The received account blames the longbow and gunpowder. Neither explanation survives contact with the evidence.

Infantry beat knights long before firearms mattered. Flemish townsmen destroyed a French cavalry army at Courtrai in 1302. English longbowmen did it again at Crécy in 1346 and at Agincourt in 1415 — but so did Swiss pikemen, and so did dismounted men-at-arms, and the decisive factor was usually discipline and ground rather than any single weapon.

The revealing detail is that plate armour reached its technical peak in the late fifteenth century, precisely as the knight's battlefield dominance ended. Armour did not fail. What changed was economic and political: states could afford standing professional armies, and no longer needed to grant land in exchange for service. Knighthood survived the change, but as a title, a court honour and a chapter of heraldry rather than a way of making war.

There is no medieval document titled "the code of chivalry". What we call the chivalric code is a reconstruction, assembled from three kinds of source that often disagree.

The romances came first: Chrétien de Troyes in the 1170s and 1180s gave Europe Lancelot, Perceval and the Grail, and with them an ideal of the knight as lover, penitent and seeker. Then came the didactic treatises — Ramon Llull's Book of the Order of Chivalry around 1275, and Geoffroi de Charny's Book of Chivalry in the 1350s, written by a working soldier who died holding the French royal banner at Poitiers. Finally the heralds and chroniclers, Froissart above all, recorded what knights actually did, which was frequently at odds with what the treatises said they should do.

Read together, the three tell you something more interesting than a code: they tell you what a violent warrior class wished to believe about itself.

That effort was deliberate. From the late tenth century the Peace of God and the Truce of God tried to limit private warfare, forbidding attacks on clergy, peasants and merchants, then forbidding fighting on certain days altogether. When the First Crusade was preached in 1095, it offered a further solution: send the violence elsewhere and make it meritorious.

The military orders were the logical endpoint. Founded around 1119, the Templars and their successors produced something genuinely new — the monk-knight, bound by vows of poverty and chastity, whose vocation was to kill. Medieval writers found the combination as strange as we do, and argued about it for centuries.

Tournaments, meanwhile, were repeatedly condemned. The Second Lateran Council banned them in 1139 and refused Christian burial to those who died in them. The ban failed completely.

The path was long and began early. Around seven, a boy of good family left home to serve as a page in another household, where he learned to ride, to serve at table and to read a little. Around fourteen he became a squire, attached to a knight, responsible for his arms and horses, and increasingly present on campaign.

Dubbing came in the early twenties, when it came at all — many squires never advanced, for lack of means. The ceremony evolved from a purely military gesture, the colée, a blow to the neck or shoulder, into an elaborate liturgy: a night's vigil, a bath, a blessing of the sword on the altar, an oath. The Church spent two centuries turning an act of war into a sacrament.

The knight begins as a job, not a rank. In the eleventh century a miles was simply a man who fought on horseback for a lord — often of modest birth, sometimes barely distinguishable from a household servant with a sword. Only across the twelfth and thirteenth centuries did knighthood fuse with the aristocracy and become hereditary, closed and prestigious.

What made that fusion possible was money. A trained warhorse, a mail hauberk, a helm, a shield, a sword and a lance represented a fortune, and the horse had to be replaced. Maintaining a single knight in the field required the revenue of an estate, which is why knights were always few: an army of several hundred of them was a serious force.

The great collections are now open. Gallica, run by the Bibliothèque nationale de France, holds the Arthurian cycle and the Book of Tournaments of René of Anjou. The British Library has digitised much of its illuminated collection, including the fight manuals and the bestiaries. The Metropolitan Museum publishes its arms and armour holdings under an open-access licence.

The material is free and extraordinary. It is also scattered across a dozen interfaces, undescribed, and mostly uncatalogued in any language a non-specialist reads — which is the gap this book was made to fill.