The Roman, the Hun and the sun

I chose a fortuitous time to read Kyle Harper’s The Fate of Rome: Climate, Disease, and the End of an Empire. This is a great book, and a nice compliment to Bryan Ward-Perkins The Fall of Rome: And the End of Civilization. Where Ward-Perkins attempts to convince you that Rome did indeed fall, and that that fall mattered, Harper takes it as a given that you accept this position. Rather, he tries to show you in The Fate of Rome that a series of contingent and necessary causal factors set the Roman system up for its fall. The fall of Rome is not just an idea, but a material event that was given a strong push by material factors.

As the The Fate of Rome was published in the fall of 2017, so it was written well before recent work which highlights both the nature and role of steppe barbarians in triggering the changes which we dramatically term the “fall of Rome” and the “barbarian migrations.” A few months ago I wrote about a paper which reported that post-Hunnic people of the Balkans were genetically different from typical Europeans in that they exhibited some East Asian admixture. Harper does assume that the Huns were barbarians whose ultimate provenance was somewhere in the region of modern Mongolia, but emphasizes that their peregrinations transformed them.

As so it did. A new paper in Nature, 137 ancient human genomes from across the Eurasian steppes, nails the overall dynamics. As illustrated in the figure above the early steppe was dominated by peoples of a West Eurasian provenance, while the latter steppe shifted toward a more East Asian shifted population.

These early groups go by various names. But the Cimmerians, Scythians, and Sarmatians have origins on the Pontic steppe. Flourishing in the first millennium before Christ, I should precisely label them “Iranian,” but that might mislead readers a bit since some of these groups were never resident within Iran. The Scythians were a presence across a huge zone of Inner Asia and were a force in Eastern Europe, West Asia, South Asia, and in Eastern Asia. Likely emerging out of the Andronovo culture, genetically the results from the paper confirm early work that Scythians mixed with the local substrate where they went. In this way, they prefigure later steppe populations. Being a nomad was a lifestyle, the genetic correlates to some extent an accident.

In The Fate of Rome  the Huns have a role to play as a push for the migration of Goths into the Roman Empire, which eventually leads to their rebellion and a collapse in both the prestige and military manpower of the Roman state. The genetic evidence above and elsewhere is strongly indicative of the likelihood that the Huns were originally part of the Xiongnu confederacy. As they moved west they mixed with post-Scythian and other Iranian and Siberian elements, and presumably by the time they arrived on European frontier of Rome they had picked up some Germanic and proto-Slavic ancestry. In 137 ancient human genomes from across the Eurasian steppes the authors also report that the East Asian gene flow was somewhat “male-mediated” in the later steppe. Similarly, earlier work on proto-Iranian peoples in the Altai region is strongly suggestive of male-mediation in West Eurasian gene flow.

The obligate and exclusive Eurasian nomad lifestyle was one dominated by men, though as one can see the importance of Genghis Khan’s wives and daughters women maintained independence as well.

For whatever reason, full-blown nomadism only became a feature of the landscape north of what became China in the last few centuries before Christ. The mobile and militarized nomadic lifestyle that emerged in western Eurasia in the years around 1000 BC seem to have taken five centuries to penetrate the far eastern fringes. Until the crushing of the Dzhungar’s by the Manchus in the 18th century, 2,000 years later, the dynamic between nomad and settled was a defining feature of Chinese statecraft and political culture. And, it was also a major feature of nomad culture, because the wealthy Chinese state was an almost irresistible attraction to steppe elites as a source of plunder and tribute.

But human action is not the only relevant parameter in human history.  The Fate of Rome  is fundamentally a work of history, but it also takes ecology and evolution seriously. In fact, it foregrounds them. Kyle Harper makes the argument that the expansionary phase of the Roman Empire was not necessarily coincidental, or at least it was lucky indeed because there was a climatic optimum, similar to the one which preceded the demographic expansion of medieval Europe. In contrast, in the 6th century, the world went through some of the coldest years in the Holocene because of a combination of fluctuations in solar radiation and volcanic explosions. I assume that the likelihood of the latter is Poisson distributed, so the combination of decreased radiation and several successive volcanic events can be chalked up to randomness. But its consequences were not random at all.

The climatic changes can have demographic and social consequences obviously. Desperate armed pastoralists can overwhelm states, and change the course of history, just as peasants can rebel from taxes and subordination. And, pastoralists can also bring Yerisina pestis, the plague. Climate is an abiotic pressure which is to some extent an exogeneous shock which occurs randomly, and does not react to human feedbac k.* Disease though is a biotic pressure, and though it may relate to abiotic forces, human interaction and agency matter quite a bit.

The Fate of Rome clearly hinges on abiotic factors as initial drivers: a good harvest is good for the state. But the biotic factors, disease, are partly under the control of the state. The Romans did not have germ theory, and were under constant stress due to the high pathogen load, especially of the cities. Harper presents the evidence of high mortality within Roman society well. Because of the endemic ubiquity of disease even elites were impacted by it. But Rome was not just affected by endemic ailments, it was subject to pandemics and plagues. Three loom large in  The Fate of Rome:

  • The Antonine Plague, which ended the expansionary phase of the High Empire in the middle to late 2nd century.
  • The Plague of Cyprian of the middle 3rd century which ushered in a period of state collapse.
  • And finally, the Justinian Plague which marked the end of Late Antiquity and the beginning of the “Dark Ages.”

One of the major insights that Kyle Harper reiterates is that these plagues, these pandemics, are a feature/bug of the Roman imperial system. They are not just the consequence of simply settled agricultural society. As described in books such as Pandora’s Seed, agriculture and settled society transformed the lifestyles of human groups, and many diseases which were rare in hunter-gatherer populations probably became common among farmers. But The Fate of Rome the author argues that pandemics were a novel outcome of complex imperial state-systems with long-distance trade-networks. Small-scale pre-state Neolithic chiefdoms did not have the scale and interconnections to foster plague.

Mass pandemics of smallpox, plague, and influenza are then aspects of civilized life, not, settled agricultural life. This puts the argument of Charles C. Mann in 1493 into greater focus. It wasn’t just more extensive and intensive agriculture in the Old World which left Amerindians vulnerable, it was also that the Old World had thrown up several massive imperial systems which had incubated pandemic producing pathogens (smallpox and influenza epidemics were a major issue in New World societies). These were unleashed at once upon New World societies.

It also suggests to us why adaptation seems to be occurring in the last few thousand years. Bouts of plague which persisted for generations may have driven immunological responses.

Kyle Harper also seems to agree with the general thesis in The Fall of Rome that this period in European civilization was in some ways proto-modern, with economic specialization resulting in a modicum of affluence in ways unimaginable in times before, or after. Trade and some level of mass production allowed British peasants to eat off tableware that was standardized, and not homemade. In contrast after Britain’s post-Roman regression a more local economy had to step in. The most curious fact from The Fall of Rome is that pollution in British ponds did not attain Roman levels until the early modern period, with the rise of industrialization. Again and again  The Fate of Rome emphasizes that social and economic complexity achieved in the Roman Empire was not attained in Europe again to the same scale as the early modern period.

Roman wealth was fundamentally due to the returns on scale and specialization that are the hallmark of Smithian growth. Though the Romans did invent a few things, Roman prosperity was not fundamentally driven by innovation. Rather, the Roman peace was a framework for trade and exchange that took advantage of abiotic clement conditions (the Roman climatic optimum highlighted in The Fate of Rome).

But this political system had biotic costs, as well as being subject to biotic shocks. Though Romans may have been wealthier than their Iron Age predecessors in things, and also wealthier than their early medieval successors, they were also a smaller people. Using isotope data Harper suggests that this is not due to Malthusian immiseration as the imperial population pushed up against food supply. Apparently Romans did not subsist on gruel alone, but ate a fair amount of meat, especially pork. Rather it was the high pathogen load enabled by the advancement of Roman urban life and its scale. Rome was a world of intense morbidity.

Unlike physical/abiotic forces biological/biotic pressures on human existence are adaptive. Moderns know this with the rise of antibiotic resistance, it’s the eternal race. The Romans were not aware of the consequences of their means of prosperity, and were not ready for the exogenous shocks of climate and disease which were to perturb their state system.

But The Fate of Rome is not just a story of exogenous factors, climate and disease. Rather, Harper puts into stark relief the variables which might push an empire over the edge, or eat into its seed corn of human capital. That does not negate the fact that endogenous variables matter. The Roman elite of the early centuries exhibited some level of asabiyyah, social cohesion. The Empire was fundamentally not a strong state in comparison to modern ones. It was a thin skein of cities and fortifications binding together an overwhelmingly rural population of villages. Its achievement of peace and prosperity was bound up in an ideology and identity focused project which bound together an elite (or bound together elites).

The origins of this elite were not always arbitrary. Though the Empire was famously cosmopolitan, The Fate of Rome crystallizes something that anyone who had sat back and thought about could see: certain groups bound the imperial state together as a ruling caste. Harper observes that between the reign of Claudius and Phocas, from 268 to 602, 75% of the Emperors were of Illyrian/Balkan stock. That is, 75% of the Emperors were drawn from 2% of the Roman Empire’s territory. The exception being the Theodosian dynasty, which was of Iberian origin and jumped into the breach after the defeat of Valens at Battle of Adrianople.

This is a fascinating fact in and of itself. Harper points out that these Emperors from the Danube frontier did not enrich their own region to the detriment of others. They were ideological heirs of the earlier Roman project, and their identity was as Romans first, Illyrians and Thracians of Latin stock second (or third, after Christianization). But they brought particular skills of administration and an overall martial attitude which served to lead the Empire through a period of greater stress than it had been subject to during the earlier climatic optimum.

The Fate of Rome does not plumb the depths of ideological and social change but emphasizes their interaction with biotic and abiotic factors. Harper observes that public temple building decreases sharply after the Cyprian plague. Why? Perhaps there was a loss of faith in the old religious institutions. Though popular paganism remained dominant, new elite religious ideologies such as the cult of the Invincible Sun and later Christianity came out of the shadows during this period.

These cultural and political aspects remain bit players and mostly offstage in The Fate of Rome. If you are interested in political narrative, then something like Peter Heather’s The Fall of the Roman Empire may be more to your taste. If culture, then Mary Beard’s SPQRBut ultimately social, political, economic, biological, and climatological factors are critical and interconnected. The rise of plague is hard to understand outside of the context of trade, which was enabled by political power and unity. Ecological factors may have driven Yerisina pestis out of its Central Eurasian reservoir, and those ecological factors may have been triggered by climatic variables.

The fall of Rome is a huge topic. I’m just glad that we’re beyond the revision of the previous generation which denied that it happened in the first place. The reason that it occurred is probably contingent in the details, though inevitable over the long-term. All things must end, even the Roman peace.

* This is not totally true, but over the time-scales we’re talking about probably mostly true.

7 thoughts on “The Roman, the Hun and the sun

  1. Regarding the Fall of Rome, I quite like Ko and Koyama’s hypothesis contrasting Rome and China:

    In this Rome faces multiple small threats on multiple fronts, each of which Rome draws small proportions of income from via levies and taxes. Germanic barbarians, Huns, Sassanid Persians, piracy in the Mediterranean. (The pastoralist nomads themselves are far less important, relevant in the thinly populated Balkans where Rome’s presence is relatively marginal and wealth is not high, and maybe more effectual through to cause disruptions on the borders of other groups that send threats Romes way. As you state, the Danubian frontier!) This raises incentives for the central power to pull out of each province, and for each province to manage its own defense, all of which pushes towards fragmentation. (Also gives more competing identities for people to throw their loyalty around).

    In contrast China has a single heartland directly facing a single, much more overwhelming threat from the North. No competing civilizations on the borders (a la Sassanids) and no groups like the Germanics, sedentary but not civilized and militarized or “mobilized” to an extremely high degree (as all small scale and small population societies are, whether their basis is pastoral or agricultural). Indochina and Tibet are far from settled by any competing powers, at least at this point in history. There’s no incentive for anything but holding together the central heartland together against the single threat.

    Within that fundamental dynamic, you have things that Rome did poorly on and the Han were more capable on, and vice versa, outbreaks of disease and climatic challenges, but fundamentally something like above as what secured survival of China and the end of Rome (as political powers).

  2. From the OP: … between the reign of Claudius and Phocas, from 268 to 602, 75% of the Emperors were of Illyrian/Balkan stock. That is, 75% of the Emperors were drawn from 2% of the Roman Empire’s territory.

    From Matt’s comment above: … in the thinly populated Balkans where Rome’s presence is relatively marginal and wealth is not high.

    An interesting juxtaposition: an obvious hypothesis is that the only career for ambitious youth was in the army, where the best/luckiest became Imperators. If so, there are parallels with the role that boxing (and sports more broadly) has played for poor American youth for much of the last century and a half, as well as (even more broadly) military service since at least WW2. The phrase “thinly populated” makes the 75% figure even more impressive.

    This area included the Macedonian heartland: did it have a very much stronger martial tradition than elsewhere in the empire?

    On a personal note, I am in the middle of The Fall of Rome and The Fate of Rome is on my night table, patiently waiting its turn. Fall is by turns very interesting and very dry.

  3. Razib: In 137 ancient human genomes from across the Eurasian steppes the authors also report that the East Asian gene flow was somewhat “male-mediated” in the later steppe. Similarly, earlier work on proto-Iranian peoples in the Altai region is strongly suggestive of male-mediation in West Eurasian gene flow.

    Connected to this, some big points on the uniparental lineages in these two papers look like this to me:

    – Presence of y R1b (R1b1a1) and N in Botai. Despite lack what looks like a lack of any geneflow from any culture like the Yamnaya with Near Eastern ancestry. N also present in the Baikal Early Neolithic groups, despite these groups not being too close to Botai autosomally (Botai look pretty modest in their East Eurasian ancestry, mostly essentially ANE).

    Strengthens possibility than N lineages did indeed move fairly back and forth between mostly East Eurasian groups like the Baikal EN and mostly ANE groups like Botai, and that modern N1a1 in Northeast Europe may have originated most recently in an ANE rich group with little East Eurasian ancestry. (The ultimate origin of N becomes more shadowy, though East Eurasia seems likely based on distribution of O).

    – On the mtdna side Botai seems to show a bit more diversity than WHG or EHG (who are strongly U, particularly U5), with lineages at Z (Native American), R and K (frequent in Near East).

    – As we get later in time, after the Okunevo_EMBA, who have a good amount of Q, steppe lineages seem almost homogenously R1, with few exceptions, until C and O emerge in the Turkic period.

    Even among the Xiongnu samples, and discounting the Xiongnu West Eurasian ancestry male (R1b), one sample (DA39) who looks pretty secure as a senior member of Xiongnu confederation (“aristocratic burial”) with an autosome profile on Eurogenes that is fairly close to present day Mongolic/Tungusic groups, and of low if any West Eurasian ancestry is R1.

    (Two “Xiongnu” samples found in a mass grave in Mongolia, DA43 and DA45, autosomally almost exactly match Northern Han Chinese on the Eurogenes tools; correspondingly they are both O3a, one of them O3a3b2, which I believe is downstream from O-M7 which today has a fairly Chinese specific distribution).

    Central Asian autosomal turnover probably lesser in a sense than the degree of y turnover, as so many people prior to those major male led expansions of Turkic and Mongolic language were already admixed between largely R1 West Eurasian males from proto-Iranic / Scythian groups and East Eurasian females. (Corollary, autosomal:y ratio probably underrates the recent male led character of Turkic / Mongolic expansion?).

  4. One other thing I’d add (taking up the whole comments here, I promise I intend this to be my final long comment), that I haven’t seen remarked on anywhere yet, is that the “137 ancient human” paper has some nice potential in terms of resolving the Indo-Iranian connection to Balto-Slavic.

    There’s been this connection in language, particular derived innovations shared between Balto-Slavic and Indo-Iranian, that linguists have been clear about, but it’s been a puzzle in population genetics, as present day Balto-Slavic speakers don’t seem to have any of the Iranian Neolithic or East Eurasian ancestry (certainly not consistently or more than other Europeans) in any way that actually would connect them to previous Sarmatian or Scythian samples (who *do* have such ancestry).

    In contrast, Balto-Slavs actually seem to have a connection to a Baltic late Bronze Age group, which looks like a fusion of the Corded Ware culture with an unusually heavy amount of local HG ancestry.

    (Ways to solve this linguistic puzzle have involved imaginative solutions postulating that Indo-Iranian and Balto-Slavic don’t actually share innovations and are just conservative, where other languages have lost the features.)

    However this paper looks like it could helps solve that. The Hungarian Scythians sampled look to cluster essentially with the Bell Beaker samples from Hungary (via Eurogenes analysis) and are described by the paper as like Andronovo with additional European Neolitihc ancestry. What this suggests is that despite large contacts in material culture and probably language with other Scythian-Sarmatian groups, the autosomal dna of people in the Hungarian steppe did not change!

    This leaves an obvious solution to the problem: present day Balto-Slavic speakers are could be descended largely from a mixture of the Hungarian Scythians and the Baltic Bronze Age groups. This explains both the linguistic connection to Indo-Iranian and the Baltic BA genetic signature. (Bonus: It also means that I can call the later expansion of Russian across Eurasia “The Revenge of the Scyth”).

    As an incidental, it also helps with all those bits in ancient European history where various early medieval groups claimed Scythian ancestry, which seemed strange for the same reasons given above (where’s the East Eurasian or Iranian Neolithic ancestry?). If it’s from the Hungarian Scyths, who are essentially European, that may not imply any significant amount of autosomal shift.

    Annoyingly, there is no y-dna assignment on the Hungarian Scythians beyond R1; if they were Eastern European R1a-Z282, then that would start to look really solid.

  5. Related to the Xiongnu/Hun theory, I recently read an interesting book covering the trade between China and Roman Empire. Among other things, the author the history of the Tarim Basin and the history of Chinese interactions with the Xiongnu and related steppes people, including the Chinese push of the Xiongnu and others out of the Tarim Basin, followed by trade route connections to the Roman Empire.

    The book is not long, but has a lot of interesting information, including a lot of comparing Chinese accounts vs Western accounts of the same/similar peoples and events in the region. It reads a little like academic work transcribed into a retail book, but that’s what I like.

    The author assumes Xiongnu = Huns (from memory, without acknowledging that is a theory and not a certainty).

    Title: The Roman Empire and the Silk Routes: The Ancient World Economy and the Empires of Parthia, Central Asia and Han China
    Amazon link:

  6. Correction – I wronte “a lot of comparing Chinese accounts vs Western accounts of the same/similar peoples and events in the region”. I overstated it – there is some interesting comparisons, but “a lot” is an overstatement.

    Nonetheless, the book is very interesting, including a discussion in an early section regarding Chinese vs Roman iron making technologies.

Comments are closed.