Declining intelligence in the 20th century: the case of Estonia

Soviet-era stamp. In Estonia, cranial volume shrank between the cohort of girls born in 1937 and those born in 1962, apparently because the intellectually gifted were more likely to pursue higher education and postpone childbearing.

 

Is the genetic basis for intelligence declining from one generation to the next? That’s the conclusion of several recent studies on alleles associated with high educational attainment. By adding up such alleles over the genome, we can get a person's "polygenic score." By calculating the mean polygenic score for each generation, we can then find out whether this genetic basis is declining.

The polygenic score has declined among Icelanders since the cohort born in 1910 and among Euro Americans between the 1931 and 1953 cohorts (Beauchamp 2016; Kong et al. 2017). The Icelandic study is especially interesting because that country took in very few immigrants during the period under study. The decline was thus driven by internal factors. One reason seems to be the tendency of university-educated people to delay reproduction and have fewer children. But that's not the whole story. Even among Icelanders who didn’t pursue higher education, fertility was lower among the intellectually gifted, apparently because their intelligence was associated with a desire to plan for the future and delay gratification.

Before the twentieth century, such forward-looking people were reproductively successful. They were the ones who had enough resources to survive disasters of one sort or another: famine, disease, the Little Ice Age, etc. Today, such disasters are a lot less deadly, and it no longer matters so much whether one is a grasshopper or an ant.

Moreover, because of demographic and cultural changes during the twentieth century, it’s no longer possible to count on the same degree of assistance for child-raising from relatives and grandparents. With childbearing at later ages, grandparents are either dead or too frail to help. With people moving around more, not all relatives live nearby. If you’re the sort of person who plans for the future and delays gratification, you may be a lot more intimidated than your forbears by the costs of raising a family.

Shrinking cranial volume in Estonians

Cranial volume correlates with IQ and with educational attainment, albeit imperfectly (see Frost 2020). Has it been declining in tandem with the decline in alleles for educational attainment?

In Soviet-era Estonia, cranial volume was one of several anthropometric traits that were measured in girls born between 1937 and 1962. Because the measurements were mandatory, there was no volunteer basis; mortality bias was minimal because all the participants were younger than 20. In this respect, the study is better than Western biobank studies. On the other hand, the results may be less applicable to Western populations, given the differences in demographic history. Estonia had no postwar baby boom. Fertility then rose from the late 1960s until the breakup of the Soviet Union. By the late 1980s, fertility was actually higher in Estonia than in any other major region of Europe.

Nonetheless, there were demographic similarities between Soviet Estonia and the West, particularly the rising prevalence of single mothers and the influence of education on fertility:

- Divorce rates began to rise during the interwar years, equalling or exceeding those of Scandinavia from the 1970s onward.

- Throughout the twentieth century, Estonian women with only primary education bore 0.5 to 0.75 more children on average than women with tertiary education. In the population under study, taller children and those with larger crania were more likely to go on to secondary and/or tertiary education, independently of sex, socioeconomic position, and rural vs urban origin (Valge et al. 2019).

The second factor seems to explain why cranial volume declined from the older cohorts to the younger ones:

[...] the majority of selection for smaller cranial volume acted indirectly via educational attainment, whereas the direct path of selection in the SEM model was non-significant (Figs. 2 and 4). In other words, consistent with our prior expectations, girls with larger heads were selected against because they were more likely to obtain higher education than girls with smaller heads. Lower education (Tiit, 2013) and rural origin (Kulu, 2005) have been independently and additively associated with higher fertility in Estonia throughout the past century. The reason for the link between education and fertility is that early reproduction, a major determinant of LRS, is not compatible with schooling for both cultural and genetic reasons. (Valge 2020)

 It is doubtful that this decline is due to ethnic change. All of the girls were from Estonian schools (Russian-speakers had their own schools). Nonetheless some of them were of mixed background. According to a personal communication from the corresponding author, 84% of the fathers and 93% of the mothers were Estonian. Ideally, the study should be redone without individuals of mixed parentage. The problem here is not only that one of the parents was non-Estonian but also that such individuals were disproportionately economic migrants who had trouble finding suitable work elsewhere in the Soviet Union.

Other anthropomorphic changes

Height also declined. Unlike cranial volume, this decline was not wholly explained by educational/socioeconomic differences:

Notably, higher reproductive success of shorter girls in Estonia could not be entirely ascribed to indirect selection via educational attainment, nor via other measured socioeconomic variables such as rural/urban origin, although indirect selection via education did account for a large portion of total selection (Fig. 4). The finding that selection against height remains after controlling for education or income (that favours less-educated individuals who are generally shorter than highly-educated ones) is consistent with findings of studies reviewed by Stulp and Barrett (2016).

 

Female hips and female jaws became narrower even after controlling for educational/socioeconomic differences. There seems to have been selection for rounder female faces, but this selection is significant only if one allows for nonlinear effects. Finally, there was no direct selection on two markers of overall health and nutritional status: handgrip strength and lung capacity.

In general, "direct selection favoured shorter, slimmer and lighter girls with smaller heads, more masculine facial and body shapes and slower rates of sexual maturation."

Conclusion

The genetic basis for intelligence has declined in European populations, apparently since the early twentieth century. This decline is attested by two "hard" measures: 1) alleles associated with educational attainment; and 2) cranial volume. Furthermore, it is attested in two relatively homogenous societies, i.e., Iceland and Estonia.

In Estonia, the decline seems entirely due to the intellectually gifted going to university and postponing family formation. In Iceland, this factor explains only part of the decline: the intellectually gifted chose to postpone family formation even when they didn't go to university. Perhaps the Soviet system was better at steering gifted individuals into higher education.

On a final note, this problem will not go away on its own. If we wish to have large numbers of intellectually gifted people who plan for the future and delay gratification, we will need to reverse certain social and cultural changes of the twentieth century.

Comments by Peeter Horak

In an email, Peeter pointed out that the decrease in height due to natural selection might be offset by an increase in height due to lower pathogen load (as a result of vaccination and antibiotics, see Hõrak and Valge 2015). In addition, we currently don't know the direction of selection on boys. It may entirely cancel out natural selection on girls if men's income and education correlate positively with their reproductive success. In the sample under study, taller boys and those with larger heads went on to obtain more education; if they were reproductively successful, there would be sexually antagonistic selection: selection would favor larger boys and smaller girls at the same time.

References

Beauchamp, J.P. (2016). Genetic evidence for natural selection in humans in the contemporary United States. Proceedings of the National Academy of Sciences 113(28): 7774-7779 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4948342/

Frost, P. (2020). Did women jumpstart recent cognitive evolution? Evo and Proud, July 1 https://evoandproud.blogspot.com/2020/07/did-women-jumpstart-recent-cognitive.html

Hõrak, P., and M. Valge. (2015). Why did children grow so well at hard times? The ultimate importance of pathogen control during puberty, Evolution, Medicine, and Public Health, 2015 (1): 167–178, https://doi.org/10.1093/emph/eov017

Kong, A., M.L. Frigge, G. Thorleifsson, H. Stefansson, A.I. Young, F. Zink, G.A. Jonsdottir, A. Okbay, P. Sulem, G. Masson, D.F. Gudbjartsson, A. Helgason, G. Bjornsdottir, U. Thorsteinsdottir, and K. Stefansson. (2017). Selection against variants in the genome associated with educational attainment. Proceedings of the National Academy of Sciences 114(5): E727-E732 https://core.ac.uk/download/pdf/154416179.pdf

Valge, M., P. Horak, and J.M. Henshaw. (2020). Natural selection on anthropometric traits of Estonian girls. Evolution and Human Behavior in press. https://doi.org/10.1016/j.evolhumbehav.2020.07.013

Valge, M, R. Meitern, and P. Horak. (2019). Morphometric traits predict educational attainment independently of socioeconomic background. BMC Public Health 19: 1696. https://link.springer.com/article/10.1186/s12889-019-8072-7

British hipsters

Why have British women become broader-hipped over the past three thousand years? (Wikicommons: Niek Sprakel)



British women have become broader-hipped over the past three thousand years or so. That's the conclusion of a recent study of alleles that influence female hip circumference, using data from the UKBiobank. 

Audrey Arner and her colleagues at Penn State identified 148 SNPs associated with female hip circumference and 49 SNPs associated with first child birth weight. Nine of them influence both women's hip circumference and first child birth weight. The SNPs associated with female hip circumference seemed to influence first child birth weight but not vice versa. There also seems to have been selection over approximately the last three thousand years for women with broader hips.

The baby's head is the biggest challenge during childbirth:

Human birthing is difficult owing to a tradeoff between large neonatal brain size and maternal pelvic dimensions, which are constrained by aspects of bipedal biomechanics. The net effect is that human neonatal head size closely matches maternal pelvic dimensions, unlike in our closest living relatives, the great apes, whose pelvic dimensions are larger than neonatal head sizes. (Franciscus 2009)

Have female hips become broader over the past three thousand years because the birth canal has had to accommodate babies with larger brains? That hypothesis would be consistent with an analysis of ancient DNA by Michael Woodley of Menie and others, who showed that alleles for educational attainment gradually increased in frequency between 4,560 and 1,210 years ago in Europeans and Central Asians. That increase may have been due to gene-culture coevolution: as societies grew larger and more complex, the average person had to perform mental tasks that likewise became larger in number and more complex. Such an environment would have favored the survival and reproduction of individuals with higher cognitive ability. Mean IQ thus rose over time, as did cranial capacity.

On the other hand, Henneberg (1988) showed that cranial capacity steadily shrank from the Mesolithic to modern times, becoming 9.9% smaller in men and 17.4% smaller in women. His conclusion was based on a large sample: 9,500 male skulls and 3,300 female skulls.

So we have a contradiction. Perhaps cranial capacity didn't really shrink from the Mesolithic to modern times. Perhaps smaller skulls are more likely to decompose faster. The skulls we unearth would therefore be a biased sample, and this bias toward preservation of larger skulls would gradually increase for skulls that have been in the ground longer.

The problem of "preservation bias" has already been noted with respect to female and infant remains:

There are nearly always more males than females in skeletal collections from archeological sites [...]. This has been explained in part by the comparatively rapid disintegration of lightly built female skeletons.

[...] The burial records show that most of the people buried in the Purisima cemetery were either infants, children, or elderly adults. The skeletal remains excavated from the cemetery, in contrast, are predominately those of young adults. The underrepresentation of young children in the skeletal collection is most likely a result of the comparatively rapid disintegration of their incompletely calcified bones.

[...] If, on the other hand, infants or elderly people are more common in a skeletal collection from a recent cemetery than they are in an ancient one, much less can be inferred about differences in the original age structure of the two burial populations. Such a difference would be expected due to differential preservation, even if the age structures of the two burial populations were identical. (Walker et al. 1988)

The same preservation bias might cause an overrepresentation of larger skulls among older remains.


References

Arner, A., H. Reyes-Centeno, G. Perry, and M. Grabowski. (2020). Pleiotropic effects on the recent evolution of human hip circumference and infant body size. The 89th Annual Meeting of the American Association of Physical Anthropologists (2020), April 17
https://meeting.physanth.org/program/2020/session26/arner-2020-pleiotropic-effects-on-the-recent-evolution-of-human-hip-circumference-and-infant-body-size.html  

Franciscus, R.G. (2009). When did the modern human pattern of childbirth arise? New insights from an old Neandertal pelvis. Proceedings of the National Academy of Sciences 106(23): 9125-9126.
https://www.pnas.org/content/106/23/9125.short  

Henneberg, M. (1988). Decrease of human skull size in the Holocene. Human Biology 60: 395-405.
https://www.jstor.org/stable/41464021  

Walker, P.L., J.R. Johnson, and P.M. Lambert. (1988). Age and sex biases in the preservation of human skeletal remains. American Journal of Physical Anthropology 76: 183-188.
https://onlinelibrary.wiley.com/doi/abs/10.1002/ajpa.1330760206

Woodley of Menie, M.A., S. Younuskunju, B. Balan, and D. Piffer. (2017). Holocene selection for variants associated with general cognitive ability: Comparing ancient and modern genomes. Twin Research and Human Genetics 20: 271-280.
https://www.cambridge.org/core/journals/twin-research-and-human-genetics/article/holocene-selection-for-variants-associated-with-general-cognitive-ability-comparing-ancient-and-modern-genomes/BF2A35F0D4F565757875287E59A1F534 

Not getting the point

Samuel George Morton, an early American anthropologist. He fudged his data to suit his preconceived ideas on race, according to Stephen Jay Gould. It later turned out that Gould was the fudger.

Stephen Jay Gould (1941-2002) is still seen as a great evolutionary biologist, if not one of the greatest. Yet the years since his death have steadily tarnished his memory. This is especially so for his best known book, The Mismeasure of Man, which focused on an early American anthropologist, Samuel George Morton (1799-1851). In this book and in an earlier Science article, Gould showed how Morton had fudged his measurements of a collection of skulls to make Europeans seem bigger-brained than Africans.

Gould didn’t re-measure any of the skulls. He reanalyzed Morton’s data … and in the process did far more fudging than Morton had ever done. When a team of physical anthropologists, headed by Jason E. Lewis, located and re-measured half of the skulls, they found only a few randomly distributed errors in the original measurements. Morton had in fact tended to overestimate African skull size (Lewis et al., 2011).

Interestingly, the same conclusion had been reached almost a quarter-century earlier by John S. Michael, a senior at Macalester College in Minnesota. This discovery has been recounted on John’s blog:

I re-measured the Morton skulls in 1986 as part of my undergraduate thesis, which was limited in scope and conducted without the rigor of graduate research. Nonetheless, I determined that my measurements more or less matched Morton’s, and so I described his overall results to be “reasonably accurate.” (Michael, 2013a)

Troubled by his finding, he got in touch with Gould:

In 1986, I mailed my results to Gould, who requested we meet after he gave a lecture in May at the University of Minnesota. Our meeting lasted perhaps five minutes. He told me that I “missed the point,” and abruptly ended the conversation, ignoring me and instead speaking to the man next to him. My recollection is that he did not say goodbye, so I simply walked away. […] After I published my paper in 1988, I sent Gould a copy but got no response. When I wrote him again, he replied that he had lost it and requested another copy, which I sent. I never heard back from him.

Sometime later, Gould gave a lecture at the University of Pennsylvania, after which he was asked a question about my paper. His response was simply that he would not discuss it, and he did not. Gould never mentioned my paper in any of his prolific writings. In 2011, Lewis wrote that, “were Gould still alive, we expect he would have mounted a defense of his analysis of Morton.” Soon after that, Prothero noted, “I’m sure if Steve were alive, he would be able to counter these accusations in his own inimitable way.” And yet these two statements conflict with the fact that Gould actually had two opportunities to counter such accusations, and instead chose to silently disengage. (Michael, 2013b)

Although John Michael’s paper appeared in Current Anthropology, a leading journal in its field, the response was largely silence (Michael, 1988). As recently as five years ago, a science historian had only this to say:

Gould’s interpretation of Samuel George Morton’s cranial data have been questioned by John S. Michael, who, as an undergraduate student at Macalester College, re-measured the skulls as part of an honors project (Michael, 1988). It is not entirely evident that one should prefer the measurements of an undergraduate to those of a professional paleontologist whose own specialist work included some very meticulous measurements of fossil snails. (Kitcher, 2004)

Some people were more supportive, but they were the wrong kind:

Because my findings refuted the writings of Gould, a left-leaning anti-racist Jew, I was celebrated in hate-filled white supremacist web pages, such as davidduke.com and stormfront.org. My work was grossly misquoted in a series of papers by J. Philippe Rushton, a proponent of eugenics from University of Western Ontario. In 2002, he served as the president of the Pioneer Fund, which the Southern Law Poverty Center designated as a “White Nationalist” group because it continues to fund the study of “breeding superior human beings that was discredited by various Nazi atrocities.” I have written this paper in part to document my strong displeasure that my work was used to promote eugenics or racist ideology, which I in no way support. (Michael, 2013a)

Yes, supporters can be as problematic as detractors. But a scientific finding is not invalidated because its supporters are the wrong kind of people. It stands or falls on its own merits. Also, the Southern Law Poverty Center is hardly an impartial source.

John Michael was ultimately vindicated when the Lewis et al. paper came out two years ago. Yet, even then, he never got the credit he deserved, as may be seen in a Nature editorial that raised the possibility of an improper relationship between Lewis’ research team and the University of Pennsylvania:

Of course, Lewis and his colleagues have their own motivations. Several in the group have an association with the University of Pennsylvania, and have an interest in seeing the valuable but understudied skull collection freed from the stigma of bias (Anon, 2011)

No evidence is provided for this curious accusation. In any case, Gould’s fudging had originally been revealed by someone from another university and from another state.  But who remembers?

Conclusion

John Michael found himself in an unequal battle. As a graduate student he was challenging not only an Ivy League professor but also a leading antiracist crusader. It didn’t really matter whether Gould was telling the truth or not. There was something bigger at stake—the war on racism. And that war had to be won.

John indeed “missed the point.” By insisting too much on truth, he was being at best naïve and at worst a willing accomplice of racism—by far a greater evil than falsehood. This was how many well-meaning people saw things in the late 20th century.

References

Anon. (2011). Mismeasure for mismeasure, Nature, 474, 419.

http://www.nature.com/nature/journal/v474/n7352/full/474419a.html

Gould S.J. (1981). The Mismeasure of Man. New York: W. W. Norton and Company.

Gould, S.J. (1978). Morton’s ranking of races by cranial capacity: unconscious manipulation of data may be a scientific norm, Science, 200, 503–509.

Kitcher, P. (2004). Evolutionary Theory and the Social Uses of Biology, Biology and Philosophy, 19, 13-14.

Lewis, J.E., D. DeGusta, M.R. Meyer, J.M. Monge, A.E. Mann, and R.L. Holloway. (2011). The Mismeasure of Science: Stephen Jay Gould versus Samuel George Morton on Skulls and Bias, PLoS Biology, 9(6) e1001071

Michael, J.S. (2013a). Stephen Jay Gould and Samuel George Morton: A Personal Commentary, June 7, Michael1988.com

http://michael1988.com/?p=114

Michael, J.S. (2013b). Stephen Jay Gould and Samuel George Morton: A Personal Commentary, Part 4, June. 14, Michael1988.com

http://michael1988.com/?p=203

Michael, J.S. (1988). A new look at Morton’s craniological research, Current Anthropology, 29, 349–354.

An African community in Roman Britain?

Figurines of Nubian archers (from Egypt)

The Roman conquest of Britain brought not only cultural change but also profound ethnic change, i.e., an influx of soldiers, officials, and traders from elsewhere. Until recently, historians placed this influx mainly in the first century of Roman rule. As the native British became Romanized, they would have increasingly filled local positions in the army and the administration. This view now appears to be erroneous: the process of ethnic replacement continued unabated and may have even intensified. In particular, the last century of Roman rule saw many soldiers arrive from North Africa and sub-Saharan Africa.

Leach et al. (2010) argue that the north of England, with its many garrison towns, became heavily multi-ethnic through the stationing of foreign soldiers and the subsequent settlement of veterans on local estates:

Inscriptions from York attest to the presence of Gauls, Italians and a possible Egyptian at York, although much of the epigraphic evidence is significantly earlier than the burial discussed here (Ottaway 2004). Artefactual evidence also suggests that ‘the Roman north was a cosmopolitan place with a great mixing of people from all over the empire’ (Cool 2002: 42). For example, Swan (1992) argued for the presence of North Africans in York on the basis of braziers and other vessels typical of North African food-ways but made in local fabrics.

Previous studies of the physical remains of the people of Eboracum (Buxton 1935; Warwick 1968) suggested that the males exhibited heterogeneous craniomorphometric traits suggesting migration from a variety of geographic locales. However, both argued that the female population of Roman York was indigenous. In their view, the diversity noted in the female crania represented genetic admixture between local women and migrant males (Warwick 1968: 155).

Leach et al. (2009) provide evidence for intense foreign settlement. At one burial ground near Roman York, craniometric analysis revealed that 66% of the individuals clustered most closely with Europeans, 23% with sub-Saharan Africans, and 11% with Egyptians. At another, the proportions were 53% European, 32% sub-Saharan, and 15% Egyptian (Leach et al., 2009).

In a subsequent article, Leach et al. (2010) focus on one burial: a young woman 18 to 23 years old who had been buried between 350 and 400 AD. The authors dubbed her the ‘Lady of York’ because of her stone coffin and its rich array of grave goods, apparently a sign of high status. Nonetheless, her skull showed little or no affinity to any European population, the closest match being a sample of African-American women. Various facial indices showed a mix of sub-Saharan African and European traits, suggesting a person of mixed parentage or perhaps a North African. An African origin is also suggested by the presence of elephant ivory among the grave goods. The authors conclude that this burial “contradicts assumptions that may derive from more recent historical evidence, namely that immigrants are low status and male, and that African individuals are likely to have been slaves.”

Undoubtedly, many Africans rose to high positions in the Roman army. This was especially so for the Nubians, who were prized for their skills in archery. This being said, the Nubians came from the Egyptian culture area and, as such, attached great importance to human burial. In comparison to Europeans of the same socioeconomic status, they would have been more inclined to provide grave goods and to use tombs made from impervious materials.

High status is also inconsistent with a Latin inscription placed in the coffin: SOR AVE VIVAS IN DEO [Sister, hail, live in God]. This is Vulgar Latin, the kind used by the lower classes. An educated citizen would have inscribed the more classical soror instead of sor, especially in the formal context of a burial.

An African community in Roman Britain?

Leach et al. (2010) conclude that the Romans laid “the foundation of a multicultural and diverse community.”

The word ‘foundation’ suggests permanence and is perhaps inappropriate. If the Lady of York had lived a full life, she would have seen the end of Roman Britain—less than a half century later. She would have certainly witnessed the pullout of the legions and, perhaps, the arrival of the Dark Ages … with the breakdown of law and order after 430.

What happened to her community? Did it play a role in the struggles between the Romano-British and the Anglo-Saxons? Did it survive into the early Anglo-Saxon period?

I will try to answer these questions in my next post. For now, let me ponder another question: why had Roman Britain been so intensely militarized? More legions were stationed there than in Lower Germany, which was closer to the Empire’s core and faced a much larger barbarian population. If Britain had been ruled indirectly, through client states, the Romans would have been free to concentrate more soldiers along the crucial Rhine-Danube border.

These questions have answers … more or less. The Empire had annexed Britain during the early days of seemingly endless expansion, when military leaders worried much less about spreading themselves too thin. Britain was also a Celtic homeland. Its people were closely related to those of Gaul and had openly sympathized with them during the rebellion led by Vercingetorix. Even as a collection of client states, Britain would have remained a threat to Roman rule in Gaul. Finally, more legions were needed in Britain because the barbarians to the north (Scots and Picts) were less interested in being friends of Rome than those along the frontier of continental Europe.

Well, perhaps. In any case, history would have turned out the same either way. The main barbarian threat—the one that led to the sack of Rome in 410—came not from the barbarians outside the Empire but from those who had been allowed to settle within. This was an eventuality that Roman military strategists had never foreseen and for which they were unprepared. When the end came, the border fortifications proved to be of no help.

References

Leach, S., H. Eckardt, C. Chenery, G. Müldner, & M. Lewis. (2010). A Lady of York : migration, ethnicity and identity in Roman Britain, Antiquity, 84, 131-145.

Leach, S., M. Lewis, C. Chenery, G. Müldner, & H. Eckardt. (2009). Migration and diversity in Roman Britain: A multidisciplinary approach to the identification of immigrants in Roman York, England, American Journal of Physical Anthropology, 140, 546-561

P.S. The link and reference information for my 2008 article have changed. They are now:

Frost, P. (2008). Sexual selection and human geographic variation, Special Issue: Proceedings of the 2nd Annual Meeting of the NorthEastern Evolutionary Psychology Society. Journal of Social, Evolutionary, and Cultural Psychology, 2(4), pp. 169-191.

http://www.jsecjournal.com/articles/volume2/issue4/NEEPSfrost.pdf

Was Roman Britain multiracial?

Historians often assume that the Romans changed Britain politically but not demographically. The indigenous elites adopted Roman culture while the mass of the population remained Celtic. When the Anglo-Saxons arrived in the fifth century, much of this population fled to Wales and Cornwall, where they would retain their language and traditions. Meanwhile, those who remained behind were obliterated through a process of ethnic cleansing and coerced assimilation.

This historical account may be false. First, the Roman occupation seems to have brought profound demographic change. This has been suspected for some time on the basis of unusual burial objects and epigraphic inscriptions that record the presence of individuals from throughout the Roman Empire. Now, after analyzing remains from two burial grounds near Roman York, a research team has concluded that the buried individuals had diverse geographic origins (Leach et al., 2009). In particular, the craniometric data revealed many of sub-Saharan or Egyptian origin. At the ‘Trentholme Drive’ burial ground, 66% clustered most closely with Europeans, 23% with sub-Saharan Africans, and 11% with Egyptians. At the ‘Railway’ burial ground, the proportions were 53% European, 32% sub-Saharan, and 15% Egyptian.

York was a legionary fortress, so these individuals may have been legionnaires. There are, in fact, epigraphic references to African soldiers and even a written account about one in a history of the Emperor Septimius Severus (146-211 AD) (Scriptores Historiae Augustae, p. 425).

On another occasion, when he was returning to his nearest quarters from an inspection of the wall at Luguvallium (Carlisle) in Britain, at a time when he had not only proved victorious but had concluded a perpetual peace, just as he was wondering what omen would present itself, an Ethiopian soldier, who was famous among buffoons and always a notable jester, met him with a garland of cypress-boughs. And when Severus in a rage ordered that the man be removed from his sight, the Ethiopian by way of jest cried, it is said, “You have been all things, now, O conqueror, be a god.”

Why were these Africans so far from home? In the case of the Egyptians, Rome thought it unwise to station soldiers among people of the same ethnic background. The temptation would be strong to side with the locals if a rebellion occurred. In the case of the sub-Saharan Africans, they were recruited into the army for the same reason that Germanic barbarians were recruited: Rome could not meet its manpower requirements solely from within its empire. There was also a perception that the Romans had become soft and that barbarians made better soldiers.

Finally, Rome, like many multi-national empires, had a policy of moving people around in order to promote a common identity and to eliminate ethnic distinctiveness. The Assyrians had perfected this policy, e.g., the deportation of the Jews to Babylon and their replacement by other peoples. The Roman authorities used their army to this end. They wished to create an atomized society where regionalism or ethnicity could not mobilize resistance to imperial rule.

It is likely that these legionnaires had a major demographic impact wherever they were stationed, especially if we include the many officials, petty functionaries, traders, and others who came in their wake. Much of Roman Britain thus seems to have been Romanized in culture and multiethnic in origin.

This, in turn, calls for a few other reinterpretations. Wales and Cornwall are not Celtic-speaking today because they took in Romano-British refugees fleeing Anglo-Saxon invaders. They were simply those parts of Britain that had remained Celtic in language, culture, and population. The rest—present-day England—had long become heavily Romanized and cosmopolitan.

Nor do we have to postulate a process of ethnic cleansing and coerced assimilation to explain the extinction of Roman Britain in the 5th and 6th centuries. As Seccombe (1992) points out, the Roman Empire suffered from negative population growth. Not enough people married and had children to offset relatively high mortality among infants and young adults. In breaking down local collective identities, whether ethnic or regional, the Empire had created an atomized and increasingly anonymous society without the carrots and sticks that tightly knit societies use to push individuals down the path of family formation.

Once Rome had pulled its troops out of Britain in the early 5th century, there was no longer an inflow of people to offset the demographic deficit. The local population fell into decline, and the decline accelerated in the 6th century when plagues killed three out of every ten people. The Romano-British needed no help from the Anglo-Saxons to die out. They did it largely on their own.

References

Leach, S., M. Lewis, C. Chenery, G. Müldner, & H. Eckardt. (2009). Migration and diversity in Roman Britain: A multidisciplinary approach to the identification of immigrants in Roman York, England, American Journal of Physical Anthropology, 140, 546-561.

Scriptores Historiae Augustae – Septimius Severus 22:4-6, transl. D. Magie (1922-1932) Vol 1, London: Heinemann.

Seccombe, W. (1992). A Millennium of Family Change. London: Verso.