The Celtosceptic Dawn

The narrative crumbles

Already in the 1960s this Celtic narrative was being challenged. The oft-cited ‘empire’ suggested by Hallstatt and La Tène was supported by no other imperial settlements. There was no evidence of people in Austria/Switzerland speaking anything like Celtic. Nor did historians regard it as plausible for entire peoples and languages to be obliterated by D-Day-style invasions, landings and genetic replacements. Ancient civilisations did not evolve in this catastrophic way.

Prehistory now became an academic confrontation. In 1966 the Disney Professor of Archaeology and Ethnology at Cambridge University, Sir Grahame Clark, was dismissive. He described the traditional Celtic narrative as ‘suffering from invasion neurosis’. He objected to ancient Britons being seen as ‘dimwits clad in druidical garb … figures of fun if not of scorn’, needing to be taught a lesson in Bronze Age civilisation by incoming Celtic warriors. The inhabitants of the British Isles did not need conquerors to help them advance in trade and technology. There was no reason to posit any assault, wipeout or swamping just because a people appeared to acquire a new language. Clark’s Oxford contemporary, the Anglo-Saxon scholar J. R. R. Tolkien, derided the whole idea as nothing but a ‘fabulous Celtic twilight … a magic bag into which anything could be tossed and anything retrieved’. It was still one of which he made full use as a novelist.

Prehistoric archaeology was long a closed shop, confined to scholars of things, of mounds, walls, pots and jewels. In the second half of the twentieth century it was transformed by the arrival of physicists, chemists, pathologists, anthropologists and linguists. The radiocarbon dating of plants and organic matter revised chronologies. In the 1990s the application of ancient DNA (aDNA) genetics to human skeletons revealed their life cycles in extraordinary detail. Bones, teeth, skin, even faeces began to tell their tales.

The dating of Britain’s oldest resident, the Red Lady of Paviland, in South Wales, was revised 30,000 years, from early Roman to before the last Ice Age. She turned out to be a Palaeolithic young man with a taste for fish. Stonehenge’s Amesbury Archer of c.2300 bc was revealed as Swiss, with a tooth abscess and a gammy leg. His partner came from Kent. Science had revolutionised archaeology.

By the 1990s the sceptics were in full cry. The anthropologist Malcolm Chapman published his Celts: The Construction of a Myth in 1992, describing them as the mere ‘others’ of ancient Europe. They filled a gap in Europe’s story with ‘an apparent continuity and substance’. The word Celt had become so obfuscated that Chapman suggested it be banned from academic discourse. A British Museum archaeologist, Simon James, in 1999 likewise dismissed the Celts as an eighteenth-century invention. Calling them a specific people was ‘a political falsification of history, dangerous in the hands of separatists’. It was as if history lay not in the mountains and seas of Britain but in the minds of feuding scholars.

In The Discovery of France, the historian Graham Robb wrote that to gather under one name the myriad peoples who once inhabited Europe was anyway a fallacy. They did not see themselves as such. They owed loyalty to a family or clan tradition and to the territory on which it had settled. As for ‘ethnic identity’, Robb pointed out that it been ‘long eroded or at least delegated to the wizards of DNA’. That certain peoples once used a shared language to help communicate with neighbours meant nothing. English speakers do likewise without becoming English. The Celts were simply a myth.

The paths diverge: Celtic from the west

Faced with this blizzard of scepticism, the Celts’ most ardent champion, Oxford’s Barry Cunliffe, did not disown them. But he published two works, The Ancient Celts in 1997 followed in 2001 by another called Facing the Ocean, which were both firmly in the revisionist line. They drew on two influential insights into the study of prehistory. The first was the idea of the longue durée associated with the twentieth-century French historian Fernand Braudel (1902–85). He emphasised the slow passage of historical change, often unnoticed over time, warning in particular against imposing on prehistory a template of events-led upheavals familiar in modern times: a high-speed chronology of monarchs, conquests, invasions, revolutions and imported technologies. In particular Braudel objected to attributing every advance in human behaviour to an external agent. This was a reprise of Clark’s ‘invasion neurosis’.

Another insight was equally significant to the saga of the British Isles. This told ancient historians not to overrate contact by land compared with contact by sea. Before the advent of the horse, water was by far the fastest and most efficient means of communication. A journey that on foot could take weeks or months could be undertaken in hours or days by even the most primitive boat. It was by water that early contact was made and languages employed to facilitate it. This helped explain the swift advances of the riverine and insular cultures of the Egyptians and the Aegean islanders compared with their hinterlands.

Sea-borne mobility round the coast of Europe rose dramatically in the second millennium bc with the burgeoning trade in copper, tin, tools, pots, axes, ploughs, swords and coins. Transport became an industry in itself. Families became extended communities that needed to converse with strangers. This concept of ‘sea-as-land’ viewed the peoples of the British Isles not as English, Scots, Irish or Welsh, but as those of the Irish Sea, the North Sea and the English Channel. The infertile limestone ridge running from the Scottish Highlands down the spine of England should thus be seen as not a bridge but a divide.

According to Cunliffe, the peoples of the west coast of Ireland would thus have felt more akin to those sailing the Atlantic and touching the shores of Cornwall, Brittany and Spain than they would to the peoples of eastern England. He noted that western burials indicating ‘elite behaviour manifest in war gear and feasting accoutrements were similar from southern Portugal to northern Scotland’. These people were not invaders or conquerors, they were descendants of the tribes that probably inhabited these shores far back in time.

Advances in genetic archaeology were now supporting such a thesis, that the mass of the population of the British Isles was probably stable throughout prehistory. Analysis of DNA from ancient skeletons published in 1996 suggested that 70 per cent of Britons alive at the turn of the twentieth century were descended from the same people as had inhabited the British Isles in the Mesolithic or Middle Stone Age. Much publicity greeted the revelation that a local teacher in Somerset and three of his pupils shared their DNA with that of a 9,000-year-old skeleton found in the Cheddar Gorge. It also appeared that their ancestors had mostly originated from a common source, the Iberian peninsula.

This strongly supported Cunliffe’s thesis that ‘a very high percentage of the British population, both male and female, are descended from hunter-gatherer pioneers who arrived before 4,000 bc, and that the Atlantic littoral zone provided one of the major corridors of movement’. This movement occurred long before the arrival of any supposed Yamnayans or ‘Celts’.

Language remained a conundrum. Mesolithic Europe would have contained an unknown number of tongues. As in today’s Amazon or Borneo jungles, land-bound peoples needed only to communicate with their village and neighbourhood. Seagoing people had to deal with strangers and needed a common tongue. This had made the Mediterranean the prime conduit for the western diffusion of what was called Indo-European, supposedly out of Anatolia (Turkey) and the near east. A second route eventually yielded the Germanic languages that spread, it is believed more slowly, to the north across the Caucasus into eastern Europe.

In 1987 Cambridge University’s Colin Renfrew set out a proposed timeline for the Mediterranean diffusion that reached Italy, southern France and Spain, as described by the Greek geographer Strabo in the first century ad. Renfrew suggested that proto-Celtic-speaking might have reached the Iberian peninsula as early as the seventh to sixth millennium bc. Here evidence was supported by pioneering research into Iberian-Celtic inscriptions by John Koch, an American linguist at the University of Wales based in Aberystwyth. This was supported by a large database of European Celtic place names and inscriptions assembled in 2006 by Patrick Sims-Williams, also in Aberystwyth. Celtic traces were concentrated in Iberia and western France but died out going east, though Sims-Williams favours France for ‘Celtic’ language origination.

Celtic as the lingua franca of Atlantic trade now took on a life of its own, detached from any necessary link to race, tribe or population migration. Cunliffe dubbed his ‘Celts from the west’ as merely users of a language that permeated northwards up trading routes from Portugal across the Bay of Biscay to the western extremities of the British Isles. He suggested this may have begun as early as the third millennium bc, associated with the spread out of Iberia of the so-called Beaker culture with the transition to the Bronze Age. In this case, Celtic might almost be called the ‘language of bronze’. To Koch, the evidence of Iberian Celtic ‘finally loosened the long-held spell of the Hallstatt and La Tène cultures on the origin of the Celts’. Cunliffe echoed Chapman in wondering if Celtic languages might better be named ‘Atlantic’.

A 2006 survey of this genetic revolution by Stephen Oppenheimer in his compendium The Origins of the British confirmed the scepticism of prehistorians since the 1960s. There was ‘no clear direct evidence, linguistic, archaeological or genetic, which identified [central European] cultures as Celto-linguistic homelands’. A steppe origin for Celtic Europe was ‘one of the last remaining archaeological myths left over from the nineteenth century’. To Oppenheimer, ‘it might almost be regarded as a hoax’. There was a Celtic group of languages, now identified as a common tongue carried north by Cunliffe’s peoples of the sea. There were no Celts, just sociable sailors.

The steppe strikes back

Genetics had certainly transformed the debate on Celtic origins in the 1990s but it did not rest. By the late 2010s laboratories across Europe and America were studying a wealth of ancient genome material pouring in from European excavation sites. Initially this confirmed that the British Isles were originally peopled roughly two-thirds from Iberia and one-third from northern Europe. Iberian DNA was close to 100 per cent down the western half of the British Isles, but so-called steppe DNA rose to nearer 40 per cent towards the North Sea.

In 2015 a survey of the origins of the ‘Peoples of the British Isles’ provided a new X-ray of Britain’s past. The survey did not undermine the earlier divergence of Iberian versus steppe origination but it did offer it a more fine-grained account of the occupation of the British Isles. Most intriguing was the mapping of genetic ‘clusters’. This agreed with a marked difference in clusters between the west and the east. But the most noticeable feature was that those in the west seemed markedly diverse, while those in the east seemed synthesised over large areas.

It was as if genes were reflective of geography, perhaps not surprisingly. The 2015 survey found that, in the west, ‘an Atlantic population that dates back to the Neolithic or even Mesolithic remains unchallenged’. These Celtic speakers appeared fragmented and isolated, forming tightly differentiated communities constant over time. They did not mix genetically or linguistically right down to the twentieth century. Other distinctions were even more specific. The Cornish clusters changed at the Devon border. The north Welsh were different from the south Welsh, Lancashire from Yorkshire and the Scottish Highlands from the Lowlands. Anyone who knows these regions today will recognise such divergences. A Devonian acquaintance told me he never felt comfortable in ‘foreign’ Cornwall, while a South Walian once described north Wales as ‘Taliban country’.

Archaeogenetics now began to reveal patterns that were anything but straightforward. The distinction of male Y-chromosomes and female mitochondrial DNA suggested separate backgrounds, possibly between horse-bound ‘alpha male’ raiders going in one direction and female slaves in the other. Then in 2018 new material from the Reich genetics laboratory at Harvard began issuing a mass of new data. This appeared to confirm Gimbutas’s thesis in the 1950s of a large-scale incursion of steppe males, her Yamnayans, across east Europe in the early Bronze Age, c.3300 bc.

More serious was the indication that this incursion subsequently led to substantial population replacement in Britain, some suggested by as much as 90 per cent. Generalisation was complicated by so much evidence coming from burial sites possibly of ‘mobile elites’, unrelated to underlying populations. This was somewhat like saying that the Amesbury Archer’s Alpine origin proved that the Swiss built Stonehenge. But it revived a flurry of controversy, reinforced by a further Reich cache of skeleton material in 2021 supporting a later incursion into Britain in c.1200 bc, possibly from France. It was suggested in addition that these migrants could well have brought a form of Celtic language with them that was to emerge as Brythonic.

The debate was abruptly blown open. There were now advocates of Cunliffe’s Celtic language reaching Britain from the ocean, but others of its arriving from the east and from the centre of Europe. In 2021 Cunliffe and Koch prepared a compendium of essays titled Exploring Celtic Origins. They accepted that much was changing. ‘What seems now to have been an age of innocence,’ they wrote, ‘is overtaken by a deluge of aDNA … a glimpse of tantalising complexity.’

It still seemed implausible that there had been a 90 per cent replacement of Britons in the early Bronze Age or in the later Iron Age of the first millennium bc. To Cunliffe this would have required ‘a genocide far outstripping the worst excesses of the twentieth century’. It was invasion neurosis gone mad. But he did admit to a new scepticism surrounding the evolution of Britons in the third to second millennium bc. New evidence was replacing old certainties, at least as far as the occupation of the eastern side of the British Isles was concerned. Prehistorians felt they were hanging on the words of lab technicians, awaiting each new email with trepidation.

The conundrum of language

What seems clear to a lay observer is that Europe in the Bronze Age of the third and second millennia bc was slowly emerging from a patchwork of hundreds of long-established tribal settlements. Links were formed by seagoing and then land-borne trade as people and their lifestyles were transformed by the arrival of metals. The population enclaves on the western side of Britain would have been hubs of this great economic awakening, possibly centred on the tin resources of Cornwall. It seems more than likely that these peoples would have become users of the Celtic-root languages spreading up the Atlantic coast with the trade in metals.

Linguists tell us that Celtic was once divided into some dozen tongues. So-called ‘continental’ Celtic may have covered northern Italy, much of France, Spain and Portugal. It has all but vanished. The six ‘insular’ Celtics still found in the British Isles and Brittany were divided into two distinct groups, to the west Goidelic, and to the east Brythonic. A reasonably robust consensus is that Goidelic travelled the Atlantic routes north from Iberia, touching the outer rim of Ireland and western Scotland. Brythonic kept closer to the French shore, crossing the Channel to Cornwall and south-west Britain and extending north to Wales and much of the north. Whatever their trajectories, Goidelic and Brythonic each fragmented into different languages and dialects, mostly incomprehensible to each other, though Cornish and Breton claim to be as close as Danish and Swedish.

It remains extraordinary that this fragmentation never healed. Centuries of English-speaking evolved a common tongue, but ‘Celtic’ never did. This was despite its core language having apparently served as a trading lingua franca, as later did Latin and English. To be ‘Celtic-speaking’ was merely a classification, never to have felt the need for a linguistic bond. To speak a Romance language is not to be a member of a Romance people. The same is surely true of ‘Celts’.

If you find an error or have any questions, please email us at admin@erenow.org. Thank you!