When Did the World Get So Polarized? Or, A Short History of the Latitudinal Gradient, 1820-2000

My main beef with the Pomeranzian modern economic history literature is that it gets the explanandum wrong. What needs to be explained is not why the world was polarized in 1800, with Europe taking the lead over other core regions of Eurasia (India, China, Japan). But rather why did the world become polarized from the late-nineteenth century onwards. Another misunderstanding relates to latitude, often thought of as a constant conditioner. It is anything but. Latitude, rather effective temperature for which latitude is a good proxy, is a dynamic variable whose conditioning of societal paths has itself been conditioned by technology and knowhow. To wit, it mattered that industrial work could not be performed in low latitude nations with the same intensity as high latitude nations because of the thermal balance equation. A fruitful strategy to towards answering Yali’s question is to ask, What is the history of the latitudinal gradient? When precisely did high latitude nations pull away decisively from low latitude nations?? The answers one obtains are consistent with a very late departure.

Auke Rijpma has developed a panel dataset of an index of composite well-being derived from life expectancy, urbanization rates, stature, maternal mortality and so on. In this dispatch, we’ll look at the evolution of the latitudinal gradient in this measure. The data is obtained from Clio-Infra.

Our task is simple. We project the index onto latitude for each decade from 1820 to 2000 and examine the diachronic or temporal pattern that emerges. That will allow us to establish the fact of Recency. We begin with the slope coefficient. The structural break appears to be 1860-1880. But let’s dig further.


What happens when we control for volatility? The next figure displays the t-Statistics. The structural break in 1860-1880 now appears to be located precisely in 1870.


But how much variation does latitude actually explain? The next figure displays the percentage of variation explained. From 1830-1860, the latitudinal gradient explains around 10 percent of the variation. That jumps to 25 percent in 1870. It continues to rise intermittently to 40 percent in 1980. That’s the full hockey-stick right there.


Again, let me emphasize that this is merely the explanandum. We have established the fact of a structural break in polarization along latitude around 1870. What is required is a compelling account of how and why this obtained then. And most importantly, why the Industrial Revolution failed to spread beyond the northern core. We can’t do that if we obsess over the British textile boom in 1760-1830. It is time to abandon Whiggish discourses of British self-congratulation.

Postscript. We can also test the Matthew Effect. (“For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.”) We simply project change in the wellbeing index during a decade onto the level of well being at the beginning of the decade. Positive values denote divergence (the already well-off becoming even more well-off relative to the less well-off) while negative values denote convergence.


The pattern that emerges is very interesting. The coefficient is extremely unstable in the ancien regime, ie until the structural break in 1870. In that decade the Great Divergence begins. There are three major waves of divergence: turn of the century, interwar, and postwar. Divergence stops in its tracks after 1980. That’s a very tight representation of the hockey-stick and the attendant divergence in everyday living standards between the core and periphery of the modern world economy.


The Boas-Chomsky Universality Theorem; Or, Cutting Molecular Anthropology Down to Size

Advances in genomics in the late-twentieth century inflated an expectations bubble. DNA was supposed to solve all medical problems before long and locate the real, molecular sources of human behavioral variability. DNA has been mobilized to answer Yali’s question — “Why is it that you white people developed so much cargo … but we black people had little cargo of our own?” If one wants to be strict about Mendelian genetics we can date it precisely at the turn of the century with the rediscovery of Mendel’s work in 1900. But appealing to biological reductionism to answer Yali’s question goes back to at least the early nineteenth century. (More on that soon.)

The most refined, if somewhat esoteric, recent attempt may be found in Gregory Clark’s A Farewell to Alms. Clark posits that the reason why the Industrial Revolution began in England (not true) was because the upper classes, who had more industrious, patient, disciplined, ie more bourgeois, genotypes, were more fecund than the lower classes (with their inferior genotypes). He literally has equations to track the slow regression to the mean of all the world’s upper class genotypes and tries to show that such sustained differential fecundity or differential survival of types did not happen elsewhere until much later. He never bothers to motivate the extremely polarized initial conditions — the Normans had to have been extraordinarily superior genotypes relative to the Anglo-Saxons (the Angles and the Saxons, or whatever you want to call the biological populations of England at the time) at the time of the Conquest in 1066 if a thousand years of regression to the mean has not eliminated their naturally higher rates of getting into Oxbridge and leaving large estates. In other words, Clark is the most refined neoracialist of the day: ‘the natural hierarchy of the races’ is posited as the controlling initial condition, since the rest follows from differential fecundity and survival.


An even more sophisticated sleight of hand may be found in the work of Galor and Maov; published, astonishingly enough, in American Economic Review. Here, ‘the natural hierarchy of the races’ is folded into the second moment — what explains global polarization is a quadratic or hump-shaped relationship between ethnogenetic diversity and economic potential. Societies with too much diversity are often dysfunctional; societies with too little genetic variability often fail to produce real breakthroughs except once in a blue moon; in the Goldilocks zone of genetic diversity we find the most advanced societies, the most innovation and all the other goodies economists love. There may be some truth to the matter. Ethnic diversity may indeed condition societal paths. But whether this is the controlling variable is hard to decide. Are high latitudes already privileged over low latitudes (with their high tropical diversity) and their Goldilocks diversity compounds or adds to their privilege, or does diversity trump latitude? They have established the former by showing that the partial correlation coefficients are significant after controlling for latitude (and a slew of other conditioners). They can hardly be expected to establish the latter for latitude itself controls diversity, and exceptions working against the grain of the Heliocentric geometry of our life world are few and far between.

The question of whether mere code is a good place to place causal vectors in human evolutionary biology is rarely posed. To be sure, we can explain certain patterns in physical characters and archeological assemblages etc by mobilizing population history (eg, depigmentation in northern Indians reflects their northern origin, as does the fact that they speak Indo-European). And molecular anthropology has a fine-scale resolution on population history. But even population history merely moves the explanandum to the initial conditions; we need to explain the initial conditions themselves by appealing to deeper principles (respectively, Gloger’s rule is why northern Indians are relatively more depigmented than southern Indians; Indo-Europeans descend from a people whose ethnogenesis was triggered by the introduction of advanced Sumerian technology in the steppe c. 5ka).

Genetics may prove useless to solve disease or explain phenotypic variation but molecular anthropology can certainly do phylogenetics. It is so good that other physical anthropologists test the phylogenetic signal of traditional trackers (craniometric and craniodental metrics, nonmetric cranial and dental polymorphism frequencies, slower-moving postcranial variables and so on) against molecular evidence; as indeed I have been doing without explanation.

But the fact that we have fine-scale or high-dimensional data does not land the causal structure of the world in our lap. It is the explanandum; not the explanation. More generally, DNA, brain scan and surveillance data, what we have is the curse of high-dimensional, fine-scale data that one can interrogate with machine learning algorithms. But Google can never predict whether I will have soup or salad with my lunch tomorrow; at the very least Google is forbidden from telling me the prediction (say soup) for then I can say, screw you Google, today I will have salad just to fuck with you. Not just ego but every awake human is capable of doing the exact same thing to Google. (We should make Google tell people its everyday predictions to screw with their predictive algos!)

To say that something is under genetic control, as I said about pelvic bone width and femur head diameter, is merely to place the explanandum at a specific level of analysis; in the realm of evolutionary biology and really slow-moving climatic and social transformations. In other words, it is just a genetic representation of systematic phenotypic variability that can then be explained though logics of bottlenecks, founder effects, population history, and selection. Put another way, Mary’s genes may explain her wide hips, but the fact that Europeans populations have wider hips requires explanation at a different level of analysis. A satisfactory explanation would reduce the specific explanandum of European pelvic bone widths to the general logic of body size for (say) our genus, order or kingdom.

So we can say that there has been strong selection in the genus Homo for bigger bodies and bigger brains constrained by recoverable secondary productivity, fertility rates, packing thresholds, and labor productivity. Put bluntly, hominins could afford their more energetically expensive bodies for any length of time only when sufficient protein was recoverable from the environment given the state of knowhow/level of foraging productivity, fertility rates and packing thresholds did not bind in the Malthusian sense, epidemiological burdens were light enough to allow high net nutritional standards, and natural selection continued to favor bigger bodies. We know that bigger bodies are still under selection for any reduction in health insults in any human population results in bigger bodies. This implies that the explanation of the Early modern collapse in body size in 1450-1650 must be sought in sharply reduced living standards (ie lower net nutritional status) that may themselves be  explained by say climate stress, epidemiology, war and societal instability.

Whatever the cause of the reduction, the reference frame we are working with allows us to establish the fact of an Early modern catastrophe that can be read off European phenotypes (and maybe genotypes some day). That this fact could be read off the genetic code is not explained by Mendelian laws. The notion that it is so is based on a misunderstanding of the role of the Mendelian component of the tripartite modern evolutionary synthesis. (The other components are natural history and natural selection.)

So, we may take the evidence we have marshalled as proof against the refined neoracialism of Gregory Clark. He has established that the upper classes were taller, bigger, and more fecund than the lower classes in England in the centuries after the Conquest. But the inference that English and European genotypes converged over time to the upper class genotype through this process is not borne out by European phenotypes which deteriorated sharply in the crucial phase at issue, 1450-1700 CE. For if he was right, we should expect the opposite movement in the slowest moving variables from what we see: European pelvic bone width should increase after the Medieval period.


Boasian antiracists know that there is a short step from worship of genes to worship of race. But they have failed to articulate the precise problem with biological reductionism. The problem lies in conflating levels of analysis. What looks like an explanation at the level of the individual or cohort, is the explanandum at the level of populations over many generations. To wit, DNA is just code. To say that vertical or global polarization (ie the differential economic performance of populations) is due to genetic differences — ie racialism as a scientific hypothesis — is at best like saying that rich Americans face lower tax burdens than their secretaries because of specific articles in the tax code. It is simply the wrong level of analysis — for a more satisfactory explanation of differential tax burdens we need an account of American political economy. Likewise, even if it can be shown that something is under genetic control, we still need an account of why the genetic code says what it does. And that takes us straight back across the Nature-Society dichotomy.

I owe a better, more general, explanation of the level of analysis problem to J. Dmitri Gallow. Think of a pole in a field whose shadow lands on May 14 on every year at a specific position where a stone has been in place since the pole was erected. A reductionist explanation would be: Here are the laws of optics, here’s the height of the pole and the distance to the stone, here’s the position of the sun that time of the year. One can seemingly hardly argue with the reductionist account. But wait. What if I told you that the pole was erected by Mr. Dawson because his wife came to grief precisely where the stone now sits on May 14, 1947, and that Dawson erected the pole to do precisely what it does every year on that sad day (in obedience to the laws of optics) precisely to commemorate the day she came to grief? That changes everything. The reductionist account turns out to be actively misleading in discerning the chain of causation for it turns out that human agency was the cause of the fact we were seeking to explain; that the nuts and gears of the reductionist account were involved in merely passive proximate processes.

A more fundamental critique of racialism and the self-congratulation of molecular anthropology must begin with the fundamental insights of Boas and Chomsky. The key observation is that the language faculty is exclusive to our species and uniform across it. To wit, a Japanese child growing up in Bengal would speak Bengali like a native. That can only be if the capacity for generative grammar (ie the capacity to generate a countable infinity of structured sentences from a finite lexicon) is truly universal across our species — and by implication under the tightest genetic control. The import is intellectually and politically radical: The Boasian universal — all hunter-gatherers were found without a single exception to have music, dancing, gossip, mythology, ritual, foraging expertise, and the full package of the drama of human life — trumps whatever minor systematic differences exist between populations. Put bluntly, not only the Romans and not only us moderns, but all societies of fully modern humans (ie endowed with the Boasian universal), whether they had high civilization or not, can identify with Greek drama. For even if we admit genetic disadvantages as handicaps, it may be retorted that even a severely handicapped person may turn out to make extraordinary contributions to the human project — think of Stephen Hawking. The human universal thus furnishes a powerful weapon against biological reductionism and the politics of biological differences.

We gotta cut molecular anthropology down to size. If only to get a better handle on physical anthropology, and ultimately, to make progress in answering Yali’s question.








The Incredible Shrinking European, 1450-1850

The stupendous rise in Western living standards from the late-nineteenth century can be read off the stature of populations over time. As late as 1850, Chinese people were taller than Britons. More significantly, Anglo-Saxons in settler colonies were taller than Britons well into the twentieth century. Britons became taller than Americans for the first time only in 1930. We have seen how the settler premium in stature vanished in the early-twentieth century.


Is the above graph evidence of an exit from the Malthusian Trap posited by modern economic historians? For a long time that has been the received wisdom. But only because economists studiously ignored accumulating evidence of much higher living standards before the Columbian era of world history. Seen in the really long run, it really has been downhill from the days of the Gravettian mammoth hunters. But that is far from a story of uniform decline. The LGM bottleneck generated a dramatic decline in body size. The invasions of central Eurasian farmers and Yamnaya pastoralists witnessed massive sex-biased partial population replacement. So the changes between the Late Upper Paleolithic and Iron/Roman periods cannot be interpreted as changes in living standards. Fortunately, there were no further massive population pulses coming to Europe so that we may indeed interpret the collapse in body size between the Medieval and Early modern eras as evidence of a severe fall in living standards.


But hang on, I showed in the previous dispatch that latitude is a strong correlate of body size variables. So how do we know for sure that these shifts are not measuring rebalancing along the latitude gradient? Perhaps the Early Modern collapse is simply a compositional effect due to differential population growth along latitude? In order to make sure that our results are robust to population history, sex, and latitudinal effects, we shall control for sex and latitude and restrict our attention to the period after 250 BCE. We plot Pearson residuals to examine the diachronic pattern in size variables unexplained by sex and latitude.

We begin with the main result. Table 1 reports the P values for Student’s T-tests. Smaller values indicate that the change between periods is significant. Recall that stature is a function of femur length and body mass is largely a function of femur head diameter or pelvic bone width. The bone lengths are more reliable than body mass and stature estimates because they are actual measurements.

Table 1. T-statistics.
Transition Iron/Roman – Early Medieval Early Medieval – Late Medieval Late Medieval – Early modern Early modern – Very recent
Transition date (rough) 500 CE 1000 CE 1500 CE 1900 CE
Transition date (range) 350-650 CE 1025-1050 CE 1450-1680 CE 1850-1900 CE
Body mass 0.176 0.552 0.000 0.191
Stature 0.005 0.241 0.000 0.013
Femur length 0.023 0.848 0.000 0.068
Femur head diameter 0.014 0.320 0.000 0.869
Bi-iliac width 0.131 0.423 0.000 0.291
Souce: Ruff et al. (2018); author’s computations. Estimates in bold are significant at the 10 percent level. 

We see that the collapse in body mass associated with the Late Medieval-Early modern is extremely significant. Let’s look at what this looks like. We begin with stature, perhaps the easiest to interpret anthropometric measure of living standards. Interestingly, the fall of Rome was followed by a significant increase in stature. The rise in stature in the very recent period is also statistically significant.


With body mass, only the Late Medieval/Early modern collapse is significant. Body mass is an excellent measure of the scale of the nutritional needs of an animal, so the fact that we see no rise in the Early modern/Very recent transition suggests that our most recent sample is indeed quite limited. The rise in European body mass in the twentieth century has been much more dramatic than that evident in these graphs. But again, we are not going to change anything by hand. Even if the sample is unrepresentative, let the data speak!


As mentioned before, direct bone measurements are more reliable than allometric estimates of stature and body mass. Of all our measures, the one that contains the most information on living standards is femur maximum length (the length of the thigh bone), which controls stature. Table 1 shows that the increase after the fall of Rome and the Early modern collapse are significant at the 1 percent level. The recent rise is significant at the 5 percent level.


Femur head diameter and pelvic bone width are less plastic or slower moving than femur length (ie they are under tighter genetic control). Nonetheless they are excellent indicators of overall body size; at any rate better than body mass estimates due to estimation error. Bi-iliac width or pelvic bone width is particularly important because it controls natural BMI. Populations with wider pelvic bones have higher BMI, meaning that we should really not be using BMI as a rough measure of fitness for that penalizes people with wider pelvic bones for no good reason.

Since femur head diameter and pelvic bone width are very slow-moving (canalized) we should really not expect to see much movement in these variables in the few generations that make up the Early modern-Very recent transition.

In line with the stability of pelvic bone width, we find that only the Early modern collapse is significant. Even more than stature, the collapse in pelvic bone width reflects a prolonged, multi-generational deterioration in living standards that worked at least partly through selection against bigger, more expensive bodies.


Femur head diameter is also heavily canalized (ie under tight genetic control). The Early modern collapse in this variable is significant at 1 percent. The increase in femur head diameter after the fall of Rome is also significant at the 5 percent level, suggesting that the improvement in living standards during late antiquity were prolonged enough to change European genotypes over centuries.


Zooming into the past 2500 years helps us see the major shifts in European living standards. The big picture is that living standards improved after the fall of Rome and collapsed outright in the early modern transition. The latter in particular demands explanation. I have suggested previously that the collapse may be related to epidemiology (due to both the epidemiological unification of the world and early modern urbanization that resulted in greater incidence of communicable diseases). The strong selection against big bodies suggests that this factor may have been crucial. That would also explain why net nutritional standards rose dramatically after the introduction of indoor plumbing, modern medicine, urban sanitation, and public health measures from the late-nineteenth century onwards.

Is it not astonishing that in the very centuries than Europe was spreading its dominion over the world Europeans were literally shrinking?


An Illustrated Guide to European Living Standards in the Really Long Run

Part of the problem with modern scholarly discourses on global polarization is that they get the explanandum wrong. Contrary to the emplotment of modern economic historians, global polarization did not begin with the Industrial Revolution. More precisely, the latitudinal gradient in living standards became steeper in the century after 1870 as high latitude nations pulled away from low latitude nations. But the gradient already existed before the nineteenth century. When did it come into existence? I will show that it has always existed; that high latitudes have been favored for at least forty thousand years.

No matter what numbers are being thrown around, we do not have data on effective incomes going back more than a few centuries. And even the national economic statistics that we do have are not entirely reliable. Much more reliable are actuarial measures like life expectancy. But these too aren’t available beyond a few centuries. We have no choice but to rely on anthropometric measures like stature and body size. Fortunately, these are faithful indicators of living standards, particularly when averaged over long periods. The reason is very simple. Whether or not there were Malthusian cycles before the breakthrough to industrial modernity, long run differences in height and body size between chrono-demes indicate differences in net nutritional status (otherwise they couldn’t possibly last for so long). This true both diachronically and synchronically; that is, it holds across the panel data. We’ll focus on the former in this dispatch. But we’ll get the latter out of the way first.

We will interrogate variables from the Ruff et al. (2018)’s osteometric dataset that are known to correlated with everyday living standards from modern data. Table 1 reports the estimates. The latitudinal gradient is estimated by OLS after controlling for fixed effects for sex and period. We can see that it is extremely significant for all variables except pelvic bone width (which may have to do with small sample sizes for specific periods). Note that this is within Europe. Across the globe the tyranny of the isotherms is even more manifest. The parameter estimates for the dummies in Table 1 measure how much the periods diverge from the Bronze Age.

Table 1. Linear Regressions (tStat). 
Stature Body mass Femur maximum length Pelvic bone width Femur head diameter
Intercept 134.56 38.29 96.22 70.70 80.45
Male 34.67 32.82 33.31 8.31 43.89
Latitude 7.79 7.02 6.70 0.69 6.96
Early Up. Pal. 5.70 3.48 5.57 0.23 4.32
Late Up. Pal. -3.02 0.23 -4.33 -2.69 1.82
Neolithic -2.12 -2.25 -2.89 -3.56 -2.10
Iron/Roman -1.52 0.91 -0.01 -0.36 2.21
Medieval 0.62 2.27 2.30 -1.61 4.72
Early modern -5.46 -3.48 -3.39 -4.69 -2.07
Very recent -2.67 -2.03 -1.33 -3.71 -1.90
Adj R^2 0.39 0.36 0.38 0.08 0.08
N 2120 2053 2040 1195 2031
Souce: Ruff et al. (2018); author’s computations. Estimates in bold are significant at the 5 percent level.

The following tables report the means for the periods.

Table 2. Anthropometric measures for European males. 
Min age (ka) Max age (ka) Femur max length Femur head diameter Pelvic bone width
Early Upper Paleolithic 26,406 33,700 482 49 279
Late Upper Paleolithic 6,025 21,922 435 48 270
Neolithic 3,975 7,300 449 47 270
Bronze 2,950 4,350 453 47 275
Iron/Roman 1,650 2,250 449 48 276
Medieval 550 1,350 458 49 274
Early modern 150 320 436 47 263
Very recent 10 100 448 47 267
Souce: Ruff et al. (2018).
Table 2. Anthropometric measures for European females.
Min age (ka) Max age (ka) Femur max length Femur head diameter Pelvic bone width
Early Upper Paleolithic 26,406 32,285 432 46 268
Late Upper Paleolithic 5,928 19,013 410 43 260
Neolithic 3,975 7,300 415 42 259
Bronze 2,950 4,350 413 42 272
Iron/Roman 1,650 2,250 421 43 266
Medieval 550 1,350 421 43 265
Early modern 150 320 417 42 262
Very recent 10 100 412 41 263
Souce: Ruff et al. (2018).

The basic pattern is hard to escape. Pre-LGM populations were bigger and taller than populations at the other end of the great bottleneck associated with the Last Glacial Maximum. The decline between the Early and Late Upper Paleolithic is highly significant. During the Neolithic, Europe saw the arrival of slightly taller but otherwise smaller populations of farmers from the Near East. This is expected since height is more plastic than other measures of body size and these populations were adapted to the warmer climes (and thus smaller in accordance with Bergmann’s rule). The same biological populations survived into the Bronze Age. Appropriately, we see little change between the Neolithic and the Bronze Age. Although female pelvic bone width does widen. The Iron/Roman period is after the arrival of the Yamnaya pastoralists. These were bigger people but not so wide at the hips. Stature increased between the Iron Age and the Medieval Age, but only for males. Body mass barely increased. There was a major fall in living standards in the early modern transition. Was this related to the epidemiological unification of the world or the transition to urban life? At any rate what is clear is that there was a severe decline in living standards between the Medieval and the Early Modern eras. In the last transition, that between Early Modern and Modern eras, we see that male living standards improved but female living standards did not. But we should note that male measures are more reliable since some basic parameters of female skeletons, especially pelvic bone width, may be more plastic due to the rapid feedback from maternal mortality. BodymassheightFemurFHDPelvic

So I was not kidding about it being downhill since the days of the Gravettian mammoth hunters. Do not miss the force of the argument. Averaged over thousands or hundreds of years, anthropometric metrics like stature and body mass contain a very strong signal of living standards. The big picture is clear from the index displayed next. It adds up the values of femur maximum length, pelvic bone with and femur head diameter after standardizing them to have mean 0 and variance 1 (and adding 100 to the result). It could not be clearer that pre-LGM populations were dramatically bigger than later Europeans; that there was a significant shrinking of European body size between 1450 and 1720 CE, followed by a significant improvement in the last 100 years or so.


The notion that European living standards were stuck in a Malthusian trap is inconsistent with the evidence marshalled here. The evidence indicates that European nutritional standards were higher for two thousand years before 1450, when they began to decline, not to improve until the late-nineteenth century. Is the Malthusian Trap a real thing or a phantasmagoria of economic historians seeking a structural break on the wrong side of 1800? And what on earth were the Gravettians eating? (Ha.)



The Great LGM Bottleneck; Or, Has It Been Downhill Since the Days of the Gravettian Mammoth Hunters?

Because Europeans have been obsessed with ‘the races of Europe’ — the title of Ripley (1899); later echoed by Coon (1939) — and craniologists prevailed against philologists in identifying them after the professionalization of science in the late-nineteenth century, all human taxa (specific and subspecific) are still identified through craniology. This is as it should be since there is a stronger population history signal in craniometrics relative to linguistics (although the strongest population history signal is in dental polymorphisms). However, what physical anthropology (molecular or otherwise) of the living can tell us is the phylogenetic relationship of living populations. It cannot tell us about population history, or about the histories of vanished paleo-demes. For that we need molecular or cranial paleontology. That is, we need fossils.

Contemporary populations of European ancestry are admixtures of Yamnaya pastoralists who arrived from the steppe 5ka and Neolithic farmers who arrived from the Near East around 9ka. Reich et al. assumed that under these Neolithic populations was a uniform substratum of Upper Paleolithic hunter gatherers presumably descended from the Cro-Magnons, although they did distinguish between western and eastern paleo-demes.

That presumption turns out to have been badly wrong. The reason is that the great ice age that lasted from 30-20ka — the last glacial maximum (LGM) is dated to 26ka — generated a very severe population bottleneck. The impact was particularly drastic in Europe where faunal populations, including human demes, became isolated in refugia (largely in southern Europe). The drastic reduction in population size meant a major loss of genetic diversity. For instance, the diagnostic haplogroup for Asians that present-day Europeans do not have, has been found in pre-LGM populations. It was lost during the LGM-bottleneck.

Replicating the work of Brewster et al. (2014), I will show that pre-LGM paleo-demes were physically different from post-LGM paleo-demes. These paleo-demes were the authors of remarkable Upper Paleolithic cultures, particularly the Gravettian. Indeed, the LGM discontinuity is so drastic that we may speak of the vanished civilization of the Gravettian peoples. We’ll look at fossil skeleton metrics for pre-Neolithic paleo-demes in Europe. Unlike Brewster et al. (2014), we’ll distinguish between three chrono-groups: pre-LGM/Early Upper Paleolithic 35-20ka, late glacial/Late Upper Paleolithic 18-10ka, and Holocene/Mesolithic 10-5ka. We’ll label them with the predominant cultures in these sample: Gravettian, Magdalenian, and Mesolithic. All cranial data is from Brewster et al. (2014), and all postcranial data is from Ruff et al. (2018), Skeletal Variation and Adaptations in Europeans: Upper Paleolithic to the Twentieth Century.

We begin with postcranial size variables. There was a steady decline in stature over time. The Gravettians were considerably taller than later paleo-demes. Stature.png

They also weighed a bit more. Although the t-test is not significant (p=0.152).

body_mass.pngRecall that these are allometric estimates. We have much more reliable measures of body size. Femur head diameter and pelvic bone width are not only the main weight carrying parameters of the human skeleton, they are also under tighter genetic control.


Gravettian and Magdalenian femur head diameter and pelvic bone width are not significantly different. However, Gravettian and Mesolithic populations are significantly different by both measures (p=0.078 for femur head diameter and p=0.016 for pelvic bone width).


Gravettian and Magdalenian femur length differences are highly significant (p<0.001). Recall that femur lengths and stature are controlled by net nutritional status. This suggests that the Gravettians had much higher nutritional standards than later paleo-demes.

femur_max_length.pngMoving onto craniometrics, we look first at skull size calculated as the geometric mean of skull length, height, and width. The difference between Gravettian and Magdalenian skull size is not statistically significant (p=0.706). In sum, the Gravettians were taller than Magdalenians but not otherwise bigger. Both were however absolutely bigger than later Mesolithic Europeans.


We now look at cranial variables after adjusting them for skull size. So these are shape variables that contain information largely orthogonal to the size variables we have so far examined. Of the 10 cranial measurements in the Brewster et al. dataset, only in the following four is there a statistically significant difference between the pre-LGM Gravettian and the post-LGM Magdalenian skulls. Gravettian skulls were longer.


Gravettian foreheads were broader.


They had taller noses …


… that we narrower.


Table 1 reports the results of our t-tests for size-adjusted Gravettian-Magdalenian craniometric differences.

Table 1. Gravettian vs. Magdalenian craniometric characters. 
Martin number Description pVal
M1 Maximum cranial length 0.026
M8 Maximum cranial breadth 0.970
M9 Least frontal breadth 0.000
M17 Basibregmatic height 0.194
M45 Bizygomatic breadth 0.454
M48 Nasoalveolar height 0.846
M51 Orbital breadth 0.968
M52 Orbital height 0.171
M54 Nasal breadth 0.003
M55 Nasal height 0.016
Source: Brewster et al. (2014), adjusted for skull size.

Who were these people with long heads, narrow and tall noses, and broad foreheads? Why were they so tall? People have tried to trace the tall stature of contemporary Bosnians to inheritance from this paleo-deme. That patent nonsense for stature is controlled by health insults. Whatever the correlations with haplotype, this variable is just too plastic. Moreover, the genotype of the Gravettians largely vanished in the LGM bottleneck. All later demes, including and especially contemporary populations, inherited little of their genotype. As far as present-day populations of European ancestry are concerned, they might as well be Neanderthals.

But I would go much further. Ancient paleo-demes, particularly pre-LGM ones like the authors of the Gravettian, cannot be projected onto contemporary classifications. Not only do they have different genotypes and different cranial morphologies, they also had different skins and eye colors than contemporary Europeans. For instance, we know that the genetic basis for depigmentation came under selection in Europe only 7ka; while that for blue eyes came under (probably sexual) selection 14ka. We don’t know yet if the Gravettians looked fair or if some of them had blue eyes. Possibly. But if so, it had a different genetic basis from contemporary morphotypes.

What is clear is that they had much higher standards of living than later paleo-demes. Why? The answer is the great mammoth steppe.


Source: Clive Gamble, Settling the Earth.

During the Upper Pleistocene, vast herds of woolly mammoth and other megafauna roamed across the northern Eurasian steppe, extending from the great European plain all the way to north America. How was it possible to sustain such high secondary productivity during the ice ages? As Gamble reports,

The mammoth steppe comes down to a simple observation by zoologist Dale Guthrie – that today’s boreal and tundra vegetation of Beringia could not have supported the large herds of bison, horse, reindeer and above all mammoth whose bones, and sometimes carcasses, are preserved by their millions in its frozen silts. Tundra plants and boreal trees produce a toxic litter that affects the soil, leaving very little for animals to feed on. Furthermore, they act as a blanket so that the level of annual freeze–thaw is small. The factor which changed this balance between plants and soil was the aridity of the Pleistocene. Creating the mammoth steppe depended on high evaporation and a deeper thaw in summer that released nutrients from lower down in the soil. This, according to Guthrie, broke the cycle of low soil nutrients and toxic plants. It produced a richer soil and the conditions for the abundant growth of grass. These grasses were resistant to grazing pressure, grew quickly and formed a rich mosaic of vegetation conditions, likened to the weave in a plaid, and contrasted with the stripes, or bands, of vegetation found in the warmer Holocene. Isotope studies suggest that mammoths consumed higher quantities of dry grass than the other grazers such as reindeer, horse and woolly rhino. The mammoth steppe developed across Western Europe to Beringia during a major glaciation, MIS12, 500ka. It supported a diverse, high-biomass animal community, compared to the present tundra and boreal forests. Besides the major herd animals — mammoths, woolly rhinos, bison, horse, red deer, reindeer, musk ox and saiga antelope — there were major carnivore guilds of lions, hyenas, wolves, leopards and foxes, and omnivores such as bears.

The Gravettian civilization was based on the western extremity of the mammoth steppe. As the mammoth steppe contracted east, the Gravettians moved with it. The west-east movement is very marked in Gravettian sites. While some Gravettian populations specialized in mammoth hunting, the notion that they were all specialized mammoth hunters has been discredited. The Gravettians certainly hunted mammoths — they were probably the first humans to do so — but they also hunted other game. Their kitchen middens display significant variability. They did however collect mammoth bones at scale. They build shelters with them, used them to make all sorts of tools. Mammoth bones very likely also had a religious function in Gravettian civilization. For as Soffer (1993) reported and was recently confirmed for a Russian site, the spatial patterns of Gravettian camp sites suggest that mammoth bones were segregated from those of other animals. 

So the Gravettians not only had a highly distinct morphology (and genotype), they also had a very distinctive way of life that was highly advanced. They were certainly better fed than later populations. We are just beginning to appreciate the ways of life, and indeed, histories, of this ancient civilization.

This foray into the Gravettian mammoth hunters will help us think about combined and uneven development in the Upper Pleistocene. More on that later.



A Simple Test for the Existence of Continental Races in Homo Sapiens

Neoracialists insist that if we apply the same scientific standards to our species as we apply to other species we will find that like most species in the animal kingdom, humans too have subspecies. It is true that systematists recognize subspecies in many animal species. Chimps are a famous instance. But systematists have not recognized subspecies in Homo sapiens for decades. Is it because of they have let politics interfere with science, as the neoracialists claim? Or is it because scientists are convinced that our species does not have subspecies? And do they have good reasons for thinking that? After all, whatever the politics, that there may be subspecies in Homo sapiens is a scientific hypothesis.

The dominant process of speciation in the animal kingdom has been well understood since Ernst Mayr’s work at mid-century. Despite the promise of the title, Darwin’s The Origin of Species had failed to provide a compelling mechanism of speciation. Mayr showed that speciation is essentially a process whereby geographically isolated populations diverge from each other until they acquire isolating mechanisms (eg, different mating calls). This is called allopatric speciation. This is a process; not an event. The test is what happens when the ranges of such isolated populations overlap. If they do not interbreed and occupy different niches so that they can exist sympatrically, they are described as good species. If they show marked differentiation but still interbreed in the wild, they are described as allopatric subspecies or continental races. If they can interbreed but rarely do so and cannot exist sympatrically (say if they occupy the same niche) they are called allospecies. So allospecies is the halfway house between allopatric subspecies and good species. With these definitions, Upper Pleistocene hominin taxa are best regarded as subspecies or at worst allospecies. They were certainly not good species since we know Sapiens interbred with them and coexisted with them for at least ten thousand years (40-30ka in Europe).

What about the alleged extant races of man? The usual proof of nonexistence offered by scientists is that racial taxonomy explains a negligible portion of human genetic and phenotypic variation. The game was initiated by Lewontin’s famous intervention in 1972, “The Apportionment of Human Diversity“. He showed that 6.3 percent of human genetic variation is accounted for by racial taxonomy. Moreover, 85 percent is within ethnic groups (racial anthropologists’ “subraces”). So it’s pretty much a useless fiction. More recently, Templeton (2013) revisited the question and explicitly compared humans and chimps. Table 1 reproduces his estimates. The numbers make it obvious that there is ample reason to recognize subspecific taxa in chimps but not in humans.

Table 1. Apportionment of genetic diversity in humans and chimpanzees. 
Species Number of “Races” Number of populations Among individuals within populations (%) Among populations within races (%) Among races (%)
Chimpanzees 3 5 64.2 5.7 30.1
Humans 5 52 93.2 2.5 4.3
Source: Templeton (2013).

In what follows, I will offer a more compelling test of the existence of continental races. In the previous dispatch I demonstrated that dental polymorphisms contain an extraordinarily strong population history signal. If there are continental races in man, then these must map onto the phylogeny obtained from physical anthropology. Figure 1 shows the first and second principal components obtained from Turner’s data on dental polymorphisms. It’s clear that Turner’s Sundadonts (Southeast Asians and Polynesians) and Sinodonts (Northeast Asians and Americans) cluster together, as do the Africans and west Eurasians (Europe, India and the Middle East). Also, Melanesians and Australians cluster together but far from New Guineans. These roughly correspond to the big continental races proposed by racial taxonomists from Coon to Wade: Caucasoid, Negroid, Mongoloid, Australoid, although Sundadonts have not been recognized as such by any authority to my knowledge. Still, we will not change anything by hand. Let the data speak!


Figure 1. First two principal components of variation in dental polymorphisms.

The phylogram obtained for these (admittedly broad) “races” is displayed in the next figure. It’s a reduced form version of the more fine-grained one we derived in the previous dispatch and consistent with all the evidence from archaeology, linguistics, and physical anthropology. What it reflects is the peopling of the world and genetic isolation over thousands of years. If these aren’t races, there aren’t any.


Figure 2. High level phylogeny. Source: (Turner 2018), author’s computations.

So, the test. If this “racial” taxa obtained from the slowest-moving and least confounded variables deserve subspecific status, then other systematic variation in human biology should map onto this classification. More precisely, phenotypic characters cannot be properly described as “racial” if they yield maps that are inconsistent with each other. For instance, skin reflectance does not map onto this classification since northern populations across these taxa are depigmented while low latitude populations are strongly pigmented. Okay, so skin reflectance is not diagnostic and that’s been understood for a long time. Like many other species, Sapiens obey Gloger’s Rule — pigmentation falls in high latitudes and rises in low latitudes. Even Aristotle knew this!

I am instead going to look at RBC polymorphisms. Blood groups are known to vary systematically. Famous neoracist Rushton made a big fuss about small differences in mean frequencies of blood groups between his three races (White, Black, and Asian; obviously). Well, let’s see if there is anything to it.

If these taxa are deserving of subspecific status then this systematic variation should map onto our classification, particularly since we are working at such an absurdly coarse level of analysis. If it does not even work here, then the case for awarding subspecific status to these taxa is really tenuous. Table 2 displays the frequencies of blood types for our taxa.

Table 2. Frequencies of blood groups.
Africa Western Sundadont Sinodont Sahul
O+ 0.45 0.36 0.39 0.50 0.40
A+ 0.27 0.32 0.26 0.28 0.31
B+ 0.18 0.14 0.28 0.14 0.08
AB+ 0.04 0.05 0.06 0.04 0.02
O- 0.03 0.05 0.00 0.03 0.09
A- 0.02 0.05 0.00 0.01 0.07
B- 0.01 0.02 0.00 0.00 0.02
AB- 0.00 0.01 0.00 0.00 0.01

We can see that there is significant systematic variation in the frequency of blood groups. If the taxa deserve to be recognized as subspecies then the distances obtained from RBC polymorphisms should yield similar phenotypic distances to that obtained from dental polymorphisms (which we know contains the strongest phylogenetic signal). The basic test for the similarity of distance metrics is the Mantel test. If the distance matrices are similar the P-value of the Mantel test statistic should be close to zero (0.05 is the usual standard of significance). Table 3 displays the test results.

Table 3. Mantel tests. 
Pearson Spearman
Statistic -0.569 -0.405
P-value 0.958 0.913
Source: Turner (2018), author’s computations.

So our racial taxonomy fails catastrophically. This is generally the case for most so-called racial characters. There is plenty of systematic variation in phenotypic characters. But they don’t map onto each other. (Are gingers a race?) If there were actual continental races in Homo sapiens, the world would look very different. And this is no hypothetical. For there were in fact continental races of man for most of our evolutionary history (Neanderthals, Denisovans, and so on). We mercilessly wiped them off the face of the earth like all the other megafauna of the Upper Pleistocene. Make no mistake: Our race is the most dangerous species to have ever walked on this planet.


Out-of-Africa Phylogeny from Dental Polymorphisms

I discovered the solution to the problem I was working on. Rather I discovered that Hanihara was struggling with the same problem and that he had figured out the solution by 2008. The population history signal is contained in the neutral drift component of systematic variation in human phenotypes. By neutral, I mean derivation due not to diversifying selection but rather to founder effects and isolation-by-distance. There is definitely a population history signal in cranial and craniofacial characters. A weaker signal is found in phonemic data and postcranial skeletal morphology. All of the above, with the exception of linguistic data, are to a greater or lesser extant confounded by natural selection. And linguistic data has a weak population history signal, certainly at great time depths.

So I have been looking at craniodental data. Metric craniodental data (teeth diameter etc) contains as strong a population history signal as craniometrics. But as it turns out we can do much better. The real gold is in dental polymorphisms. Polymorphisms are discrete variants of phenotypes, eg, blood groups (called RBC polymorphisms), hair color (we get the ginger from Neanderthals), eye color, and so on. Such polymorphisms can usually be found in all populations. The differences in their frequency contains some information on population history. This is usually badly confounded by, say, sexual selection. Think of hair color or eye color.

In the case of dental polymorphisms, there are good reasons to believe that the population history signal is not confounded at all so that frequency differences between populations reflect shared inheritance. The basic operating logic is that of Mayr’s founder effect. All founder populations carry with them only a subset of the phenotypic diversity. Moreover, their smaller effective population size means even greater loss of diversity over time as lineages die out randomly.


Figure 1. Founder effect. Source: Wikipedia Commons.

In a recent paper in Nature, Hanihara and others showed that dental morphology contains an extraordinarily strong population history signal that is highly correlated with the best estimates from molecular anthropology. So this is a very promising line of investigation.

Let me illustrate with some frequency distributions for some of these dental polymorphisms. Figure 2 displays the incidence of shoveling of the first Upper Incisor (UI1) by macro region. Why has this morphotype reached near fixation in northeast Asia and the Americas? Why do the western old worlds and eastern old worlds cluster together and on opposite sides of Africa which is right in the middle? The right answer lies in the details of the Out-of-Africa dispersals. Phylogeographer Stephen Oppenheimer argued forcefully in Out of Eden for a thick version of the southern route or beachcomber hypothesis whereby there a single exit from Africa to India, and it was from the subcontinent that Homo sapiens heading west to Europe, south and east to Sahul, a branch of whom instead went up the coast to China. The morphotype achieved near fixation in northeastern Asia before the founder populations migrated to the Americas across Alaska. Moreover, Oppenheimer argues that the big freeze of the Last Glacial Maximum about 20ka decimated northern populations and reduced their genetic diversity. The actual fixation of this morphotype in northern America may reflect this natural history.


Figure 2. UI1 shoveling incidence rate by macroregion.

In the case of the awkward-to-pronounce Hypoconulid polymorphism, we can see the eastern and western branches going out from India both became derived; in the opposite direction. The morphotype became more frequent in the west and less frequent in the east, with a particularly severe decline in frequency in Australia and northern America. LGM strikes again?


Figure 3. Frequency of the Hypoconulid morphotype.

The premolar accessory cusp morphotype is rather rare outside Australia and Melanesia, suggesting that this is a derivation in these populations after they split from the others. This ancient clade also stands out in metric craniodental traits. The derivation is largely a function of the great time-depth of the split. It is no surprise that geographic isolates display marked derivation in the human species. It holds across the animal kingdom.


Figure. Frequency of premolar accessory cusp morphotype.

Almost simultaneously as their paper in Nature, Hanihara and gang published an extraordinary study in Current Anthropology, where they showed that of all the competing Out-of-Africa dispersal scenarios, Oppenheimer’s beachcomber single-dispersal Out-of-Africa via India scenario (“BSD” in the next figure) is the most consistent with the phylogeny obtained from dental polymorphism frequencies.


Figure 4. Out-of-Africa dispersal scenarios. BSD = beachcomber arc single dispersal; EE = eastward expansion single dispersal; MD = multiple dispersals; MDI = multiple dispersals and Australo-Melanesian isolation.

I like to check things with my own two hands. I used the Hanihara (2008) data on dental polymorphisms to extract a phylogram that reflects our population history. In order to compare the frequency data from different polymorphisms, I converted them to percentile scores. Then I used the Euclidean metric to obtain pairwise phenotypic distances, from which I obtained the phylogram using the standard neighbor-joining algorithm. The accuracy of the resulting phylogram is simply astonishing. I hit pay-dirt alright.


Figure 5. Phylogeny from dental polymorphisms.

One can read off our population history from this diagram. Homo sapiens left Africa via the Red Sea route to India 100-80ka. While many founders stayed behind on the subcontinent, some beachcombers kept going further east. These populations reached Sahul by 60ka. The Negrito populations of the Andaman Islands and Melanesia, and the Australian aboriginal population are actual relict populations from that original dispersal. (BTW, the Negritos aren’t black and fizzy haired because they are close to Africans. Their dark skins reflect directional diversifying selection under similar tropical conditions; like the different origins of depigmentation on either extremity of Eurasia, it’s the classic case of a homoplasy.) On their way to Sahul, some of these founders stayed back in southeast Asia which at time was a large continent called Sunda. Some of them would later go on to become Polynesians and the greatest sea-farers the world has ever seen. Others went up north to northeast Asia and later from there on to the Americas soon after the Last Glacial Maximum. Before all these developments in the far east, pioneer Sapiens ventured forth north and west from India, taking the trans-Caucasus route to Europe where they would go on to “replace” Neanderthals. The move back into Africa (the Ethiopians are the descendents of this reverse migration) followed after southwest Asia was finally peopled (it had been a forbidding desert before).

So here’s the kicker. When Dravidians (9ka) and later Indo-Europeans (4ka) reached India, it was a reunion long in the making. What is astonishing is how this turn table between India and Europe has been turning throughout deep history. In a sense, the paleogeographic logic of this pattern puts the flesh on the bone on Diamond’s insight — that the east-west Eurasian axis was more advantageous than the north-south axes of Africa and the Americas for the simple reason that people, ideas, and technology (and germs) could move faster against the latitude gradient than along it. What is amazing is the emerging picture of how Diamond’s logic worked in practice — at least where pots were moving with people (one can hardly demand more of physical anthropology).

Oppenheimer argues that the Upper Paleolithic revolution in Europe (it is to paleoanthropologists what the Industrial Revolution is to modern economic historians and the Neolithic revolution is to prehistorians) was the handiwork of founders from the Indian subcontinent. He traces both the people who authored the Aurignacian culture c. 43ka as well as the later arrivals who authored the Gravettian c. 35ka, the great mammoth hunters of north-central Eurasia, to the subcontinent. But that deepens the paradox of the Upper Paleolithic. If the authors of the Aurignacian came from the subcontinent and so did those of the Gravettian, they why does the Upper Paleolithic not reach India until tens of thousands of years later? Was it combined and uneven development structured by our Heliocentric geometry? Or was something even deeper at play. That’s the big open question of paleoanthropology.