Celts to Anglo-Saxons, in light of updated assumptions

By Razib Khan | June 23, 2011 5:40 pm

Over the past week there have been three posts which I’ve put up which are related. Two of them have a straightforward relation, Britons, English, Germans, and collective action and Britons, English, and Dutch. But the third might not seem related to the other two, We stand on the shoulders of cultural giants, but it is. When we talk about things such as the spread of language through “elite emulation” or “population replacement” they’re rather vague catchall terms. We don’t decompose them mechanistically into their components to explore whether they can explain what they purport to explain. Rather, we take these phenomena for granted in a very simplistic black box fashion. We know what they’re describing on the face of it. “We” here means people without a background in sociolinguistics, obviously.

To give an example of the pitfall of this method, in much of Rodney Stark’s work on sociology of religion (the production before his recent quasi-apologetic material) his thinking was crisp and logical, but the psychological models were intuitive and naive and tended to get little input from the latest findings in cognitive science. In One True God he actually offers an explanation for why Christian Trinitarianism is psychologically more satisfying than the starker monotheism of the Jews and Muslims, or the more elaborated diffuse polytheism which predates monotheism.  All I will say is whether you are convinced or not, Stark’s argument has some logical coherency and a level of plausibility, until you explore the literature in cognitive science on conceptualization of supernatural agents. The psychological literature as outlined in Theological Incorrectness indicates quite clearly that no matter the explicit philosophical nature of God as outlined in a given religion, cognitively the human mind has strong constraints in terms of how it represents abstractions, so that the vast majority of believers conceptualize the godhead in an invariant manner. To be more clear about it, even though Jews and Muslims are strict monotheists and some Hindus conceive of themselves as polytheists,* their concrete mental image of the divine doesn’t vary much from person to person and religion to religion. As a practical matter Hindus who may accept the reality of a nearly infinite number of gods on paper still exhibit personal devotion to only a few. Jews and Muslims who are strict monotheists nevertheless may have cults of saints and lesser supernatural agents in their mental universe. There is a difference between saying you accept the reality of millions of gods, and actually being able to mentally focus on millions of gods. The latter is not possible, and that has real world consequences, in that the concrete difference between an avowed polytheistic and monotheist in terms of mental state is minimal at most. So the psychological contrasts which Stark assumes motivate higher order social differences turn out to be superficial word games in pure cognitive terms.**

Similarly, we know intuitively what “elite emulation” means. It’s self-evident in that the mass of the population emulates the elite in terms of their folkways. But how does this really play out? The description is just a description, it doesn’t elaborate on the process of how you get from A to Z. When I try and find references in the ethnographic literature, generally what I encounter is in as an aside. What has been gnawing at me are cases like the Bulgar assimilation into the Slavic substrate, and the Magyar assimilation of their own Slavic and Latinate substrate. What distinguishes these two cases? They’re two instances of mobile populations from the western margins of Inner Asia erupting into the eucumene. Even if they were not pure horse-nomads in the vein of the Huns, they were clearly amongst the last of these class of peoples to force themselves into the heart of Europe after the fall of Rome because of their obligate male militarization and mobility. In the case of the Bulgarians all that remains of their distinctive identity as a mobile Turkic population is their ethnonym.  In contrast in the case of the Magyars they imposed their Ugric language language upon the population which they dominated. Modern Hungarians don’t seem to be any genetically different from what you’d expect based on geography. This is in contrast with Anatolian Turks, who do seem to have a minority East Asian element. The emergence of a dominant Magyar ethnicity on the Hungarian plain in the early medieval period then is clearly an instance of elite emulation if there ever was one, in contrast to the absorption of the Bulgars into their substrate. But this is just a description, it doesn’t tell us why elite emulation worked in one zone, but not in another.

The reason that the three posts above are related is that mass cultural shifts don’t just happen on the individual level. A lot of one’s world view is absorbed implicitly from socialization in a group setting. And as noted in the paper on the transmission and evolution of culture you have very little conscious and reflective understanding of how different discrete elements cohere into a functional whole. This doesn’t mean that wholesale adoption of another culture is impossible, but one has to be aware of parameters which make such a transformation plausible. Consider a few examples. There are individuals who move from one nation to another as adults, and over time they assimilate by and large to the nation to which they have moved. It is a process which takes decades, but it does occur. But this requires total immersion in an unfamiliar set of folkways, until what was once familiar becomes alien, and what was once strange becomes second nature. In contrast, if you move from one nation to a Diaspora of your own nation in another nation, then the shift in values is likely to be far less. You have critical cultural mass and can self-select a social environment which doesn’t perturb your cultural presuppositions.

Let’s shift to another example which is somewhat different in its parameters. One can argue that the culture of black Americans is not predominantly African in origin, but a melange of Anglo influences. Remember that the vast majority of the ancestors of black Americans were likely in the United States of America well before 1800. The slave community of the early republic was already indigenous, with minimal ties to the African lands from which their ancestors had been transported. But how exactly did cultural change happen so rapidly to so many? In the case of black Americans their ancestors were brought over from different regions with little in common aside from being African. It has long even been argued by some that the slave owners were careful not to allow for the concentration of particular groups amongst the set of their human property lest they mobilize based on common ethnic bonds (apparently there were concentrations of Igbo in Virginia, which planters attempted to diminish by exportation to the Deep South). The black American culture emerged as a creole culture, drawing upon common African tendencies, but also English speaking and Christian, two traits derived from the society into which they were thrown. The main exception in the United States to this tendency of only superficial African culture traits are the Gullah people of Low Country South Carolina. Their culture seems to have some genuine connection to the folkways of the Guinea coast. It is also notable that the relationship of the slaves who were rice farmers of coastal South Carolina to their owners was very different from that of the norm across most of the South. The ancestors of the Gullah people were more like serfs than slaves. They were tied to the land, but were still given the space and liberty to have a modicum of normal family and social life, which was not necessarily the case with most American slaves. Much of this is based on the raw economics of rice farming, which was not as amenable as cotton and sugar production in turning human slaves into pure labor units of value which were perfectly interchangeable and expendable. Whatever the details of the economics, it is noteworthy that the American blacks who preserved the greatest proportion of African cultural traits are those who were the least dehumanized by the grinding logic of early modern cash-crop slavery.

The point of the two examples above is that we can see plainly how individuals in some contexts absorb by and large the values of other societies. If you are extracted from your society of origin and thrown into a new context, you slowly can absorb new cultural norms, explicitly and implicitly. If you are enslaved and thrown together into an undifferentiated mass with individuals with whom you share little culturally, then a common creole culture will emerge. It  turns out that the creole is often, but not always, derived in its fundamentals from the enslavers, who often impose their language and religion upon those whom they enslave. How about other cases?

First, let’s consider the case of the Muslim world. The conquests between 650-750 pushed the dominion of Islam from the Atlantic to Sindh. Within the Afro-Asiatic zone there was a shift from non-Arab to Arab ethno-linguistic identity, concomitant with a conversion to Islam. In the Persian world there was a switch from Zoroastrianism to Islam. There are differences of detail here. We do know that Greek remained a language of administration in the Arab Caliphate into the early 8th century, at least two generations after the conquest. After the shift to Arabic the process of Arabicization of language and Islamicization of religion proceeded much further. It seems likely that the majority of individuals in the Levant and Mesopotamia identified as Arab Muslims sometime during the prime of Abbassid Caliphate, 150-200 years after the conquest. Nevertheless, substantial minorities in the Fertile Crescent zone remained Christian or Jewish, and preserved non-Arabic languages for centuries. A similar shift in Egypt can probably be pushed back by about 100 years. What you see is a gradual and synchronous shift toward identification as Arab and Muslim. Over time this marginalized the language of ancient Egypt, what became Coptic, and non-Arabic Semitic languages in the Fertile Crescent. The most recent phylogeography seems to suggest that this transformation was not purely cultural, insofar as there seem to be small, but consistent, differences between Muslim and non-Muslim populations (the Muslim populations are invariably more cosmopolitan, exhibiting signs of being impacted by gene flow across the Islamic international, from Inner Asia down to Sub-Saharan Africa).

Persia, what became Iran, is a different case. What occurred here is that there was a gradual shift in the peasantry to Islam, but by the mid-9th century there were no elite lineages which patronized Zoroastrianism. The last Zoroastrian principalities in the mountains of northwest Iran submitted by the middle of the 9th century (though there were Zoroastrian inspired rebellions drawing upon folk religion as late as the 10th century). Without elite patronage and protection Zoroastrianism seems to have lacked the institutional robustness to withstand Islamicization. The material I have seen suggests that Islamicization in the core Persian lands Islamicization occurred at around the same pace as Egypt, with a Muslim majority by 1000 at the latest. But unlike the Christians of Egypt the Zoroastrians of Iran almost disappeared, preserving themselves as relict populations in very isolated regions such as Yazd. One can give many explanations for this, but the connection between the Iranian elite and the Zoroastrian religion was traditionally very close, and once that connection was severed with the Islamicization of the Iranian-speaking elite it seems that Zoroasrianism simply withered for lack of patronage. This may have an analogy with the Church of the East, which unlike the Jacobites of Syria and the Copts of Egypt had no theological “sister churches” which were not under Muslim dominion. Interestingly the Persians did not abandon their language. One thesis proposed is that language shift was relatively easy for Syriac speakers, and perhaps even other Afro-Asiatic groups such as Coptic and Berber populations, but not for Indo-Europeans.

Another example which might give us further insight is the case of Latin America. By and large Iberian language and religion have superimposed themselves upon the region. How? In the southern cone of Latin America, reaching up into much of Brazil, it is through simple demographics analogous to the North American model. But in much of the rest of Latin America there has been a great deal of racial amalgamation. In particular, the mestizo populations tend to have male Iberian ancestors and female native ancestors. But culturally they are identified more with their Iberian ancestors than their Amerindian ones; at least in relation to civilizationally salient markers of note such as language and religion.  Why is there such a disjunction between the genetic parity in terms of ancestry and cultural skew toward their Iberian forebears? Because genes and memes have different inheritance constraints. In particular, memes are far more flexible in terms of how they transmit, allowing for asymmetric vertical transmission (identify with the culture of one parent, not the other), as well as horizontal transmission (across peer groups). Consider for example a toy example: most socialization occurs in peer groups, which develop their own norms. But, those norms themselves cue on the modal cultural pattern of the previous cohort. What does this mean concretely? That peer groups are homogeneous evolutions from the dominant, but not exclusive, culture of the parental generation. To give a real world example, American children develop their own idioms, accents, and slang. But these linguistic tendencies themselves need the starting point of the dominant language of the society. The parental generation might be 10% non-native English speakers with all the peculiarities which that might entail, but the offspring generation might be impacted not at all by the 10% non-native English speakers, because they take their cue from the 90% of parents who are native speakers. Of course given enough demographic heft immigrants can change a language. So in Argentina the influx of Italians was so overwhelmingly that it reshaped the Spanish spoken in that nation. And yet Spanish is spoken in Argentina, English in the United States, and Portuguese in Brazil, despite the fact that the majority of the ancestry in these regions may not be from Castile, England, and Portugal, respectively.

A final example I want to cover is that of the spread of the Turks. There seems a rough consensus that circa 2,000 years ago the progenitors of the modern Turkic languages was localized roughly to the region of western Mongolia and its environs. Much of western Eurasia which is now dominated by Turks was the domain of Indo-European speakers. With the expansion of the Turks between 500 and 1500 the zone of Indo-European speech from Inner Asia to the Mediterranean fragmented. By this, I mean that Indo-European did not disappear, but it was marginalized or absorbed by Turkic groups across much of its old zone. In Central Asia the Tajiks are the remnants of the dominant Iranian populations, along with isolates like the Yaghnobis who preserved themselves in mountainous redoubts. On the plains north of the Black Sea the Turks cleared out most of the Iranian populations. Groups such the Ossetians took refuge in mountainous zones as well. Of course this particular example of the spread of Turks has been somewhat erased by the later expansion of Slavs, and assimilation of many of the Turkic groups on the frontier of demographic expansion (though groups like Tatars and Chuvash remain). In Iran and Anatolia Turks and Indo-European speakers, whether it be Greeks, Armenians, and Persians, existed in symbiosis for centuries. Over time the northwestern Persian zone was fully Turkicized. But if you looked at a map of language distributions circa 1900 you’d see Turkic interlaced with various pre-Turkic Indo-European languages from the Aegean to Khorasan. In other words, the shift from Indo-European to Turkic has been halting, fragmentary, and incomplete, in many regions. But what about the genetic impact? Because Turks were originally East Asian that is not too difficult to ascertain. It seems that in Central Asia, what was once termed Turan by the Iranians, the impact has been substantial, and some cases dominant (e.g., the Kazakhs). This stands to reason because the migration of some groups, like the Kyrgyz, occurred in historical time. On the other extreme the Anatolian Turks seem to be ~5% or so East Asian. This seems a strong argument for elite emulation. But, it needs to be qualified by the fact that the arrivals of Turks to Anatolia occurred five centuries after their initial peregrinations out of Mongolia. In other words, the dilution of the genetic signal of the nomads who arrived after the defeat at Manzikert has to be kept in perspective insofar as these populations had already spent a substantial period of time amongst Iranian populations in Central Asia and Iran. That being said, it seems that Anatolia Turkish identity did emerge to a great extent out of assimilation of Greek, Armenian, Slavic, and Kurdish populations and individuals into a Turkish identity.

I review all this in detail because these are the sorts of scenarios and dynamics which go through my head whenever I attempt to evaluate inferences made from genetics, historical texts, and archaeology.  From what I can tell there isn’t a good theoretical construct for how Magyars and Anatolian Turks could be created through assimilation over the period of centuries. This makes good sense, as how many anthropological field programs could afford to track the ethnogenesis of a group over centuries? (not to mention the fact that many disciplines were only invented in the 20th century!) In lieu of a robust theory we have to make recourse to looking over the empirical distribution on a case by case basis so as to make analogies to the underlying parameters which might align across two cases. In that way we can interpolate from one known case to an unknown one.

This is especially important when textual records are thin, and the genetic inferences are shaky. So when it comes to the case of the Anglo-Saxon conquest and assimilation of the British and the emergence of an English identity I attempt to compare it to other cases I’m aware of. There are two primary issues for me in this specific instance:

- The rapidity of the shift, from a predominantly non-German landscape in the 5th century to a German one in the 7th

- The cultural regress from a “high civilization” to a “low” tribal one

In two of the cases above, the transformation of Greek, Kurdish, and Armenian Anatolia, into Turkish and Kurdish Anatolia, as well as the evolution from the non-Arab non-Muslim world to the Arab Muslim world, the shifts occurred much more slowly than the British to English one. Focusing on the core Fertile Crescent zone I think a conservative estimate is that it took three centuries to accomplish the cultural transformation which occurred in Britain in one century. In the Turkish case it took somewhat longer, and wasn’t truly completed until the population exchange with Greece and the Armenian Genocide. The large Kurdish minority also indicates its relative lack of completion, though perhaps one can make an analogy between Wales and Kurdistan in terms of being peripheral. In the case of Hungary the issue is more confusing, especially because of nationalist biases among scholars. The area of Hungary, what was once Roman Pannonia, had been barbarized for centuries before the arrival of the Magyars. It was likely populated by Slavic tribes, the remnants of the Avars, and also a the Latinate residual of the post-Roman period (what became Vlachs and Romanians across the inland Balkans). It is a key point to remember that modern Hungarians seem to occupy the genetic position which isolation-by-distance models would predict. The intrusion by a population from the Urals had very little impact. This is the perfect case of elite emulation.

There are three papers of particular interest in relation to the “Anglo-Saxon question.”

- Evidence for an apartheid-like social structure in early Anglo-Saxon England

- Is it necessary to assume an apartheid-like social structure in Early Anglo-Saxon England?

- Integration versus apartheid in post-Roman Britain: a response to Pattison

Overall I would say that Mark Thomas, who makes the argument for a major demographic impact, gets the better of the exchange. I can’t really evaluate the linguistic and much of the archaeological evidence, but the overall big picture supporting papers which argue for major continuity across Europe from the Paleolithic which are appealed to in the second paper have now been thrown into sharp doubt. I do think that using “North Wales” as a “Celtic reference” may not be optimal, but Thomas et al. basically suggest that Y chromosomal lineages are sharply differentiated between some English towns and Wales, where the former cluster with Frisian samples. In 2011 I suspect we could get better geographic coverage, as well as more precise phylogenies. Thomas’ argument is rather simple insofar that demographic parameters can be easily modulated to explain how a minority of German males could have a quick impact on the dominant Y chromosomal signal among the English. Think the “Genghis Khan haplotype” writ small.

But it isn’t the genetics which really motivates me to explore the topic of the Anglo-Saxon conquest of Britain. It is the cultural parameters. In the case of the Magyars you have a situation where the Ugric ruling caste was overlain on a predominantly barbarized substrate. By this, I mean that these populations had lost their contact with Romanitas, which by the 5th century included Christianity. The Avars and Slavs who arrived after the collapse of Roman power had never been in contact with this cultural complex, but the Latinate population of the Balkans and Pannonia had. What happened to it? We know that it persisted linguistically, because the Romanians and Vlachs are the descendants of the Balkan Latins, who contributed many emperors, from the 3rd to the 6th centuries (e.g., Justinian the Great was of Latin speaking background from Macedonia). But there is some suggestion that these populations lost their connection to Christianity in the interregnum. This is not totally shocking, insofar as Christianity’s hold on Roman identity was newfound and tenuous when the Empire withdrew from Pannonia and ceded it to barbarians.

Interestingly there is one dynamic which I do want to mention in relation to the “eastern front” of the Roman Empire, and that is the de-Germanization of much of this region between the 5th and 10th centuries, before the push during the medieval period of German settlers in their “drive to the east.” In Empires and Barbarians the historian Peter Heather suggests that these sorts of ethnic shifts were a matter of the translocation of militarized mobile elites. In other words, when the free warrior peasants atop the German status hierarchy emigrated to the lands of the collapsing Roman Empire to serve in the retinues of the Franks, Vandals and Goths, they left the more marginal Germans without a leadership class. These individuals were eventually assimilated into the simpler subsistence lifestyle cultural complex of the Slavs who were pushing in from the east, and eventually dominated by Inner Asia populations such as the Avars and Magyars. Without a mobilized elite it may be that elite emulation occurs rather easily as pre-literate subsistence level groups shift identities.

This may be at work in post-Roman Britain. In preparation for this post I read some archaeological papers. I won’t repeat what I read because it is difficult for me to make coherent sense of much of the data, except to say that there did seem to be a collapse of long distance economic ties which characterized the Late Roman Empire. This is most evident to archaeologists in terms of ceramic styles and standardization. It is known that some of the same Saxon tribes which presumably arrived in Britain also served in the armies of Theodoric the Ostrogothic warlord who ruled Italy in the 6th century. In fact there is evidence that communication networks were robust enough that when Ostrogothic power collapsed some Saxons decamped for Britain via Germany, in search of greener pastures.

What Britain may have been subject to was not the influx of amorphous Saxon hordes, but well armed and coherently mobilized groups of free warriors. And who did they encounter once they arrived on British shores? There are some conjectures that security and safety in the Roman Empire had made the indigenous British soft and weak. This is pejoratively put, but the reality is that once cultural folkways are lost, it may not be easy to reconstruct them de novo. What I’m referring to is the process whereby militarily robust groups which are “raw” are “cooked” by civilized conditions, and so rendered less easily mobilizable in subsequent generations. The armies of the Roman Empire were overwhelmingly Italian in the 1st century, but were predominantly non-Italian, with a substantial number of barbarian “federates,” by the 3rd century. A similar shift occurred in the Islamic world, where an Arab military caste became civilian rentiers, ceding ground to a Turkic slave caste, who eventually took power. Though Ibn Khaldun examined the rise and fall of asabiyyah in the Maghreb, perhaps the best illustration of this phenomenon is in Inner Asia. The Xiongnu were pacified by the Chinese not through military defeat, but gifts of luxury goods which became more salient markers of elite status than martial prowess. The Turks were a barbaric military elite across much of Asia, until many settled down and became sedentary farmers, as in Uzbekistan or in the Tarim Basin. They were easily smashed by the Mongol military formations, who were themselves only recent adopters of full-blown nomadism, having left a partially hunter-gatherer lifestyle on the Siberian fringe recollected in the insults hurled against them by their enemies. Finally, the descendants of Genghis Khan were assimilated into the Manchu Empire through promises of status and security, and turned against the more barbaric and militarily vital Oyrat Mongols of the Dzunghar Confederacy, who were not Genghisides, and were reputed to be more “raw” than the eastern Khalkha.

As the German military bands arrived on the British scene they certainly did confront a Romano-British elite, as remembered in the legends of King Arthur. But the existence of Brittany indicates that many of these individuals may not have been inclined to stand and fight, because they had other options. Like the German peasants of the east abandoned by their military elites, the British peasants, whether Latin or Celtic speaking, may have had no armed defenders who exhibited cultural solidarity with their folkways. An analogy here may be made to many Christian notables who fled the Islamic conquests for the Byzantine Empire, which offered them opportunity without abandoning their cultural identity as Christians. The emperor Leo the III may have been from such stock.

And yet as noted by Mark Thomas, and reiterated by the large genetic footprint of Turks and Mongols in Central Asia, an elite male lineage can have an outsized impact over the medium term because of reproductive differences across classes. In the modern developed world on average the highest fertilities are correlated with lower socioeconomic status. This seems to not have been the case in the pre-modern world, where those who were more prosperous tended to be more well fed, and so more likely to survive diseases. Note that despite the fact that ~1/3 of the European population reputedly died during the Black Death, only one reigning monarch seems to have succumbed. The medieval mind may have attributed this to divine providence, but it was probably more a function of higher nutritional levels which very elite individuals could take for granted.

So why am I being so long-winded in this post? Because when I write about a lot of issues I explore here I really don’t explicitly expose you or myself to the internal calculations which I’m implicitly running to evaluate probabilities. Part of it this simply that I don’t have conscious access to a lot of the internal cognitive logic, but in this case I can touch upon issues and the dynamics which are always on my mind, because it’s rather cut & cried.

To me the collapse of Christianity among Britons makes sense only if all elite Christian institutions disappeared, in particular the sub-elites which would have patronized the local parish structures from the Roman period. It looks like that without such elite instruction the illiterate rural peasantry has a bias toward reverting to common universal pagan superstitions. One can see ethnographic evidence of this from Protestant European nations, which took a greater detailed interest in the 18th century of these instances of religious devolution and heresy which tended to crop up when there was at least one generation of lack of pastoral oversight. A newly, and perhaps nominally, Christianized population in the British countryside may easily have acceded to the gods of the Germans because their understanding of Christianity may have been little different from German paganism on the conscious level (I would argue that this was the universal case among peasants before the Reformations, Protestant and Catholic, and the spread of literacy). From what I have read on the transition of English peasants to Protestantism in the 16th century is that the major counter-response to the Reformation generally occurred when the pageantry and pomp associated with late medieval religion was stripped away by Puritan Reformers. The nominal switch of institutional allegiance from Rome to the monarch was of little concern to the subsistence farmer.

So we need to differentiate between the more coherent sense of national or ethnic identity which was existent as far back as antiquity among elites, to the more local self-conception of subsistence farmers. The dissolution of Christianity can then be explained by the institutional collapse combined with the novelty of the religion even during the 5th century (Rome itself was a pagan stronghold as late as 400 A.D., Christianity having deeper roots in the Eastern Mediterranean). Here you can make an analogy with what occurred to Zoroastrianism in Iran.

So can an elite emulation model explain the de-Christianization and Germanization of 6th century Britian? I think it could, but, we need to add in a few extra parameters to our model. From above we know that in a pre-modern environment elites, especially elite males, have higher fertility and reproductive value. A relatively small and coherent German military class could quickly have an outsized genetic and demographic impact. From what I can gather the German tribes at this period practiced partible inheritance, which would have pushed the elites outward to expand their domains. The example of Latin America shows the power of consistent male migration and polygny in changing the genetic landscape of a region.

Finally, we need to go back to the insight that structures above the level of the individual matter. If Peter Heather is correct it seems plausible that the German warrior bands arrived with women. This would have preserved German culture in a relatively pure form for several generations. Here we need to go to the Turkish example: it seems that a great number of Rumelian (Balkan) Turks may have been Islamicized Slavs and Albanians who shifted to a Turkish identity. This is classic elite emulation, as an ethnic and religious change were necessary for social advancement. But, this occurred centuries after the initial admixture of coherent Turkish tribes in Anatolia with the local substrate. By the time time that the Balkans were conquered a relatively undiluted Turkic identity existed in Anatolia into which one could assimilate.

Back to the Anglo-Saxon case, what may have occurred is that a relatively unadmixed German population was rooted on the old Saxon Shore. This population recreated in totality the folkways of northern Germany. With the evacuation and emigration of the Romano-British elites to the greener pastures of the continent this German warrior caste may have expanded outward in the 6th century, and then started to absorb the remnant British elites as well as the majority peasant population into its cultural complex. In other words, for Germanization to proceed in the fashion in which it did proceed one needed a relatively pure “anchor” group into which others could assimilate. This does not seem to be likely if the German conquest was a matter of a few tens of thousands of men. If that was the case, what probably would have occurred was what did happen with the Mongols in Central Asia and on the plains north of the Black and Caspian Seas: they would have been absorbed in all but name into the substrate.

If my model is correct then the majority of the ancestry of the people in some eastern English localities should cluster with Frisians, while very little of the ancestry of English people in regions like Devon may be German at all. As I have noted before the two parental populations are rather close, and we have the confounds of migrations before and after this admixture event. Hopefully with large data sets of British and German males we can get a sense of the mutational landscape with enough depth and precision to resolve the relationship of some of these lineages.

Update: If you’re read this far, please see this very interesting response/reaction from Brown Pundits.

* I am aware that most Hindus are monistic.

** To be clear, I think there are reasons that the “god of the philosophers” has become culturally hegemonic across the world. But I don’t think that that reason has much to do with innate individual psychology. Stark operates with on a “rational actor” model in mind, so it makes sense why he’d try to frame in such a manner. But it’s kind of like trying to build biochemistry on alchemy. The lower units of organization are not coherent.

CATEGORIZED UNDER: Anthroplogy, History
  • http://worthwhile.typepad.com/worthwhile_canadian_initi/ Nick Rowe

    Nothing useful to add. But this is very impressive, Razib.

  • Brent Michael Krupp

    Yes, very impressive. Thank you so much for spelling it all out in such great detail. It fits in with much that I’ve read in the past but really illuminates and adds to it.

  • Chris

    Sorry I am probably being a bit slow, but I am finding it hard to get what your overall conclusion is. Are you saying that the majority of the ancestors of current English people date from pre-roman Britain? You say:
    “If my model is correct then the majority of the ancestry of the people in some eastern English localities should cluster with Frisians, while very little of the ancestry of English people in regions like Devon may be German at all.”
    But can you give a rough percentage of what proportion of the current genes in England as a whole is from pre-Roman Britain and what proportion is from Germany?

  • Teapot

    Well worth a long article, its a fascinating subject. You can see how “elite emulation” can work when people join a military ruling class, or when middle-easterners adopt islam for tax reasons, or when literacy starts to take off, but otherwise its very tricky. I’m always reminded of the lines in Holy Grail:

    “I am Arthur, King of the Britons!”
    “The Britons? Who are they?”
    “We all are, we are all Britons”

    In many periods the peasants will only have cared about the fact that every now and then a load of armed men (who claimed to “own” the land they farmed) would turn up and steal most of the food they’d grown. Whether these men had the same skin colour, the same religion or spoke the same language would have been a minor detail.

  • Steve C

    I think the difference between the Bulgars and the Magyars is that the Magyars were actually a large folk migration. I’ve read estimates of 400K. At the time of their migraion, they had been in the Pontic Steppes for 400-500 years and took on the general European characteristic that existed there. The elite warriors may that could have maintained a Uralic/Turkic element numbered only 20K.

    The Bulgars, on the other hand, were probably only a warrior elite.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    But can you give a rough percentage of what proportion of the current genes in England as a whole is from pre-Roman Britain and what proportion is from Germany?

    of english whites, 20 anglo-saxon, 10 post-anglo-saxon (dutch, danes, flemmish, french protestants, normans, etc.), 70 roman era or earlier. my key point is that this is an average, and there’s probably lots of variance.

  • Josh L

    Very informative and comprehensive article. I would love to read up more on how these dynamics were at play in other historical situations. In particular, I wonder what were the mechanics of Romanization in western Europe, as it is often assumed that everybody simply adopted the “superior” Roman culture and Latin language, though the truth was undoubtedly more complex. One detail which was off, though- Persia was already called “Iran” well before the Arab conquest.

  • Gav

    So, more or less what Bede said.

  • http://entitledtoanopinion.wordpress.com TGGP

    I second Rowe.

    Janes Scott wrote in “The Art of Not Being Governed” that irrigated rice farming is ideal for labor exploitation, but his focus was on southeast asia and I doubt he had cane/cotton (which aren’t food staples) in mind as a comparison.

    Speaking of which, accounts I’ve read of cane planatations in the caribbean depict it as uncommonly harsh and involving a very high death rate that needed replenishment with slave importations. Southern american cotton slaves, in contrast, underwent population growth and were exported further into the south. Does anyone know to what extent that was a matter of the geographic location vs the crop?

  • Eurologist

    I like your description of a “warrior-farmer” elite. Coming back to my idea that perhaps there already was a foothold settlement to make it all work so smoothly in the beginning, one should note two things: firstly, it was very common in Roman times that higher-level officers were given land upon retirement, and secondly, soldiers from NW German/Frisian/Dutch “tribes” were often selected to lead difficult missions in adjacent regions, including the Isles. These people were known to be highly organized and were well-versed in Roman military and leadership know-how (which later enabled them to drive out the Romans and create the Frank empire). So, many of them were very familiar with the geography and country as a whole, and I don’t find it far-fetched that some of them settled there (were given land). And they probably clustered to remain in contact. I also agree that communication pathways across central Europe seem to have been very strong at the time.

  • authun

    The small male dominated military elite model of the anglo saxon adventus model was formulated as the numbers required for a single event mass invasion are not credible. One to two million north germanics moving into England, where Millet estimated the population of Britain as 3.9 million, is simply too high a number. Yet Britons remain more or less archaeologically invisible in contrast to the ample evidence for germanic settlement in the east and the language is replaced. The question then arose, if a smaller number of germanic settlers came, what were the mechanisms by which their culture became dominant?

    Heinrich Härke proposed that Britons may be found in germanic graves and that we should not assume that all germanic grave assemblages imply the presence of germanic settlers. Furthermore he suggested that the presence of weapons in graves should not imply warriors, pointing out that these were often found in burials of infants, women and men too old to fight. They should therefore be seen as a badge of rank, an indication of the status of their family within the community.

    There are parts of England which do not produce germanic archaeology for the years 450 – 600 AD but which show virtually no evidence of Britons either. We don’t know why this is but it seems clear that production of goods, hair pins, ceramics, clasps etc either ceased or were made of of materials which were of poorer quality and which simply perished. In contrast, germanic settlements in the east show the presence of imported goods. These are not confined to jewellery or other metallic artefacts but include household items such as quernstones, sometimes imported from the Vulkaneiffel between the Mosel and Rhine and we also see differences in dietary practices.

    Identifying Britons living amongst the Germans is therefore difficult. We don’t have anything which is typically British to use an an indicator. Nor do we know the fate of Britons who were not living under germanic rule. With a few exceptions such as Tintagel in the far west, their situation appears to be one of complete social and economic collapse.

    Capelli’s 2003 study broadly confirmed Weale’s earlier study fregarding the high frequency of continental y chromosomal haplotypes in England, ie ~54% but showed that the distribution was far from even. The map in the article by Der Spiegel, http://www.spiegel.de/images/image-226190-thumbflex-fwgo.jpg has been redrawn by them but is based on one produced by Mark Thomas sometime around the publication of the Capelli study. The hotspots for continental haplotypes on this map are not, as often claimed, completely obscured by later Danelaw settlement as the most western hotspot is in the upper Thames valley, outside of the the Danelaw. Capelli’s admixture analysis suggested a presence of 70.8% of continental haplotypes for Chippenham. The map corresponds nicely with Page’s distribution of pre 650AD runic finds in England, again well before Danelaw settlement, http://cpt.co.uk/maps/page_pre650_runes.gif

    If the question of how these early settlements led to a complete cultural transition remains unanswered, so too does the question of when these early settlements started. The sort of land apportionment Eurologist refers to above was not limited to high level officers but was given to large groups too:

    276 – 279 (Zosiumus): Probus defeats alemannic insurection around the Rhine and along with their leader Igillus “sent [them] to Britain, where they settled, and were subsequently very serviceable to the emperor when any insurrection broke out”.

    372 (Ammianus Marcellinus): Valentinian sends the Alamannic Fraomarius of the Bucinobantes [an alamannic canton], along with other Alamannic troops commanded by Bitheridius and Hortarius, to Britain.

    The Romans had started a policy of resettling tribal groups within the Limes and giving them land in return for military service during times of need, known as laeti. One of the better researched groups are the Saxonnes Biocassini, or the Saxons of Bayeaux. ‘Lats’ appear as a special grouping in the early law codes in Kent and may refer to Laeti whose priviledges and responsibilities were still recognised at the start of the 7th cent.

    Heinrich Härke is publishing an article entitled ‘Estimating demographic parameters in Early Anglo-Saxon England’ in the journal Medieval Archaeology towards the end of this year. It promises to be a good read especially as it will be published at around the same time as the first publication of the People of the British Isles genetic study.

  • marcel

    The medieval mind may have attributed this to divine providence, but it was probably more a function of higher nutritional levels which very elite individuals could take for granted.

    Probably? You think? Any actual evidence to back up this statement? Until I see any, I’ll give equal weight to divine providence and higher nutritional levels!

    (Mild) apologies for the snark. This is very impressive. Trying to synthesize the genetics with other fields to make sense or figure out something about a part of history where the details are not clear is fascinating. Because of my greater familiarity with European history than that of Africa or East or South Asia, I find this very interesting.

    2 questions:
    1) In the 2nd paragraph from the end, you refer to an unadmixed German population. Does this mean one that has not (yet) crossed with another (e.g., Britain or Roman)?

    2) Your conclusion is not clear to me. Are you saying that the reason that the Britains assimilated to the Germans and not vice-versa is, most likely, that the Germans came with their families/women so were able to maintain their culture long enough to incorporate the locals? And that, in contrast, the reason that the Magyars and Turks assimilated to the local cultures was that they were basically victorious armies who, having conquered, settled down with local women and eventually assimilated into the local culture though of course they did so as a ruling caste?

    Thank you for this post and the massive effort behind it. I third Rowe.

  • PD Shaw

    It’s interesting how these theories correspond with those of John Morris’ The Age of Arthur: A History of the British Isles from 350 to 650.

    This book emphasized a point that this modern American had not grasped which is the geography of Britain. The Southeast contains the high productive land, and the terrain of the island also offers key military advantages to control the core of the island. Thus, there is a tipping point where control of the Southeast will result in control of the island.

    Morris gives credence to the Anglo-Saxon story that German warriors were invited by a British lord to defend his lands from Pictish raids. In exchange they were given land for their people to settle along the Saxon shore. These were not large settlements, but created a productive base to which later Germans would be invited (I believe in response to British efforts to push out the Germans). In the ensuing wars, a tipping point was reached that reduced the native population to peasantry or into the hills of Wales or the North.

  • chris w

    I read on Wiki that the earliest male lineages in the British Isles now regarded as originating from a Balkan migration instead of an Iberian migration, as was earlier hypothesized. That hypothesis is based upon Y-DNA analysis though — I don’t know if there’s an estimate for ancestry percentage from autosomal DNA. Do you? You gave the 70 percent figure, but it would be interesting to see that disaggregated for the different waves of pre-Roman migration (e.g, Celts, Balkans, etc)

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    Southern american cotton slaves, in contrast, underwent population growth and were exported further into the south. Does anyone know to what extent that was a matter of the geographic location vs the crop?

    economics of scale rice ~ tobacco > cotton > sugar. IOW, sugar was more easily transformed into a factory cash crop where you needed to mix some capital with male replaceable labor.

    Your conclusion is not clear to me. Are you saying that the reason that the Britains assimilated to the Germans and not vice-versa is, most likely, that the Germans came with their families/women so were able to maintain their culture long enough to incorporate the locals? And that, in contrast, the reason that the Magyars and Turks assimilated to the local cultures was that they were basically victorious armies who, having conquered, settled down with local women and eventually assimilated into the local culture though of course they did so as a ruling caste?

    first, you can impose your culture without bringing women. to some extent you see this in latin america. but, latin american societies exhibit hybridization in some areas. e.g., mexican cuisine. presumably this seems lacking among the english. my model posits that there wasn’t an early period of hybridization where british elements leaked into the initial deme. that hybridization came later, but by that point the english had enough critical mass that they simply assimilated without hybridizing.

    also, part of my point is that the maygars did not assimilate. they assimilated the substrate into their language. the magyar connections to inner asia even persisted after their christianization. turks of all religions fleeing the mongols in the 13th century were allowed to resettle in hungary granted that they converted to christianity.

    i’m giving you an autosomal estimate. the Y would be german biased i think. some mtDNA i britain is rooted to the paleolithic. google cheddar man.

  • ackbark


    I’d suggest they remain culturally self-coherent because they were situated between a lot of not-especially coherent German populations and not-especially coherent Slavic populations (at least compared to themselves), neither of which had any particular cultural cachet, so a cultural tug in one direction is immediately counter-balanced by a tug in the other leaving them preserved in a default state of Magyar. Or is that just simplistic?

    Saxon Shore

    Perhaps not so much about defense as tax collecting and trade management and has the effect of instituting Saxon-Frisian as the default trade language of the entire coast of Britain and northwest Europe, so when the Romans collapse the Germans have these established business and family ties to fall back on and the language quickly filters over Britain as the standard in business opportunity. And Celts, who may be divided among themselves, then find it convenient to speak in Saxon rather risk a quarrel, with one another, over nothing.

    Perhaps Britain wasn’t so much conquered as bought.

  • authun

    The Saxon Shore forts don’t correspond well with the earliest anglo saxon pagan cemeteries in the Yorkshire Wolds, Sancton, Newbald etc. and the Upper Thames valley, eg. Berinsfield. These areas also show the highest frequencies of haplotypes judged to be of continental origin. The forts in Norfolk do correspond with early cemeteries and the area is similar in terms of genetic evidence but the majority of the forts are along the south east coast, from the Isle of Wight to Essex, where the genetic signal is weaker. I think it would be wrong to equate the shore forts with early settlement and a mechanism for achieving control. They may be, but only in some parts.

    The trading function of these forts at the end of the roman period is likely unimportant as the economy had collapsed. Their importance may be that they were garrisoned by germanic auxilliaries or that laeti, settlers who were granted land in return for miliary service in times of need, were relocated from the continent to these areas. This appears to have happened along the gallic saxon shore for example, around Boulogne.

  • http://washparkprophet.blogspot.com ohwilleke

    Nice, thick, nuanced analysis with excellent examples.

    Another interesting example to add to the mix would be the case of Mongol efforts to incorporate Koreans into their domains by forcing local notables to marry Mongol princesses and then giving those princesses real power politically with the Mongol elite and via religious patronage.

    While it made a dent, in the end it basically failed. When the Mongol war machine’s impact became less dominant, the Mongol princesses were purged, assassinated or marginalized by forces led by an emergent bourgeois and class of courtiers and lesser aristocrats that hated them with a passion. The religious institutions were looted and appropriated by the state. The Mongol princesses also didn’t manage to stay in power long enough (just a few generations) to consolidate Mongol culture and the fact that it was princesses rather than princes that were brought into the Korean culture may also have been too great a leap for the Korean society to accept even though the Mongols were somewhat more accepting of powerful women. The relatively small number of political subdivisions in Korea at the time and the peripheral location also probably helped.

    One can also look more recently at the persistence of Western institutions in newly independent colonial powers. British institutions survived in India after almost five hundred years of British colonial rule before independence (although arguably penetrated less deeply in Pakistan than in India or even Bangladesh), but places where colonial powers had ruled less long (e.g. much of Africa) seem not to have developed the critical mass necessary to make that approach work.

    Then again, that could be a function of the theory that political complexity must advance sequentially. India has a fairly well developed system of regional states in place when the British arrived and the British colonial rule relied on playing one regional state against another. In the African case, there were few full fledged states, more big man run villages and chiefdoms, when colonial powers arrived, so political units had to be built from the gound up, often without success.

    Put another way, while “barbarism” may reduce resistance to assimilation, it also means that the number of “chunks” of organized people who have to be assimilated one group at a time is greater if the society is too fragmented.

    Along the same lines, one important factor in Islamic expansion was rapid assimilation of existing elites. This was also the case in the rapid Nazi expansion where local business and political leaders were co-opted. (I’m not, by the way making an Islamo-Fascist suggestion, adoption of local elites, also Holy Roman Empire, is a general phenomena). The ideal state for assimilation may be one where the substate people are divided, but into chunks just a single level of political complexity below the one aspired to by the rulers.

    If the substate people are truly unified and do not have co-opted elites (as the Roman Catholic church did for the Irish in the face of British rule), it can be almost impossible even after a millenium to secure assimilation, while in the case where the elites are co-opted (e.g. Mary Queen of Scots), it can happen much more quickly.

    Tibetan assimiliation hasn’t been too effective without heavy handed demic replacement in part because the Tibetans have maintains political unity through a government in exile.

    The point that “unlike the Christians of Egypt the Zoroastrians of Iran almost disappeared”, of course, has at least one very obvious source: Islamic tolerates “People of the Book” (i.e. Jews and Christians) as a matter of religious doctrine (perhaps to defuse early resistance from these locally powerful groups), while it officially does not tolerate other religions (whose Mediterranean influence was already weak due to the active suppression of these religions by the late Roman emperors).

    The classic case of a superstate succumbing to a substate which also bears examination is the case of the Romans in the Eastern empire adopting the language and culture of the Greeks that they conquered (and the early Arabs as well, as noted).

    One way to think about it is that co-opted local elites may only be sincere in their adoption of superstate norms when the superstate can somehow demonstate that they have something that is worth emulating. Actual “barbarism” in the substate people may be less important than a self-perception that the leadership and culture of the conquered society is rotten. The cultural self-esteem of the elites may be at work as a powerful factor.

    The Sumerian replacement by Akkadian elites, the demise of Harappan culture that opened the door to Indo-Europeans, and the weakness of the C-T culture that opened the door to Indo-Europeans in the greater Balkans involved rather advanced cultures for their day, but elites that were discouraged about the value of their own traditions as a result of the climate setbacks of the drought/climate shift of ca. 2000 BCE.

    The climate shift producing leadership failures that accompanied Bronze Age collapse ca. 1200 BCE, may similarly have undermined the self-confidence of the Bronze Age elites of that time making them open to the new post-Bronze Age collapse regimes. For example, it appears that the megalithic culture saw a massive collapse in metal trade goods and trade in general before the way was opened for a Celtic replacement of that culture. Classes of leaders who can’t deliver lose confidence in themselves and are open to alternatives.

    Perhaps there is something to the unending cycles of self-doubt and condemnation of corrupt leadership regimes that pervade the Hebrew Bible. Perhaps it is that kind of self-image of one’s society (which is seen again in the hand wringing classical literature about the late Roman empire’s corruption), that cause existing substrate regime leaders to lose faith and become vunerable to momentarily militarily strong superstrate leader’s alternatives.

    Would the Spanish culture have become so dominant if the Aztecs and Incas hadn’t experienced such spectacular defeats that make the old rulers and culture look inferior to the new, even if culturally neutral germs may have been as important as guns and steel and cultural advantages in fact?

    Was the willingness of the elites of Britain to assimilate into Germanic culture, and the Hungarian to assimilate into Urgic culture, to a great extent a product of the fact that the Roman-Christian approach to which local elites owed residual allegiance been so deeply discredited by the fall of the Roman empire, rather than simply the loss of ties to it?

    The longest lived non-Indo-European societies of Europe (Basque and Etruscans) were also those that had the least to gain from conversion to Indo-European society because they had adopted many of the non-linguistic technological innovations (e.g. art technique, metal working, more advanced farming, architectural and military technologies) that their neighbors had not. Survivors and sustainers of old regimes tend to be those that are early cultural adopters than use their imitation to hold their own.

  • antonio

    hi razib, thanks for the superberb post: do you have any nice reference on magyars/hungarian case?

  • AG

    Xiongnu were pacified by the Chinese through divide and conquer. Southern xiongnu became part of Chinese force to fight against northern xiongnu. Southern xiongnu elites were awarded with noble and military titles. At end of Jin dynasty (After Han and three kindoms), southern xiongnus were first rebel force uprising against Jin rulers since xiongus felt they had been treated like `slaves’.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar

Login to your Account

E-mail address:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »