How the English abolished their British (Celtic) ancestors

Reading both Bryan Ward-Perkins’ 2000 paper Why did the Anglo-Saxons not become more British? and The fine scale genetic structure of the British population, published in 2015, is interesting. To date, this second paper is probably the “best of breed” when it comes to estimating Anglo-Saxon admixture into the British population in the 5th to 7th centuries (confirmed with a smaller sample ancient DNA publication). The authors conclude that:

Two separate analyses (ancestry profiles and GLOBETROTTER) show clear evidence in modern England of the Saxon migration, but each limits the proportion of Saxon ancestry, clearly excluding the possibility of long-term Saxon replacement. We estimate the proportion of Saxon ancestry in C./S England as very likely to be under 50%, and most likely in the range 10%-40%.

The ancient DNA paper gives an estimate of ~38% Anglo-Saxon (German) for the “East English.” So the two seem roughly in line. The C./S. England cluster refers to the genealogical network of the lowlands of central and eastern England.

There are several ways we can look at this. First, the majority of the ancestors of the modern English were British. That is, Brythonic people of various levels of Romanization. They became Anglo-Saxon. Even on the “Saxon Shore” in the far east of England it is likely that the majority of the ancestors of the natives derive from post-Roman Britons (if barely).

A second way to look at it this that this validates Peter Heather’s model in Empires and Barbarians: The Fall of Rome and the Birth of Europe. The model being that the post-Roman barbarian migrations were coherent “folk wanderings,” and large numbers of Germans moved into the collapsing Roman Empire. In post-Roman Britain, a large number of Germans clearly arrived and demographically marginalized many Britons. To be sure, it is unlikely that in the year 550 AD the census size of Germans to Britons in East Anglia was ever 38 to 62 in ratio. Rather, I suspect that in the centuries after the rise of the Anglo-Saxon kingdoms German elites had higher reproductive rates than the Britons due to their superior access to resources. Over time this resulted in their contribution being more prominent in the genealogies of people alive today.

Ward-Perkins’ paper is an interesting complement because he comes to roughly similar demographic conclusions well before the DNA evidence was available (though he does allude to ancient DNA!). He reiterates that the English seem to have subjected their British (i.e., Brythonic) heritage to the dustbin. English notoriously has very few loanwords from Celtic languages, and the Latinate that it does have seems to have arrived later. Not from sub-Roman Britons (who likely exhibited some level of bilingualism in Latin, which their elites still patronized for a period). Second, the Christian religion, which seems to have been well established when the Empire abandoned Britain, went into sharp regression in Anglo-Saxon areas. Third, despite attested instances of intermarriage between British and Anglo-Saxon elites, the royal genealogies of the English and the Welsh remained very distinct. Even the House of Wessex, which has clearly Celtic names early in the genealogy (they may have been Britons who assimilated to a Saxon identity), claims descent from Woden!

The conclusion from Ward-Perkins’ paper, and confirmed now by DNA, seems to be that the post-Roman Christian Britons assimilated to a German pagan identity. Through comparative methods, he makes suggestions for why this might have happened, and how it happened. Ward-Perkins observes that the Germanization of Slavic peoples during the medieval period exhibited analogs to the Germanization of the British: there are very few Slavic loanwords in eastern German dialects. He also suggests that the Islamicization of the Near East and North Africa may present an analogy, as subjugated peoples escaped their subordinate positions through religious conversion. A milder version of this also occurred in Frankish Gaul, where Gallo-Roman aristocrats donned trousers and adopted German forenames.

In short, Ward-Perkins seems to be arguing that the strongly subordinate position of the Britons due to their differences in culture from the German ruling class motivated them to shift their identities. The reality is that in the Late Antique period much of Europe was superficially Christianized in the first place, so adopting paganism may not have been that difficult for semi-literate warlords, let alone the peasantry.  The spread of Slavic identity in the Balkans (among with migration), the Magyarization of much of Pannonia, which had previously been Slavic and Latinate, are examples within Europe of this pattern.

But, there is another dynamic between Briton and Anglo-Saxon that Ward-Perkins thinks is salient: the centuries of vicious battles between the two groups. With hindsight, the slow defeat of the Britons and their marginalization to the “Celtic Fringe” seems inevitable, but as late as the first half of the 7th century A.D. this was not clear. Cadwallon ap Cadfan, king of the northwest Welsh kingdom of Gwynedd, conquered Northumbria in 633, so that rule by Britons all the way to the North Sea was a reality for a short period. While in Gaul the Franks quickly conquered the whole region and began integrating the Gallo-Roman aristocracy into their own power structure, in Britain, the native population retained independent geopolitical power for centuries, and so were perceived as a clear and present threat. Which they were deep into the early medieval period. To be a Briton in Anglo-Saxon territory then was to be behind “enemy lines.”

Because we have good genetics for this region we can quantify the likely genetic contribution of the new culture that came to dominate the land of Britain. What became England. But also due to the interest in archaeology in England, we know that there was a wholesale human geographic change in this territory. The manner in which peasants tilled the land transformed with the arrival of the Germans. In other words, a wholesale transformation of folkways, and the adoption of an alternative ancestral framework. The nominally Christian Britons having taken up tribal German paganism with the shift in language was the least of it.

There is a narrow and specific insight here. The unwinding of the complex Roman state, which was pacified and economically complex, and centered around cities, and the switch to a more militarized and small-scale, “primitive”, the social system almost certainly allowed for the adoption of German folkways by sub-Roman Britons. As the villas of yore decayed and collapsed the semi-Romanized inhabitants of Britain discarded their Classical raiments and moved into the misty forests of their primordial tribal past. To survive the shock of the German folk migration the Britons reverted to a way of life more akin to their pre-Roman ancestors, including the resurrection of hillfort culture. But in the process, the ability of the post-Roman Britons to absorb the Germans into an aspirational “high culture,” the fragments of Romanitas, disappeared. The Anglo-Saxons saw in their British enemies just another coalition of tribes as themselves. And so it reverted to a Hobbesian scenario. And here the British lost.

There is also a general insight. How can a people “forget” their past origins? How can they create entirely fictive genealogies? As economists would say: incentives matter. Many human populations emerge through a process of genetic amalgamation, but cultural identity is not governed by the laws of segregation. Cultural identity and memory can rupture and shift far more rapidly because the laws of cultural inheritance are more plastic and protean. There was a clear folk migration of massive numbers of Germans into what became England, but their history and folkways were adopted wholesale by the native peoples whom they conquered. This sort of process likely has occurred many times across human history.

21 thoughts on “How the English abolished their British (Celtic) ancestors

  1. One of the things pointed out in “Empires of the Word” is that the adoption of the English language by Britons is an unusual example of a Germanic volkwanderung imparting their language on a subject people. Frankish Gaul wound up speaking a derivative of Latin rather than German, as did the areas of Italy conquered by the Lombards. When Vikings conquered Normandy, they picked up the French language of their subjects, and then when the Normans conquered England the same process happened again, this time with their French eventually being replaced by their subject’s English. It seems that the Celtic languages didn’t tend to persist, generally being replaced by Romance languages (as in the case of Gaulish to French). The collapse of the Roman empire didn’t result in them reverting to Celtic or switching over to German for the most part.

  2. English geology is probably important here too. The productive agricultural land is South and East of the Tees-Exe line. Once the Anglo-Saxons were able to seize or control the most productive lands (and harry or threaten the ability of the British to farm the rest), the conquest was essentially complete. But the mountains and hills to the North and West would be immune to centralized control and remain refuges for peripheral kingdoms and sources of raids (at least so long as the Anglo-Saxons were unable to employ an occupying army as in Roman Wales and Northumbria).

    Still, the small paper’s estimate of 30% Anglo-Saxon ancestry in Cornwall, Wales and Scotland seems too large (and the paper recognizes a large spread), but it is suggestive of a large Anglo-Saxon migration coupled with better access to resources.

  3. tggp, the ratio of conquerors/to conquered seems to have been much closer in britain. probabky 10x greater for conqueror-class (most of these had to be peasants in their day to day life) than in continental europe.

  4. Was religion a mediator ? In Gaul, the Franks converted to Christianity very early, which linked them to the Latin “high culture” you mention. Same for the Vikings in Normandy. Perhaps that explains why they eventually assimilated, with little impact on the local languages.

  5. toto, frankish groups were part of the roman scene liminally for several centuries as federates. so i don’t think it was “early.”

    in general i tend to take religion as an indicator variable, not causal. the frankish conversion was part of the broader allure of high culture, which included nonchristian antiquity (the frankish general arbogast was a greo-roman hellenist, not a christian, around 400 AD, and clovis’ father had roman titles).

  6. I’ve heard there were studies that showed much greater Briton ancestry on the female than male line in current English genetics. Is that accurate?

    ‪Since it’s thought Christianity in Roman Britain was concentrated in the towns, might they have abandoned towns in Anglo-Saxon dominated areas in favor of the Celtic fringe?‬

  7. Razib: very interesting. Has Patrick Wyman seen this and reacted to it?

    BTW: I couldn’t get anything from the link to Ward-Perkins paper.

  8. Razib,

    Not sure if my memory is correct, but doesn’t what you described for late antique Britain could also apply to the regions of Gandhara (now Pakistan) and Bengal (now Bangladesh) under the Islamic conquests of the 11th to 13th centuries? i.e. modern inhabitants there no longer preserve any remnants of their once rich pre-Islamic culture.


  9. riordan,

    the problem there is most bengali muslims are aware their ancestors were hindu if you ask them enuf. for many parts of my family they even know when they converted and take pride if it happened earlier 😉 and bengali hindus still exist and are a witness to the dharmic culture of bengalis.

    punjabis are a little more in denial, but even they will usually admit that various tribes in that region of all religion share common ancestry.

    I’ve heard there were studies that showed much greater Briton ancestry on the female than male line in current English genetics. Is that accurate?

    yeah, Y chromosome work from mark thomas. no follow up, but seems plausible.

    ‪Since it’s thought Christianity in Roman Britain was concentrated in the towns, might they have abandoned towns in Anglo-Saxon dominated areas in favor of the Celtic fringe?‬

    wark-perkins points out st. alban cult in kent indicative of celtic christian continuity in that area. so some memory of christian customs persisted. but yes, christianity in the roman world was an urban religion. when urban civilization collapsed, so did christianity.

  10. Razib/Riordan:

    When it comes to India, perhaps the Aryan arrival in the Indus Valley is a better parallel to the Anglicization of Britain, compared to the Islamic conquests? Or are the dynamics of the Aryan invasion/immigration in India totally different in your view?

  11. History = his story.

    Story can be true or fictional, which carries a lot of subjective components. Genetic analysis is objective data which is hard to fabricate.

    I was indoctrinated by my family as pure blood Northern Han who are proud of people as origin of Han culture. I believe the family tale without any doubt until modern ancestry genetic analysis comes. DNA analysis indicates my 90% mongol ancestry.

    When I studied the hometown village history, it all made sense now. The very village was established right after fall of Mongol Yuan dynasty. Many Mongol soldiers served under Ming (Han) emperor were settled in the region and converted into farmers as Han people. Mongol names and intermarriage were forbidden under Ming laws.

    Mongols are not very ethnic centritric in the first place. We are happily embracing inter-ethnic marriage and new ethnic identity.

  12. When it comes to India, perhaps the Aryan arrival in the Indus Valley is a better parallel to the Anglicization of Britain, compared to the Islamic conquests? Or are the dynamics of the Aryan invasion/immigration in India totally different in your view?

    15% of east bengali ancestry is east asian. no memory.

  13. Long time reader here Razib. I know this is completely off-topic so sorry but I don’t trust anyone else on the blogosphere…I would really appreciate if you could enlighten me as to why most published work about the African genetic landscape and African-Americans fails to identify what you did in your ‘africa in 12 admixture chunks’ piece a while ago. Why are Yorubas consistently used as THE modal West-african population when Yorubas are clearly 60-65% ‘bantoid’ or west-central african compared to the almost fully west-african Mandenkas [and assumed mande peoples all together]. Geographically your original piece makes the most sense. Is this some sorta researcher bias wherein a greater affinity is projected between Nigerian populations and African Americans?

    Also why is it that in eastern + southern bantu speakers [luhya west kenya/zulu etc] the mandenka-oid component found in western bantu speakers like the fang and kongo, is seemingly swallowed up by the nilotic/san component? this shit makes no sense as presumably the original bantu peoples that spread east+south also carried a minority west/mandenkoid component? why are southern nigerian pops painted by researchers as being almost alien when compared to bantoid groups…not to mention tiny fst distance between westafrican/mandenkoid and westcentral-african bantoid? Do they just not wanna admit that precise ancestry and ‘cultural’ roots for AAs is almost impossible…would kinda damage ancestry/23andmes marketing model I assume

  14. AG – LOL! My wife the same (proud ‘pure Han’), except in her case it turned out she has 10% of something totally unsuspected, at a time depth where the old family oral historians must surely have known. It really shocked her and she still doesn’t know what to think about it, but she’s pretty much over it, just puzzled. I still think 23andMe might not have nailed it – 10% of something for sure, but maybe misattribution – Mongol or Manchu would make more sense. My daughter who understands Shandong dialect tells me the old family oral historians used to mutter to each other in dark corners about some ‘Xiongnu’ ancestry in the family (so evidently regarded as something ‘shameful’), but that makes no sense at all in terms of time depth, so they must have meant something else – some other brand of northern ‘barbarian’.

  15. Razib,

    Thanks for the reply. I can’t resist a related but totally “deeper” question: how many Bengalis you know are able to remember a (very rough) time period when their ancestors used to be Buddhists (as opposed to being Hindus), and then converted to Islam?

  16. riordan, unless their ancestors were chakma (there are some) it is zero. ppl reading histories know ancestors were buddha, but the bengali buddhist tradition is very marginal. only the barua in the far southeast in chittagong. though many i think migrated to west bengal.

  17. Thank you Razib for that explanation. I guess this is another small piece to the puzzle of just how abrupt and complete the Pala Empire collapsed during the 12th century

  18. Razib –
    I think a really powerful parallel to what you lay out in this article is in Anatolia (and the Levant). It appears from the genetic evidence recovered from pre-Turkic burial sites and modern population studies, the change from the pre-11th Century population of what had been the Eastern Roman Empire to the Seljuk and then the Ottoman domain was like what you described: a religious and cultural adaptation to meet new circumstances. Most of the Anatolian “Turks” (largely devout Muslims) likely do not realize that they are actually descendants of a multi-ethnic (but not Turkish) Greek-speaking, Christian populace. The Eastern Roman Empire lived on, with new rulers and a different religious/social carapace imposed on top of it.

  19. 1) The Anglo-Saxons may always have been present in small numbers on the North Sea coasts as fishing communities (Oppenheimer). Also, the Belgae were certainly immigrating to Britain at the time of J Caesar’s incursion. Fishing villages were self sufficient and survived the plagues?

    2) The South and East have large navigable rivers penetrating deep inland. Easy to conquer/dominate if you have boats.The Severn is far more difficult to access from the sea.

    3) very few Brythonic words are in English but Welsh grammar can account for many differences between English and German,in particular verb forms. So there was some cultural interaction. Also the English word for Father – Dad is Welsh (but it is Russian for Grandad too).

    4) Different folkways seem to include A-S use of heavy ploughs (invention of horse collar). Thus A-S could expand into mostly empty land with heavy clay soil especially Mercia, the least understood place in this discussion. The Britons lost England when Mercia broke through into the Severn Valley. West of the Severn, conscripts were 2 inches shorter at the time of WW1.

    5) But muh King Arthur muh. The British cultural areas that did survive traded by sea as far South as Spain (Tintagel) and were relatively wealthy. Perhaps in these areas Romanitas survived? British Christians made almost no attempt to Christianize A-S kingdoms except Northumbria (unlike Mercia & Wessex, comparatively well recorded & slow take over). So there was a very clear divide.

    So A-S genes survived the plagues of 540 AD etc better and expanded onto heavy clay soils with support of German relatives from less favoured coastal locations.

Comments are closed.