For most of the medieval and early modern period, King Arthur was not considered a legend. He was a king of the Britons who had reigned in the sixth century, conquered much of western Europe, and died of wounds received at Camlann. Chroniclers recorded his deeds alongside those of Alfred and William the Conqueror. Monarchs invoked him in diplomatic correspondence, and monks exhibited his bones. For roughly four centuries, his historicity was settled fact.

That consensus eroded slowly. Humanist critics questioned it in the sixteenth century. In the twentieth, scholars such as David Dumville argued that there was simply no historical evidence about Arthur at all. Yet the figure has never fully returned to the category of fiction. He occupies an unstable middle ground: too poorly attested to be history, too persistently historicised to be dismissed as pure invention.

The more productive question is therefore not whether Arthur existed, but how he came to exist as history. Following Maurice Halbwachs, we can treat the Arthurian tradition as a case study in how communities construct a usable past through the social frameworks of memory. The process runs from scattered early references, through Geoffrey of Monmouth's audacious synthesis, to modern archaeology's search for Camelot. Tracing it reveals the mechanisms by which myth acquires the texture of fact, and why that texture proves so durable.

The earliest source for the period in which Arthur supposedly flourished is Gildas's De Excidio Britanniae, a sixth-century jeremiad. It mentions the British victory at Mount Badon and names Ambrosius Aurelianus as a leader of resistance. It does not mention Arthur. Neither does Bede, writing in 731 and depending on Gildas. The silence is not conclusive, since Gildas was writing a sermon rather than a chronicle. But it establishes the evidentiary baseline: the contemporary record contains no Arthur.

The earliest possible reference appears in Y Gododdin, a Welsh elegy for warriors who fell at Catraeth. In it, one hero is praised for feeding ravens on the rampart, 'though he was no Arthur.' The line is tantalising and deeply problematic. The poem survives in a thirteenth-century manuscript. Its core may be early, but its text accumulated over centuries of oral and written transmission, and the Arthur line may be a later interpolation.

The decisive turn comes with the Historia Brittonum, compiled around 829–830 and later attributed to 'Nennius.' Here Arthur appears as a dux bellorum, a leader of battles who fought twelve engagements alongside the kings of the Britons and killed 960 men in a single charge at Badon. The tenth-century Annales Cambriae then supplies dates. Badon appears in an entry corresponding to 516. The 'strife of Camlann,' in which Arthur and Medraut fell, appears under 537.

These texts perform a crucial operation. The Historia Brittonum's battle list has the cadence of heroic verse, and several of its sites cannot be identified. The same compilation includes among its marvels a cairn bearing the footprint of Arthur's dog and the tomb of Arthur's son, whose length changes each time it is measured. Yet the annalistic format of the Annales Cambriae converts narrative into chronology. A dated entry carries an authority that a praise poem cannot, regardless of when the entry was actually written.

This is why Oliver Padel's argument has proved so influential. Padel proposed that Arthur may originally have been a figure of folklore, a mythical protector associated with wonders in the landscape, who was only later historicised, much as Irish tradition historicised Fionn mac Cumhaill. Whether or not one accepts the model, it inverts the conventional question. Rather than asking what historical kernel lies beneath the legend, it asks how a legend acquired a historical shell.

Takeaway

Chronology is itself a rhetorical form. Once a figure is assigned a date, the burden of proof quietly shifts from those who assert existence to those who doubt it.

Around 1136, Geoffrey of Monmouth, a cleric active in Oxford, completed the Historia Regum Britanniae. It traced British kingship from Brutus, a Trojan exile, to the Saxon conquest, and its climax was Arthur. Geoffrey's Arthur defeats the Saxons, conquers Ireland, Iceland, Norway and Gaul, and marches against Rome itself. He is recalled by Mordred's treachery and carried, mortally wounded, to the isle of Avalon.

Geoffrey claimed to be translating 'a very ancient book in the British tongue' supplied by Walter, archdeacon of Oxford. No such book has ever been identified. Whatever Welsh materials he drew upon, the narrative architecture is overwhelmingly his own. It is modelled on classical historiography, complete with speeches, genealogies, and a coherent chronology that synchronises British kings with biblical and Roman history. The form of history lent the content its credibility.

The appeal of this structure helps explain its success. Geoffrey wrote during the succession crisis known as the Anarchy and dedicated versions of the work to Robert of Gloucester and other magnates. He offered the Anglo-Norman elite a prestigious insular past, one that could rival the Carolingian legend of Charlemagne and that placed the Anglo-Saxons beneath an older, more glorious British sovereignty. The work survives in more than two hundred medieval manuscripts, an extraordinary circulation for a secular history.

Dissent existed. William of Newburgh, writing in the 1190s, accused Geoffrey of fabricating out of 'an inordinate love of lying.' But the tradition quickly acquired institutional anchors. Wace's Roman de Brut (1155) added the Round Table. In 1191, the monks of Glastonbury were rebuilding after a devastating fire, and they announced the discovery of Arthur's grave, complete with a lead cross naming him. Edward I presided over the reburial of the remains in 1278 and later drew on Geoffrey's history to assert English overlordship of Scotland.

This is the mechanism in its mature form. Geoffrey's text supplied narrative coherence. Glastonbury supplied material proof. The crown supplied political endorsement. Each layer cited the others, producing a closed circuit of corroboration. By the time Henry VII named his eldest son Arthur in 1486, the figure was a dynastic asset. When Polydore Vergil's sceptical Anglica Historia (1534) provoked John Leland's indignant defence, the dispute was no longer really about evidence. It was about national honour.

Takeaway

A fabrication becomes history not when one reader believes it, but when institutions such as churches, crowns, and archives acquire a stake in its truth.

Modern scholarship abandoned Geoffrey but not the hope that something lay behind him. In the twentieth century, that hope migrated from the library to the trench. Archaeology promised an independent form of evidence, free of medieval embellishment. In practice, Arthurian archaeology has repeatedly shown how legendary frameworks shape the questions excavators ask and the meanings they assign to what they find.

Between 1966 and 1970, Leslie Alcock excavated South Cadbury in Somerset under the auspices of the Camelot Research Committee. He found substantial late fifth- or sixth-century refortification of the Iron Age hillfort, a discovery of real significance for post-Roman Britain. But the site's Arthurian identity derived from John Leland's sixteenth-century identification, not from anything in the ground. The project's very name presumed its conclusion. Alcock himself later grew markedly more cautious about the Arthurian framing.

Tintagel follows a similar pattern. Its Arthurian association originates with Geoffrey, who made it the site of Arthur's conception. Ralegh Radford's excavations in the 1930s interpreted it as a Celtic monastery. Later work recast it as a high-status secular centre with extensive Mediterranean imports. In 1998, a sixth-century slate inscribed with the name Artognou was found there. The press hailed it as an 'Arthur stone,' despite scholarly insistence that the name had no demonstrable connection to Arthur.

Radford also excavated at Glastonbury in 1962 and reported evidence of an earlier disturbance consistent with the monks' 1191 dig. The finding confirmed that the monks had dug, a fact never seriously doubted. In popular reception, however, it lent a residue of plausibility to the claim that they had found Arthur. Here, archaeology did not test the legend so much as inherit its questions.

David Lowenthal's distinction between history and heritage is illuminating. History seeks to understand the past on its own terms; heritage selects and celebrates the past to serve present identity. Arthurian archaeology has often operated between the two, applying scientific method within a frame inherited from Geoffrey. A bronze figure installed on Tintagel's headland in 2016 was officially not Arthur, yet it was widely read as him. Each ambiguous find is measured against a figure whose existence the evidence was supposed to establish.

Takeaway

Method does not guarantee independence. An excavation framed by a legend's geography will tend to find the legend's period, then mistake that for the legend's confirmation.

The making of Arthur's historicity was not a single deception but a cumulative process. Annals supplied dates, Geoffrey supplied narrative, monasteries supplied relics, monarchs supplied authority, and archaeologists supplied the language of scientific inquiry. Each stage built on the credibility of the last.

What changed across the centuries was not the evidence but the needs it served. Welsh tradition sought a deliverer. Anglo-Norman elites wanted lineage, and Tudor monarchs wanted legitimacy. Post-war Britain, uncertain of its place in the world, found in Somerset's hillforts an origin story of resilience. In Halbwachs's terms, Arthur has been continually reconstructed within the social frameworks of each present.

The lesson for memory studies is uncomfortable. Arthur's persistence suggests that historicity is partly a social status, conferred by repetition and institutional investment. Of any founding figure, it is worth asking not only whether the evidence supports them, but who has needed it to.