One thousand and five hundred years ago innumerable Germans, Saxons, Angles and Jutes, came to the shores of Britain, and transformed it into England. One thousand and five hundred years ago the trunk of the English language was grafted upon a fundamentally British ethnic root. Post-Roman Britain was subject to a massive migration of Germans in the 6th century, and, the majority of the ancestry of the people of the British Isles derives from the period before the Anglo-Saxon migrations. Both these facts are implied by a recent paper published in Nature, The fine scale genetic structure of the British population (the link should work for those without academic access, though the supplement is probably really what you should read!). This paper is close to the final answer as to the question of whether Gildas was right, that the British* were driven into the sea by hordes of German barbarians, or, whether post-World War II historians were right, that change occurred through cultural emulation from elites on high. Gildas’ model of ethnic replacement was formulated in the context of his being a British cleric who was admonishing his flock, for they had clearly angered God to lose so much ground to the pagan Germans, who were on the march in the 6th century. In contrast, over the past few decades the thesis that the change in language and ethnic identity in Britain, what became England, must have occurred predominantly through cultural diffusion of Anglo-Saxon norms, has been the mainstream position. In Norman Davies’ book The Isles he wrote that the ancestor of English was of “pure Germanic character,” but also observed that “Modern genetic research is showing quite convincingly that the Germanic invasions, like the Celtic invasions before them, were insufficient to transform the existing gene pool to any major degree.” Writing in the year 2000, Davies is on strong ground when it comes to the genetics, at least for what he had on hand. But 15 years is a long time in science, and the post-genomic era has blossomed in the intervening period.
In the figure to the right you see an illustration of the genetic clusters inferred from the new Nature paper extant across modern England. They utilized fineSTRUCTURE to extract these components. If you want to understand the guts of the methods, I recommend Inference of Population Structure using Dense Haplotype Data by Lawson et al. As repeatedly emphasized within the paper the population genetic structure across the British Isles is very subtle. Another way to say this is that the British Isles is very homogeneous genetically. This is true to some extent of Northern Europe as a whole. When looking at Fst values across very distinct European populations they are quite modest. For example, the value comparing North-Wales and Kent samples is 0.002. That means 0.2% of the genetic variation in a pooled sample of North Welsh and Kentish individuals is partitioned across the populations. In contrast, a European population compared to the Han Chinese will give an Fst value of ~0.10, so that 10% of the genetic variation in a pooled sample is partitioned across the populations. With such small genetic distances it is difficult to tease apart historically informative structure with conventional methods, such as PCA and ADMIXTURE, which rely on genotypes. In contrast, fineSTRUCTURE leverages more information by looking at haplotypes, which encapsulates not just genotypes which vary across populations, but the structure across genotypes in individuals. This gives one a crisper snapshot of more recent patterns of relatedness.
But even with this method you can see that a vast swath of England proper can not be broken apart into local regions with great confidence. That suggests that there’s just little regional genetic structure to be found. The authors argue that this pattern is indicative of the fact that this was the core zone of Roman dominance, and later, of the commonwealth of Anglo-Saxon kingdoms (the Heptarchy). As such it was unified culturally and economically in a manner which likely facilitated gene flow, which prevents the divergence across populations which allow one to infer population history more easily. To make matters worse, the genetic distance between source populations on the Saxon Shore, assorted Germanic groups, and culturally Brythonic groups**, is very low. Around the year 2000 I was curious about the peopling of Iceland, and at that stage the genetics was not very robust when it came to giving definitive answers as to the contribution of Irish and Scandinavian people, because the two groups are genetically very similar unless you use genome-wide methods.
This brings up a very curious angle on the paper: because it took so long to bring into press its background framing is already a touch anachronistic. Last month Nature also published Massive migration from the steppe was a source for Indo-European languages in Europe. And it is this paper which suggests why Northern Europeans are genetically homogeneous: there was massive demographic disruption in the late Neolithic and early Bronze Age. Contrary to Norman Davies asserts above it is quite possible that massive demographic changes ensued with the arrival of the Celtic languages. And, the people who were replaced or marginalized were not the descendants of the Mesolithic hunter-gatherers fleeing Doggerland, but what David Reich’s lab would term “Early European Farmers” (EEF), a compound of populations from West Asia bearing agriculture and a group of Mediterranean hunter-gatherers.
Within the supplements the authors mention that there is a curious enrichment of Spanish-like ancestry in Northern Wales. They state that “while our data supports some low level of ancestry from southern France/Spain in ancient British populations it is hard to reconcile with major contributions to modern British ancestry from these regions.” While the dominant cluster in England is ~1 percent Spanish-like, in certain Welsh groups it groups it goes above 5 percent. Though the authors are good in other areas to clarify that their clusters are not “real”, but reifications, that is, abstract mappings upon genetic variation which is distilled and concentrated in a human digestible form, it gets muddled here. There is a good amount of ancient DNA to suggest that the first farmers across much of Northern Europe genetically resembled modern Southern Europeans, due to the demographic migration and expansion of farmers from those regions. Later, a second wave of migrants from the east, possibly Indo-Europeans, replaced and assimilated a great deal of this component, introducing the exotic “Ancestral North Eurasian” (ANE), as well as reintroducing higher fractions of hunter-gatherer ancestry. So the elevation of the Spanish-like cluster in western Britain may be a function of the fact that the Indo-European newcomers had less of an impact in those regions.
How does this dovetail with the idea that the original hunter-gatherers of Europe had a modest impact on the modern genetics of this region? One has to go back to the start of the Holocene. A conventional framework dating back decades is that after the last Ice Age Europe was repopulated by a single group of hunter-gatherers who had retreated to “refugia” in the south. This group, expanding rapidly across an empty landscape, had gone through long periods of low effective population, and was genetically homogeneous. Before the arrival of the farmers from the ancient Middle East then the whole of Europe from Spain to the Urals, fading into Siberia, may have been inhabited by a people who exploded out of their Mediterranean fastness. This explains the extremely high representation of Y chromosomal haplogroup I and mtDNA haplogroup U5 in ancient hunter-gatherer remains across the continent. The arrival of intrusive farmers on the southern and eastern edge of the continent from West Asia resulted in a synthetic population, where the farmers amalgamated with a group of European hunter-gatherers, to produce the “Early European Farmers,” EEF. Expanding in a rapid explosive sweep across Europe this group brought both West Asian and hunter-gatherer ancestry across the continent. Later the proto-Indo-European group was formed via amalgamation between disparate elements, and including the far eastern branch of European hunter-gatherers.
With that model in mind, what this paper using the PoBI data set has done is map out the pattern of variation as generated predominantly by the EEF and Indo-Europeans, as well as a later contribution of Germanic people, who themselves are compounds of EEF and Indo-Europeans! But it seems to frame the results in the context of the older model, whereby there is a much larger role for cultural change from hunter-gatherer to agriculturalist, and modern Northern Europeans are the descendants of hunter-gatherers who were long resident in their present locales. Using that lens the people of North Wales are depicted as the most pure descendants of the first hunter-gatherers to arrive in Britain, rather than the group with the greatest affinity to EEF.
This is a minor matter, but, it highlights the fact that this paper took a long time to come to press. It was rather notorious in fact within the human population genetic community for how tardy it was. Joe Pickrell points out to me that the PoBI paper in Nature was submitted in November of 2013, while the Indo-European migration paper was published one month earlier, but submitted in December of 2014. The Lazaridis et al. preprint which has reshaped much of the current thinking on the peopling of Europe during the Holocene was put on biorXiv in December of 2013.
But the biggest headline finding of this paper, that 10 to 40 percent of the ancestry of the English who are distributed across the expanse of southern and central England, have ancestry derived from German barbarians who arrived in the 6th century, is not much affected by this strange lacunae due to the vicissitudes of publication which I cover above. Reading through the supplements I am moderately confident that this value will stick. The issue that the source populations here are rather close genetically pops up even with the methods and data they have at hand, but it seems likely that they’re converging on the correct value. It would explain why older less powerful methods gave conflicting results.
It also illustrates two facts. First, the English are genetically more like their neighbors of the “Celtic Fringe” than they are like the Germans over the sea. Second, the English nevertheless have a substantial dollop of ancestry which indicates a genuine folk wandering occurred of massive proportions (at least by pre-modern standards) across the North Sea after the fall of Rome. The results are in perfect alignment with Peter Heather’s argument in Empires and Barbarians, where he suggests that German tribes who burst into the interior of the Roman world in the last centuries of the Empire, and then took over huge areas after the fall of the Western Empire, were coherent national-tribal entities. This is in contrast to the thesis that they were ad hoc social constructions of motley mercenaries of small numbers, invented de novo in the wake of the emergent post-Roman order. Heather never argues for a predominant replacement of indigenous populations anywhere, rather, he makes the case that many of these migrations were substantial, including women and children, and can be understood in nationalistic terms.
Though the modern English are not predominantly German, ancestry fractions on the order of 10 to 40 percent indicate considerably heft to the migration. And, it serves to explain to us I think why the archaeology and cultural history indicate a rupture of massive proportions, reflecting in fact the apocalyptic tones of Gildas, rather than the more sanguine theses propounded by the doyens of the study of Late Antiquity and the evolution of the Roman world into the medieval one. Gildas was a highborn man, and this holds the clue to why and how east and south Britain became England. The whole of the post-Roman Brythonic elite of these regions was defenestrated in a manner reminiscent of the flight of the earls, except more significant by orders of magnitude. In contrast in what became Francia the German newcomers did not totally marginalize the Roman elite, who retained influence in the Church, and, temporal power across the domains of the lenga d’òc. The Frankish elite in Neustria quickly became Romanized. Why did the German leaders in Britain not accept the norms of the Romanized Brythonic elites in analogous manner?
The ultimate causes can be chalked up to historical contingency. But the large genetic contribution suggests that the German nobility could recreate a Saxony-over-the-Sea in toto. There was no need to abandon the old gods and old ways, because their reshaped their new land in the imagine of their old one. The archaeological evidence seems to be that the agricultural system of post-Roman Britain was radically transformed, indicating wholesale transfer of skills and communities across the North Sea. But for me the key issue is that the Christian Church collapsed in eastern and southern Britain, only to reappear around the year 600 under both Brythonic and Continental missions. Despite the influence of the Celtic Church in the early decades, English Christianity was not an organic outgrowth of a religion which was submerged in the intervening century. Rather, it was a fresh planting of what had died. I have made an analogy before of what happened to Christianity in Britain to what happened to Christianity in the Balkans. While a regression occurred in post-Roman Gaul, what became Francia, it pales in comparison to the cultural devolution and atavism in Britain and the Latin-speaking world of the Balkans.*** It is fashionable in some quarters to declare that the Christian Church saved European civilization after the collapse of Rome, that the Church was the ghost of Rome. There is some truth in this, but, the example of Britain and the Balkans suggests to me that institutional and formal religion of the sort which we see in Christianity necessarily needs a minimal level of social and economic complexity, and concomitant “buy in” from the elites. Without the support of the powerful these sorts of institutional religions decay rapidly back toward primal animism and folk paganism. The old gods of the Celts and Romans were memories, but the new gods of the Germans were living and vital. It was a natural fit for the small scale economies which arose in the post-Roman landscape of proto-England.
Gildas seems to have been wrong, or, frankly simply lying as to the facts, when it comes to the British folk as a whole across what became England. They were still there, centuries after Gildas flourished. The Law of Ina indicate this clearly. But, they were not the great and powerful, the heirs of Gildas’ hero, Ambrosius Aurelianus. The substantial genetic impact of the Germans probably did not occur simply through a few generations of replacement. Rather, the majority local population may have been marginalized to such an extent that they did not replace themselves, and the German gentry may have experienced downward mobility in Malthusian conditions. The land that they expanded into was not empty, but, without the complex institutional supports of the Roman system which was maintained in degraded form by the Brythonic warlords, many fled or diminished. The same could be said about the Balkans, where the Roman system collapsed with the predations of the Huns in the early 5th century. As outlined in The Geography of Recent Ancestry Across Europe, the Slavs did have a major impact across Eastern Europe across this period, as their culture replaced what had been there before. But in the Balkans these Slavs seem to have absorbed a considerable number of the local population if genetics is a guide, those which were previously Latin or Illyrian in ethnicity and identity. And just as in Britain the Christian religion disappeared, only to reappear only in later centuries due to missionary activities.
If Gildas was talking only of the Brythonic elites, his language may in fact not have been much of an exaggeration. The collapse of the Christian religion in post-Brythonic Britain-becoming-England also highlights to us the difference between modern mass culture, and that of the ancient and medieval world. The Roman Empire was nominally Christianized by the 5th century. Though there were elite holdouts among the high aristocracy and philosophers, they were a waning force. The Roman identity transitioned to a Christian identity. To be Roman was to be a Christian, Romanitas sanctified by Christ. To be a barbarian was to be a heathen (or, heretic in the case of Arians). The populace as a whole became Christian, because they followed the cultural identity of the elites. But, it can be argued that until the Reformation period the peasantry of Europe to a great extent were de facto pagans. Mass religious identity only took root with the spread of literacy, and confessional competition induced by the emergence of Catholic-Protestant divisions. The principle of cuius regio, eius religio, allowed for the solidification of confessional-national boundaries under the leadership of ruling elites in the 16th century. But a century later the conversion of the House of Hohenzollern to Calvinism from Lutheranism would not result a change in the dominant religion of their subjects. In England the Catholicism of the Duke of York rendered him unfit to rule in the eyes of his subjects, who welcomes the Protestant William and Mary (note, here “subjects” probably really means the nobility and gentry). In the 18th century the previously Protestant Electors of Saxony converted to Catholicism. And yet their domains remained almost wholly Lutheran.
The British peasants who lived under the rule of German elites in 6th century England were likely often descended from nominally Christian ancestors. But Christianity during that period was a religion of the mighty, at least in its full liturgical grandness. That was the legacy of Constantine. Marriage was solemnized in the Church for elites, not the peasants. Though the peasantry no doubt had a vague Christian identity, their understanding of the details of the faith were modest at best (we know this because there were debates by Churchmen whether ignorant conversions were true conversions). There was often a baptize-first and inculcate-later policy, as indicated by continued persistence of folk pagan practices across Christian Europe among the peasantry which had to be suppressed by the Church. The removal of an elite which would patronize the Church resulted in its withering, and the reversion of the peasantry which remained to a pagan identity. Consider what happened to the Secret Christians of Japan, who were more obviously zealous than Late Roman peasants likely were, but nevertheless could only maintain a syncretized religious identity.
These results show that a minority, but dominant, population can nevertheless culturally replace a majority in an inferior position. Many of the aspects of a given culture, such as institutional religion and confessional identity, are predicated on particular economic and social preconditions, which once removed result in very rapid change. Additionally, the fact that modern English has little Brythonic Celtic or Latin**** influence from this period (the Latin influence came with the Normans) indicates that upward and horizontal assimilation could exhibit only marginal reciprocity. The majority fraction of non-German ancestry in southern and eastern England could also be a function of later gene flow, equilibrating across the broader English realm.
The caricatures of extreme genetic and culturalist positions in this case misled us, and only resulted in confusion. A true understanding of the dynamics of post-Roman cultural change must take into account the results from genetics, which are the best demographic data we have in light of the collapse of the tax collecting apparatus of the Late Roman state.
* When I say British here I mean the Brythonic speaking Celtic and post-Celtic people who inhabited the British Isles before the arrival of Germans.
** I am being vague here because there is not great clarity whether the people of Roman Britain were Latinized by this time, though they retained some sort of distinctive identity clearly since they reemerged in other parts of Britain as a post-tribal Celtic people.
*** Latin, not Greek, was the dominant language of the hinterlands beyond the coast across the Balkans during the period of the later Roman Empire. The existence of Romanians and Vlachs is a testament to this.
**** A nod to those who argue that the post-Roman British world was one where the peasantry spoke vulgar Latin, as was the case in Gaul.