerful sees in the empire, tried to turn Emicho's band away and took Jewish residents into his own palace for safety. It was not enough: sympathizers inside Mainz opened the gates, and on 27 May 1096 the crusaders overran the archbishop's residence. Traditional accounts describe over a thousand deaths in a single day, including parents who, faced with a choice between forced baptism and their children's lives, are said to have killed their own families rather than submit — an act the chroniclers frame within the Jewish concept of kiddush ha-Shem, sanctification of God's name through martyrdom. Historians treat the chronicles' more dramatic details, including specific numbers and reported speeches, with some caution, since they were composed to commemorate martyrs rather than to document events neutrally, but the scale of the killing at Mainz is corroborated across independent accounts.
\nEmperor Henry IV, in Italy at the time and no part of the violence, later accused Ruthard of profiting from the murdered Jews' property — a charge almost certainly sharpened by the fact that Ruthard was Henry's opponent in the ongoing Investiture Controversy over who controlled the appointment of bishops. Whether the accusation was accurate or simply convenient is still debated.
\nCologne's bishop took a different approach when Emicho's band arrived around 30 May: rather than holding a single fortified refuge, he dispersed the city's Jewish residents into smaller nearby towns, including Neuss, Wevelinghoven, Eller, and Xanten, hoping the crusaders would not track them down. Some survived; over the following weeks, though, armed bands located several hiding places and killed the Jews sheltering there, so Cologne's diocese recorded a death toll comparable to Mainz's, even without a single massacre matching the scale of 27 May. Smaller-scale violence also touched Trier and Metz, farther up the Rhine and Moselle.
\n| Date | \nEvent | \n
|---|---|
| 27 November 1095 | \nUrban II preaches the First Crusade at the Council of Clermont | \n
| Spring 1096 | \nIndependent crusading bands, including Count Emicho's contingent, assemble in the Rhineland | \n
| 3 May 1096 | \nAttack on Speyer's Jewish community; Bishop Rüdiger Huzmann shelters survivors | \n
| 18 May 1096 | \nEmicho's band besieges Worms; most Jewish residents take refuge in the bishop's palace before it falls | \n
| 27 May 1096 | \nMainz overrun; Archbishop Ruthard's protection fails and the city's Jewish community suffers the campaign's heaviest losses | \n
| ~30 May 1096 | \nCologne's bishop disperses the Jewish community to nearby towns | \n
| June 1096 | \nCrusading bands hunt down dispersed Jews around Cologne; smaller-scale violence reaches Trier and Metz | \n
| 1097 | \nHenry IV permits Jews who were forcibly baptized to revert to Judaism | \n
| 1120 | \nPope Calixtus II issues Sicut Judaeis, reaffirming papal protection for Jewish communities | \n
Modern estimates for the total killed across the Rhineland range from roughly two thousand to twelve thousand, a spread reflecting genuine uncertainty rather than careless disagreement. The most detailed accounts come from Hebrew chronicles — the Mainz Anonymous, written closest to the events, and the longer Solomon bar Simson chronicle, compiled four decades later — both composed to memorialize the dead as martyrs, not to produce a dispassionate body count. Some historians read their numbers as roughly indicative, while others treat more consistent figures, such as the reported toll at Mainz, as reliable because they recur across independent traditions. What is not disputed is that the Rhineland massacres were the largest coordinated violence against Jewish communities in Latin Christendom to that date, carried out by crusaders acting entirely outside the chain of command Urban II had set in motion.
\nHenry IV was in no position to prevent the massacres, but his response afterward mattered. In 1097 he issued an edict permitting forcibly baptized Jews to return openly to Judaism, overriding church doctrine that baptism, however coerced, could not be undone — a position Pope Clement III's circle protested but could not reverse in the empire's own courts. Henry also opened an investigation into the plundering of Jewish property, a process that touched Ruthard's own conduct. A generation later, in 1120, Pope Calixtus II issued Sicut Judaeis, formally reaffirming papal protection for Jewish communities — a bull popes would reissue repeatedly over the following centuries, in part because the pattern set in 1096 kept recurring whenever new crusades were preached.
\nThe Rhineland massacres are a reminder that the First Crusade's violence did not begin at the walls of Nicaea or Jerusalem, and it did not always follow the instructions of the men who preached it. Bishops in Speyer, Worms, Mainz, and Cologne tried, with very different success, to shelter communities that had lived under their protection for generations; the outcome depended less on doctrine than on how much armed force a bishop could bring against a mob that answered to no one. The episode set a grim precedent: crusade fervor turning against local Jewish communities would recur across the two centuries of holy war fought by orders like the Knights Templar, through settlements like the Treaty of Jaffa, and on to the crusader states' collapse at the fall of Acre in 1291. It also forced emperor and papacy to state, more explicitly than before, that forced conversion had no legitimate place in Christian practice, even as Jewish communities' vulnerability to religious violence stayed unresolved for centuries.
\nDid Pope Urban II order the attacks on Jewish communities? No. Urban II's preaching at Clermont in 1095 called for an armed expedition against Muslim powers controlling Jerusalem and made no mention of Jewish communities. The massacres were carried out by independent bands, particularly one led by Count Emicho, without papal or episcopal authorization.
\nWhich cities suffered the worst violence? Mainz saw the heaviest single death toll, with traditional accounts describing over a thousand killed on 27 May 1096 after Archbishop Ruthard's protection failed. Worms and the towns around Cologne also suffered severe losses; Speyer's community survived largely intact because Bishop Rüdiger Huzmann sheltered them successfully.
\nWhy do death toll estimates vary so widely? The main sources are Hebrew chronicles composed to commemorate the dead as martyrs, not to produce precise casualty counts, and written years to decades after the events. Historians treat the overall scale — several thousand dead across multiple cities — as well established, while individual numbers remain harder to verify.
\nWhat happened to Jews who were forcibly baptized during the massacres? In 1097 Henry IV issued an edict allowing forcibly baptized Jews to return openly to Judaism, breaking with standard church teaching on the permanence of baptism and drawing protest from Pope Clement III's circle. Many who had converted under threat of death used this permission to formally rejoin their communities.
\nIn the autumn of 9 AD, three Roman legions marching through the forests east of the Rhine walked into an ambush that erased them from the Roman order of battle forever. Publius Quinctilius Varus, the governor of Germania, led roughly twenty thousand men into a running fight against a Germanic coalition organized by a Roman-trained officer named Arminius, and over three days the column was cut to pieces. The disaster, remembered as the Battle of the Teutoburg Forest, stopped Roman expansion east of the Rhine dead and left a boundary that would hold, with only brief exceptions, for the rest of the empire's existence.
\nArminius was a young noble of the Cherusci, a tribe whose territory lay near the middle Weser, and by the time of the battle he held Roman citizenship and equestrian rank, honors earned through service as an auxiliary officer in Rome's own army. He had likely fought for Rome in the brutal Pannonian revolt of 6 to 9 AD, gaining direct knowledge of Roman tactics, discipline, and the vulnerabilities of a legionary column on the march. Varus, by contrast, arrived in Germania in 7 AD with a reputation built in Syria, where as governor he had administered a settled, thoroughly Romanized province. Ancient writers agree that Varus treated the still-unconquered tribes east of the Rhine as though they were already a tax-paying province, and that this misjudgment, combined with his trust in Arminius as a loyal ally, opened the door to catastrophe.
\nArminius drew Varus's confidence over months, sitting at his council table while secretly building a coalition of Cherusci, Marsi, Chatti, Bructeri, and other tribes. In the late summer of 9 AD, with Varus's army stationed at a summer camp likely near the Weser, Arminius reported a local uprising and urged the governor to march his column through unfamiliar forest terrain to suppress it, a detour that took the legions away from the safer, more open routes they usually followed. Arminius rode ahead, ostensibly to rally allied troops, and instead joined the ambush he had spent the summer preparing.
\nVarus commanded three legions, the seventeenth, eighteenth, and nineteenth, along with six cohorts of auxiliaries and three cavalry squadrons, encumbered by a long baggage train of wagons, families, and camp followers, since the march was arranged more like a peacetime relocation than a campaign. The terrain chosen for the ambush, likely a narrow corridor between wooded hills and boggy lowland near modern Kalkriese in Lower Saxony, forced the column into a thin, strung-out line that could not form the close ranks and defensive squares that made a legion formidable in open battle. Rain and wind added to the chaos, and earthen ramparts built by the Germanic warriors funneled the Romans into killing zones where javelins could reach them from cover before they ever saw an enemy formation to fight.
\nThe fighting stretched across three days as the column tried repeatedly to break out and reach safer ground, digging in each night behind hastily built fortifications and losing more men with every renewed attempt to move. Some historians argue the Romans came close to escaping the worst terrain on the second day before a final storm and a fresh ambush finished the destruction; the ancient sources disagree on the precise sequence, and no continuous eyewitness account survives. Varus, seeing the battle lost, took his own life rather than face capture, following the example of several of his senior officers, as Cassius Dio's later account of the disaster's final hours records.
\nEstimates of the dead vary, but ancient and modern historians generally put the toll at somewhere between fifteen and twenty thousand men, a substantial share of Rome's entire military presence on the Rhine frontier. As Suetonius records, when word reached Augustus in Rome, the emperor was so shaken that he reportedly went weeks without cutting his hair or beard and would strike his head against a doorpost, crying "Quintili Vare, legiones redde," Quinctilius Varus, give me back my legions. The three destroyed legions, the seventeenth, eighteenth, and nineteenth, were never reconstituted, and their numbers disappeared permanently from the Roman army list, a mark of disgrace unique in the empire's history.
\nThe immediate strategic effect was the collapse of Rome's ambition to hold a province of Germania stretching to the Elbe, a goal Drusus and his son Germanicus had spent two decades pursuing. Germanicus later led punitive campaigns across the Rhine, recovering two of the three lost legionary eagles and visiting the battlefield itself to bury the scattered bones of Varus's dead, but Tiberius recalled him before the territory could be permanently reoccupied. The Rhine, not the Elbe, became the empire's fixed boundary for the remainder of Roman rule in the West, a limit that shaped how Roman roads and frontier infrastructure developed for centuries afterward, oriented around defense of an established line rather than continued conquest.
\nFor nearly two thousand years, the exact site of the battle was unknown, and antiquarians proposed dozens of candidate locations across northern Germany, disagreements that still surface in scholarship today. That changed in 1987, when a British army officer and amateur archaeologist named Tony Clunn began finding Roman coins and sling bullets near Kalkriese, north of Osnabruck, using a metal detector during off-duty exploration. Systematic excavation over the following decades uncovered thousands of Roman military artifacts, mass graves of unburied human and animal remains, and the remnants of the earthen rampart the Germanic warriors had built to channel the Roman column into ambush.
\nThe Kalkriese finds are now widely, though not unanimously, accepted as marking the Teutoburg Forest battlefield or at least a major episode within the running fight, since the geography of a three-day retreat likely spanned many kilometers across more than one site. The museum built at Kalkriese today interprets the coins, weapons fittings, and skeletal remains recovered there, evidence that corroborates the broad shape of Tacitus's later description of Roman bones left scattered where they fell, even where the literary and archaeological records cannot be reconciled in every detail.
\n| Date | \nEvent | \n
|---|---|
| 7 AD | \nVarus becomes governor of Germania after serving as governor of Syria | \n
| 6-9 AD | \nArminius likely serves as a Roman auxiliary officer during the Pannonian revolt | \n
| Late summer, 9 AD | \nArminius reports a false uprising and persuades Varus to divert his column | \n
| c. 8-11 September, 9 AD | \nThree days of ambush and running battle destroy Varus's army | \n
| 9 AD | \nVarus takes his own life; the seventeenth, eighteenth, and nineteenth legions are annihilated | \n
| 14-16 AD | \nGermanicus campaigns across the Rhine, recovering two of the three lost eagles | \n
| 1987 | \nTony Clunn's finds at Kalkriese begin the modern rediscovery of the likely battlefield | \n
The Teutoburg Forest disaster did not end Rome's power, but it drew a line the empire never again seriously tried to erase. For the next four centuries, the Rhine marked the practical edge of direct Roman rule in the region, a boundary that shaped everything from the artistic and political image Augustus cultivated in the years after the defeat to the military reforms his successors pursued along the northern frontier. Set beside Crassus's earlier catastrophe at Carrhae, Varus's defeat shows how a single miscalculated campaign could redraw an empire's ambitions for generations, and how the long, slow contraction that eventually reshaped the Western Roman Empire had roots in frontiers Rome chose, after Teutoburg, never to cross again.
\nWho was Arminius, and why did he betray Rome? Arminius was a noble of the Cherusci tribe who held Roman citizenship and equestrian rank after serving as a Roman auxiliary officer, likely during the Pannonian revolt. Ancient sources give no single explanation for his betrayal, but modern historians generally point to a combination of tribal loyalty, resentment of Roman taxation and legal authority, and personal ambition to lead a Germanic coalition rather than serve within it.
\nHow many Roman soldiers died at Teutoburg? Ancient and modern estimates vary, but most put the toll at somewhere between fifteen and twenty thousand dead, including the three destroyed legions, six cohorts of auxiliaries, and three cavalry units, along with camp followers traveling with the column. The precise figure cannot be established with certainty, since no ancient source gives a verified count.
\nIs Kalkriese definitely the site of the battle? Kalkriese, near Osnabruck in Lower Saxony, holds the largest body of physical evidence linked to the disaster and is widely accepted as marking at least part of the battlefield. Because the fighting stretched across three days and likely covered a broad area, some historians argue the ambush unfolded across multiple sites rather than at Kalkriese alone, and the question remains open.
\nDid Rome ever try to reconquer Germania after the defeat? Yes, briefly. Germanicus led campaigns across the Rhine between 14 and 16 AD, defeating Arminius's coalition in open battle and recovering two of the three lost legionary eagles, but Tiberius recalled him rather than hold the territory permanently. After that, the Rhine remained Rome's practical frontier for the rest of the empire's history in the West.
\nOn 2 September 1192, a truce was sworn at Jaffa that ended the Third Crusade without a final battle and without Jerusalem changing hands. Richard was too ill to attend and swore by proxy; Saladin, by custom, did not swear at all, and his relatives and emirs gave the oath in his place. The Third Crusade is remembered as a failure, and the document that closed it is the reason. That reputation rests on a single measure — the city — and the agreement itself is more interesting than the verdict.
\nBy the summer of 1192 both sides were finished. Richard had twice advanced to Beit Nuba, within about twelve miles of Jerusalem, and turned back both times without giving battle. His reasoning was consistent and is recorded on both sides: the city could probably be taken, and could not be held once the army sailed home, because holding it required a garrison in a fortress cut off from the sea by hostile country. In July his own council of Templars, Hospitallers and Syrian barons told him the same thing.
\nSaladin's position was no better. His coalition had been in the field for years, his emirs wanted to go home to their own lands, and the loss of the Egyptian caravan in June had shown how exposed his supply lines were. When Richard, already preparing to leave, was drawn back to relieve Jaffa in August and beat a much larger force in front of the town, the message was unmistakable to both: neither could finish the other.
\nRichard was seriously ill at Jaffa through late August. Negotiations that had run intermittently for a year — including the extraordinary proposal that Saladin's brother al-Adil marry Richard's sister Joanna and rule Jerusalem jointly — came to a conclusion in the first days of September.
\nThe agreement was a truce, not a peace, and it ran for three years and eight months from September 1192. Its substance was territorial and practical:
\nAscalon is the term to watch. It is the clearest evidence that this was a negotiated settlement between two exhausted powers rather than a favour granted by a victor — Saladin conceded the coast, Richard conceded the fortress that would have made the coast dangerous.
\nThe usual explanation is that Richard lacked the strength. The better one is that he lacked the afterwards. A crusading army was a temporary thing: men had taken a vow, discharged it, and went home. Whoever took Jerusalem would hold it with whatever garrison stayed behind, in a city with no port, surrounded by territory Saladin's successors would continue to control.
\nThe full record of Richard's operations makes the shape of this plain. He fought and won every engagement he commanded in person, from Limassol to the relief of Jaffa. Two of the eighteen operations are advances on Jerusalem; neither is a siege. That is not a record of a commander who was beaten in front of the objective. It is the record of a commander who declined the objective twice, on the same reasoning, and was proved right by what happened to the kingdom that came after him.
\n| Date | \nEvent | \n
|---|---|
| Sept 1191 | \nBattle of Arsuf; Richard reaches Jaffa | \n
| Dec 1191 – Jan 1192 | \nFirst advance to Beit Nuba; withdrawal without battle | \n
| Jan – May 1192 | \nAscalon rebuilt at Richard's expense | \n
| June 1192 | \nEgyptian caravan captured near Tell al-Khuwaylifa | \n
| Jun – Jul 1192 | \nSecond advance to Beit Nuba; withdrawal without battle | \n
| 27 July – 5 Aug 1192 | \nSaladin takes Jaffa; Richard's counter-attack retakes and holds it | \n
| 2 September 1192 | \nTruce sworn at Jaffa | \n
| 9 October 1192 | \nRichard sails from Acre | \n
| 4 March 1193 | \nSaladin dies at Damascus | \n
Saladin died six months after the truce was sworn, and his empire divided among his relatives almost immediately. Richard was shipwrecked on the way home, taken by Leopold of Austria, and held for over a year while England raised a ransom that has been estimated at several times the crown's annual revenue. Neither man ever returned to the question.
\nThe truce held. It was renewed and adapted by their successors, and the coastal kingdom it preserved lasted until 1291. Judged against the promise that launched the crusade — the recovery of Jerusalem — the Treaty of Jaffa is a failure. Judged against what either side could still enforce in September 1192, it is close to the best outcome available to both, which is why both could present it at home as a success.
\nDid Richard and Saladin ever meet? No. All the negotiations went through intermediaries, most often Saladin's brother al-Adil, and the two exchanged gifts and messages without ever being in the same place.
\nHow long was the truce meant to last? Three years and eight months from September 1192. It was intended as a breathing space, not a settlement, and both sides expected the war to resume.
\nWhy did Richard not visit Jerusalem when pilgrims were allowed in? The chroniclers give no single reason, and the explanations offered since — that he would not enter as a guest a city he had failed to take, or that he was still unwell and anxious to sail — are inferences rather than evidence. What is recorded is that parties of his men went and he did not.
\nWas giving up Ascalon a defeat? It was the price of keeping the coast. Richard had rebuilt it at his own cost and it was militarily valuable, which is why Saladin insisted; that both sides fought hard over this one clause is the best indication that the treaty was a genuine negotiation rather than a dictated peace.
\nImage: "Saladin the Victorious" (coloured), after Gustave Doré. Public domain, via Wikimedia Commons.
\n" }, { "id": "https://writesidesofhistory.com/the-third-generation-twenty-years-of-attacks-and-almost-no-convictions/", "url": "https://writesidesofhistory.com/the-third-generation-twenty-years-of-attacks-and-almost-no-convictions/", "title": "The Third Generation: Twenty Years of Attacks and Almost No Convictions", "summary": "From 1984 the Red Army Faction killed bankers, industrialists and officials with military precision — and almost nobody was ever convicted for it. What the unsolved cases say about the group's last phase.", "date_published": "2026-09-08T00:00:00.000Z", "image": "https://writesidesofhistory.com/images/generated/stammheim-as-a-building-a-courtroom-built-for-a-trial.webp", "tags": ["modern"], "content_html": "Between 1985 and 1991 the Rote Armee Fraktion shot or blew up a company chief executive, a research director, a senior diplomat, a banker and the head of the agency privatising East Germany's economy. Almost nobody was ever convicted for any of it. The German federal agency for civic education puts it plainly: despite an intensive search, most members of the third generation were never identified.
\nThe group that carried out these attacks was not the one that had been tried at Stammheim. Its predecessors were dead or in prison; its cadre was small, technically capable, and — this is the point — invisible in a way the first generation never was.
\nThe strategy changed with the personnel. From the early 1980s the RAF pursued what it called an internationalisation of the struggle, aiming at NATO and at the men who ran the West German economy rather than at the domestic security apparatus. The first attempt under that strategy failed: in December 1984 an attack on the NATO school at Oberammergau came to nothing.
\nWhat followed did not fail. On 1 February 1985 attackers entered the house of Ernst Zimmermann, chief executive of the engine builder MTU, at Gauting, tied up his wife, took him to the bedroom and shot him in the back of the head. On 8 August a car bomb at the Rhein-Main air base killed an American soldier, Frank Scarton, and a civilian employee, Becky Bristol, and injured twenty-three people; the identity card used to get onto the base had been taken the day before from a serviceman, Edward Pimental, who was shot for it in Wiesbaden.
\nOn 9 July 1986 a charge left at the roadside at Straßlach killed the Siemens manager Karl-Heinz Beckurts and his driver, Eckhard Groppler. On 10 October Gerold von Braunmühl, a department head at the Foreign Office, was shot outside his house in Bonn — with a weapon that, in all probability, had already been used to murder Hanns-Martin Schleyer nine years earlier. On 30 November 1989 a shaped charge at the roadside destroyed the armoured car of Alfred Herrhausen, the chairman of Deutsche Bank, and killed him instantly. On 1 April 1991 Detlev Karsten Rohwedder, president of the Treuhandanstalt, was shot through the window of his house: the last person the RAF murdered.
\nTwo attacks on officials failed. A bomb intended for Hans Tietmeyer, then state secretary in the finance ministry, missed him in 1988; an attack on Hans Neusel, state secretary in the interior ministry, failed in 1990.
\nAnd in March 1993 the group destroyed a newly built prison at Weiterstadt that had not yet received a single inmate. The damage ran to more than a hundred million marks, and it was, as Christopher Daase puts it, the RAF's last spectacular action. Nobody was killed. Five years later the group dissolved itself.
\nThree things separate this period from the 1970s.
\nThe group had learned. The first generation was caught because it lived in the open, kept documents and moved in a recognisable milieu of supporters. The third generation operated with far fewer people and far less paper.
\nThe technique changed. Remotely triggered roadside charges and long-range shooting leave less behind than a bank robbery or a kidnapping. The attacks were described at the time as having military precision, and the description was accurate in the way that mattered forensically.
\nThe evidence that did exist was contested. The Herrhausen and Rohwedder investigations both ran for years and produced no conviction for the killings themselves. Where our dataset records "no perpetrator was identified", it is recording an outcome, not a gap in the research.
\nThis is where a dataset earns its keep or fails. It would be easy to write "RAF" in an attribution column for all eight of these rows and move on. It would also be wrong: the attacks were claimed in the group's name, and the group's authorship is not seriously doubted, but that is a different claim from saying who did them.
\nOur register of RAF attacks therefore keeps three separate columns — what was claimed, whether the act was criminally attributed, and what the legal outcome was — and for this period the third column mostly says that no perpetrator was identified. That is the finding of the federal agency's own account, and it is stated affirmatively rather than as an absence, because it is the single most important fact about the group's last decade.
\nThe register also does not name suspects. A number of individuals have been publicly linked to these attacks over the years without ever being convicted of them, and naming them in a table that exists precisely to separate accusation from proof would be a self-defeating thing to do.
\nThere is a live argument about whether the third generation was the same organisation at all. One reading treats it as continuous: the same name, the same commando convention, the same targets in a different register. Another reads it as a small successor group that inherited a brand and a set of contacts and had little organisational connection to what came before.
\nThe evidence that would settle it is exactly the evidence that does not exist — the identities of the people who carried out the attacks. That is an unusual position for a subject so recent and so heavily documented, and it is the reason the third generation remains the part of the RAF's history that is argued about most and known least.
\nThe attacks, the victims and the failure of the investigations here follow the RAF dossier of the Bundeszentrale für politische Bildung — the article on the third generation (49299) and the roll of the dead (49319) — together with Christopher Daase's article in Aus Politik und Zeitgeschichte (30202) for Weiterstadt and for the argument about why the group failed. The companion piece on the Kommando statement explains how the claims themselves work as evidence. Our dataset of RAF attacks records all twenty-seven incidents with the article that carries each one.
\nOn 12 September 1631, off the shoals called the Abrolhos on the Brazilian coast, a Dutch squadron of sixteen ships attacked a Spanish and Portuguese fleet twice its number. By nightfall about three thousand men were dead, the Dutch admiral had drowned himself, and both sides had reason to call it a victory.
\nThe Dutch West India Company was chartered in 1621 as an instrument of war as much as of trade. Its target was the Iberian Atlantic — the sugar of Brazil, the silver route through the Caribbean, the slaving stations of West Africa — and its method was a fleet. By 1630 it had taken Recife and Olinda, and it held the coast of Pernambuco against a Portuguese population that had not surrendered with the towns.
\nThat is the position the Spanish fleet of 1631 was sent to change. Its commander, Antonio de Oquendo, had twenty ships of war, twelve caravels carrying troops for Pernambuco and Paraíba, and twenty-four merchantmen loaded with Bahian sugar sailing under his convoy. His orders were to make Bahia first — instructions Southey calls ill-judged, because they gave the Dutch commander Adriaen Janszoon Pater time to reach Recife, land his own troops and supplies, and put to sea in search of the enemy with sixteen good ships.
\nOquendo had the numbers; Pater had the bigger ships and the heavier guns. Before the fighting began Oquendo was advised to take the soldiers out of the transports and distribute them among his warships and larger merchantmen. He judged himself strong enough already and ordered the transports to fall to leeward — a decision that kept the troops safe and kept them out of the battle.
\nWhen the fleets met, Oquendo took the weather gage of Pater's flagship and grappled with it. What followed was the kind of engagement that decides itself by small acts. The Dutch tried to break free; a Spanish captain, Juan Costelho, got aboard the enemy and passed a cable around her foremast to hold the two ships together — a service, in Southey's phrase, that cost him his life. A second Dutch ship bore up to attack Oquendo from the other side; Cosme do Couto Barbosa ran his own small vessel between them and was sunk for it, and was picked up a prisoner. That manoeuvre appears to have saved the Spanish commander.
\nThen Pater's ship caught fire. The Spaniards kept up so continuous a discharge that the flames could not be fought. Oquendo cut the cable that held him alongside, because the Dutch ship's destruction was now certain and his own was so disabled that he had to be towed clear by Juan de Prado.
\nPater could probably have lived. The Spaniards were picking up survivors and many of his men swam to them. Instead he took his colours, wrapped them round his armour, and went over the side.
\nThe Spanish flagship's counterpart, the almiranta, engaged another Dutch ship and was sunk. At nightfall the two fleets were still in sight of one another.
\nThe answer depends on what was being contested.
\nBy the measure of the day itself, the honours were even and the losses were equal: about three thousand men in all, on both sides together. Southey notes that the Portuguese accounts relate Pater's death to the honour of their enemy and do not accuse the Dutch of misconduct, while De Laet, on the Dutch side, complains that few of the Dutch captains did their duty.
\nBy the measure of the campaign, the Iberians achieved what they came for. Oquendo detached the Conte de Bagnuolo with the relief for Pernambuco and told him to stand in for the coast and make the first port he could. The troops landed. The Dutch had failed to stop the convoy, which was the only thing that could have made the battle worth its cost to them.
\nBy the measure of the war, it changed less than either side hoped. The Company kept Recife, took Paraíba in 1634, and held Dutch Brazil for another two decades before it was lost — not to a relief fleet from Europe, but to a revolt of the Luso-Brazilian planters.
\nThis is why our record of Dutch Golden Age sea battles treats outcome as a judgement rather than a fact, and marks the certainty of each row accordingly. Abrolhos is entered as a Spanish and Portuguese victory, because the strategic object — getting the relief through — was achieved, and because the Dutch lost their admiral and his flagship. A Dutch account can be written from exactly the same facts in which sixteen ships fought thirty-two to a standstill.
\nThe record's other convention matters here too: ship and casualty counts are not tabulated. The figure of three thousand dead comes from Southey, who took it from the accounts available to him; other traditions give other numbers, and putting one of them in a column would give the disagreement a precision it does not have.
\nThe Dutch side of this battle is in De Laet's Iaerlijck Verhael, the West India Company's own annual narrative, printed in 1644. It survives as a black-letter book whose scanned text cannot be read mechanically with enough confidence to cite a page, which is why the dataset row points to Southey's History of Brazil, where the same material is set out in English with the Dutch and Portuguese versions side by side. Southey is a nineteenth-century writer working from seventeenth-century books, and he is cited here for what those books say, not as a modern authority on the Atlantic war.
\nJebel Sahaba, on the east bank of the Nile in what is now northern Sudan, has been described for over fifty years as the earliest evidence of warfare anywhere. In 2021 a team went back to the same bones, examined all of them under a microscope, and came away with a different story — not a quieter one, but a longer one.
\nThe cemetery was excavated in the mid-1960s by Fred Wendorf, as part of the salvage archaeology carried out before the Aswan High Dam flooded a stretch of the Nubian Nile. It was published in 1968 as Site 117: a burial ground of the Final Palaeolithic, with stone points found in and among the bodies, and a striking number of individuals showing injuries.
\nThe dating puts the cemetery between roughly 13,400 and 18,600 years ago, which makes it the earliest funerary complex known from the Nile valley. The collection went to the British Museum, and the interpretation went into every survey of prehistoric warfare written since: a massacre, and the oldest one on record.
\nThe 2021 re-examination in Scientific Reports did three things.
\nIt raised the count. With supernumerary bones and teeth taken into account, the site can now be regarded as holding the remains of at least sixty-four individuals — three of which are now missing from the collection. The older figure of sixty-one was a floor, not a ceiling.
\nIt found far more injuries. More than a hundred previously undocumented lesions were identified, on individuals already known to be injured and on others who had not been recognised as such. Several had lithic artefacts still embedded in the bone. Most of the trauma is consistent with projectile weapons, which in this period means arrows and darts.
\nIt found healing. This is the finding that matters. A quarter of the skeletons with lesions carry both healed and unhealed trauma — that is, injuries survived long enough for bone to remodel, and injuries that were not survived. A single massacre cannot produce that pattern. People who were shot, recovered, and were later shot again were living through something that lasted.
\nThe distinction is not academic. A massacre is an episode: a raid, a rout, a collapse of relations between two groups on a single day. Repeated raiding is a condition — a way of life in which violence recurs at intervals over years, and in which the same individuals can be wounded more than once.
\nThe Late Pleistocene Nile valley offers a plausible setting for the second reading. The river was a corridor of concentrated resources in an increasingly arid landscape, and groups whose subsistence depended on the same fishing and fowling grounds had reason to contest them repeatedly. Embedded stone and healed fractures have also been documented at the Wadi Halfa cemetery associated with the Qadan industry, which suggests the pattern is not confined to one graveyard.
\nNone of that turns Jebel Sahaba into something gentle. It turns it into something more durable, and rather more disturbing: not the memory of a single terrible day, but the ordinary background of several generations.
\nOur register of collective violence before 3000 BCE records Jebel Sahaba with a minimum number of individuals that names both figures — the sixty-one in the collection and the at-least sixty-four of the restudy, with the three now missing stated — and with a certainty of medium rather than high.
\nMedium is the honest rating. The bones are real, the injuries are real, the count is as firm as anyone can make it. What is contested is the shape of the event, and a register that quietly picked one reading would be doing the work a reader ought to be able to do themselves.
\nThe same problem recurs at the other end of the record. Nataruk in Kenya, published in Nature in 2016 as the earliest evidence of inter-group conflict among foragers, was challenged within months by specialists who argued that the cranial damage is consistent with sediment pressure after burial. Our row for Nataruk carries low certainty and says so.
\nThere is a temptation to treat a re-examination as a debunking, and this one was reported that way in places. It is better understood as what the discipline is supposed to do: the material stays, the questions change, and the second look asks things the first look could not.
\nWendorf's team worked against a rising reservoir with the methods of the 1960s, and produced a record good enough that a team fifty years later could go back to the same collection and find a hundred injuries they had missed. That is not a failure of the original work. It is the reason the original work still matters.
\nThe findings described here are from Crevecoeur and colleagues, "New insights on interpersonal violence in the Late Pleistocene based on the Nile valley cemetery of Jebel Sahaba", Scientific Reports 11 (2021), which is open access. Wendorf's original 1968 report remains the description of the excavation itself. For the comparison with a European site where the same single-event question arises, see the 2023 study of San Juan ante Portam Latinam in the same journal. Our dataset of prehistoric collective violence records thirteen sites on the same terms, each with the reference that carries it.
\nIn the first week of May 1191, a storm scattered Richard the Lionheart's fleet off the southern coast of Cyprus, and three of his ships went aground near Limassol. Within a month the island had a new ruler, and its old one was in silver chains. The conquest of Cyprus takes up a few pages in most accounts of the Third Crusade, filed under preliminaries — the detour before the real business at Acre. It deserves better. Of everything the Third Crusade achieved, Cyprus was the only territorial gain that lasted, and it outlived the crusader states on the mainland by nearly three centuries.
\nIsaac Doukas Komnenos was a great-nephew of the Byzantine emperor Manuel I. He had been captured by the Armenians in Cilicia, ransomed, and had arrived on Cyprus around 1184 with forged imperial documents appointing him governor. Once installed he stopped pretending, declared himself emperor, and held the island independent of Constantinople for seven years. A Byzantine fleet sent to recover it in 1187 was wrecked by a Norman-Sicilian squadron before it could land, and after that no one in Constantinople had the ships to try again.
\nSeven years of independent rule left almost no defenders. Neophytos the Recluse, a Cypriot hermit writing within five years of the conquest, calls Isaac's rule a calamity and describes confiscations and violence against the island's own population — and Neophytos had no reason to flatter the Latins who replaced him, since he condemns them too. That double condemnation is the most useful thing a historian has here, because both surviving Latin narratives were written to justify Richard, and a Greek source that dislikes everyone is a better check on them than either is on the other.
\nRichard had sailed from Messina on 10 April 1191 after wintering in Sicily, where his relations with the local population had already produced two armed confrontations. The fleet ran into heavy weather off Crete. Three ships were driven onto the Cypriot coast near Limassol; one of them carried Joanna, Richard's sister and the widowed queen of Sicily, and Berengaria of Navarre, whom he would marry within the month.
\nWhat Isaac did next decided everything. The survivors of the wrecked ships were, by the Latin accounts, stripped and imprisoned, and the ship carrying the two women was refused fresh water in harbour. Whether Isaac was asserting a coastal ruler's customary rights over wreck or committing an outrage depends entirely on which account is being read, and no Cypriot account of the incident survives. What is not in dispute is the effect: when Richard's main fleet arrived on 6 May, he had both a grievance and a pretext, and a hostile ruler sitting astride his supply line to Acre.
\nThe campaign that followed was short and one-sided, and the dataset behind this article records it as two operations rather than one: the landing and battle at Limassol, and the conquest of the island that followed.
\nRichard's men made an opposed landing from ships' boats at Limassol and drove Isaac's force off the beach. Isaac withdrew inland, then agreed terms, then repudiated them within a day and fled — the Latin chroniclers treat this as proof of bad faith, though a ruler with no fleet, no allies and a hostile population had few options that were not humiliating. Richard then split his forces and worked round the island, taking the coastal castles while Guy of Lusignan, the displaced king of Jerusalem who had joined him, led a second column. Isaac surrendered at the end of May at Cape Saint Andreas, on the condition — famously — that he not be put in irons. Richard had him fettered in silver.
\nThat last detail is the kind of thing that survives because it is memorable, and it should be treated with care: it appears in the Latin tradition, which had every reason to enjoy it. Its persistence in every retelling since says more about how Richard's reputation was built than about what happened at Cape Saint Andreas.
\nHe sold it. Within weeks Richard had passed Cyprus to the Knights Templar for 100,000 Saracen bezants, of which the order paid 40,000 down. The Templars garrisoned the island thinly, provoked a revolt in Nicosia at Easter 1192, put it down with a sortie that Neophytos describes as a massacre, and handed the island back as a bad investment. Richard then transferred it to Guy of Lusignan, who had lost the kingdom of Jerusalem to Conrad of Montferrat and needed somewhere to rule. The Lusignan dynasty held Cyprus until 1489.
\nThe financial logic is worth stating plainly, because it is usually skipped. Richard needed money and a base; he got both, and disposed of the administrative problem twice. Our record of his campaigns counts Sicily and Cyprus alongside Arsuf and Jaffa for exactly this reason: treating them as preliminaries makes the campaign look like a failed march on Jerusalem, when the ledger of what was actually gained and held points somewhere else entirely.
\n| Date | \nEvent | \n
|---|---|
| c. 1184 | \nIsaac Komnenos arrives in Cyprus with forged imperial letters | \n
| 1185 | \nIsaac declares himself emperor of Cyprus | \n
| 1187 | \nA Byzantine recovery fleet is destroyed before it can land | \n
| 10 April 1191 | \nRichard sails from Messina | \n
| early May 1191 | \nStorm wrecks three ships near Limassol; survivors detained | \n
| 6 May 1191 | \nRichard's fleet reaches Limassol | \n
| 11–12 May 1191 | \nRichard marries Berengaria of Navarre at Limassol | \n
| late May 1191 | \nIsaac surrenders at Cape Saint Andreas | \n
| June 1191 | \nCyprus sold to the Knights Templar | \n
| Easter 1192 | \nTemplar garrison suppresses a rising in Nicosia | \n
| 1192 | \nCyprus transferred to Guy of Lusignan | \n
Isaac Komnenos died in captivity, probably around 1195, and left almost nothing behind: no coinage tradition, no administration that survived him, no partisans. He is remembered because he was in the way of a better-documented man. That is a thin sort of immortality, and it is also why he is worth writing about — the Third Crusade is usually told through the two figures on its poster, and the island that changed hands in fifteen days did more for the Latin East than either of them managed on the mainland.
\nJerusalem was never attacked. Acre was retaken and lost again a century later. Cyprus was still a Latin kingdom when Columbus sailed.
\nWhy did Richard attack Cyprus at all? The immediate cause was Isaac's treatment of shipwrecked crusaders and his refusal of harbour rights to the ship carrying Richard's sister and betrothed. The strategic cause was that Cyprus sat on the sea route to Acre, and a hostile ruler there threatened the supply line of an army that could not feed itself from the Levantine coast.
\nWas Isaac Komnenos really a Byzantine emperor? No. He was a member of the imperial family who seized the island using forged credentials and then claimed the imperial title. Constantinople never recognised him and tried once, unsuccessfully, to remove him by force.
\nWhy did the Templars give Cyprus back? They had paid 40,000 bezants of a 100,000-bezant price and garrisoned the island with fewer than a hundred men. When the Greek population rose at Easter 1192 the order concluded it could not hold Cyprus and administer it profitably at the same time, and returned it to Richard rather than commit more resources.
\nHow long did Latin rule last? The Lusignan kingdom lasted until 1489, when Cyprus passed to Venice; the Ottomans took it in 1571. Latin rule on the island outlasted the last mainland crusader stronghold, Acre, by almost two centuries.
\nImage: "C. 1140 Asia Minor and Crusader States" by William Robert Shepherd. Public domain, via Wikimedia Commons.
\n" }, { "id": "https://writesidesofhistory.com/the-kommando-statement-how-the-raf-wrote-its-own-record/", "url": "https://writesidesofhistory.com/the-kommando-statement-how-the-raf-wrote-its-own-record/", "title": "The Kommando Statement: How the RAF Wrote Its Own Record", "summary": "Almost every Red Army Faction attack came with a claim of responsibility signed with a dead comrade's name. What those statements were for, and why they make a dataset of the group's attacks possible at all.", "date_published": "2026-09-04T00:00:00.000Z", "image": "https://writesidesofhistory.com/images/stock/memorial-stone-siegfried-buback-karlsruhe.jpg", "tags": ["modern"], "content_html": "Almost every attack the Rote Armee Fraktion carried out arrived with a text. It named the operation after a dead member of the group, explained the target in the group's own vocabulary, and was sent to a news agency. Those statements are the reason a register of RAF attacks can be compiled at all — and the reason it has to be read with care.
\nThe convention was established early and kept for twenty years. A commando was named for someone the group counted as its own dead: Kommando Petra Schelm, after the first RAF member shot by police; Kommando Thomas Weisbecker; Kommando Holger Meins, after the prisoner who died on hunger strike in 1974; Kommando Ulrike Meinhof; Kommando Siegfried Hausner, after the man who died of injuries from the Stockholm embassy attack; later Kommando Mara Cagol for an Italian, Kommando Patsy O'Hara for an Irish hunger striker, Kommando George Jackson for an American.
\nRead as a series, the names are a claim about lineage and about internationalism, and they were meant to be read that way. Each attack was presented as an answer to a specific death, and the deaths chosen from outside Germany asserted that the group belonged to a worldwide movement rather than a West German one.
\nThree things at once.
\nThey were an attribution device. A bombing without a claim is an event; a bombing with a claim is a position. The May offensive of 1972 — the United States Army's V Corps headquarters in Frankfurt on 11 May, the police headquarters at Augsburg and the Bavarian criminal police office in Munich on 12 May, a federal judge's car on 15 May, the Springer building in Hamburg on 19 May, and the headquarters of the US forces in Europe at Heidelberg on 24 May — was a sequence, and the statements are what made it legible as one.
\nThey were an argument. The texts explained targets in the group's own terms: American military installations as the war in Vietnam brought home, the Springer press as the manufacture of consent, a judge as the machinery of the state. Whether one finds the argument coherent is a separate question from whether it was made, and it was made every time.
\nThey were a discipline. A group that names its operations after its dead is telling its own members what the operations are for. The naming survived every change of personnel, which is one of the few threads running unbroken from 1972 to the 1990s.
\nThe statements create a specific problem for anyone trying to compile a record. A claim of responsibility is evidence that a claim was made. It is not evidence that the claim is true, and it is not a legal finding.
\nThat is why our register of RAF attacks keeps claimed by and criminally attributed in separate columns, and why the second column has three values rather than two: yes, partly, and no. Partly is the interesting one. It covers the attacks where a court convicted people of membership or of participation without establishing who did the specific act — which is the ordinary condition of terrorism prosecutions, and which a single "solved / unsolved" column would erase.
\nThe legal column, in turn, records only what an official account states. For the May 1972 offensive that is the Stammheim judgment: life sentences for the three surviving defendants, pronounced in April 1977. For the offensive of 1977 — the killing of the Federal Prosecutor General Siegfried Buback and his companions in April, of the Dresdner Bank chairman Jürgen Ponto in July, and the kidnapping and murder of the employers' president Hanns-Martin Schleyer from September — it is Brigitte Mohnhaupt's conviction as ringleader for nine murders and nine attempted murders, five life sentences and fifteen years.
\nTwo kinds of row do not fit the pattern, and both are marked in the dataset.
\nThe first is the attack that was never claimed. Some incidents were attributed to the group by investigators or by courts without any Kommando statement, and the claim column says so rather than inventing one.
\nThe second is the reverse, and it belongs to the group's last decade: attacks that were claimed in the group's name and never attributed to anybody. From the mid-1980s the federal agency's own account is blunt about the result — despite an intensive search, most members of the third generation were never identified. A statement arrived; nobody was ever convicted for what it described. That is the subject of the companion article on the third generation.
\nThere is an editorial line under all of this that is worth stating plainly. Quoting a terrorist group's justification at length gives it the platform it was written for; refusing to describe it at all makes the record incomprehensible. The dataset takes the narrow path: it names the commando, because that name is a fact about the attack and the way the group organised its own memory, and it does not reproduce the texts.
\nIt also names no individual who was never convicted. Victims are named, because their identity is public and is the identity of the event. Suspects are not, because a register that distinguishes claims from findings would be a poor thing if it made an accusation of its own in the process.
\nThe sequence of attacks and the legal outcomes here follow the RAF dossier of the Bundeszentrale für politische Bildung: the articles on the first generation (49256), the Stammheim trial (49264), the second generation (49268) and the clemency debate (49324). Our dataset of RAF attacks records twenty-seven incidents on these terms, each with the article that carries it.
\nOn 17 August 1606 eleven Dutch ships attacked a Portuguese armada in the Strait of Malacca. Eight days later the Dutch sailed away and the Portuguese Viceroy held the field. It is recorded here as a Portuguese victory, and it is also the moment the Portuguese position in the Strait began to fail.
\nMalacca was the hinge of the eastern trade. Whoever held it controlled the narrow water through which the spice of the Moluccas, the silk of China and the cloth of Coromandel had to pass, and the Portuguese had held it since 1511.
\nIn 1606 the Dutch admiral Cornelis Matelieff de Jonge laid siege to it in alliance with the ruler of Johor. Danvers puts the combined force at a hundred and fifty sail and sixteen thousand men, against a garrison of little more than a hundred Portuguese under André Furtado de Mendonça. The siege ran four months, and the besiegers are said to have thrown more than fifty thousand cannonballs into the city without making much impression on it.
\nWhat broke the siege was news rather than force. The Viceroy of the Estado da Índia, Dom Martim Afonso de Castro, had left Goa in May with twelve galleons, four galleys and some seventy other vessels, and he was on his way. Matelieff raised the siege and withdrew to meet him at sea.
\nThe Viceroy reached Malacca after a stormy passage in which some of his ships were lost. Shortly after his arrival, on 17 August, eleven Dutch vessels attacked him.
\nThe first day's fighting lasted until dark with about equal loss on both sides. The next morning it resumed, and the balance began to tell: the Dutch ships were the larger and the more heavily armed, and they did considerable damage to the Portuguese. Vessels on both sides burned. The running fight went on, in Danvers's account, for eight days altogether, at the end of which the Portuguese ships were very much disabled.
\nThen, on 24 August, the Viceroy offered battle again — and the Dutch declined it and sailed away.
\nBy the only test that a seventeenth-century commander would have recognised, that settles it. The side that offers battle and is not met has won the day, and the Viceroy entered Malacca with his fleet, where he found a city badly knocked about by the siege but still Portuguese.
\nThe Viceroy then divided his fleet: seven galleons under Dom Álvaro de Meneses were sent towards the Nicobar Islands to meet the ships expected from Portugal, and five more under Nuno Álvares Pereira went to escort provision ships coming from Java through the Singapore Strait.
\nThe Dutch, having refitted, learned that the armada had been split. On 22 October they came back to Malacca and fell on Pereira's five galleons alone.
\nThat is the shape of the whole contest in miniature. The Portuguese could win an engagement and still lose the sea, because their strength was a single fleet that had to be everywhere, and the Company's strength was a permanent presence that could wait, repair and return. Within a generation the Dutch held the approaches; in 1641 Malacca itself fell to them.
\nOur record of Dutch Golden Age sea battles enters this action as a Portuguese tactical victory with medium certainty, and the certainty rating is doing real work.
\nThe facts are not much disputed: the dates, the eleven Dutch ships, the eight days, the withdrawal. What is disputed is the scoring. The Dutch tradition has always read Cape Rachado as a strategic success — the blockade was resumed within weeks, the Portuguese fleet was damaged beyond quick repair, and the campaign that mattered was the one for the Strait rather than the one for that week. The Portuguese tradition reads it as a relief expedition that did exactly what it set out to do.
\nBoth readings are supportable from the same sequence of events, which is precisely why the dataset carries an outcome column and a certainty column rather than one column pretending to be a fact.
\nCape Rachado is one of four Asian actions in the record, and the reason there are only four is worth stating. The wars of the Republic in Europe were fought between states that kept navies and printed accounts of them; the wars of the Company east of the Cape were fought by a chartered corporation whose records went into its own archives and stayed there.
\nThe other three Asian rows come from Danvers's narrative of the Goa blockades: two engagements before the bar of Goa on 21 January and 11 February 1637, in which a Dutch squadron of ten vessels was met by a Portuguese fleet and the Portuguese claimed the advantage, and the destruction of a Portuguese fleet lying at anchor off Puludindin by seven Dutch ships in the same year.
\nThe narrative here follows Danvers, The Portuguese in India, volume II, pages 136 and 137, which gives the siege, the eight-day battle and the division of the fleet in sequence. Matelieff's own journal, in Peter Borschberg's scholarly edition, is the indispensable Dutch account and the place to go for the admiral's reasoning; it is a modern scholarly edition and not openly available, which is why the dataset row cites Danvers for the page a reader can check.
\nSometime around 2750 BC, potters near the Tagus estuary in what is now Portugal began shaping a distinctive inverted-bell-shaped cup, decorated in horizontal bands of cord and comb impressions, and within a few centuries variants of that same pot were being buried with the dead from Sicily to Scotland. Archaeologists call it Bell Beaker pottery, and for more than a century it has posed one of prehistory's stubbornest questions: did a drinking vessel simply become fashionable across a continent with no writing, no roads, and no political unity to spread it, or did the people who made it actually move? Ancient DNA research over the past decade has supplied a partial answer, and it is more complicated, and more interesting, than either extreme.
\nThe classic Bell Beaker is a wheel-less, hand-built ceramic vessel with a flaring rim, a rounded belly, and a narrow base, its surface covered in comb- or cord-impressed zones often filled with white paste to make the pattern stand out against the fired clay. It is usually found as a single, largely intact pot placed in an individual grave rather than the collective tombs and settlement middens of the preceding Neolithic. That shift in burial custom is itself significant: across much of Atlantic and Central Europe, communal megalithic burial, in which the bones of many generations mingled in one chamber, gave way to single burials furnished with a small, standardized kit of personal grave goods, of which the beaker itself was the most consistent element.
\nThe style first appears in Copper Age Iberia in what specialists call the Maritime tradition, and from there it moved fast. By roughly 2500 to 2400 BC, Beaker-style pottery had reached the Atlantic coast of France, southern Germany, and the Rhine basin, and by 2450 to 2300 BC it had crossed to the British Isles, the Low Countries, and southern Scandinavia, a Channel crossing made by boat, since the land bridge of Doggerland had already vanished beneath the North Sea. In under three centuries a specific decorative grammar had become recognizable from the Algarve to the Orkney Islands, a reach unmatched by almost any other Copper or early Bronze Age artifact type.
\nBell Beaker pottery rarely traveled alone. Graves across the culture's range share a recurring set of objects archaeologists call the "Beaker package": the beaker itself, a stone wrist-guard (worn to protect the forearm from a bowstring), barbed-and-tanged flint arrowheads, small tanged copper knives, and V-perforated bone or shale buttons for fastening clothing. Where metal was available, Beaker burials are some of the earliest clear evidence of individual metalworking skill in prehistoric Europe, marking the tail end of the Copper Age in many regions.
\nBeaker Europe was not the only part of the continent changing fast in these centuries; further east, the giant proto-city settlements of the Trypillia culture had already been abandoned, a reminder that Copper Age Europe was a patchwork of trajectories. This reach across thousands of kilometers is what made the Beaker phenomenon so puzzling: a shared taste in drinking vessels does not, by itself, prove that anyone moved. For much of the twentieth century, archaeologists split over whether to imagine "Beaker folk" migrating outward from Iberia, or whether the pottery and its status symbols spread the way any prestige fashion spreads today, carried by trade, marriage, and imitation rather than by settlers.
\n| Date (BC) | \nDevelopment | \n
|---|---|
| c. 2750–2700 | \nMaritime-style Bell Beaker pottery appears in the Tagus estuary region of Portugal | \n
| c. 2500–2400 | \nStyle spreads along the Atlantic façade into France and the Central European "Eastern Group" | \n
| c. 2450–2300 | \nBell Beaker pottery reaches the British Isles, the Low Countries, and Denmark | \n
| c. 2400–2300 | \nBurial of the Amesbury Archer near Stonehenge, southern England | \n
| c. 2200–1800 | \nBell Beaker as a distinct pottery style fades out across Europe, giving way fully to early Bronze Age traditions | \n
| 2018 | \nGenome-wide ancient DNA study documents large-scale population change accompanying the Beaker complex in Britain | \n
No single burial illustrates the debate's stakes better than the Amesbury Archer, found in 2002 during housing development work near Stonehenge and excavated by Wessex Archaeology. Dated to roughly 2400 to 2300 BC, the grave held one of the richest sets of Beaker-period grave goods ever found in Europe: five beaker pots, three copper knives, sixteen barbed flint arrowheads, a pair of stone wrist-guards, a shale belt ring, boar tusks, and a kit of flint-knapping and metalworking tools, including cushion stones used as a portable anvil, which led excavators to identify him as an early metalworker. Isotope analysis of his tooth enamel suggested he grew up on the continent, possibly the Alpine region, before spending his adult life in Britain, making him one of the clearest examples of the long-distance movement the Beaker complex has long been suspected of carrying. His grave sits only a few centuries after the main construction phases at Stonehenge, in a landscape whose earlier monument-building traditions echo, in spirit if not method, the collective ambition seen at sites like Gobekli Tepe millennia earlier.
\nEarly twentieth-century archaeologists, working in a culture-historical tradition that equated a distinct artifact style with a distinct migrating "people," were comfortable speaking of an invading Beaker folk. By the 1970s and 1980s, processual archaeologists such as Colin Renfrew pushed back hard against that model, arguing that a shared drinking vessel and its prestige goods could spread as an ideology among existing populations without any large-scale movement of people. For a generation, "pots, not people" became something close to orthodoxy in Beaker studies, and the older migrationist reading was treated as a relic of an earlier, more credulous archaeology. Ancient DNA has since forced a more nuanced middle position, and the evidence differs sharply by region.
\nA landmark 2018 study in Nature, based on genome-wide data from 170 Neolithic, Copper Age, and Bronze Age individuals, found strikingly little genetic connection between Beaker-associated people in Iberia and those in Central Europe, despite their sharing the same pottery style and grave-good package. In Iberia, the population that adopted Beaker pottery around 2750 BC stayed genetically continuous with earlier Copper Age Iberians, much as the isolated Alpine farming community behind Otzi the Iceman shows almost no steppe ancestry a few centuries earlier, supporting the idea that the style itself, not migrants, arrived in Iberia first. In Central Europe, by contrast, Beaker individuals already carried substantial "steppe" ancestry inherited from the Corded Ware culture, itself descended from earlier migrations out of the Pontic-Caspian steppe.
\nThe clearest case of migration came from Britain. The same study found that the arrival of the Beaker complex there coincided with the replacement of around 90 percent of the pre-existing British gene pool within a few hundred years, one of the most dramatic population turnovers documented anywhere in ancient European genetics. Migrants carrying steppe-related ancestry, equipped with the Beaker package, did not simply introduce a new pottery fashion to Neolithic Britain; within a handful of generations they became the overwhelming majority of the ancestors of the people living there. Researchers caution that sample sizes remain modest, and the exact mechanism of that replacement is still debated.
\nThat regional contrast is the real lesson of Beaker research: neither "it was all migration" nor "it was all an idea" survives contact with the genetic evidence. The same pot traveled by two different means depending on where it landed, carried by traders and imitators along the Atlantic coast, and carried in the bodies, and eventually the majority ancestry, of migrants who crossed into Britain.
\nBell Beaker pottery endures as one of prehistoric archaeology's best case studies in how much can ride on a single, humble artifact type. A cup that fits in one hand became the thread connecting Copper Age Iberia, Corded Ware Central Europe, and Bronze Age Britain, and following that thread with radiocarbon dates, isotopes, and now whole genomes has repeatedly overturned assumptions about how ideas, objects, and people moved through a Europe with no writing to record any of it. The debate is not closed; regional chronologies are still being refined, and the mechanisms behind Britain's turnover remain contested. But the outline is clearer than at any point since the style was named: sometimes a shared fashion really was just a fashion, and sometimes it was a migration wearing a fashion's clothes.
\nWhat is Bell Beaker pottery? It is a distinctive style of hand-built, inverted-bell-shaped drinking vessel, decorated with cord- or comb-impressed bands, that spread across Copper Age and early Bronze Age Europe between roughly 2750 and 1800 BC. It is usually found in individual graves alongside a recurring set of grave goods known as the Beaker package.
\nDid the Beaker culture involve migration, or just a shared fashion? Both, depending on region. Ancient DNA shows Iberia's population stayed largely continuous even as it adopted Beaker pottery, suggesting the style spread there as an idea, while Britain saw roughly 90 percent of its gene pool replaced within a few centuries of the complex's arrival, a clear sign of migration.
\nWho was the Amesbury Archer? A man buried around 2400 to 2300 BC near Stonehenge with one of the richest Bell Beaker grave assemblages ever found, including five pots, copper knives, and metalworking tools. Isotope evidence from his teeth suggests he grew up on the continent before living out his adult life in Britain.
\nWhy did archaeologists stop believing in a Beaker migration for a while? In the mid-to-late twentieth century, processual archaeologists reacted against culture-historical models that too readily equated pottery styles with migrating peoples, arguing instead that Beaker pottery spread as a prestige fashion among existing populations. Ancient DNA research since 2015 has shown that migration did play a major, if regionally uneven, role after all.
\nIn 1983 a farmer near Talheim in Baden-Württemberg found human bone while digging. What came out of the ground was a pit containing thirty-four people — men, women and children — killed with axes and adzes and thrown in without arrangement. For a generation of prehistorians the find was decisive, because it arrived in the middle of an argument about whether farming societies in Neolithic Europe fought at all.
\nThe bodies were not laid out. In a period whose burials are otherwise careful — a body on its side, knees drawn up, a pot and a few tools at hand — the Talheim dead were dumped. The injuries are concentrated on the head, and they are the injuries a stone adze or an axe makes: rounded or wedge-shaped depressions in the vault of the skull, most of them at the back.
\nThat last detail carries most of the argument. Wounds to the back of the head are not what a face-to-face fight produces. They are what killing produces, when the person struck is not resisting, or is running, or is already down.
\nThe minimum number of individuals is thirty-four. That figure comes from the excavation report, and it is repeated in the comparative literature — the 2023 restudy of San Juan ante Portam Latinam tabulates the three known Linearbandkeramik massacre sites together, and gives Talheim thirty-four, Asparn/Schletz sixty-seven and Schöneck-Kilianstädten twenty-six.
\nUntil the 1980s the standard account of the European Neolithic ran roughly like this: farming arrived, populations grew, land was plentiful, and organised violence belonged to the later world of chiefdoms, metal and fortification. Prehistoric warfare was treated as a late development, and the absence of evidence was read as evidence of absence.
\nTalheim broke that reading, because it could not be explained away. A single pit with thirty-four bodies, all bearing the same class of injury, deposited at once, is not an accumulation of accidents. And the Linearbandkeramik — the first farming culture of Central Europe, recognisable by its incised pottery — was supposed to be exactly the sort of society in which this did not happen.
\nTwo more sites followed. At Asparn/Schletz in Lower Austria the dead lay unburied in the ditch of an enclosure, with carnivore gnawing and partial disarticulation showing that nobody came back to bury them. At Schöneck-Kilianstädten, excavated much later and published in 2015, twenty-six people were killed by blunt force and arrow injuries and deposited in a commingled mass grave. The excavators of that site noted something that recurs at all three: young women are missing from the dead, which is usually read as their having been taken.
\nIt is worth being precise about the limits, because Talheim is often made to carry more than it can.
\nIt does not date a war. A single event leaves a single deposit. Whether the killing at Talheim belongs to a long conflict, a raid, or a one-off collapse of relations between two settlements is not recoverable from the bones.
\nIt does not establish who did it. Isotope work on the teeth of the Talheim dead has been used to argue about where the victims grew up, not about the attackers, who left nothing behind but their weapon signatures.
\nIt does not make massacre the norm. The Halberstadt grave, published in 2018, contains nine individuals, seven of them male, with cranial injuries clustered so tightly at the back of the head that the excavators read them as executions rather than as a fight — captives killed in a controlled way, not a community wiped out. Two very different events, three hundred kilometres apart, in the same culture.
\nThat is the honest position: the Linearbandkeramik ended in a period that included at least three destroyed communities and at least one set of executions, and the relationship between those events is not settled.
\nThe word that does the work in all of this is perimortem — an injury inflicted around the time of death, distinguished from an older healed injury by the absence of healing, and from later damage by the way fresh bone breaks. It is the hardest judgement in the discipline, and it is where reasonable specialists disagree.
\nThat is why our register of collective violence before 3000 BCE carries a certainty column rather than a simple list of massacres. Talheim is rated high: the count, the injuries and the single deposition are all settled in the literature. Herxheim, where the same culture's remains were processed in ways that some read as cannibalism and others as a secondary mortuary rite, is rated low, and the row says why.
\nSet beside the rest of the record, Talheim is neither the largest nor the oldest. Jebel Sahaba in Sudan is far older and, on the 2021 restudy, records repeated raiding rather than one event. San Juan ante Portam Latinam in Spain holds the remains of more than three hundred people. Potočani in Croatia has forty-one, killed in a single episode around 4200 BC.
\nWhat Talheim has is clarity. Thirty-four people, one pit, one class of injury, one moment. In a discipline that argues about almost everything, that combination is rare enough to have become the reference point against which every other contested site is measured — including the ones this register rates as uncertain.
\nThe Talheim count and the comparison with the other Linearbandkeramik sites are taken from the 2023 study of San Juan ante Portam Latinam in Scientific Reports, whose table 3 sets the three sites side by side. The Kilianstädten excavation report in PNAS (2015) is the fullest account of a third massacre site and of the missing young women; the Halberstadt paper in Nature Communications (2018) is the counter-case. Our own dataset of prehistoric collective violence records each site with its minimum number of individuals, the trauma reported and the reference that carries it.
\nOn the morning of March 25, 1957, delegations from six European states climbed the steps of the Capitoline Hill in Rome to sign two treaties that neither the assembled photographers nor most of the signatories themselves fully expected to reshape a continent. In the Sala degli Orazi e Curiazi of the Palazzo dei Conservatori, ministers from Belgium, France, West Germany, Italy, Luxembourg, and the Netherlands signed instruments creating the European Economic Community (EEC) and the European Atomic Energy Community (Euratom). What began as a technical arrangement to remove tariffs and coordinate nuclear research became, within a generation, the legal foundation of the European Union. The Treaty of Rome did not create a superstate, and it did not end national rivalry in Europe. It did something narrower and more durable: it bound six economies together through law, making withdrawal too costly for any of them to attempt.
\nThe road to Rome began with a collapse. In August 1954, the French National Assembly refused to ratify the European Defence Community, a plan for a joint European army that would have folded a rearmed West Germany into a supranational force rather than a national one. The vote effectively killed the idea of political and military union as the next step after the 1951 Treaty of Paris, which had created the European Coal and Steel Community. For advocates of integration, the defeat was a serious setback, but not a fatal one. If defence and high politics were too sensitive to pool, economics might not be.
\nThat calculation drove the foreign ministers of the ECSC's six member states to meet at Messina, Sicily, from June 1 to 3, 1955. There they agreed to explore a general common market and a joint organization for atomic energy, and they appointed Belgian Foreign Minister Paul-Henri Spaak to chair an intergovernmental committee that would turn the idea into a workable plan. The shift from a defence community to a customs union was itself telling: governments already accustomed by Marshall Plan aid to coordinating economic policy simply changed terrain rather than giving up on unity.
\nSpaak's committee worked for roughly nine months, and its report, approved by the six governments on April 21, 1956, became the blueprint for what followed. Rather than integrating industries one sector at a time, as the ECSC had done with coal and steel, the report recommended a horizontal approach: a general customs union with a common external tariff, built up gradually so that no single national economy would be shocked into collapse. The Venice conference of foreign ministers endorsed the report in late May 1956 and authorized full treaty negotiations, which opened at the Château de Val Duchesse outside Brussels.
\nThose negotiations were not a formality. French negotiators, wary of exposing a protected economy to German and Benelux competition, pressed for safeguard clauses, a transitional timetable, and the inclusion of France's overseas territories in an association arrangement preserving colonial-era trade links. France also tied its support for the common market to a parallel Euratom treaty, since planners hoped pooled research would deliver nuclear technology for both civilian power and an eventual independent arsenal. West Germany's Konrad Adenauer treated economic integration as the surest path to full political rehabilitation for a state only a decade removed from total defeat. Britain, invited to the Messina talks, chose not to join, preferring to keep control of its own trade policy and Commonwealth ties — a decision that shaped the Six's early years and foreshadowed the long dispute over British membership that would follow.
\nBy March 1957 the text was settled, and the six governments arranged a signing ceremony designed to carry symbolic weight: Rome, the city that had once ruled the same territories the new Community now spanned. On March 25, in the Palazzo dei Conservatori, Christian Pineau and Maurice Faure signed for France, Konrad Adenauer and Walter Hallstein for West Germany, Antonio Segni and Gaetano Martino for Italy, Paul-Henri Spaak and Jean-Charles Snoy et d'Oppuers for Belgium, Joseph Bech and Lambert Schaus for Luxembourg, and Joseph Luns and Johannes Linthorst Homan for the Netherlands. Two separate treaties were signed that day, one founding the EEC and the other Euratom, though both are conventionally remembered together as the Treaty of Rome. Ratification followed through the winter, and the treaties entered into force on January 1, 1958.
\nThe EEC treaty committed its six members to a customs union and a common market built around the free movement of goods, services, capital, and labor — the "four freedoms" that remain the legal core of the European Union today. Article 8 set a transitional period of twelve years, divided into three four-year stages, during which internal tariffs would be phased out and a common external tariff phased in. In practice the customs union was completed faster than planned: internal tariffs and quotas were eliminated by July 1, 1968, eighteen months ahead of schedule. The treaty also set out objectives for a future common agricultural policy, though the detailed mechanism — price supports and market intervention — was not agreed until 1962.
\nThe treaty established a Commission to propose and administer policy, a Council of Ministers to decide, a Common Assembly (forerunner of the European Parliament) to provide oversight, and a Court of Justice to settle disputes between members. Walter Hallstein, the West German jurist who had led his country's delegation at Val Duchesse, became the Commission's first president in January 1958 and spent nearly a decade pushing the young institution to act with more independence than some member governments were comfortable granting it. That tension between supranational initiative and national control resurfaced decades later, for instance when the Commission's first attempts at computerized governance ran into resistance from member states guarding their own administrative sovereignty. The Euratom treaty pooled research funding and nuclear materials among the Six, but never delivered the joint reactor program French sponsors wanted, as national programs — especially France's own weapons effort — diverged from its cooperative aims.
\n| Date | \nEvent | \n
|---|---|
| August 30, 1954 | \nThe French National Assembly declines to ratify the European Defence Community, ending plans for a joint European army. | \n
| June 1–3, 1955 | \nForeign ministers of the ECSC's six members meet at Messina and agree to explore a common market and an atomic energy community. | \n
| April 21, 1956 | \nThe Spaak Report is approved, setting out the blueprint for a general customs union built up in stages. | \n
| Late May 1956 | \nForeign ministers meeting in Venice endorse the Spaak Report and open formal treaty negotiations at Val Duchesse. | \n
| March 25, 1957 | \nThe Treaties of Rome establishing the EEC and Euratom are signed at the Palazzo dei Conservatori. | \n
| January 1, 1958 | \nThe treaties enter into force; Walter Hallstein becomes the first President of the EEC Commission. | \n
| July 1, 1968 | \nInternal tariffs and quotas among the Six are eliminated, completing the customs union eighteen months ahead of the treaty's own schedule. | \n
The Treaty of Rome was not written as a founding constitution, and its authors made no claim to have ended the nation-state in Europe. Some historians, following Alan Milward, read it instead as six governments using shared institutions to strengthen their own states and economies, not as a first step toward a federal Europe dissolving national sovereignty. Whichever reading one prefers, the treaty's practical effect was to lock France, West Germany, Italy, and the Benelux countries into a legal relationship that outlasted every government that signed it. The customs union it created was completed ahead of schedule; the Commission it established grew into a genuine executive; and the "four freedoms" it wrote into treaty law became the template against which every later step in integration, from Britain's 1973 accession to the passport-free travel later built by the Schengen system, would be measured. Six delegations on the Capitoline Hill could not have promised the European Union that followed, but they wrote the clauses that made it possible.
\nWhy is it called the "Treaty of Rome" if two treaties were signed? On March 25, 1957, the same six governments signed two separate instruments in the same ceremony: one establishing the European Economic Community and one establishing the European Atomic Energy Community, or Euratom. Because they were signed together in Rome, both are conventionally referred to as the Treaty of Rome, though lawyers and historians sometimes distinguish them as the "Rome Treaties" to be precise.
\nWhy didn't Britain join the EEC from the start? British policymakers were invited to the Messina talks but declined to join the negotiations that followed, preferring to protect their trade policy independence and their trading relationships within the Commonwealth rather than commit to a customs union and common external tariff. That choice left European integration to the original Six and set up Britain's later, repeatedly frustrated attempts to join, which did not succeed until 1973.
\nDid the Treaty of Rome create the euro or the European Parliament? No. The 1957 treaty established a Common Assembly with advisory and oversight powers, which only gained direct elections in 1979 and the name European Parliament later still; a single currency was not seriously negotiated until the Maastricht Treaty of 1992, decades after Rome.
\nWhy did France insist on the Euratom treaty alongside the common market? French negotiators made the customs union conditional on a parallel atomic energy treaty because French planners wanted pooled European resources to support both a civilian nuclear power program and, eventually, an independent French nuclear weapons capability. In practice Euratom's joint research ambitions faded quickly as national nuclear programs, particularly France's own, diverged from the treaty's cooperative framework.
\nOn 14 July 1789, several hundred Parisians armed with muskets looted from the Invalides the day before turned on the Bastille, a royal fortress-prison in eastern Paris, and after four hours of fighting that killed roughly a hundred of them, broke inside and freed the seven prisoners they found there. The fortress held almost nothing of what the crowd expected: no torture chambers, no mass of political victims, just four forgers, two men held at their own families' request, and an elderly would-be assassin of Louis XV. What the Bastille held mattered less than what it stood for. Within a generation, its fall was celebrated as the birth of French liberty, and within a century, 14 July had become the republic's national holiday.
\nThe Bastille was never designed as a prison. Built between 1370 and 1383 to defend Paris's eastern gate during the Hundred Years' War, it was an eight-towered stone fortress with walls nearly a hundred feet high, ringed by a moat and armed with cannon. Only in the seventeenth century, under Cardinal Richelieu and later Louis XIV, did the crown convert it into a state prison for holding people by royal warrant, the notorious lettres de cachet that let a minister or an angry father have someone confined without trial. Voltaire spent time there, twice. By the 1780s the Bastille was an anachronism, expensive to garrison and holding only a handful of inmates; some historians argue the crown might have closed it within a few years regardless.
\nWhat survived was its reputation. Decades of pamphlets and prison memoirs had turned the Bastille into shorthand for arbitrary royal power, a building whose thick walls supposedly hid the crown's enemies indefinitely. That imagined Bastille, built by the printing press as much as by medieval masons, is what the crowd marched on in July 1789.
\nFrance entered the summer of 1789 in genuine crisis. A poor 1788 harvest had driven bread prices to their highest levels in decades, royal finances were exhausted by war debt, and the Estates-General, summoned for the first time since 1614, had spent weeks deadlocked over how the three orders should vote. In late June the Third Estate broke away, declared itself the National Assembly, and swore not to disperse until France had a constitution. Troop movements around Paris and Versailles in early July convinced many Parisians the king was preparing to dissolve the Assembly by force.
\nThe spark came on 11 July, when Louis XVI dismissed Jacques Necker, his finance minister and, to much of Paris, the one royal official who seemed to sympathize with reform. News reached the capital the next day and set off open unrest, with crowds clashing with royal troops near the Palais-Royal. On 13 July a mob broke into the Invalides and seized roughly 30,000 muskets, along with cannon, but almost no gunpowder. The obvious place to find powder was the Bastille, which stored several tons for the royal arsenal. By the morning of 14 July, thousands of armed Parisians and mutinous French Guards had gathered outside its walls, demanding the powder and the withdrawal of cannon they believed were aimed at the neighborhood.
\nThe Bastille's governor, Bernard-René Jourdan de Launay, had a garrison of only around a hundred men, mostly elderly Invalides veterans reinforced by a detachment of Swiss troops, against a crowd that grew through the morning into the thousands. Negotiations over the powder stalled, and around one in the afternoon fighting broke out after a delegation crossed into an outer courtyard and the defenders, fearing an assault, opened fire. The battle lasted roughly four hours. Protected by the walls and a raised drawbridge, the garrison inflicted far heavier losses than it suffered: ninety-eight attackers were killed against a single defender, a disparity that shows how well the old fortress still worked defensively even against a force many times its size.
\nDe Launay briefly considered blowing up the powder magazine rather than surrender, which would have leveled much of the neighborhood, but his own officers talked him out of it. Around five in the afternoon, with no relief coming from Versailles, he lowered the drawbridge. The crowd poured in, freed the seven prisoners, and turned on the garrison and its commander. De Launay was marched toward the Hôtel de Ville, beaten along the way, and killed before he arrived; his head was cut off and paraded through Paris on a pike, as was that of Jacques de Flesselles, the city's provost of merchants, accused of sabotaging the crowd's efforts to arm itself.
\n| Date (1789) | \nEvent | \n
|---|---|
| 5 May | \nThe Estates-General opens at Versailles for the first time since 1614 | \n
| 17 June | \nThe Third Estate declares itself the National Assembly | \n
| 11 July | \nLouis XVI dismisses finance minister Jacques Necker | \n
| 12–13 July | \nRiots spread through Paris; a crowd seizes muskets from the Invalides | \n
| 14 July, morning | \nArmed Parisians gather outside the Bastille demanding gunpowder | \n
| 14 July, afternoon | \nFighting breaks out; after roughly four hours, Governor de Launay surrenders | \n
| 14 July, evening | \nDe Launay and provost Jacques de Flesselles are killed by the crowd | \n
| 16 July | \nLouis XVI reinstates Necker and withdraws troops from Paris | \n
| 17 July | \nDemolition contractor Pierre-François Palloy begins tearing down the fortress | \n
| 1880 | \nThe French Republic formally adopts 14 July as its national holiday | \n
The most striking fact about the Bastille's fall is how little it actually liberated. Records list exactly seven prisoners inside on 14 July: four men convicted of forging financial documents, two confined at the request of their own families, and the elderly Comte de Solages, held on a warrant obtained by his relatives. There was no crowd of political dissidents chained in dungeons, no torture apparatus for the mob to discover, though rumor soon supplied one anyway. Some historians argue the Bastille's real significance lay less in what the crowd found than in what the assault proved possible: that armed Parisians could defeat a royal garrison and that the king's authority in his own capital was no longer absolute.
\nThat gap between the modest reality and the enormous symbol did not go unnoticed at the time. Gouverneur Morris, the American merchant then living in Paris, wrote skeptically in his diary about the disorder and violence he witnessed around the fall of the Bastille, a reminder that eyewitnesses did not uniformly read the event as a triumph of liberty. Demolition began within days under the contractor Pierre-François Palloy, who turned stones from the ruins into medallions and miniature Bastilles sold across France, manufacturing the fortress's afterlife as a revolutionary relic even as its walls came down.
\nThe fall of the Bastille did not by itself create a new government, but it settled, for the moment, whether the crown could reverse the political changes of the preceding weeks by force. Louis XVI traveled to Paris on 17 July, wore the revolutionary cockade, and effectively recognized the new order in the capital. The National Assembly, no longer facing military dissolution, continued drafting what became the Declaration of the Rights of Man and of the Citizen weeks later. News of the fall also helped trigger the rural panic historians call the Great Fear, which spread through the countryside in late July and fed directly into the Assembly's decision to abolish feudal privileges on the night of 4 August.
\nParisian crowd politics of this kind had precedent in earlier confrontations between subjects and unpopular rule, from the largely bloodless deposition behind England's Glorious Revolution of 1688 to the pamphlet wars that had already been shaping French opinion for years, as detailed in accounts of how the crown had once tried to control political messaging through the printing press. Paris would see the same pattern of crowd action against fortified state power recur, on a bloodier scale, in the Paris Commune of 1871.
\nThe Bastille that fell on 14 July 1789 was already a fading institution, garrisoned by old soldiers and holding a handful of forgers and family embarrassments rather than the political prisoners of legend. What made the day revolutionary was not what the crowd found inside, but the demonstration that royal authority in Paris could be broken by force and that the king would, for now, accept it. The Bourbon monarchy that had built its Versailles around rituals of control and commissioned generations of court portraiture to project unshakeable majesty could not hold a fortress with a hundred veterans against several thousand armed Parisians. The stones went to Palloy's workshops, and 14 July became the date the French Republic would eventually choose to mark the beginning of its modern political life.
\nHow many prisoners were actually freed from the Bastille? Seven: four men held for forgery, two confined at their own families' request, and an elderly nobleman detained on a warrant obtained by relatives. None were the political prisoners the crowd's rhetoric had led many to expect.
\nWhy did the crowd attack the Bastille specifically? The fortress stored a large supply of gunpowder the crowd urgently needed, having already seized thousands of muskets from the Invalides the previous day. Its reputation as a symbol of arbitrary royal imprisonment made it a natural target.
\nWhat happened to the Bastille's governor after he surrendered? Bernard-René de Launay was marched toward the Hôtel de Ville, beaten along the way, and killed before he arrived. His head was carried through Paris on a pike, along with that of the city's provost of merchants, Jacques de Flesselles.
\nWhy did 14 July become France's national holiday? It offered a single, dramatic day that later republican governments could celebrate without endorsing the Revolution's more divisive later phases. The French Republic formally adopted it as the national holiday in 1880.
\nOn 25 September 1066, an English army that had covered nearly 200 miles in under a week fell on a Norwegian invasion force outside a small Yorkshire crossing, killed a king, and ended the last serious Scandinavian bid to rule England - only to march south again within days to lose everything at Hastings. The Battle of Stamford Bridge rarely gets top billing in the story of 1066; it is remembered, if at all, as the warm-up act before William the Conqueror's landing. But on its own terms it was one of the most complete tactical victories of the English Middle Ages, and the exhaustion it left behind may have cost Harold Godwinson his crown three weeks later.
\nWhen Edward the Confessor died childless in January 1066, the English throne passed to Harold Godwinson, the most powerful nobleman in the kingdom. His claim was immediately contested from two directions. Duke William of Normandy asserted that Edward had promised him the succession years earlier, and began assembling an invasion fleet on the Channel coast. Far to the north, King Harald Sigurdsson of Norway - known to history as Harald Hardrada, "the Hard Ruler" - had his own claim, inherited from an earlier agreement between Norwegian and Danish kings, and a formidable reputation as a mercenary commander who had fought from Sicily to Constantinople.
\nHarold Godwinson spent the summer of 1066 watching the Channel, where William's army sat waiting for a favourable wind. It was the wrong coast to watch. In early September, Harald Hardrada's fleet - later Norse tradition, admittedly written generations afterward, put it at some 200 ships, with other estimates running higher - sailed down the English coast and up the Humber estuary toward York, a city and hinterland still marked by two centuries of earlier Scandinavian settlement.
\nHardrada did not sail alone. Waiting to join him was Tostig Godwinson, Harold's own younger brother. Tostig had been Earl of Northumbria until a rebellion against his harsh rule drove him out in 1065; when Harold, as king, upheld the Northumbrian rebels' demands rather than restore his brother, Tostig went into exile nursing a grievance that curdled into treason. He tried and failed to raise support from the King of Denmark before finding a willing partner in Hardrada, offering local knowledge and a claim to legitimacy in exchange for help retaking his earldom.
\nThe alliance made for an unusual invasion: a Norwegian king pursuing an old dynastic claim, reinforced by an English earl settling a family score. It also meant the invaders knew exactly where England's northern defenses were weakest.
\nThe local earls, Edwin of Mercia and Morcar of Northumbria, gathered what forces they could and met the invaders at Fulford, just outside York, on 20 September. The battle was a disaster for the English earls; their army was broken and York's citizens, with no relief in sight, agreed to surrender the city and provide hostages and provisions to the Norwegian-led force. It looked, for a few days, as though northern England had simply changed hands.
\nNews of Fulford reached Harold in the south while he was still watching for William. He responded with a march that historians still treat as a logistical feat: gathering troops as he went, his army is generally reckoned to have covered close to 185 miles from London to Yorkshire in well under a week, an extraordinary pace for an eleventh-century army moving on foot and horseback. It was a very different kind of mobility from the mounted knights William the Conqueror would field weeks later; Harold's strength was speed on the march rather than shock combat, closer in spirit to the traditions of horse-breeding and campaign logistics that shaped medieval European warfare more broadly.
\nHarold's speed caught the Norwegians completely off guard. Confident that no relief force could arrive so quickly, and with York already surrendered, Hardrada and Tostig had left a substantial part of their men - and much of their armor - back at the ships moored at Riccall, several miles from the agreed hostage-exchange point at Stamford Bridge, a crossing of the River Derwent about eight miles east of York.
\nOn the morning of 25 September, Harold's army appeared on the road from York, catching the Norwegian force strung out and lightly armored on both banks of the Derwent. According to later English tradition, a single Norwegian warrior held the narrow bridge alone long enough for his comrades on the eastern bank to form a shield wall, before an English soldier floated a boat or barrel beneath the bridge and speared him from below - a story vivid enough to have become the battle's defining image, though it appears in the English chronicle tradition rather than in Snorri Sturluson's saga of Harald's life.
\nOnce the crossing was forced, the fighting became a grinding, hours-long struggle. Harald Hardrada, fighting in the front rank without the mail he had left at the ships, was killed by an arrow to the throat. Tostig took command of the Norwegian line and, when Harold reportedly offered him peace and his earldom back if he would abandon the fight, refused. The battle resumed and Tostig, too, was killed. A relief force under Eystein Orre, who had force-marched from the ships on hearing the news, arrived late in the day for a last, desperate assault - remembered in Norse tradition as "Orre's Storm" - but it broke against the English line as well.
\nThe scale of the defeat is best measured by the ships. Of the several hundred vessels that had carried Hardrada's army to England, only a small fraction - traditionally put at around two dozen - were needed to carry the survivors home. It was, by any measure, the end of Norway's ambitions to rule England, and it is often described, with the usual caveats that attach to symbolic turning points, as marking the close of the Viking Age of large-scale Scandinavian conquest in Western Europe.
\nHarold had no time to enjoy the victory. Word reached him within days that William had landed at Pevensey on the south coast on 28 September, while Harold's army was still in Yorkshire. He marched his exhausted, depleted force back south, covering the distance to London in roughly a week and pressing on to meet William near Hastings on 14 October without waiting for reinforcements from the earls who had lost at Fulford. Whether fatigue from Stamford Bridge decided the outcome at Hastings is a question historians still debate rather than settle; what is not disputed is that the same army fought two major battles against two different invaders inside a single month, an ordeal that would strain any medieval force. The crown that Stamford Bridge had briefly secured for Harold would pass out of English hands entirely within weeks - and the vulnerability of an English throne to invasion by a rival with an old dynastic claim was a lesson the kingdom would relearn more than once, down to the battlefield deaths that ended the Wars of the Roses at Bosworth Field four centuries later.
\n| Date (1066) | \nEvent | \n
|---|---|
| Early September | \nHarald Hardrada's fleet sails from Norway and joins Tostig Godwinson | \n
| 20 September | \nNorwegian-led force defeats Earls Edwin and Morcar at the Battle of Fulford | \n
| c. 24 September | \nYork surrenders; Harold's army, having marched from the south, reaches Yorkshire | \n
| 25 September | \nHarold's army destroys the Norwegian force at Stamford Bridge; Hardrada and Tostig are killed | \n
| 28 September | \nWilliam of Normandy lands at Pevensey on the south coast | \n
| 14 October | \nHarold's exhausted army is defeated by William at the Battle of Hastings | \n
Stamford Bridge deserves to be remembered on its own terms, not just as a prelude. It was a rare instance of an Anglo-Saxon army out-marching and out-fighting a professional Scandinavian force on ground of its own choosing, and it broke, in a single afternoon, an invasion that had already taken York. That it is remembered mainly as the battle before Hastings says more about the scale of what followed than about the battle itself: England's crown changed hands twice within three weeks in the autumn of 1066, and the first of those changes was a genuine English triumph before it became, in hindsight, the opening act of a national catastrophe.
\nWhy did Tostig Godwinson fight against his own brother? Tostig had been driven out as Earl of Northumbria in 1065 after a rebellion against his rule, and King Harold sided with the rebels rather than restore him. Exiled and embittered, Tostig sought allies willing to invade England on his behalf, eventually finding one in Harald Hardrada of Norway.
\nHow large was Harald Hardrada's invasion fleet? Later Norse saga tradition put the fleet at around 200 ships, though other estimates run higher; precise figures are disputed because the surviving accounts were written well after the event. Whatever the exact count, it was large enough to defeat a regional English army at Fulford and occupy York before Harold arrived.
\nDid the victory at Stamford Bridge cost Harold the crown at Hastings? It is a long-debated question rather than a settled one. Harold's army marched roughly 185 miles north, fought a major battle, then marched south again to face William within three weeks, and some historians argue that fatigue and losses from Stamford Bridge weakened his force at Hastings, while others stress that Harold's core mistake was attacking before his full strength had assembled.
\nIs Stamford Bridge really "the end of the Viking Age"? It is a common shorthand rather than a precise historical boundary. Scandinavian raiding and dynastic ambitions in the British Isles did not vanish overnight, but Stamford Bridge marked the last time a Norwegian king led a full-scale invasion aimed at conquering England, which is why many historians treat it as symbolically closing that era.
\nA medieval craft guild was an association of people who did the same work in the same town, given standing by the town or the crown to set the terms of that work. It decided who could practise the craft, how long they had to train, what counted as an acceptable piece of work, and what happened to a member who fell short. Between roughly 1100 and 1500 that arrangement spread across urban Europe, and the ladder it is remembered for — apprentice, journeyman, master — became the ordinary route into skilled work in hundreds of towns.
\nWhat follows is how that ladder worked in practice: what an apprenticeship contract committed both sides to, what a journeyman was paid and how far he could move, what it actually took to become a master, and how quality was checked once someone was one. It also sets out how much of this varied. There was no single European guild system, and the sections below name the towns where the rules ran the other way.
\nThis page covers the medieval centuries. For what became of these institutions between 1550 and 1750 — their entanglement with the state, chartered companies and eventual abolition — see what the guild system became in the early modern period. Both pages sit in the Medieval Economy & Guilds collection.
\nGuilds followed towns. As urban populations grew from the eleventh century onward, crafts that had been practised in a handful of workshops became trades with dozens or hundreds of practitioners competing in the same market, buying from the same suppliers and selling to customers who could not easily judge what they were buying. A buyer can weigh a loaf. A buyer cannot see whether a silver cup is sterling or whether a cloth was fulled properly, and will find out only after paying.
\nCraft associations answered several problems at once, which is why they are hard to summarise. They gave the town a body that could be held responsible for the quality of a trade's output. They gave masters a way to limit competition among themselves. They gave workers a burial fund, a chapel and someone to appeal to in a dispute. And they gave rulers a unit of the urban population that could be taxed, licensed and, when necessary, blamed.
\nThe earliest surviving craft charter in England is a royal grant to the weavers of London in 1155, under Henry II. That it came from the crown rather than from the weavers is characteristic: guild privileges were typically granted, confirmed and sometimes revoked by an authority above the guild, and they were very often paid for.
\nThe craft ladder as statutes described it — and the exits most people took
\nThe ladder is not a staircase. Most people who entered it did not reach the top. Apprenticeships were abandoned, masters died mid-term, and the costs of mastership — fees, tools, a shop, citizenship — stopped many journeymen permanently. Journeyman was a destination as often as a stage, and much urban craft work was done by people who stayed there for life.
\nThe three ranks are a convenience of English. Contemporaries used their own words, and the words did not always divide the same way: Zünfte and Ämter in German-speaking towns, arti in Florence, métiers in Paris, gremis in Barcelona, Gaffeln in Cologne, misteries and companies in England. Some crafts had intermediate grades; some had none worth the name.
\nApprenticeship was a contract, and its central feature is that it bound both parties. The master undertook to teach the craft — not merely to use the labour — and to provide food, lodging and usually clothing. The apprentice's family undertook to pay a premium and to leave the child in the master's household, under the master's authority, for the agreed term. Breach was actionable, and town courts heard cases in both directions: masters suing runaways, and families suing masters who had taught nothing.
\nTerms varied by craft and by town, and the variation was not random. Trades where the skill took long to acquire, or where the master's investment in tools and materials was high, tended to demand longer terms; trades that needed hands more than judgment demanded shorter ones. Around 1268 the Paris register compiled for the provost, Étienne Boileau, set down for each of roughly a hundred trades its own apprenticeship term, its hours and its inspection arrangements — the point being precisely that they differed from one another. Anyone who quotes a single European apprenticeship length is quoting one town, one trade and one century.
\nTwo things the contract did not guarantee. It did not guarantee that the apprentice would ever practise the craft independently: completing a term made you a trained worker, not a master. And it did not guarantee equal access. Places were limited by statute in many crafts, premiums excluded poorer families, and the sons and daughters of existing masters routinely entered on easier terms — sometimes formally, by a statute that halved their fee, and sometimes simply because their father was the master who took them.
\nThe word records the arrangement. A journeyman was paid by the journée, the day's work — not by the piece, not by the year, and not by a share of the shop. He had finished his term, he could be trusted with the work, and he worked in another man's premises with another man's tools for a daily wage.
\nThat made journeymen the mobile part of the workforce. A master's need for hands rose and fell with the season and the order book, so journeymen moved between workshops, and frequently between towns, in a way apprentices could not and masters generally did not. Skill travelled with them, which is one of the routes by which techniques crossed borders — and one reason towns that wished to keep a technique tried to stop the people who carried it from leaving. Lucca bound its silk workers to secrecy under oath and heavy penalty, as the case of the Lucchese silk guilds shows; when political upheaval drove those workers out after 1314, the technique went with them to Venice and Florence.
\nJourneymen also organised. Where their numbers were large and their prospects of mastership poor, they formed their own associations for burial, drink, mutual aid and — visibly by the fourteenth and fifteenth centuries — for bargaining and stoppages. Masters' guilds and town councils regarded these bodies with suspicion and legislated against them repeatedly, which is itself evidence that they existed and had force.
\nMastership was an economic threshold before it was a test of skill. A master needed premises, tools and working capital for materials, and had to pay an entry fee to the guild. In most towns he also needed to be a citizen — a burgher — and citizenship itself was commonly bought. The practical consequence is that the decisive question for a competent journeyman was usually not whether he could do the work but whether he could raise the money, and the ordinary solutions were inheritance, marriage into a master's family, or a widow's workshop.
\nThe masterpiece belongs to this picture, but not at the beginning of it. The requirement to submit an approved specimen of work spread unevenly through the later medieval centuries and hardened afterwards. Where it existed it did real work: it demonstrated competence in a trade where a bad piece could injure a customer or discredit a town's export. But the historiography divides sharply on what it mainly accomplished. Sheilagh Ogilvie reads masterpiece requirements, alongside fees and waiting periods, principally as instruments for restricting entry and protecting the incomes of those already inside. The essays collected by Stephan Epstein and Maarten Prak give more weight to what guild training transmitted, and treat the requirement as part of a system that did in fact produce and preserve skill. The disagreement is genuine and unresolved, and a reader should know that rather than be handed one side of it.
\nFamily and network mattered throughout. Reduced fees for masters' sons, marriage between craft households, and the inheritance of a workshop and its custom all meant that the ladder was climbed most easily by people already standing near it.
\n| \n | Apprentice | \nJourneyman | \nMaster | \n
|---|---|---|---|
| Entry | \nContract with one master; premium paid by the family; places often capped by statute | \nCompleted term, attested by the master | \nEntry fee, capital for a workshop, usually town citizenship; in some crafts an approved masterpiece | \n
| Work | \nLearns the craft while working in the master's shop and household | \nSkilled work in another master's shop, with the master's tools and materials | \nRuns a shop; takes apprentices; answerable to the guild for what it produces | \n
| Payment | \nBoard, lodging, clothing and training; little or no wage | \nA wage by the day | \nThe profit of the shop, less fees, dues and levies | \n
| Obligations | \nObedience and residence for the term; secrecy in some crafts | \nGuild and civic dues; often barred from taking apprentices | \nUpholds standards, submits to inspection, pays dues, may serve in guild office | \n
| Possible next step | \nJourneyman — if the term is completed and attested | \nMaster — if capital, citizenship and, where required, a masterpiece can be had | \nGuild office; civic office in towns where the crafts governed | \n
| Important regional variation | \nTerm length set per craft and per town; the Paris register of c. 1268 gives roughly a hundred different ones | \nMovement between towns was normal in German-speaking lands, more constrained where a technique was protected, as at Lucca | \nIn Nuremberg after 1349 there was no autonomous guild to admit anyone: the council licensed and policed the crafts itself | \n
Enforcement is where guild regulation stops being a set of rules and becomes visible in surviving objects. The standard machinery had four parts.
\nInspection by sworn officers. Guilds appointed wardens or searchers with the right to enter workshops, examine work in progress and seize what failed. The English statute of 1300, under Edward I, required that silver reach the sterling standard and gold the touch of Paris, and directed the guardians of the craft to go from shop to shop to assay the work.
\nMarks. The same statute introduced the leopard's head as a compulsory assay mark on silver — an early instance of a public sign that a piece had been tested by someone other than the person selling it. Marking systems spread across trades and towns: cloth seals, town marks and makers' marks all did the same job of attaching a traceable claim to an object after it left the workshop.
\nStandard weights and measures. Guilds and town governments maintained standards for the units their trade sold by, and inspected against them. The offence in most surviving cases is not poor craftsmanship but short measure.
\nSanctions, graded. Fines were the ordinary penalty. Above them lay confiscation and destruction of the defective goods, public exposure of the offender, suspension from the trade and, at the extreme, expulsion — which in a town where the guild controlled practice meant the end of a livelihood.
\nThese systems protected buyers, and they protected the reputation of a town's exports, which was often the more pressing motive. They also gave existing masters a lawful instrument to use against competitors, and the records show it being used that way.
\nThe same rules face two directions, and any honest account has to say so. Limiting the number of apprentices maintained the quality of training and limited the future supply of competitors. Restricting practice to guild members guaranteed that work was done by trained hands and excluded everyone else: women in most crafts and most towns, migrants, Jews, and in many places anyone not born within the walls. Fixing prices or terms of sale stabilised a trade and suppressed competition within it.
\nWhether the net effect helped or hindered European economic growth is the central dispute in the modern literature, and it remains open. What can be said without taking a side is that guild privileges were legal instruments obtained from an authority, that they were valuable, that they were paid for, and that they were defended in court by the people who held them.
\nReading guilds only as economic institutions misses most of what their members would have said they were for. A craft association typically maintained an altar or a chapel, paid for candles and masses, buried its dead and prayed for them. It supported members who fell ill, and widows and orphans after a death. It held a feast, processed on its saint's day and enforced a standard of conduct on its members that had little to do with workmanship.
\nAt York in 1415 the Ordo paginarum recorded which of some fifty crafts was answerable for which pageant in the Corpus Christi cycle. It is a religious document and an administrative one at the same time: to know which guild owed which play is to have a register of the city's trades and their standing. Craft, worship and civic obligation were not separate spheres that guilds happened to straddle. They were the same membership seen from different angles.
\nParis, c. 1268 — regulation written down. Étienne Boileau, provost of Paris, had the customs of the city's trades compiled into the register known as the Livre des métiers. It records, trade by trade, apprenticeship terms, working hours and inspection arrangements for roughly a hundred crafts. It is the earliest systematic record of its kind for a major European city — and it is a record made by royal authority, so it shows what the crown wished fixed in writing as much as what workshops did.
\nNuremberg, 1348–49 — the crafts without a guild. After a craftsmen's rising, the patrician council of Nuremberg banned autonomous guild organisation outright. The crafts themselves did not disappear: for the next four centuries they worked under sworn ordinances issued and policed by the council itself. Nuremberg is the standing refutation of the idea that medieval urban crafts were necessarily guild-governed. The regulation was there; the self-governing guild was not.
\nFlorence, 1293 — the guild as constitution. The Ordinances of Justice tied eligibility for political office to membership of a guild, making the arti the constitutional basis of the city. In Flanders the crafts entered the government of Bruges and Ghent in the settlement of 1302, and at Cologne the Verbundbrief of 1396 handed the council to twenty-two Gaffeln, craft federations, an arrangement that held for four hundred years. Where guilds governed, guild membership was the route to civic power — and exclusion from a guild was exclusion from politics.
\nThe strongest single claim this page makes is a negative one. There was no uniform European guild system in these centuries, and the apprentice–journeyman–master ladder is a serviceable summary rather than a rule that was everywhere in force.
\nTerminology varied, as the names above show. Apprenticeship terms were set per craft and per town. Access varied: some crafts admitted women in their own right, most did not; some admitted outsiders, many did not. The masterpiece was required in some crafts and unknown in others. Governance varied most of all — Florence and Cologne were governed through their crafts, while Nuremberg forbade autonomous guilds and Lübeck kept its Ämter firmly under a council that never let them govern. Leiden had guilds like any Holland town, but its great export trade, the drapery, was run by the city as a municipal nering under sworn officials rather than by a craft guild at all.
\nAnd the absence of evidence is not evidence of absence. The towns named here are places where the institution happens to be unusually well documented. There were hundreds of others.
\nNothing about 1500 ended any of this. What changed over the following centuries was the setting: growing state power that wanted to license and tax crafts directly, chartered companies operating at a scale no town guild could regulate, and new ways of organising production — putting-out into the countryside, and later the factory — that worked outside guild jurisdiction altogether.
\nGuild rules also became national law in places where they had been local custom. The English Statute of Artificers of 1563 fixed a seven-year apprenticeship for the realm; the apprenticeship clauses were not repealed until 1814. The institutions themselves were dismantled unevenly and late: Tuscany in 1770, France in 1791, the Dutch Republic in 1798, with compulsory membership surviving in the Habsburg lands to 1859. That story — adaptation, entanglement with the state, and abolition — is the subject of the early modern guild system, 1550–1750.
\n| Date | \nEvent | \n
|---|---|
| 12th century | \nCraft associations multiply as urban populations grow across western Europe. | \n
| 1155 | \nHenry II grants a royal charter to the weavers of London, the earliest surviving craft-guild charter in England. | \n
| 1271 | \nVenice sets down the statute of the glassmakers in the Capitolare dei fioleri. | \n
| c. 1268 | \nThe Livre des métiers records the customs of roughly a hundred Paris trades for the provost, Étienne Boileau. | \n
| 1291 | \nThe Republic of Venice orders the glass furnaces moved to Murano, enclosing the craft on an island. | \n
| 1293 | \nFlorence's Ordinances of Justice tie political office to guild membership. | \n
| 1300 | \nThe statute of Edward I sets the sterling standard and introduces the leopard's head assay mark, with wardens assaying shop to shop. | \n
| 1302 | \nAfter the Battle of the Golden Spurs, the craft guilds of Bruges and Ghent take a formal share in city government. | \n
| 1308 | \nThe statute of Lucca's Corte dei Mercanti codifies oversight of a silk industry already exporting across Europe. | \n
| 1327 | \nThe London Goldsmiths' Company receives its royal charter. | \n
| 1332 | \nA guild-based city constitution replaces patrician rule at Strasbourg. | \n
| 1348–49 | \nAfter a craftsmen's rising, Nuremberg's council bans autonomous guild organisation; the crafts continue under council ordinance for four centuries. | \n
| 1368 | \nA guild constitution gives the crafts a majority on the Augsburg council. | \n
| 1382 | \nThe Guild of Saint Luke is chartered at Antwerp, registering the city's painters and printers. | \n
| 1396 | \nCologne's Verbundbrief hands the council to twenty-two Gaffeln. | \n
| 1415 | \nYork's Ordo paginarum records which craft answers for which pageant of the Corpus Christi cycle. | \n
| 15th century | \nGuild statutes are recorded at Kraków under Magdeburg town law; Leiden's drapery is run as a municipal nering rather than by a craft guild. | \n
For teachers and students
\nLearning objectives
\nKey terms
\nDiscussion questions
\nHow long was a medieval apprenticeship? There is no single answer, and that is the historically accurate reply. Terms were set craft by craft and town by town: the Paris register of about 1268 records roughly a hundred trades each with its own arrangements. Longer terms went with crafts where skill took longer to acquire or the master's investment was greater. The familiar figure of seven years comes from later English law — the Statute of Artificers of 1563 — and applying it to the medieval centuries across Europe is an anachronism.
\nDid every journeyman become a master? No, and most did not. Mastership required capital for premises and tools, an entry fee, and in most towns citizenship, which was itself usually bought. A competent journeyman could be blocked indefinitely by cost alone. Many remained journeymen for life, and journeyman work accounted for a large share of urban craft production.
\nWas a masterpiece always required? No. The requirement spread unevenly through the later Middle Ages and became more general afterwards; plenty of crafts never used one. Where it did exist, historians disagree about its main function — a genuine test of competence, or a device for limiting entry alongside fees and waiting periods. Both readings are argued in the literature cited below.
\nCould women join medieval guilds? In most crafts and most towns, not on the same terms as men. There were real exceptions, particularly in textile trades and in the position of masters' widows, who could sometimes continue a workshop and its rights after a husband's death. But exclusion was the norm, and it was one of several — migrants and religious minorities were commonly barred as well.
\nWere all medieval towns governed by guilds? No. Nuremberg banned autonomous guild organisation after the rising of 1348–49 and had its council regulate the crafts directly for the next four hundred years. Lübeck kept its craft offices under a council that never allowed them to govern, and Leiden ran its main export trade as a municipal administration rather than as a craft guild. Guild-governed towns such as Florence and Cologne were one pattern among several.
\nWritesides of History, "How Medieval Guilds Worked: Apprentices, Journeymen and Masters, 1100–1500", 2026. https://writesidesofhistory.com/how-medieval-guilds-worked-apprentice-journeyman-master/
\nIn the summer of 53 BC, the Roman proconsul Marcus Licinius Crassus led some seven legions across the Euphrates into Parthian territory, and within weeks most of that army was dead, captured, or scattered across the Mesopotamian desert. The defeat at Carrhae cost Rome an estimated 20,000 dead and 10,000 prisoners, along with Crassus's own life and the legionary eagles of his shattered legions. It remains one of the heaviest single defeats in Roman military history, and it fell not on a frontier garrison but on one of the three men who, alongside Julius Caesar and Pompey the Great, effectively ran the Roman state.
\nCrassus owed his standing to wealth and political management rather than to any real military reputation, though he had helped crush the Spartacus revolt in 71 BC. In 60 BC he joined Caesar and Pompey in the informal power-sharing arrangement historians call the First Triumvirate, renewed at the conference of Lucca in 56 BC. That renewal secured Crassus a second consulship in 55 BC alongside Pompey, and through the tribune Gaius Trebonius's law it gave him a five-year proconsular command of Syria. Caesar was carving out a reputation in Gaul; Pompey already had the glory of his eastern campaigns against Mithridates. Crassus, the financier of the group, wanted a conquest of his own, and Parthia — the powerful state that had inherited much of the old Persian world beyond the Euphrates — offered the kind of prize that could match his partners' achievements, the same elite competition for status traced in Cicero's account of the late Republic's patronage networks.
\nRome had no formal grievance against Parthia. The war was one of choice, launched for prestige and plunder rather than defense, and several senators, including Cicero, are recorded as uneasy about its legality and odds. Crassus nonetheless arrived in Syria in 54 BC and spent the winter probing Mesopotamia and gathering supplies, aiming to strike deep into Parthian territory the following season.
\nIn the spring of 53 BC, Crassus crossed the Euphrates with roughly 35,000–40,000 infantry organized in seven legions, several thousand auxiliaries, and about 4,000 cavalry, including a contingent of Gallic horsemen his son Publius Crassus had brought from serving under Caesar, where he had commanded two of the operations listed in our record of Caesar's campaigns in Gaul and Britannia. The cautious route followed the river, keeping the army supplied by water and shielding its flank. Instead, on the advice of Ariamnes, an Arab chieftain from the client kingdom of Osroene who had previously supplied Rome and now acted as guide, Crassus struck out across the open desert toward Parthian-held Mesopotamia. Ancient writers treat Ariamnes as either bribed or coerced into Parthian service, and modern historians remain divided on how much genuine treachery, as opposed to bad judgment, shaped the decision — the same caution about single, late-written sources that applies to Drusus and Germanicus's campaigns on the Rhine.
\nThe desert crossing left the Roman army without reliable water, without local support, and without the terrain cover that had protected Roman infantry in prior eastern campaigns. It also delivered them onto ground that could not have suited the enemy better: a flat, open plain near the town of Carrhae, ideal for cavalry and useless for the close-order infantry tactics on which Roman success depended.
\nThe Parthian commander, a nobleman known by the title Surena, met Crassus with a force built almost entirely around horses: roughly 9,000 mounted archers and about 1,000 heavily armored cataphracts, with no infantry at all. Roman commanders expected cavalry to run short of arrows in a sustained fight, but Surena had brought a baggage train of camels loaded with replacement ammunition, letting his horse archers maintain a punishing rate of fire for hours.
\nCrassus formed his legions into a hollow square, the standard defensive response to encirclement, but the formation simply turned the battle into a slow bombardment. Roman shields could not fully stop the heavier Parthian arrows at close range, and the legionaries had no effective way to close with an enemy that simply rode away from any advance. Publius Crassus led a mixed force of cavalry, infantry and archers in a sortie meant to drive the horse archers off; the Parthians feigned retreat, drew the smaller Roman column away from the main body, then encircled and destroyed it with cataphract lancers. Publius was killed, and Surena's men reportedly paraded his head on a spear before the main Roman line to break its morale.
\nAs night fell, the surviving Romans abandoned their wounded and retreated toward the town of Carrhae. Crassus's second-in-command, Gaius Cassius Longinus — later one of Julius Caesar's assassins — managed to extract a substantial detachment and reach Roman Syria safely, a rare episode of competent leadership in an otherwise catastrophic campaign.
\nTrapped and starving at Carrhae, Crassus agreed to negotiate with Surena. Both Plutarch and Cassius Dio, writing generations later, describe him being killed during the parley, reportedly while being helped or forced onto a horse in a scene the Roman side read as an ambush. Cassius Dio adds that the Parthians afterward poured molten gold into Crassus's mouth or severed head as a mocking comment on his legendary wealth; this detail is a moralizing story rather than a verified event, and it should be treated as the kind of legend that grows up around a famous death rather than as fact.
\nRoman losses were catastrophic: ancient figures put the dead at around 20,000 and the captured at around 10,000, against a Parthian force a fraction of the Roman army's size. Among the captured matériel were the legionary eagles, the sacred standards whose loss was a mark of institutional disgrace rather than mere battlefield defeat. They were not recovered by force; Augustus negotiated their return diplomatically in 20 BC and celebrated the settlement as a triumph, commemorating it with an arch in the Roman Forum — the same instinct for image management behind Rome's road-building propaganda under Augustus.
\n| Date | \nEvent | \n
|---|---|
| 60 BC | \nCrassus, Pompey and Caesar form the First Triumvirate | \n
| 56 BC | \nThe conference at Lucca renews the alliance and sets its members' next commands | \n
| 55 BC | \nCrassus's second consulship; the Lex Trebonia grants him Syria for five years | \n
| 54 BC | \nCrassus arrives in Syria and begins probing operations in Mesopotamia | \n
| Spring 53 BC | \nCrassus crosses the Euphrates and is guided into the open desert by Ariamnes | \n
| 53 BC | \nBattle of Carrhae: Publius Crassus's cavalry sortie is destroyed; the Roman square is worn down by Parthian horse archers | \n
| 53 BC | \nCrassus is killed during a parley with Surena after the retreat to Carrhae | \n
| 40 BC | \nParthians overrun Roman Syria; further Roman standards are lost under Decidius Saxa | \n
| 36 BC | \nMark Antony's own invasion of Parthia fails, adding further Roman losses to the account | \n
| 20 BC | \nAugustus negotiates the diplomatic return of the standards lost at Carrhae | \n
The military disaster had political consequences that outlasted the battlefield. The First Triumvirate had already been an uneasy balance between three ambitious men; Crassus's death removed the figure who had, however imperfectly, buffered the rivalry between Caesar and Pompey. Within a few years that rivalry hardened into civil war, a collapse that political historians of the late Republic — including those who study the Senate politics of Cicero's oratory — treat as one of the pivots toward the end of republican government.
\nCarrhae also set the terms of Rome's long, inconclusive relationship with Parthia. No later Roman emperor managed a lasting conquest east of the Euphrates, and the diplomatic recovery of the standards became a byword for how much prestige, rather than strategic gain, was really at stake on the eastern frontier — a theme echoed in the fiscal strain behind Diocletian's later economic reforms and in a Republican elite that, like Crassus, measured status partly in visible wealth, as seen in Rome's sumptuary laws.
\nCarrhae was not a frontier skirmish that went wrong; it was the destruction of an army led by one of the three most powerful men in Rome, undertaken for reasons of personal ambition rather than necessity. Crassus's poor route, his underestimation of a purely cavalry army, and the desert terrain combined to produce a defeat whose scale shocked contemporaries and whose consequences — the unraveling of the Triumvirate, decades of unfinished business with Parthia, and a stain on Roman military prestige that persisted until Augustus's diplomacy — outlived the man himself by generations.
\nWhy did Crassus invade Parthia when Rome had no formal war with it? Crassus wanted a military triumph to match Caesar's conquests in Gaul and Pompey's earlier campaigns in the east, and Parthia was the nearest rich target within reach of his proconsular command in Syria. Ancient writers describe the war as a matter of personal ambition rather than strategic necessity, and some Roman senators reportedly questioned its legality before it began.
\nHow could a smaller Parthian force defeat seven Roman legions? The Parthian army fought entirely on horseback, combining fast-moving mounted archers with heavily armored cataphract lancers, and it operated on open, flat terrain that neutralized Roman infantry tactics. A resupply train of camels let the horse archers keep firing far longer than the Romans expected, wearing down a Roman square that had no effective way to force close combat.
\nWhat happened to the legionary standards captured at Carrhae? The Parthians kept the eagles as trophies for over three decades. Augustus negotiated their return through diplomacy in 20 BC rather than through renewed conquest, and he presented the peaceful recovery as a major triumph, commemorating it with a new arch in the Roman Forum.
\nDid Crassus really die from having molten gold poured down his throat? That detail comes from Cassius Dio, writing about two and a half centuries after the event, and it reads as a moralizing legend built around Crassus's reputation for greed rather than a verified fact. Plutarch's earlier account instead describes Crassus's head and hand being sent to the Parthian king after his death during a parley with Surena.
\nWriting did not begin as poetry, prophecy, or law: it began as a receipt. The earliest texts anywhere in the world, scratched onto small clay tablets in the city of Uruk in southern Mesopotamia around 3200 BC, are not stories or hymns but tallies of grain, sheep, textiles, and labor moving through temple storerooms. Before that, for thousands of years, the same information had been tracked with small clay counters — tokens — pressed into wet clay or sealed inside hollow clay balls. The invention of writing, one of the handful of transformations historians agree changed the trajectory of civilization, grew directly out of the unglamorous problem of keeping accounts in an economy built on trade, tribute, and surplus.
\nLong before any script existed, small geometric objects of fired or sun-dried clay — cones, spheres, disks, cylinders — were already in use across the Near East to represent quantities of goods. Archaeologist Denise Schmandt-Besserat, who catalogued thousands of these objects from sites across Iran, Iraq, the Levant, and Turkey, argued that a cone might stand for a measure of grain and a sphere for a larger measure, with other shapes tracking sheep, oil, or textiles. The simplest tokens appear as early as roughly 8000 BC, coinciding with the spread of farming and herding economies that made tracking surplus necessary in the first place. For millennia these tokens did the work later handed to writing: they let someone absent from a transaction still know what had been counted, promised, or delivered.
\nBy around 3500 BC, as settlements like Uruk grew into some of the first true cities, a more elaborate set of tokens appeared alongside the old plain shapes — perforated, incised, and more finely made, apparently tracking finished and traded goods such as textiles and metal rather than raw agricultural staples. Some historians read this shift as evidence that long-distance trade and specialized craft production, not farming alone, pushed accounting technology toward greater complexity.
\nThe decisive step happened when tokens representing a transaction began to be sealed together inside a hollow clay envelope, known to archaeologists as a bulla, stamped on the outside with the cylinder seal of the official responsible. Excavations at Susa, in what is now Iran, uncovered bullae still containing their original tokens, confirming how the system worked. A bulla solved one problem — it made a record tamper-evident — but created another: once sealed, nobody could see what was inside without breaking it. The solution, as Schmandt-Besserat's influential theory has it, was to press each token into the envelope's soft outer surface before enclosing it, leaving a two-dimensional impression that recorded the contents without opening the seal. Not every specialist accepts every step of this account; some argue the shift from sealed administrative practice to script was less linear than a single token-to-sign chain implies. But the broad outline — an economy needing portable, checkable records of trade, arriving at a way to mark that in clay — is well established.
\nOnce the impressed symbols on an envelope's surface did the job the tokens inside had done, the hollow envelope became unnecessary. Sometime before 3200 BC, scribes in and around Uruk began flattening the clay into solid tablets, marked with impressed pictographs and numerals rather than tokens at all. That flattened tablet, not any single dramatic breakthrough, is usually treated as the birth of writing.
\nThe earliest true written tablets, from a phase of Uruk's history archaeologists call Uruk IV, date to roughly 3350 to 3200 BC, with a slightly earlier layer of purely numerical tablets, lacking pictographic signs, going back to around 3500 BC (the Uruk V phase). German excavations at the site, underway since the 1920s, have recovered several thousand of these tablets, overwhelmingly administrative in content. They record barley rations, the movement of livestock, deliveries of beer and textiles, and labor obligations owed to temple institutions — the kind of information a large economy relying on both local production and distant trade for timber, stone, and metal could not keep in memory alone.
\nThis period's script, known as proto-cuneiform, combined pictographic signs for objects and professions with an unusually complicated set of numerical systems: rather than one universal way of counting, scribes used different systems for different goods, so the symbol for "ten" measuring grain was not the symbol for "ten" measuring people or land. Decoding that layered numerical logic, opaque for decades, was the achievement of a computer-assisted study published in 1993 by Hans Nissen, Peter Damerow, and Robert Englund, who showed the earliest scribes of Uruk were not fumbling toward literature but running a purpose-built accounting system.
\nIt is worth being precise about what these first texts were not. They carried no verbs, no grammar, and almost no continuous sentences; a typical tablet lists commodity signs next to numerals, occasionally with the name of an official or institution, in the format of an inventory rather than a narrative. Full sentence structure, and with it connected prose, developed only gradually over the following centuries, through the Uruk III phase and into the Early Dynastic period. One especially early example, a small limestone tablet excavated at the Sumerian city of Kish and dated broadly to the Late Uruk period, is sometimes described as the world's oldest known written document; that claim is disputed, since the dating range is wide and comparably early tablets survive from Uruk itself, but the object is a genuine record of writing's pictographic, pre-grammatical infancy. Proto-cities like Çatalhöyük and Jericho had already shown large, organized settlements could function for millennia without any of this machinery; what changed at Uruk was a scale of exchange that finally made memory alone insufficient.
\nMesopotamia was not the only place commercial and administrative pressure produced a script. Egyptian hieroglyphs appear in the archaeological record at roughly the same time, around 3200 to 3100 BC, though historians remain divided on whether Egypt developed the idea independently or borrowed the general concept of writing from contact with Mesopotamia. Much later and apparently without external contact, the Indus Valley civilization produced its own still-undeciphered script by around 2600 BC, and Chinese writing emerged independently on oracle bones by the late second millennium BC. Trade and administration recur as pressures behind nearly every one of these separate inventions, even where the resulting signs share no common ancestry.
\n| Date | \nDevelopment | \n
|---|---|
| c. 8000 BC | \nSimple clay tokens in use across the Near East to track agricultural goods | \n
| c. 3500 BC | \nComplex tokens appear, tracking manufactured and traded goods; sealed clay bullae in use at sites like Susa | \n
| c. 3500 BC | \nEarliest purely numerical clay tablets appear at Uruk (Uruk V phase) | \n
| c. 3350-3200 BC | \nProto-cuneiform pictographic tablets appear at Uruk (Uruk IV phase), overwhelmingly administrative | \n
| c. 3200-3100 BC | \nEgyptian hieroglyphic writing appears, independently or through contact | \n
| c. 3200-3000 BC | \nUruk III phase: proto-cuneiform sign inventory expands and standardizes | \n
| 1928-1936 | \nGerman excavators recover and publish the first archaic Uruk tablets | \n
| 1993 | \nNissen, Damerow, and Englund decode the early numerical notation systems | \n
Historians disagree how much weight to put on trade, as opposed to temple and palace redistribution, in explaining why writing appeared when it did. The tablets themselves are institutional records — grain rationed by a temple, labor assigned by an authority — rather than contracts between independent merchants, which has led some scholars to argue centralized administration, not market exchange, was the immediate driver. Others note that the goods tracked, including metal, timber, and fine textiles, were themselves often obtained through long-distance trade the resource-poor Mesopotamian alluvium could not otherwise supply, making trade and temple bookkeeping two sides of one system rather than rival explanations. Either way, the underlying pressure was the same: an economy too large to run on memory and reputation alone.
\nThe written word did not arrive as an inspired leap; it accumulated, layer by layer, out of the mundane demands of an economy that had outgrown what any one person could remember. Clay tokens, sealed envelopes, impressed symbols, and finally flattened tablets each solved a narrow, practical problem of accounting for trade and temple goods, and only the cumulative result of those small fixes turned out to be one of the most consequential technologies human beings have built. Long after Uruk's scribes stopped tallying barley and sheep, the same instinct to make exchange legible would resurface in institutions as different as the Library of Alexandria and the building projects of the pyramid builders of Giza, both resting on record-keeping traditions rooted in this Bronze Age bookkeeping. It is a reminder that civilization's larger turning points, like the monument-raising of hunter-gatherers at Göbekli Tepe, often began not with an idea about the future but a practical problem in the present.
\nDid writing really begin as an accounting tool? Yes. The earliest tablets from Uruk, dating to roughly 3350-3200 BC, record allocations of grain, livestock, textiles, and labor rather than stories, laws, or religious texts. Continuous narrative writing developed only centuries later.
\nIs the clay token theory of writing's origin universally accepted? No. Denise Schmandt-Besserat's token-to-tablet theory is influential and well evidenced by finds of tokens and sealed bullae, but some scholars argue the shift from sealing practices to a true script was less direct than a single continuous chain.
\nDid other civilizations invent writing independently of Mesopotamia? Most historians think so, at least in some cases. Egyptian hieroglyphs appear at nearly the same time, though scholars disagree on whether Egypt developed the idea independently or absorbed it through contact. Indus Valley script and Chinese writing appear to have emerged independently, without Mesopotamian contact, at later dates.
\nWhat did the earliest Mesopotamian tablets actually contain? They were administrative lists: commodity signs paired with numerals, sometimes with the name of an official or institution attached, in the format of an inventory. They lacked grammar, verbs, and full sentences, which only developed gradually over the following centuries.
\nOn 3 August 1962, thirty Australian officers and warrant officers stepped off a flight in Saigon, forming the Australian Army Training Team Vietnam — the first Australian unit committed to the war and, a decade later, the last to leave. Over the following ten years the Team grew into the most highly decorated Australian unit of the conflict, its advisers embedded alongside South Vietnamese and American forces at remote border camps. Four years after the Team's arrival, a second elite formation joined it: the Special Air Service Regiment, whose small reconnaissance patrols out of Nui Dat became, by most measures, the most effective Australian fighting force in Vietnam. Together the two units — one built for teaching and advising, the other for stealth and small-unit combat — gave Australia's Vietnam commitment a reputation for elite soldiering that outlasted the war itself and still shapes how the conflict is remembered at home.
\nThe Training Team, universally known within the Army as "the Team," was raised specifically for advisory duty and placed under the command of Colonel Francis Philip "Ted" Serong, a soldier with extensive experience of jungle warfare and counter-insurgency in Malaya and Burma. Its members did not fight as a conventional unit; they were posted singly or in pairs to South Vietnamese Army battalions and to the American-run Civilian Irregular Defense Group camps that dotted the country's remote highlands and border regions, training and leading local forces against Viet Cong infiltration. This dispersed, advisory role meant Team members frequently found themselves the only Australians — sometimes the only non-Vietnamese officers at all — for miles in any direction, a position that demanded unusual initiative and exposed them to combat as often as, or more often than, soldiers in conventional battalions.
\nThe Team's strength grew steadily through the decade, reaching a peak of 227 personnel in November 1970, and almost 1,000 Australians served with it in total. Its isolation from the wider Australian contingent — which arrived only from 1965, when a battalion was committed and later grew into the 1st Australian Task Force — meant the Team operated alone for its first three years, an advisory experiment that predated Australia's larger commitment entirely.
\nThe Special Air Service Regiment's involvement began in 1966, when 1 Squadron deployed to Nui Dat as part of the newly formed 1st Australian Task Force in Phuoc Tuy province. Over the following five years, the Regiment's three sabre squadrons rotated through two tours each, acting, in the words later used by military historians, as the Task Force's "eyes and ears." Rather than fighting as a large formation, SAS patrols worked in small teams — typically five men — inserted by helicopter deep into contested territory to locate Viet Cong base areas, monitor troop movements, and report enemy strength back to Nui Dat, occasionally setting ambushes along known infiltration routes.
\nThis was reconnaissance-led warfare, closer in spirit to the intelligence-gathering that Cold War alliances relied on elsewhere — from the networks described in NATO's intelligence and technology expansion to the older logic of watching an adversary before engaging one. Roughly 580 SAS soldiers served in Vietnam, conducting close to 1,200 patrols. Regiment tradition holds that Viet Cong forces in Phuoc Tuy came to call the patrols "Ma Rung," or phantoms of the jungle, though the phrase survives mainly through unit lore rather than a documented Viet Cong source.
\nWhat made the SAS deployment distinctive was the disparity between its own losses and the casualties it reported inflicting. Official unit figures record roughly seven SAS deaths across the five-year deployment — one killed by enemy fire, one who died of wounds, three killed in accidents, one missing and one who died of illness — against 28 wounded. Over the same period, patrol reports credited the Regiment with 492 enemy killed and a further 106 "possibly killed." Historians generally treat the SAS record as genuinely exceptional, the product of superior training and the tactical advantage of choosing when to engage — but body counts from the Vietnam War as a whole are notoriously unreliable evidence, compiled under pressure to show progress, so the SAS figures are best read as the Regiment's own patrol reporting rather than an independently verified tally.
\nIf the SAS record rested on avoiding decisive contact, the Training Team's reputation rested on the opposite: advisers thrown into combat alongside the South Vietnamese units they served, with none of a conventional battalion's depth of support around them. All four Victoria Crosses awarded to Australians in the war went to Team members. Warrant Officer Kevin Wheatley was posthumously awarded the VC in 1965 after refusing to abandon a wounded comrade during a Viet Cong ambush. Major Peter Badcoe received his, also posthumously, in 1967 for a string of actions rescuing wounded advisers and rallying South Vietnamese troops under fire. Warrant Officers Ray Simpson and Keith Payne were each decorated within three weeks of one another in 1969: Simpson for leading a company out of an ambush and returning for wounded men, and Payne for spending roughly three hours, wounded himself, gathering some forty scattered and wounded soldiers under mortar fire at the border outpost of Ben Het and leading them to safety. Across its decade of service the Team received more than one hundred decorations in total, a record that, alongside the SAS, made it the most consistently celebrated element of Australia's Vietnam commitment — a status not unlike the outsized reputations small specialised forces earned elsewhere in Cold War alliances such as NATO's nuclear-sharing arrangements.
\n| Date | \nEvent | \n
|---|---|
| 3 August 1962 | \nThirty officers and NCOs of the newly formed Australian Army Training Team Vietnam arrive in Saigon under Colonel F. P. Serong. | \n
| 1965 | \nWarrant Officer Kevin Wheatley is killed defending a wounded comrade; he is posthumously awarded the Victoria Cross. | \n
| 1966 | \n1st Australian Task Force is established at Nui Dat, Phuoc Tuy province; 1 Squadron, SAS Regiment, deploys as part of it. | \n
| 1967 | \nMajor Peter Badcoe is killed leading South Vietnamese troops in a series of actions; he is posthumously awarded the Victoria Cross. | \n
| November 1970 | \nThe Training Team reaches its peak strength of 227 personnel. | \n
| 1971 | \n1st Australian Task Force withdraws from Phuoc Tuy, ending the SAS Regiment's combat deployment. | \n
| 1969 | \nWarrant Officers Ray Simpson and Keith Payne are each awarded the Victoria Cross within three weeks of one another. | \n
| December 1972 | \nThe Training Team departs Vietnam, the last Australian unit to leave after a decade of service. | \n
Australia's broader military presence wound down through 1971 and 1972 as the Task Force withdrew and Vietnamization shifted fighting back to South Vietnamese forces, a drawdown that unfolded against the backdrop of an American administration consumed by surveillance controversies of its own, as explored in the expansion of NSA surveillance under Nixon. The Training Team, true to its advisory role since 1962, stayed on after the combat units left, finally departing in December 1972 as the last Australian formation in the country — a bookend to the commitment it had opened a decade earlier, in an era when Cold War diplomacy still played out on television as vividly as on battlefields, as it had at the 1961 Vienna summit between Kennedy and Khrushchev.
\nIn the years since, both units have occupied a particular place in Australian memory of the war: not the conscript battalions whose service defined the domestic controversy over Vietnam, but small, selected forces whose casualties were comparatively light and whose acts of valour were unusually well documented. That reputation formed amid the same Cold War currents that reshaped alliances elsewhere in the period, from Western Europe's long argument over shared nuclear responsibility to the diplomatic thaw that followed, of the kind later visible in projects like Apollo-Soyuz. What set the SAS and the Training Team apart in Australia's Vietnam story was neither scale nor strategic decisiveness — a few hundred advisers and a few hundred SAS soldiers never determined the war's outcome — but how their conduct became, for a divided home front, a rare and largely uncontested source of national pride.
\nWere the SAS Regiment and the Training Team the same unit? No. The Training Team was raised in 1962 to advise South Vietnamese and allied irregular forces, with members typically posted singly to local units rather than fighting together. The SAS Regiment deployed from 1966 as part of the 1st Australian Task Force and ran reconnaissance and ambush patrols from its own base at Nui Dat. The two overlapped for years but had distinct roles, chains of command and reputations.
\nWhy did the Training Team win all four Australian Victoria Crosses of the war? Team advisers were typically the only Australians present with the South Vietnamese units they served, without the depth of support a conventional battalion could call on, which repeatedly put individual advisers in positions where they had to act alone under fire to rescue wounded comrades or rally collapsing local forces. All four VC actions — involving Kevin Wheatley, Peter Badcoe, Ray Simpson and Keith Payne — arose from exactly this exposed advisory role.
\nHow reliable are the reported SAS "kill ratio" figures from Vietnam? The Regiment's own patrol reports credited it with hundreds of enemy killed for a small number of its own dead, and most historians accept SAS patrols were genuinely effective. But body counts across the Vietnam War are widely considered unreliable, since units on all sides had incentives to report favourable figures, so the ratio is best read as internal patrol reporting rather than an audited statistic.
\nWhen did Australia's Vietnam commitment finally end? Australian combat units, including the Task Force and the SAS Regiment, withdrew from Phuoc Tuy province in 1971 as part of the broader Vietnamization of the war. The Training Team, true to its founding role as an advisory rather than combat formation, remained afterward and did not leave until December 1972, making it both the first and the last Australian unit in Vietnam.
\nIn 1682, Louis XIV moved the government of France to a former hunting lodge twelve miles outside Paris, and for the next century every French king ruled from a palace built, quite deliberately, to make the nobility easier to watch. Versailles began under Louis XIII as a modest brick-and-stone retreat for hunting. His son turned it into something without precedent in Europe: a single vast building where thousands of nobles lived and competed for the king's attention under rules he himself set. Its fountains, mirrors, and processions were not decoration bolted onto a political project; they were the project.
\nLouis XIII commissioned the original pavilion at Versailles between 1631 and 1633, a modest U-shaped structure of brick and stone built for hunting parties, not statecraft. His son had reasons of his own to want distance from Paris: the Fronde, the aristocratic revolt of his childhood years, had driven the young king from his own capital more than once, and the memory never quite left him. Beginning in 1662, Louis XIV commissioned the first of several building campaigns at Versailles, and from that date until his death in 1715 the site underwent one of the largest construction programs in European history. Louis Le Vau added a new forecourt and then, between 1668 and 1670, wrapped the old hunting lodge in a stone "envelope" that expanded the building without demolishing his father's original core. After Le Vau's death, Jules Hardouin-Mansart took over, working alongside the decorator Charles Le Brun and the garden designer André Le Nôtre to expand Versailles into the complex whose footprint survives today.
\nThe government of France moved to Versailles formally in 1682, and the palace became what one modern historian has called a miniature city, housing the king, his family, thousands of courtiers, and the machinery of ministries under one roof. This was neither hasty nor accidental. Louis XIV had watched his own minority unravel into civil war when great nobles operated from fortified estates, largely outside royal supervision. A single palace built around the king's own apartments solved that problem by architecture rather than force.
\nThe site itself was a poor choice for a palace by any ordinary standard: Versailles sat on marshy, waterless high ground with none of the natural springs or rivers that typically determined where European rulers built. Supplying the fountains meant importing water on an industrial scale, and the solution, the Machine de Marly, was among the most ambitious hydraulic engineering projects of the seventeenth century, a system of enormous waterwheels on the Seine that pumped river water uphill through miles of pipe to reservoirs feeding the gardens. Exact figures for the project's cost and workforce are disputed among historians, since the crown's own accounting obscured how expenditure on Versailles was spread across royal budgets, but the scale of labor was never in question: tens of thousands of soldiers and laborers worked the site, and disease and accidents killed workers in numbers the court did not advertise.
\nThat willingness to bend a difficult landscape to royal will was itself a political statement: a palace that could conjure fountains and canals on ground that offered none naturally demonstrated that the king's authority could overcome nature as readily as it overcame the nobility. The gardens, laid out by Le Nôtre in strict geometric perspective radiating from the king's own bedroom window, made the same argument in landscape form: every avenue in the park led the eye back to the monarch.
\nThe Hall of Mirrors, completed in 1684 under Hardouin-Mansart's direction, became the palace's most famous space and a working instrument of statecraft. Seventeen mirrored arches faced seventeen windows overlooking the gardens, an extravagant use of a material France had previously imported from Venice at high cost; the crown established its own royal mirror manufactory partly to escape that dependency, turning a luxury good into a demonstration of French self-sufficiency. Ambassadors were received in the gallery and processed its length toward the throne, the room's scale designed to intimidate before a word of diplomacy was exchanged. The link between art and political messaging ran through the whole palace, much as it did in the portrait culture of the Bourbon court, where painted images of the king served on canvas nearly the same purpose the architecture served in stone.
\nVersailles's deepest political innovation was not architectural but behavioral. Louis XIV structured the nobility's daily life around his own body: the lever, the ceremony of the king's rising each morning, and the coucher, its evening counterpart, became coveted rituals in which the right to hand the king his shirt or hold a candle marked a courtier's standing more precisely than any title. According to the duc de Saint-Simon, whose memoirs remain the fullest surviving eyewitness account of these rituals, the king used the granting and withholding of such minor privileges as continuous, low-stakes tests of loyalty. Saint-Simon wrote as an embittered participant who felt his own ducal rank diminished by a king who elevated legitimized bastards and commoner ministers over hereditary peers, so his account carries a grievance of its own, but the pattern he describes — a nobility kept too busy competing for favor to plot against the crown — matches what historians have independently reconstructed from household records.
\nThis was, in effect, a policy for neutralizing the aristocracy that the Wars of Religion and the Fronde had shown could otherwise tear the kingdom apart. A noble family that spent a fortune maintaining an apartment at Versailles, competing for court offices and pensions, had far less capacity to fund a private army on its own estates. The same logic of concentrating power through presence shaped how the crown used the printing press to control political messaging in these same decades, and it found an echo in how Prussia's Hohenzollern rulers built their own court rituals partly by copying Versailles. Louis XIV never actually said "L'État, c'est moi" — the line is a legend later writers attached to him, and most historians doubt he spoke those words — but the palace he built made a more durable version of the same claim.
\n| Date | \nEvent | \n
|---|---|
| 1631–1633 | \nLouis XIII builds a modest hunting pavilion at Versailles | \n
| 1661 | \nNicolas Fouquet's finance ministry collapses partly over the extravagance of his own château at Vaux-le-Vicomte, a cautionary tale Louis XIV absorbed before building his own palace | \n
| 1662 | \nLouis XIV commissions the first major building campaign at Versailles | \n
| 1668–1670 | \nLouis Le Vau encases the original lodge in a new stone "envelope" | \n
| 1682 | \nThe royal court and government formally relocate to Versailles | \n
| 1684 | \nThe Hall of Mirrors is completed under Jules Hardouin-Mansart | \n
| 1685 | \nLouis XIV revokes the Edict of Nantes from Versailles, ending legal toleration for French Protestants | \n
| 1715 | \nLouis XIV dies at Versailles after 72 years on the throne | \n
| 1789 | \nParisian market women march on Versailles and force Louis XVI's family back to Paris | \n
Louis XIV's successors inherited a palace that was as much a liability as an asset. Louis XV and Louis XVI kept the court there largely because moving it risked signaling weakness, even as the cost of maintaining thousands of courtiers strained finances already buckling under war debt. The isolation from Paris that had once protected the crown from urban unrest became a vulnerability once that unrest grew organized: in October 1789, a crowd of Parisian women marched to Versailles and compelled the royal family back to the capital, ending the palace's 107-year run as the seat of government. Versailles never again served as a working royal residence, though it hosted two later moments of state theater in the room built for Louis XIV's diplomacy: the German Empire was proclaimed in the Hall of Mirrors in 1871, and the treaty ending the First World War was signed there in 1919.
\nVersailles worked because Louis XIV understood something his predecessors had not: that a palace could function as a system of government rather than merely a backdrop for one. Its water engineering, mirrored galleries, and rigid daily ceremonies were not separate achievements but three faces of a single project — converting royal birth into a near-automatic machinery of obedience. That machinery outlived Louis XIV by more than seven decades, and its failure in 1789 arrived not because the architecture stopped working but because the political conditions that had made it work no longer held.
\nWhy did Louis XIV move the French court to Versailles? Louis XIV had lived through the Fronde, an aristocratic revolt during his childhood that forced the royal family to flee Paris, and he wanted a capital he controlled entirely. Versailles let him gather the nobility under one roof, where daily competition for royal favor replaced the independent power bases nobles had held on their own estates.
\nHow long did it take to build the Palace of Versailles? Construction unfolded across several campaigns from 1662 until Louis XIV's death in 1715, with major phases led first by Louis Le Vau and then by Jules Hardouin-Mansart. Work and expansion continued under later kings, though never again at the same scale.
\nWhat was the Hall of Mirrors used for? Completed in 1684, the Hall of Mirrors served as a reception space for ambassadors and state ceremonies, its mirrored arches and garden views designed to overwhelm visitors before any negotiation began. It later hosted the proclamation of the German Empire in 1871 and the signing of the treaty ending the First World War in 1919.
\nDid courtiers really compete over rituals like the king's morning dressing? Yes; according to the memoirist duc de Saint-Simon, minor privileges such as handing the king his shirt during the lever carried real political weight, since access to the king's person was one of the few currencies of status Versailles offered nobles who had lost their independent regional power.
\nWhen Pope Clement V dissolved the Knights Templar in 1312, the order did not vanish everywhere at once. In most of Christian Europe, the bull Vox in Excelso and the companion decree that followed it delivered Templar estates into the hands of the rival Knights Hospitaller. But in the kingdom of Portugal, King Dinis I refused to let the order's assets, its castles, and many of its former brothers simply disappear into another institution's coffers. Within seven years of the suppression, he had engineered something almost unique in the aftermath of the Templar trials: a new military order, built on the old one's foundations, that answered to the Portuguese crown rather than to Rome or to the Hospitallers. It was called the Militia of Our Lord Jesus Christ — the Order of Christ — and it would go on to outlive the Templars by more than two centuries, eventually bankrolling the voyages that opened the Atlantic to Europe.
\nPortugal's experience of the Templar affair looked nothing like France's. King Philip IV had launched the persecution in October 1307 with coordinated dawn arrests, and the confessions that followed, extracted under torture, gave the French crown the material it needed to press for the order's destruction across Christendom. Dinis I took a very different path. Rather than hunting Portuguese Templars as heretics, he moved in the years before the order's formal suppression to bring its properties and personnel in his kingdom under direct royal oversight, treating them as an asset to be preserved rather than a scandal to be purged. The Templar order that had built castles and financial networks across Europe had never been accused in Portugal of the crimes that damned it in France, and Dinis had no obvious motive to destroy men who had spent generations defending his kingdom's frontier with Muslim al-Andalus.
\nClement V's dissolution bull did not, by itself, decide where Templar wealth would go. That question was settled six weeks later, on 2 May 1312, by a second bull, Ad Providam, which assigned former Templar property across most of Europe to the Hospitallers. It carved out a conspicuous exception, however, for the kingdoms of Castile, Aragon, Portugal, and Majorca, where the pope left the final disposition of Templar assets to further negotiation rather than handing them automatically to a rival order already resented by local rulers for its own wealth and privileges. Aragon's King James II used that opening to found an entirely new order, the Order of Montesa, formally established in 1317 with papal approval to absorb Templar and some Hospitaller property in his territories. Dinis pursued the same strategy for Portugal, and with less friction than James had faced, since the Portuguese Templars carried no taint of confession or trial.
\nIn 1318, Dinis petitioned Pope John XXII to sanction a new military order built from what remained of the Portuguese Temple. The pope, resident at Avignon like his predecessor, agreed on terms that suited a crown eager to keep its frontier defenders intact: on 14 March 1319, John XXII issued the bull Ad ea ex quibus, formally establishing the Order of Christ and transferring to it the people, castles, and revenues of the former Portuguese Templars. Unlike the Hospitaller transfer elsewhere, this was not a rival order absorbing a defeated one; in practice, it allowed many of the same men, the same fortresses, and much the same administrative structure to continue under a new banner and a new master answerable to the Portuguese king. The order's first headquarters was modest by comparison with what it would become: the church of Santa Maria do Castelo at Castro Marim, a frontier town in the Algarve, under its first Grand Master, Dom Gil Martins, a knight who had previously served the Portuguese Order of Aviz.
\n| Date | \nEvent | \n
|---|---|
| 13 October 1307 | \nPhilip IV orders the mass arrest of Templars in France | \n
| 22 March 1312 | \nVox in Excelso dissolves the Knights Templar | \n
| 2 May 1312 | \nAd Providam grants Templar property to the Hospitallers, excepting Castile, Aragon, Portugal, and Majorca | \n
| 1317 | \nAragon founds the Order of Montesa on former Templar lands | \n
| 1318 | \nKing Dinis I petitions Pope John XXII to charter a new Portuguese order | \n
| 14 March 1319 | \nAd ea ex quibus establishes the Order of Christ, based at Castro Marim | \n
| Mid-14th century | \nThe order's headquarters moves to the former Templar castle at Tomar | \n
| 1420–1460 | \nHenry the Navigator serves as Grand Master, funding Atlantic exploration | \n
The order did not stay in the Algarve for long. Later in the fourteenth century — chroniclers and historians differ on the precise year — the Order of Christ transferred its seat north to Tomar, the fortress-town that had served as the principal stronghold of the Templars in Portugal since Grand Master Gualdim Pais founded its castle around 1160. The move put the Order of Christ in physical possession of the most important surviving monument of the old Templar presence in the kingdom: a hilltop castle built around a round church, the Charola, modeled loosely on the rotunda of the Church of the Holy Sepulchre in Jerusalem. What had been built as the Temple's Portuguese headquarters became, and remains, the Convent of Christ, a complex that UNESCO now protects as a World Heritage Site for the layered building campaigns — Romanesque, Gothic, Manueline, Renaissance — that successive masters of the order added to it over the following centuries.
\nThe Order of Christ might have remained a minor regional curiosity had it not acquired, in the fifteenth century, a Grand Master whose ambitions reached far beyond Portugal's borders. Prince Henry, later remembered as Henry the Navigator, held the position from 1420 until his death in 1460, and he used the order's revenues and its exemption from certain crown obligations to underwrite the maritime expeditions that pushed steadily down the West African coast during his lifetime. The order's distinctive emblem, a red cross pattée with a white field at its center, descended visually from the Templar cross that had marked the order's founders two centuries earlier, and it was painted onto the sails of the caravels that carried Portuguese crews into increasingly unfamiliar waters. That visual inheritance tied the voyages of the fifteenth century back to an order whose institutional memory, if not its formal mission, still stretched to the Crusader states that had given the original Templars their purpose before the fall of Acre in 1291 ended the Latin presence in the Holy Land.
\nThe Order of Christ's cross did not stay confined to a religious-military brotherhood. After Manuel, who had served as the order's governor before becoming King Manuel I of Portugal in 1495, the emblem was increasingly treated as a symbol of the crown and the kingdom's overseas ventures rather than of the order alone, paired in Portuguese iconography with the armillary sphere that also became a mark of the Age of Discoveries. Centuries later, long after Portugal's fortunes had been reshaped by episodes like the union of the Spanish and Portuguese crowns between 1580 and 1640, the cross of the Order of Christ remained visible on Portuguese military insignia, a fragment of Templar visual identity that had migrated, through the accident of Dinis's fourteenth-century diplomacy, into the symbolic vocabulary of a modern nation-state. It is a strikingly different afterlife from the one imagined by the medieval cross the Templars themselves wore as marks of penitential warfare in the Holy Land.
\nWhat happened in Portugal after 1312 was not really a survival story in the sense of secret continuity or hidden brotherhoods; it was a straightforward act of royal statecraft. Dinis I recognized that the Templars' Portuguese estates, fortresses, and experienced fighting men were too valuable to hand to a rival order or scatter through confiscation, and Pope John XXII, negotiating from Avignon with a monarch who controlled the property in question, had little practical reason to refuse. The result was an order that carried forward the Temple's castles, its cross, and much of its personnel under a new charter — one that, a century later, would help finance Portugal's opening of the Atlantic world.
\nWas the Order of Christ secretly the same organization as the Knights Templar? Not secretly, but substantially in practice. King Dinis I arranged for many of the same fortresses, revenues, and former Templar brothers in Portugal to pass into the new order chartered by Pope John XXII in 1319, so the institutional and personal continuity was real, even though the Order of Christ was a legally distinct body created by papal bull rather than a Templar order operating under an assumed name.
\nWhy did Portugal keep its Templar properties instead of handing them to the Hospitallers? The bull Ad Providam of May 1312, which assigned most former Templar assets to the Knights Hospitaller, explicitly excepted Castile, Aragon, Portugal, and Majorca, leaving their disposition open to further negotiation. King Dinis used that opening to petition for a new Portuguese order rather than let the property pass to a rival institution.
\nWhat was Henry the Navigator's connection to the Order of Christ? Henry served as the order's Grand Master from 1420 until his death in 1460, and he drew on its wealth and privileges to help finance the Portuguese voyages down the African coast associated with the early Age of Discoveries, which is why the order's cross appeared on the sails of Portuguese ships during this period.
\nWhere was the Order of Christ headquartered? It was founded at the church of Santa Maria do Castelo in Castro Marim in the Algarve in 1319, then later moved its seat to the fortress-town of Tomar, the former principal stronghold of the Knights Templar in Portugal, where the Convent of Christ still stands today.
\nIn 9 BC, a Roman general died after a fall from his horse deep in territory no Roman army had ever permanently held, and the Senate voted his family a name they would carry for three generations: Germanicus, "conqueror of Germania." Nero Claudius Drusus, stepson of the emperor Augustus, had spent four seasons pushing Roman power across the Rhine toward the Elbe, farther into Germanic territory than any Roman commander before him. His son, born Nero Claudius Germanicus in honor of that unfinished conquest, would spend his own short career avenging Rome's worst military disaster and trying to finish what his father began, only to die in Syria a decade later under circumstances Romans argued over for the rest of Tiberius's reign. Between them, father and son trace the arc of Rome's ambitions on the German frontier: how far the empire reached, and what happened to the men who reached too far.
\nDrusus, born in 38 BC to Livia and her first husband, grew up in the household of Augustus after his mother's remarriage to the future emperor. He married Antonia Minor, daughter of Mark Antony and Octavia, and by his early thirties commanded Rome's forces on the Rhine and in Gaul — a frontier Julius Caesar had crossed twice and abandoned twice, as our record of his campaigns in Gaul and Britannia sets out season by season. Between 12 and 9 BC he led four campaigning seasons against the Germanic tribes east of the river, subduing the Sicambri and Frisii, defeating the Chauci near the North Sea coast, and pushing to the Weser and then the Elbe. He also had a canal dug linking the Rhine to the Zuiderzee, the fossa Drusiana, to move his fleet against the coastal tribes, in the same tradition of infrastructure that Rome's road builders under Augustus used to bind new territory to the empire. In 9 BC he became the first Roman commander to reach the Elbe, a river Rome would never again hold as a frontier.
\nThat same year, Drusus fell from his horse while returning from the Elbe and broke his leg; the injury turned fatal within about a month. Tiberius, his older brother, reportedly rode from northern Italy to Germania in a single journey to reach him before he died. Augustus had the body carried to Rome and delivered the funeral eulogy himself, the same kind of stagecraft that shaped Augustan public image more broadly, and Drusus was buried in the imperial family's mausoleum. The Senate then voted Drusus and his descendants the honorific title "Germanicus," commemorating a conquest never finished. His son, not yet four, inherited the name and would spend his career trying to live up to it.
\nGermanicus, born in 15 BC, grew up amid the same imperial household that had shaped his father. He married Agrippina the Elder, a granddaughter of Augustus, around 5 AD, and in 4 AD Augustus arranged a two-step succession plan: he adopted Tiberius as his heir and required Tiberius, in turn, to adopt Germanicus, even though Tiberius already had a son, Drusus the Younger, securing the succession two generations ahead. Germanicus, still in his twenties, served under Tiberius during the Pannonian and Dalmatian revolts of 6 to 9 AD, gaining his first military experience shortly before Rome's frontier strategy collapsed altogether.
\nThat collapse came in 9 AD, when Publius Quintilius Varus led three legions into an ambush organized by the Germanic chieftain Arminius, remembered as the Battle of the Teutoburg Forest. Roughly twenty thousand men were killed, and Augustus, according to Suetonius, was left banging his head against a door and crying "Quintili Vare, legiones redde," give me back my legions. The disaster ended any near-term prospect of a Roman province of Germania and left the Rhine, rather than the Elbe, as the empire's boundary. The site at Kalkriese, widely accepted as the battlefield, has yielded coins, weapons, and mass graves consistent with Tacitus's later account of the slaughter.
\nWhen Augustus died in August 14 AD, the Rhine and Danube legions mutinied, demanding shorter service and better pay. Germanicus faced down the mutiny in person and sent his pregnant wife and infant son from the camp for safety, a gesture the soldiers found shaming enough to help restore order, then redirected their energy into a punitive raid across the Rhine before the year was out.
\nOver the next two campaigning seasons, Germanicus pushed deep into Germanic territory in a campaign of retribution. In 15 AD his officer Lucius Stertinius recovered one of the three lost eagle standards, and Germanicus led his army to the battlefield itself, where his soldiers buried the bleached bones of Varus's dead; a second eagle was recovered the following year. In 16 AD he forced Arminius into open battle at Idistaviso on the Weser, inflicting heavy losses. But a fleet battered by storms and the difficulty of holding ground beyond the Rhine convinced Tiberius further campaigning offered little return, and he recalled Germanicus to Rome, where the following year he celebrated a triumph displaying German captives, including Arminius's wife Thusnelda, though Arminius was never captured.
\nIn 17 AD Tiberius sent Germanicus east with authority overriding the region's other governors, to settle a succession crisis in client-kingdom Armenia and manage relations with Parthia. Tiberius also appointed Gnaeus Calpurnius Piso as governor of Syria, widely believed at the time to have been sent to check Germanicus's popularity, though what Tiberius actually intended remains disputed. The two men clashed openly, and in October 19 AD, after a lingering illness, Germanicus died at Antioch at thirty-three, reportedly accusing Piso of poisoning him with his last words. Whether he was truly poisoned, or died of illness amplified by suspicion, has never been established.
\nAgrippina the Elder carried her husband's ashes back to Rome amid extraordinary public mourning. Piso tried to retake Syria by force before being recalled and tried before the Senate in 20 AD on charges from treason to murder. He died, officially by his own hand, before the case concluded, though some suspected he was killed to spare Tiberius embarrassment. The bronze Senatus Consultum de Cn. Pisone Patre, rediscovered in fragments in Spain in the 1980s, records the Senate's staged verdict clearing the imperial household, revealing how the affair was meant to be remembered rather than what happened at Antioch.
\n| Date | \nEvent | \n
|---|---|
| 12-9 BC | \nDrusus campaigns across the Rhine, reaching the Elbe in 9 BC | \n
| 9 BC | \nDrusus dies of a riding injury; the Senate grants his family the name "Germanicus" | \n
| c. 5 AD | \nGermanicus marries Agrippina the Elder | \n
| 4 AD | \nAugustus has Tiberius adopt Germanicus as part of the succession settlement | \n
| 9 AD | \nVarus loses three legions at the Teutoburg Forest | \n
| 14 AD | \nRhine legions mutiny after Augustus's death; Germanicus suppresses it and raids across the Rhine | \n
| 15-16 AD | \nGermanicus campaigns in Germania, recovers two of the three lost eagles, fights at Idistaviso | \n
| 16 AD | \nTiberius recalls Germanicus from the Rhine frontier | \n
| 17 AD | \nGermanicus celebrates his triumph in Rome, then departs for the eastern provinces | \n
| 19 AD | \nGermanicus dies at Antioch, accusing Piso of poisoning him | \n
| 20 AD | \nPiso is tried before the Senate and dies before the verdict | \n
Drusus and Germanicus never achieved what their name commemorated. The Elbe frontier Drusus reached was abandoned within a generation, and Germanicus's campaigns avenged Teutoburg without reversing it; Rome's border settled permanently on the Rhine. What the two men left behind was less a conquered province than a case study in how the early Principate handled its most popular generals, the same tension visible later in Messalina's court under Claudius, Germanicus's own brother. Their descendants inherited the name and its danger: Germanicus's son Caligula, nicknamed by the Rhine legions who mutinied under his father, grew up to become one of Rome's most notorious emperors. The unfinished conquest stayed bound to the trajectory that carried the Western Roman Empire toward decline, a reminder that the frontier Rome could not hold shaped the imperial court as much as any battle in the city itself.
\nWho were Drusus and Germanicus? Drusus was a stepson of Augustus who campaigned across the Rhine into Germania between 12 and 9 BC, reaching the Elbe before dying of a riding accident. His son Germanicus, adopted by Tiberius as part of Augustus's succession plan, later commanded the Rhine legions and led campaigns of retribution after Teutoburg.
\nDid Rome ever conquer Germania? No. Drusus's advance to the Elbe was never held permanently, and after the Teutoburg Forest disaster of 9 AD wiped out three legions, Rome abandoned plans for a province there. Germanicus's campaigns recovered lost standards and defeated Arminius's coalition, but Tiberius recalled him rather than occupy the territory, and the Rhine remained Rome's frontier for centuries.
\nHow did Germanicus die, and was he poisoned? Germanicus died at Antioch in 19 AD after a lingering illness, and ancient sources say he accused the Syrian governor Gnaeus Calpurnius Piso of poisoning him with his final words. Whether this was genuine poisoning or illness read through contemporary suspicion has never been resolved, and historians treat it as an open question.
\nWhat happened to Piso? Piso was recalled to Rome and tried before the Senate in 20 AD on charges including treason and complicity in Germanicus's death. He died before the trial concluded, officially by suicide, though some suspected he was killed to protect the emperor from scandal. The Senate's decree on the case, rediscovered on a bronze inscription in Spain in the 1980s, records how the affair was officially resolved.
\nOn a bare limestone hill in southeastern Turkey, more than six thousand years before the first stone of Stonehenge was raised, people with no metal tools, no pottery, and no farms quarried, carved, and erected rings of pillars weighing up to several tons apiece. The site is called Göbekli Tepe, Turkish for "potbelly hill," and since systematic excavation began in 1995 it has forced archaeologists to rethink a story that once seemed settled: that monumental architecture was a product of agriculture and the surplus, hierarchy, and organized labor that farming made possible. At Göbekli Tepe, the sequence runs the other way. The monuments came first.
\nA 1963 joint survey by the University of Chicago and Istanbul University noted the mound and its scatter of broken limestone slabs, but the surveyors assumed the visible stone tips were grave markers from a medieval cemetery and moved on. The site sat largely unexamined for three decades. In 1994, the German archaeologist Klaus Schmidt, who had previously worked at the nearby Neolithic site of Nevalı Çori, went looking for other early sites in the region and was directed to the hill by local farmers who had turned up unusual stone fragments while plowing. He began formal excavations the following year under the German Archaeological Institute and the Şanlıurfa Museum, and quickly realized the buried stone tips were not grave markers at all, but the tops of monumental carved pillars standing in deliberately built stone enclosures. Schmidt directed the dig until his death in 2014; excavation has continued since under his successors at the DAI and Turkish universities.
\nRadiocarbon dating places the main monumental phase of Göbekli Tepe, known to archaeologists as Layer III, in the Pre-Pottery Neolithic A period, roughly 9600 to 8200 BC, making the oldest enclosures over eleven thousand years old. That places them thousands of years before pottery, writing, or the wheel existed anywhere, and well before domesticated cereals or livestock appear in the region's archaeological record. It is a reminder that the mobile foragers of this period were capable of far more than survival: elsewhere in the same broad era, the hunter-gatherer communities of Doggerland were building rich, if less monumental, lives across a now-drowned landscape at the opposite end of Europe. The people who built and used the earliest enclosures were hunter-gatherers, not settled farmers, subsisting on wild gazelle, aurochs, and wild cereals rather than cultivated crops. That sequence is what makes the site so disruptive to older models of prehistory, which generally assumed that only settled agricultural societies, with stored food surpluses and centralized authority, could muster the labor and coordination monumental construction requires. Göbekli Tepe suggests instead that large-scale communal building, and the social organization it demanded, may have helped drive people toward the settled life that agriculture eventually made permanent, rather than simply following from it.
\nThe enclosures Schmidt and his teams uncovered are roughly circular or oval rings of T-shaped limestone pillars, some standing more than five meters tall and weighing several tons, arranged around two larger central pillars of the same T-shaped form. More than twenty such enclosures have now been identified using geophysical survey, though only a handful, most famously Enclosure D, have been fully excavated. Many pillars carry relief carvings of foxes, boars, snakes, scorpions, cranes, and vultures, along with abstract symbols whose meaning is still debated, a tradition of symbolic animal imagery with roots stretching back tens of thousands of years earlier to the painted caves of Lascaux and Altamira. A handful of the largest pillars have stylized arms, hands, belts, and loincloths carved along their sides, strongly suggesting the T-shape represents an abstracted human figure rather than a purely architectural form. One relief in particular, Pillar 43 in Enclosure D, has drawn outside attention: a small group of researchers has proposed it encodes a record of a comet impact around 10,950 BC. Most specialists regard that reading as speculative and unsupported by the evidence, and it remains a minority position.
\nSchmidt's own interpretation, laid out in his 2012 excavation monograph, was that Göbekli Tepe functioned as a sanctuary used periodically by dispersed hunter-gatherer groups for ritual and feasting rather than as a permanent settlement. He and his colleagues found large quantities of animal bone alongside grinding stones and vessels that may have been used for brewing, evidence read as the debris of large communal feasts. Schmidt's phrase for this model, "temple before city," proposed that shared religious practice, not economic necessity, first pulled scattered foraging bands into sustained cooperation. More recent excavation by his successors, reported by the DAI, has complicated that picture. Later seasons have turned up domestic features and other traces of longer-term habitation at Göbekli Tepe and related sites, suggesting the mound was not purely ceremonial and may have supported some resident population alongside its ritual function. The debate over how to weigh sanctuary against settlement, and over whether backfilling of the great enclosures around 8000 BC was a single deliberate ritual act or the accumulated result of everyday activity and erosion, remains active among the excavators themselves rather than settled either way. Deliberately covering a finished monument in earth is not unique to Göbekli Tepe; millennia later, the Scythian royal kurgans of the Eurasian steppe were built the opposite way around, mounded up rather than filled in, but with the same instinct to seal a sacred structure beneath the ground once its purpose was served.
\n| Date | \nEvent | \n
|---|---|
| 1963 | \nA University of Chicago–Istanbul University survey notes the mound, mistaking pillar tips for grave markers | \n
| 1994 | \nKlaus Schmidt identifies the site as Neolithic while surveying the region | \n
| 1995 | \nFormal excavation begins under the German Archaeological Institute and Şanlıurfa Museum | \n
| c. 9600–8200 BC | \nMain monumental phase (Layer III enclosures) constructed and used | \n
| c. 8000 BC | \nThe largest enclosures are filled in, deliberately or through long accumulation | \n
| 2014 | \nKlaus Schmidt dies; excavation continues under new DAI and Turkish leadership | \n
| 2018 | \nGöbekli Tepe is inscribed as a UNESCO World Heritage Site | \n
| 2021–present | \nThe Taş Tepeler project expands excavation to a dozen related sites across the region | \n
Göbekli Tepe is no longer studied in isolation. Since 2021 a Turkish-led initiative called Taş Tepeler, or "Stone Hills," coordinated by Istanbul University with the DAI, has extended systematic excavation to a cluster of related mounds across the same region, including Nevalı Çori, where Schmidt first encountered T-shaped pillars, and Karahan Tepe, some 35 kilometers away. Karahan Tepe has produced its own striking finds, including a monumental carved figure discovered in 2023, one of the most complete naturalistic human sculptures known from this period anywhere in the world. Together these sites indicate that Göbekli Tepe was not an isolated anomaly but part of a wider Pre-Pottery Neolithic culture spread across the upper Euphrates and Tigris basins, one that repeatedly invested enormous communal labor in carved stone monuments before any of these communities had committed to full-time farming. That broader pattern echoes, on a much older timescale, how large early settlements such as Çatalhöyük and Jericho also organized dense communal life without writing to record it.
\nGöbekli Tepe did not just push the date of monumental architecture back by thousands of years; it undermined the assumption that farming had to come first. Hunter-gatherers with stone tools, working without draft animals, wheels, or metal, quarried and raised pillars taller than a person and heavier than a car, then filled and reburied their own creation within a few thousand years, a feat of organized labor that would not look out of place among the achievements of far later, fully agricultural societies like the builders of the Great Pyramid of Giza. What drove that effort, whether shared belief, seasonal gathering, competitive display, or some combination the surviving stones cannot fully explain, is still argued over by the archaeologists who know the site best. What is no longer in question is that Göbekli Tepe, together with its sister sites across the Taş Tepeler landscape, has permanently rewritten how early human communities are understood to have organized themselves, long before anyone had planted a field.
\nHow old is Göbekli Tepe? The main monumental enclosures date to roughly 9600 to 8200 BC, making the oldest structures over eleven thousand years old and more than six thousand years older than Stonehenge. The site was built and used by hunter-gatherer communities before agriculture was established in the region.
\nWho discovered Göbekli Tepe? A 1963 University of Chicago and Istanbul University survey first noted the mound but mistook the visible stone tips for medieval grave markers. German archaeologist Klaus Schmidt correctly identified it as a major Neolithic site in 1994 and began formal excavation in 1995, directing the dig until his death in 2014.
\nWas Göbekli Tepe a temple? Klaus Schmidt argued it functioned primarily as a ceremonial gathering place, a "temple before city," based on evidence of large-scale feasting and the absence of obvious dwellings in the areas he excavated. More recent excavation has uncovered domestic features at Göbekli Tepe and related sites, so archaeologists now debate whether it was purely ceremonial or also supported some resident population.
\nHow does Göbekli Tepe relate to Karahan Tepe and other nearby sites? Since 2021, the Taş Tepeler project has shown that Göbekli Tepe belongs to a wider network of related Pre-Pottery Neolithic sites across southeastern Turkey, including Karahan Tepe and Nevalı Çori. Several of these sites share the same T-shaped pillar tradition, and Karahan Tepe has produced its own major discovery, a large carved human figure found in 2023.
\nOn the morning of 21 February 1916, German artillery opened a bombardment on a nine-mile stretch of the French line north of Verdun so intense that some French defenders later said the ground itself seemed to catch fire. What followed was not a battle designed to break through and end the war in a single stroke, but something colder: an attempt to draw the French army into a killing ground and destroy it there, position by position, for as long as it took. By the time the fighting guttered out in December 1916, Verdun had become the longest continuous battle of the First World War, a 302-day contest fought over a few square miles of hills and forts that neither side could really afford to lose.
\nVerdun was not an obvious military target in 1916. The old fortress city on the Meuse had lost much of its garrison and heavy guns the year before, redistributed to more active fronts, and it did not sit astride any strategic prize worth the name. That, to the German Chief of the General Staff, Erich von Falkenhayn, was precisely the point. Verdun bulged into the German line and had defended France's eastern frontier since the days of Vauban; French national pride was fused to it in a way that made retreat politically unthinkable, whatever the military logic might say. Falkenhayn's plan, as he later described it, was not to capture Verdun outright but to threaten it just enough that the French high command would feed division after division into its defense, absorbing losses the French army could not sustain and the German army, firing from prepared positions with a huge concentration of heavy artillery, supposedly could.
\nThe clearest statement of that intent survives in a single document: a memorandum Falkenhayn said he sent to Kaiser Wilhelm II in December 1915, arguing that a limited offensive at Verdun would force France to defend the salient "to the last man" and that "the forces of France will bleed to death" in the attempt. No copy of that memorandum has ever surfaced in the German archives, and the only account of it comes from Falkenhayn's own memoir, published in 1919 after his dismissal from command. Some historians accept it as a genuine, if unusually cynical, statement of strategy; others argue Falkenhayn invented or embellished it after the fact to make a costly failure look like a plan working as intended. What is not in dispute is how the battle actually unfolded: an offensive that began as a narrow attack on the east bank of the Meuse and, within weeks, swallowed both banks and drew in reserves neither side had originally intended to commit.
\nThe opening bombardment fired roughly a million shells in under ten hours, and German infantry advanced behind it against French positions that had been thinned out and, in places, only partially fortified. On 25 February, a small German raiding party walked into Fort Douaumont, the most powerful of the ring of forts around Verdun, and took it almost without a fight after its skeleton garrison had already been withdrawn. The fall of Douaumont, four days into the offensive, was a shock the French high command could not absorb quietly, and command of the Second Army defending Verdun passed that same day to General Philippe Pétain, who imposed a rigid system of rotating divisions through the sector so that no unit spent too long under the guns before being relieved.
\nPétain's other contribution was logistical rather than tactical. With the main railway into Verdun cut by German fire, the French army turned a narrow departmental road running north from Bar-le-Duc into a single-purpose supply artery, later named the Voie Sacrée, the Sacred Way. Thousands of trucks moved in a continuous rotation along it, day and night, carrying men, shells, and food into the salient and the wounded and exhausted back out. It was an improvised solution to a problem the German plan had not anticipated: that a fortress city could be resupplied fast enough, by road alone, to keep absorbing an offensive meant to starve it of the will to fight.
\n| Date | \nEvent | \n
|---|---|
| 21 February 1916 | \nGerman bombardment opens the offensive on the east bank of the Meuse | \n
| 25 February 1916 | \nFort Douaumont falls; Pétain takes command of the French Second Army | \n
| Spring 1916 | \nThe Voie Sacrée sustains Verdun by continuous truck convoy | \n
| 1 May 1916 | \nPétain moves up to command Army Group Centre; General Robert Nivelle takes over the Second Army | \n
| 23 June 1916 | \nGerman phosgene gas attack marks the deepest German advance, near Fort Souville | \n
| 24 October 1916 | \nFrench counter-offensive recaptures Fort Douaumont | \n
| 2 November 1916 | \nFrench forces retake Fort Vaux | \n
| 15–18 December 1916 | \nFinal French offensive pushes the German line back; major fighting ends | \n
By the early summer of 1916, the German advance had stalled a few miles short of Verdun itself, at the shattered village of Fleury and the approaches to Fort Souville, where a gas attack on 23 June marked the furthest German penetration of the entire battle. Pétain's careful rotation policy, whatever it cost in administrative strain, kept French morale from collapsing outright, and by autumn the initiative had shifted. Under Nivelle and his subordinate General Charles Mangin, the French mounted a series of set-piece counter-offensives built around massed artillery, retaking Fort Douaumont on 24 October and Fort Vaux a little over a week later. A final French push in mid-December drove German forces back roughly to the lines they had held in February, closing a battle that had gained almost no ground for either side in ten months of fighting. It was not the swift, decisive engagement that would settle a war in an afternoon, in the way the Battle of Midway would decide the naval balance in the Pacific a generation later; Verdun was won, to the extent it was won at all, by simply outlasting the other side.
\nVerdun's casualty toll is one of the most-cited and least-settled figures of the First World War. Imperial War Museums estimates the killed and wounded on both sides at roughly 555,000 French and 434,000 German, figures the museum itself describes as necessarily approximate given the chaos of the fighting and the incompleteness of wartime records. Other tallies, factoring in the missing and men who later died of wounds, put the combined toll closer to a million. Whatever the precise number, the scale of suffering fell on medical and evacuation systems that were never designed for it; the same overwhelmed conditions that drove the wartime expansion of organizations like the Red Cross on the Western Front were nowhere more visible than along the roads out of Verdun. France's war effort would grind on for two more years, financed partly by wartime debts whose reckoning came only after the armistice, in the reparations and reconstruction politics that later produced arrangements like the Dawes Plan. Verdun itself entered French national memory almost immediately as the battle the army had not lost, a status cemented well before the United States entered the war in 1917 following the interception of the Zimmermann Telegram.
\nVerdun rarely gets remembered as a battle of maneuver, because it was not one; it was a test of which army's industrial base, medical services, and political will would crack first under sustained, grinding pressure. Falkenhayn's strategy, real or reconstructed after the fact in his own memoir, described a kind of warfare that had little to do with the breakthrough battles either side still hoped for elsewhere on the Western Front, closer in spirit to the attritional air war the Royal Air Force fought over southern England in 1940 than to any single decisive clash. What Verdun actually proved was narrower and grimmer: that a fortress could be resupplied by road alone against a besieging army with more guns, and that neither side's soldiers, having endured ten months of it, had anything left to call a victory.
\nWhy did Germany choose to attack Verdun in 1916? Verdun was a historic French fortress city whose loss would have been a severe blow to French national pride, which German Chief of the General Staff Erich von Falkenhayn calculated would force France to defend it regardless of the military cost. The salient also sat close enough to German rail lines to let the German army mass an unusually large concentration of heavy artillery for the opening bombardment.
\nWhat was the "bleed France white" strategy, and is it real? The phrase comes from a memorandum Falkenhayn said he sent the Kaiser in December 1915, arguing that a limited offensive at Verdun would force the French army to commit reserves faster than it could replace losses. No copy of the memorandum survives outside Falkenhayn's own 1919 memoir, and historians remain divided over whether it reflects his genuine prewar intent or a justification written after the offensive had already failed to achieve a breakthrough.
\nHow many soldiers were killed or wounded at Verdun? Estimates vary because wartime records were incomplete, but Imperial War Museums puts combined French and German casualties, killed and wounded, at roughly 989,000 over the battle's 302 days. Some broader counts that include the missing and men who died later of wounds run higher still.
\nDid France or Germany win the Battle of Verdun? By December 1916, French counter-offensives under Nivelle and Mangin had pushed German forces back close to their February starting positions, and Verdun itself was never captured, so the battle is generally counted as a costly French defensive success. Neither side gained any strategic advantage proportional to the losses involved, which is why historians usually treat Verdun as a demonstration of attrition's costs rather than a conventional military victory.
\nBetween 1776 and 1783, six German princes leased close to 30,000 of their soldiers to the British Crown for use against its rebellious American colonies, and Americans came to call every one of them "Hessians," whether they hailed from Hesse-Kassel, Brunswick, or four smaller states entirely. The arrangement was not slavery or kidnapping, as Patriot propaganda often implied, but a long-established European institution: a ruler's standing army was also a source of income, and a subsidy treaty could turn idle regiments into ready cash. For the German princes it was a calculated gamble on royal solvency. For the roughly 30,000 men who marched, sailed, and in thousands of cases never returned, it was eight years fighting someone else's war on someone else's continent.
\nFacing a colonial rebellion that had already produced open fighting at Lexington and Concord and would eventually require more soldiers than Parliament could recruit or the public would tolerate conscripting, King George III's ministers turned to a well-worn continental practice: hiring troops from smaller German states through formal subsidy treaties. Britain signed agreements with Hesse-Kassel, Hesse-Hanau, Brunswick (Braunschweig-Wolfenbüttel), Anhalt-Zerbst, Ansbach-Bayreuth, and Waldeck, each contract specifying how many regiments a prince would supply, how they would be paid, and what compensation the ruler would receive. Hesse-Kassel supplied by far the largest contingent, an initial 12,000 men that grew to nearly 19,000 over the course of the war, which is why American colonists lumped every contingent, Brunswickers and Waldeckers included, under the single label "Hessian."
\nThe treaties were negotiated as state business, not private contracts. Britain agreed to pay each prince at British army rates, well above what the soldiers earned at home, plus a fixed annual subsidy for the duration of hostilities. Landgrave Friedrich II of Hesse-Kassel received roughly 450,000 crowns a year for his 12,000 men, while Anhalt-Zerbst's much smaller commitment of 600 troops earned its prince only about 22,500 crowns. Critics in Britain, Edmund Burke prominent among them, attacked the arrangement in Parliament as an unworthy way to fight fellow subjects; the Continental Congress, in the Declaration of Independence, denounced George III for "transporting large Armies of foreign Mercenaries to compleat the works of death, desolation, and tyranny."
\nFor the princes, the soldier trade was less about personal enrichment than about financing the modern state. Friedrich II of Hesse-Kassel used the British subsidies to fund a standing army, expand his bureaucracy, and pay down debt without raising taxes on his own subjects, a use of "mercenary" income that historians have since treated as a case study in cameralist state-building rather than simple profiteering. That same logic of squeezing revenue from a well-organized administrative state runs through the broader story of eighteenth-century German statecraft, visible too in the Prussian legal and economic reforms of Frederick the Great and in the cameralist administrators who tried to systematize Prussia's finances in the same decades. Small German courts like Hesse-Kassel's also staged their own rituals of dynastic prestige, not unlike the ceremonial culture of the neighboring Hohenzollern court in Prussia, even as their treasuries depended on renting out soldiers abroad.
\nThe soldiers themselves had little say in the arrangement. Hesse-Kassel and the other states relied on conscription, drawn heavily from the peasantry and from men without the money or connections to buy an exemption. Regiments were filled out with a mix of long-serving professionals, conscripts, and — increasingly as the war dragged on — recruits pressed into service with little enthusiasm for a war fought thousands of miles from home.
\nThe initial German contingents landed on Staten Island in the summer of 1776, joining the British buildup for the campaign against New York, more than a year after Paul Revere's alarm riders had already turned a colonial dispute into open war. They fought at the Battle of Long Island that August and again at White Plains that autumn, and by winter several Hessian regiments held forward outposts along the Delaware River in New Jersey, including the garrison town of Trenton under Colonel Johann Rall.
\nThat posting proved disastrous. On the icy morning of December 26, 1776, George Washington's army, having crossed the Delaware the night before, struck Rall's brigade before it had finished mustering. Rall was mortally wounded rallying his men and died the next day; the attack killed or wounded roughly a hundred Hessians and took close to 900 prisoner. The defeat at Trenton, followed days later by another British setback at Princeton, reversed a string of American losses that autumn and gave Washington's army, and the Revolution itself, a reprieve it badly needed going into 1777.
\n| Date | \nEvent | \n
|---|---|
| January–February 1776 | \nBritain signs subsidy treaties with Hesse-Kassel and the other five German states | \n
| Summer 1776 | \nFirst German regiments land on Staten Island, New York | \n
| August 27, 1776 | \nHessian units fight at the Battle of Long Island | \n
| December 26, 1776 | \nWashington's dawn attack overwhelms Rall's brigade at Trenton | \n
| October 1777 | \nBrunswick troops under Baron Riedesel surrender with Burgoyne's army at Saratoga | \n
| 1778–1781 | \nGerman auxiliaries garrison New York, fight in the southern campaigns, and guard captured territory | \n
| 1783 | \nThe Peace of Paris ends the war; surviving auxiliaries are shipped home or given leave to remain | \n
Trenton was not the war's only mass capture of German troops. When General John Burgoyne's invasion from Canada collapsed at Saratoga in October 1777, the Brunswick contingent commanded by Baron Friedrich Adolf von Riedesel surrendered along with the rest of Burgoyne's army under the terms of the so-called Saratoga Convention. Riedesel's wife, Friederike Charlotte, had traveled with the army through the campaign and its aftermath, and her letters and journal, later published in translation, remain one of the fullest firsthand accounts of what the march into captivity and the long confinement that followed actually looked like for the families attached to these regiments. According to several surviving soldiers' letters, captured Hessians in Pennsylvania were often billeted among the colony's substantial German-speaking population, working on local farms while awaiting exchange or the war's end, an arrangement that blurred the line between prisoner and neighbor in ways that unsettled American and German officers alike.
\nBritain's use of manpower beyond regular redcoats was not limited to German regiments; Virginia's royal governor had already tried a different kind of leverage in late 1775 with Lord Dunmore's Proclamation, offering freedom to enslaved Patriots who joined the Crown. For the German rank and file, contact with Pennsylvania's German-speaking farmers did something quieter: it planted the idea, for a good number of them, that staying might beat marching home to conscript service and debt.
\nPeace in 1783 did not send every German soldier home. Some historians estimate that around 6,000 of the roughly 30,000 who served chose to remain in North America rather than return, settling across Pennsylvania, New York, Virginia, and the Carolinas, with a smaller number moving north into British Canada. Desertion accounted for some of that number over the course of the war, but formal discharge and simple decisions to stay after 1783 accounted for more. For princes who had leased out soldiers as a source of state revenue, this quiet postwar attrition was an unplanned cost of the arrangement, a leakage of exactly the manpower the original treaties had been designed to preserve.
\nThe word "Hessian" outlived the war by far longer than the treaties that created it, hardening into an all-purpose American term for any German soldier who had fought for the Crown and, eventually, into folklore: Washington Irving's Headless Horseman in "The Legend of Sleepy Hollow" was, by local legend, a Hessian artillery trooper decapitated by a cannonball during the war, his ghost still searching for what he had lost fighting a war that was never his own.
\nThe Hessian auxiliaries occupy an uncomfortable place in the Revolution's memory: neither the noble redcoats of British tradition nor the citizen-soldiers of the American founding myth, but conscripted subjects of small German states, leased out by their rulers to fight a war of someone else's making. Their princes profited, their labor built roads, palaces, and bureaucracies back home, and their presence in America, meant to intimidate the rebellion into submission, instead helped harden colonial opinion against reconciliation. That thousands of them chose, in the end, to stay in the country they had been sent to conquer says as much about the limits of eighteenth-century statecraft as any battle report could.
\nHow many German soldiers fought for Britain in the American Revolution? Roughly 30,000 German soldiers served with British forces between 1776 and 1783, drawn from six separate principalities under formal subsidy treaties. Some broader counts run higher once wartime replacements and reinforcements are included, since regiments were regularly refilled after losses to combat, disease, and desertion.
\nWhy were soldiers from six different German states all called "Hessians"? Hesse-Kassel supplied by far the largest single contingent, and American colonists, unfamiliar with the patchwork of small German territories, applied its name loosely to every German soldier fighting for the Crown, including Brunswickers, Waldeckers, and men from Anhalt-Zerbst. The label stuck in American usage long after the war ended.
\nWhat happened to the Hessians captured at Trenton and Saratoga? Prisoners taken at Trenton in December 1776 and with Burgoyne's army at Saratoga in October 1777 were marched into the American interior, many held or paroled in Pennsylvania and Virginia, where a substantial German-speaking population already lived. Some were eventually exchanged, while others worked local farms and, for a portion of them, never returned to formal military service.
\nDid any Hessian soldiers stay in America after the war? Yes. Some historians estimate that thousands of the German soldiers who served chose to remain in North America once the war ended in 1783 rather than return home, settling mainly in Pennsylvania, New York, and the southern colonies, with a smaller number relocating to British Canada.
\nOn the afternoon of 26 August 1346, an English army perhaps 12,000 strong stood on a low ridge near the Picardy village of Crécy-en-Ponthieu and destroyed a French force nearly twice its size. By nightfall, a third of Philip VI's army lay dead on the field, including a king, a duke, and some 1,500 knights and esquires, while English losses were counted in the low hundreds. The instrument of that slaughter was not a new weapon — the longbow had existed for generations — but a new way of organizing infantry to make cavalry irrelevant. Crécy did not just win a battle. It began to unmake the assumption, unquestioned for centuries, that mounted knights ruled the medieval battlefield.
\nCrécy was the climax of a campaign that began as plunder, not conquest. On 12 July 1346, Edward III landed a force of English, Welsh, and allied troops at Saint-Vaast-la-Hougue in Normandy, opening what historians call the Crécy chevauchée: a mounted raid designed to burn, loot, and humiliate rather than hold ground. Edward's claim to the French crown, asserted through his mother Isabella and pressed on and off since 1337, gave the invasion its legal pretext, but the immediate goal was destruction. English and allied companies stripped the Norman countryside for miles inland, and on 26 July they stormed Caen, the wealthiest city in the region, killing much of the garrison and population over several days of sack.
\nThat kind of warfare produced plunder and terror, not decisive victory, and King Philip VI mobilized to trap the raiders before they reached safety. Edward turned east and north, shadowed by a growing French army, and found the crossing of the Somme blocked at every ford until a local guide revealed a tidal crossing at Blanchetaque. The English got across just ahead of the pursuit. With the sea and the port of Calais now his only realistic refuge, Edward chose to stop running. Near the village of Crécy, on ground he had reportedly scouted days earlier, he turned to fight.
\nThe position Edward chose did most of the work before a single arrow flew. His army stood on a gentle slope, flanks anchored by the villages of Wadicourt and Crécy itself, with the small River Maye and patches of woodland further constraining French movement. Edward dismounted his men-at-arms — a deliberate break from continental practice, which prized the mounted charge as the mark of a knight — and placed them in three battles, or divisions, with the young Edward, Prince of Wales, known to history as the Black Prince, given nominal command of the vanguard at only sixteen.
\nFlanking each division, English and Welsh longbowmen were drawn up in a "herse" formation, angled wedges that let archers rake an advancing enemy from the front and both sides at once. Chroniclers describe shallow pits dug in front of the line to break up a cavalry charge, and Edward is credited with deploying a handful of primitive cannon — among the earliest use of gunpowder artillery on a European battlefield — though their effect was almost certainly more psychological than lethal. It had rained heavily that morning, and the field had turned to mud that would matter enormously for what followed.
\nPhilip's army, gathered in haste from across the French nobility and its allies, was led into action by some 6,000 Genoese crossbowmen, mercenaries with a formidable reputation. That reputation depended on equipment they did not have with them: their large shields, or pavises, had been left behind in the baggage train, and the rain had slackened their crossbow strings in a way it had not affected the English longbows, which their owners had kept unstrung and dry until the moment of battle. Advancing into low, blinding afternoon sun, the Genoese loosed a volley that fell short. The English answer did not.
\nTen thousand longbowmen, each capable of loosing ten to twelve arrows a minute against a crossbow's two or three, tore into the crossbowmen before they could reload. The Genoese broke and fled back toward their own lines — and Philip's mounted knights, contemptuous of hired infantry and impatient for glory, rode them down rather than let them pass. What followed was less a battle than a repeated collision between disciplined defense and disintegrating attack.
\nFrench chroniclers and later historians count as many as fifteen separate cavalry charges against the English line before nightfall ended the slaughter. Each followed roughly the same pattern: knights spurred forward across ground already torn by mud, fallen horses, and the bodies of the Genoese, into arrow fire that killed mounts as readily as riders, and finally into a static line of dismounted men-at-arms who had not moved from the ground they had chosen at dawn. Whatever cohesion the French army possessed dissolved with each failed assault, and command and control that depended on individual noblemen's honor rather than a unified chain proved unable to call a halt.
\nAmong the dead was John of Bohemia, blind for years yet resolved to strike at least one blow, who reportedly had his horse tied to those of his knights so they could lead him into the fray; all of them died together. Philip VI himself had two horses killed under him and was wounded by an arrow before being persuaded to leave the field after dark, a French king in flight from his own defeated army.
\nCrécy's death toll was lopsided, but its significance lay less in the casualties than in what it demonstrated. A force built around trained infantry, favorable ground, and disciplined missile fire had annihilated the mounted feudal host that had dominated European warfare since the eleventh century, and it had done so decisively rather than by lucky accident. The lesson was not lost on Edward's contemporaries, though it was absorbed unevenly: France would relearn it at Poitiers a decade later and again at Agincourt in 1415, while the system of statutes, compulsory practice, and imported yew that produced English archers kept supplying the men who made the tactic repeatable. Crécy also opened the road to Calais, which fell to a yearlong siege in the campaign's aftermath and remained an English possession for two centuries, giving England a permanent foothold on the continent.
\nThe battle's deeper cause reached back into the tangled inheritance disputes that had bound the English and French crowns together for generations. The dynastic marriages that knotted the Plantagenets to the French royal house gave Edward a claim worth fighting for, however weak in strict law, and the war it triggered would not end for over a century. French military culture, meanwhile, remained wedded to the mounted charge and the great fortified castle for generations after Crécy exposed their limits in the open field, a conservatism visible even in the fortress-building traditions of thirteenth-century French castle architecture that predated the battle's lessons about static defense against missile fire. The warhorses bred for shock combat that carried Philip's knights forward at Crécy were, in a sense, casualties of the battle too: the animal that had defined a class and a way of war proved unable to close the gap that arrows had opened.
\n| Date | \nEvent | \n
|---|---|
| 12 July 1346 | \nEdward III lands his army at Saint-Vaast-la-Hougue in Normandy | \n
| 26 July 1346 | \nEnglish forces storm and sack Caen | \n
| Mid-August 1346 | \nThe army is pursued toward the Somme, finding every ford blocked | \n
| c. 24 August 1346 | \nEdward's army crosses the Somme at the tidal ford of Blanchetaque | \n
| 26 August 1346 | \nThe Battle of Crécy is fought; Philip VI's army is routed | \n
| 4 September 1346 | \nThe Siege of Calais begins | \n
| 3 August 1347 | \nCalais surrenders to Edward III | \n
Crécy is remembered as a longbow story, and rightly so, but the weapon alone did not win the battle. What won it was the discipline to dismount proud knights and hold ground, the judgment to choose a slope that funneled the enemy into arrow range, and an army built by law and practice to deliver massed missile fire faster than an opponent could absorb it. France would take decades to relearn how to fight the English on these terms, and by then the Hundred Years' War had already been reshaped by a single August afternoon in Picardy.
\nWhy did the French lose despite outnumbering the English? Philip VI's army combined feudal levies, allied contingents, and mercenaries with no unified command, and its knights charged repeatedly and piecemeal rather than in a coordinated assault, riding down their own Genoese crossbowmen in the process and into a fixed defensive position they never dislodged.
\nDid gunpowder weapons decide the battle? No. Edward III likely deployed a handful of primitive cannon at Crécy, among the earliest recorded use of gunpowder artillery in a European field battle, but their technology was too crude to inflict serious casualties; the outcome was decided overwhelmingly by longbow fire and terrain.
\nWhat happened to Calais after the battle? Edward marched directly from Crécy to besiege Calais, which held out for nearly a year before surrendering on 3 August 1347. The town became an English possession and remained one for over two hundred years, giving England a permanent continental bridgehead.
\nHow does Crécy compare to Agincourt? Both battles pitted English longbowmen against much larger French forces on ground that restricted cavalry, and both ended in lopsided French defeats. Crécy came nearly seventy years earlier and set the pattern that Poitiers in 1356 and Agincourt in 1415 would repeat, though French commanders were slow to adapt despite the recurring lesson.
\n