Sunday Stuff – October 30th, 2011

By Razib Khan | October 30, 2011 6:49 pm


Something different today. First, an elegant international cat:

Second, reading Madagascar: A Short History prompts me to repost a very long essay I wrote ~3 years ago. I have some new ideas in the area of the evolution of religious institutions, which I want to work out in a new essay. But that’s going to have to occur when I have a long period of time to focus on something like that, and that I do not have currently.

Historical Dynamics and contingent conditions of religion


Peter Turchin’s Historical Dynamics: Why States Rise and Fall showed up a little sooner than I’d thought it would, and it was an even quicker read than War and Peace and War: The Rise and Fall of Empires (see review). There isn’t really anything new verbal in the more technical treatment, but the book is about half the length because so much of the text was condensed into simple differential equations and figures which displayed the results of simulations. The figure to the left was one that I found particularly interesting, the differential equations which this is based on are:

dA/dt = c0AS(1 - A/h) - a

dS/dt = r0(1 - A/[2b])S(1 - S)

Where A = area, c = state’s resources translated into geopolitical power, r is the growth rate, h is the spatial scale of power project, a is the geopolitical pressure from the hinterland and S is average polity-wide level of collective solidarity. You can find the elucidation of the details of the simulation in the appendix of Historical Dynamics.

Turchin was obviously pleased with how similar the dynamics of area of polity vs. time were in the simulation to what the empirical data showed. Of course, because of the sensitivity to initial parameters there isn’t going to be a real prediction of the trajectory of state rise and fall, as opposed to inferences about the likely patterns. For example, in the comments to the previous post Italy was focused in on as a weakness in many of the generalizations, and Turchin actually spends a fair amount of time admitting that he has no real answer for why Italy turned out the way it did and admits that his model can explain a lot, but not all. He’s happy with an r-squaredof 0.75.

The above was just a taste, I’m not going to go much deeper since you can get the book yourself. Mathematically oriented works are pretty straightforward and you can reject it or accept it (or not understand it). In any case, I want to focus on another issue which is emphasized in Historical Dynamics, the autocatalytic model of religious conversion. The idea here is simple; the rate of conversion is proportional to the number of converts, and the result is a logistic curve over time. Turchin draws strongly upon Rodney Stark & co’s work on the importance of transmission through social networks, and uses textual data to suggest that the growth of Christianity during the Roman Empire, and Islam in both Spain and Iran, seem to map well onto a logistic growth function.

In The Rise of Christianity Rodney Stark comes close to asserting that the conversion of Constantine, and the progression in the 4th century of Christianity becoming a state-identified cult, actually slowed the spread of the religion! Stark’s thesis is obviously derived in large part from the American experience of cult, sect and denominational rise and fall. Historically minded readers might wonder as to the generalizable nature of a supply side rational choice model for the ancient world. In The Barbarian Conversion the difference between the Roman and early medieval periods in terms of the spread of Christianity is rather clear and distinct, what was plausibly a “bottom up” dynamic quickly turned into a “trickle down” and fiat process (also see Germanization of Early Medieval Christianity).

A comparison to the Islamic case is perhaps a good analogy for what happened across much of Europe after the fall of Rome. When the elites in the German frontier, or Lithuania, or Russia, converted to Christianity, their nations were considered Christianized. That is, full members of Christendom. But the persistence of pagan practices among the populace was common, and even the newly Christian nobility often exhibited dual religious identities (e.g., public and international practice of Christianity combined with cryptic or local adherence to pagan seasonal rituals and sacrifices). I suspect that here you have a situation where autocatalytic models for the population might be appropriate to describing the dynamics of initially nominally Christian states. In Iran or Al-Andalus the elites were Muslim, and the population as a whole, who were ethnically different, lagged. From an “orthodox” Muslim perspective any state which is ruled by Muslims is by definition part of the domain of Islam (this is the rational for reconquering Spain and India by jihadists, as these lands remain Muslim in perpetuity). To some extent the Christian hierarchy seems to have taken a similar viewpoint, though there were attempts to stamp out open paganism among the peasantry, to a large extent de facto syncretism was tolerated so long as the monopoly of Christianity as the elite public religion was maintained and forms were adhered to during ritual occasions.

As I observed above, the autocatalytic model as elucidated by Rodney Stark comes close to asserting that spread of religions such as Christianity is inevitable. In One True God Stark makes this explicit. Turchin emphasizes the importance of exclusionary religions which also can assimilate outsiders in allowing for the coalescence of identity on metaethnic frontiers. In Darwin’s CathedralDavid Sloan Wilson promotes the idea that religious belief can serve functional ends in producing higher than individual left units of interest and action. Many cognitive psychologists have observed that universal religions often result in fictive kinship. Note here that the important point is not the propensity toward supernatural belief; that’s modal human cognition. Rather, it is the specific theological and institutional character of a religious organization which allows them to successfully compete with other “firms,” and if the autocatalytic dynamics are dominant these will result in the extinction of “weaker” religious organizations in the face of “stronger” ones over time via the choice of individual actors along the filaments of a social network.

A classic case study is the rise of Christianity and the late Roman Empire referred to above. It seems likely that around the year 300 about 10% of the Roman Empire’s population was Christian. Rodney Stark would hold that the conversion of Constantine and the subsequent sponsorship of the new religion by the emperors was only illustrative of the general trend at best, and possibly even detrimental. On the face of it this seems likely a ridiculous contention. Could it be that paganism was actually strengthened by state sponsorship of Christianity? That Theodosius’ forcible suppression of pagan cults around 395 was only the outcome of the relative weakness of Christianity because of its association with the Roman state? Could the fact that as the 4th century proceeded customary subsidies to pagan cults were shifted to the Christian Church have actually taken some of the thunder out of the triumph of Christianity?

Stark and company point to the anemic nature of state sponsored Christianity in Europe as compared to the free market of American religious firms. Their model is to some extent an economical one, and they hold that state enforced subsidies and monopolies do nothing but sap the vigor of any corporate entity, which the early Christian Church was to a great extent. This particular critique is not new, even if the language borrows a bit from modern economic thinking. Early Protestant radicals viewed the Roman Catholic Church as a corrupt corporation, and some of them even looked explicitly back to the “primitive” Church before Constantine as the model for how true religion should organize. The descendants of this sort of outlook are numerous in American Protestantism, though the most direct heirs are the Amish who reject the contention that the world as a whole can be saved. They are the most extreme of the Protestants who turned their backs on the concept of the Church Universal which sanctifies and saves the whole society.

But hypotheses need to be teased apart and tested. The state sponsorship of Christianity manifested in a “soft” form between 320 and 390, and in more explicit and exclusive form after 390. The subsequent identity of the Roman Empire and Christianity adds a rather large confound into the autocatalytic model. After all, though to a large extent unenforceable, the emperor Theodosius I issued edicts which banned private practice of pagan religion. There were also state approved destruction of pagan temples, as well as tacit elite approval of the vigilante violence on the part of radical priests. A good analogy for those of you who aren’t versed in this era of history would be the way Christians are treated in the Middle East, they are not forced to convert through direct violence, but there is certainly a general lack of tolerance for religious pluralism and moderate levels of intimidation directed at Christian practice on a day to day basis. The ultimate result is of course emigration and conversion in the face of strong disincentives at practice of the Christian religion. This does not show that Islam is necessarily a better “firm,” rather, state subsidy and dominant support have only expanded its operational religious monopoly. At the end of the day state support might result in such a weakened Islam that a new religion supersedes it, but that process might not come to fruition for centuries. Until then….

There are two cases I can think of which do not suffer from this direct confound of state sponsorship and subsidy. The first is Ireland, where Christianity came to dominance via diffusion across the nobility in a decentralized manner. While Ireland was being Christianized, the Roman frontier right across the Irish Sea was seeing the extinction of Romano-British Christianity aside from in enclaves in Wales. The eventual flourishing of Ireland as a center of Christian civilization in the early medieval period is well known, so I won’t belabor the point. Though no doubt prominent Irish Christians favored their own religion on their own lands, it remains that this was a decentralized society so unitary fiat could not enforce Christianity from above. In the Irish case I think it is plausible that the strengths of Christianity as a Roman religion, with the attendant associations with Romanitas, was attractive for barbarian warlords who wished to integrate themselves into the international luxury goods trade, or encourage the spread of literacy so as to rationalize their economic arrangements. These warlords likely did load the die for the Christian religion so that the consumer element might be relatively muted from a modern American perspective. But nevertheless, here you have a case where neither direct exogenous Christian force (e.g., the Germans threatening to invade Denmark unless the king converted to Christianity), nor a endogenous compulsion from the center, were operative.

The second case is more obscure, and perhaps less tenable because of the fewer facts known, but to me far more interesting. And that is Mesopotamia. Though there were some periods when what is today Iraq was part of the Roman Empire, by and large Mesopotamia was an extension of Persia before the rise of Islam. The summer capital of theParthian and Sassanid dynasties, Ctesiphon, was a successor to Babylon and Seleucia, the predecessor of Baghdad. Nevertheless, Mesopotamia was not culturally Persian, it was Semitic. Prior to the Arabicization of what became Iraq the dominant dialects were affiliated with Aramaic, though there were Arab, Persian and Greek speaking minorities. But more importantly as Peter Brown notes in The Rise of Western Christendom, Mesopotamia was a mostly Christian region (with a large Jewish minority, especially in the south). There were no Zoroastrian Fire Temples in Ctesiphon.

In the Sassanid Empire the Zoroastrian religion was very much the ethnic cult of the Persians. Though some non-Persians might espouse this religion (there are attested cases of Zoroastrian Turks, and other converts), this was not a program sponsored by the ruling caste in a proactive manner. The attempts to force Armenian nobles to convert to Zoroastrianism was the exception that proved the rule; the Armenian elite were culturally very similar to the Persian nobles with whom they fought in the armies of the Sassanids, and the Armenian ruling dynasty was even originally a cadet branch of the Parthian Arascids. Proselytising of Armenians was simply part of the project to homogenize the martial elite of the Persian Empire under the same religious ideology. In contrast, the Aramaic speaking peasantry were left to their own devices.

The relatively laissez faire attitude of the Sassanids toward the religious identity of their subjects in Mesopatamia had the expected result in terms of pluralism. Modern Haranwas usually within the orbit of the Roman Empire, but it was the only area to persist with organized paganism down into the Islamic area, and it seems likely that the Sabians of the early Muslim period emerged from this milieu (they were by the way extremely overrepresented among those involved in the preservation and transmission of classical learning). Why did they not convert to Christianity? One reason is that they were given religious tolerance because they were explicitly protected by the Shah of Persia, who could have easily intervened because of the geographic proximity of Haran to his domains. In contrast during the mid-6th century the last vestiges of institutional paganism in places like Egypt and Lebanon were blotted out under the order of the Emperor Justinian. Across the border in Sassanid Mesopotamia the majority of the population became Christians in all likelihood, but the extant presence (at least until recently) of heterodox cults such as Mandaeism and Yezidism in this region today are I believe echoes of the diversity which was the norm during late antiquity.

All that being said, it seems likely that when the Arabs conquered Iraq in the mid-600s most of the populace were Christian. It is important to note that they were Christians which the Roman Empire based in Constantinople would perceive as heretical. They were Monophysite or Nestorian in inclination, not only theologically deviant, but institutionally hostile to the Christian Church organized within the Roman Empire (those Christians in the Fertile Crescent who adhered to Roman Church were termed Melkite, which means Imperial, an allusion to their loyalties). The Nestorian Christians are often identified as the Persian Church because of that group’s almost total exclusion from the Roman Empire and prominence among ethnic Persians.

I’ve put the spotlight on Mesopotamian Christianity as it was around 600 as the dominant religion to ask this question: whatever happened to Babylonian paganism? As I said above, Roman hegemony over Mesopotamia only occurred under the religiously tolerant pagan period. The Persian rulers were interested in the religion of their Mesopotamian subjects only insofar as it had political ramifications; obviously they would encourage the anti-Roman Nestorian faction, discourage pro-Roman Melkites, and deal with the Monophysites who spanned both the Persian and Roman Empire on a case by case basis as circumstances dictated (the Persians tended to suppress socially disruptive Mazdakites and Manichaeans because these groups drew from Zoroastrianism). The case of the Sabians and the Persian protection of this pagan-descended cult against the religious cleansing which was a characteristic of Justinian’s reign in the mid-6th century also suggests that there was no hostility to polytheistic paganism as such. In fact, many scholars of Zoroastrianism contend that that religion is more monotheistic in its presentation today for two reasons. First, the period of Muslim rule of courseincentivized Zoroastrians to present the most acceptable, i.e., monotheistic, face of their religion to the majority. In India the Parsis generally escaped this, but during the period of British rule again they were faced with a monotheistically oriented group to whom they had to bend a knee, so again, an emphasis on similarities with the Abrahamic religions.

In any case, in Mesopotamia outside forces can not account from the shift from institutional polytheism to monotheistic universalist religion. Polytheistic paganism seems to have naturally withered. A quick survey of the situation in Europe after the fall of the Roman Empire seems to also show a pattern where paganism was simply not institutionally robust enough to hold off Christianization. A repeated pattern in is one where rulers who wish to cultivate ties to the civilized Roman Christian Commonwealth convert and encourage conversion among their populace, but so invite a backlash. This occurred in Scandinavia and the Slavic lands, as well among the Magyars in Hungary and the Bulgarsin Bulgaria. But, the backlash generally is only a short-term correction which only delays the inevitable. In the west Slav lands bordering Germany and in Lithuania a very robust and persistent form of paganism arose which did seem to keep Christianity at bay for several centuries, as opposed to a few generations at most as was the case above. Looking more closely one can see very specific contingent conditions which gave rise to these dynamics. The Christian assault on the Wends (ancestors of the modern Sorbs) was very much also an ethnic German one. The Christian god was identified as a German god, and the German drive to the east was one of of total ethnic and religious assimilation at best and extermination at worst. It is then no surprise that west Slavic paganism was particularly robust in terms of generating an institutional framework around which to rally against the Christian-German invasion; they were fighting total extermination as a people (if not as individuals). In contrast, the Polish who were further from the front used Christianity to buttress their independence from the expanding Germans, cultivate ties to other Christian powers, while the duke, who became a king, used the One True God and One True Church to justify his centralizing drive as the One True King. This was a rational maneuver because of their greater distance from the wave front of German expansion; Christianity was not necessarily a German religion (the Bohemians to their south of course had contact with Byzantium as well as German Christianity). The case of Lithuania is even more explicable in terms of particular geopolitical and historical conditions: with the decline of European states and the Mongol hordes the Lithuanian polity forged against the German drive to the east under the banner of the Sword Brothers and Teutonic Knights expanded to fill the vacuum. By the mid-14th century Lithuaniaincluded most of modern Ukraine, White Russia as well as the Baltic lands and parts of Poland. The majority of the subjects of the pagan Lithuanian warrior elite were Christian. Either Western Catholic or Eastern Orthodox. A conversion to Christianity would of course entail that the Lithuanian elite pick a side, Catholic or Orthodox, while persisting in their paganism allowed them to play off the two groups against each other. A substantial number of Lithuanians did convert to Orthodoxy or Catholicism, but the commanding heights remained pagan due to the geopolitical circumstances. In the late 14th century Lithuanians converted to Catholicism, cementing their alliance with Poland, naturally resulting in Lithuania becoming the marcher state aginst Orthodox Muscovy and the last frontier of the West (after the 16th century the Lithuanian nobility was totally Polonized).

The autocatalytic model does work, but I believe social and political incentives also matter. Aside from Ireland every instance of Christianity spreading and absorbing a culture in Europe after the fall of Rome was initiated from the top and down. Though most of these states had small Christian minorities, sometimes of influence, the majority of the logistic growth curve occurred while Christianity was the official religion. Many Protestants even contend that Christianization of the European peasantry was not completed until after the Reformation. But there were strong incentives to become a pious Christian in Europe after 1000, when Christianity and civilization and elite status went hand in hand, and paganism was tatamount to barbarism.

A quick trip back to late antiquity highlights the importance of the incentives and framing social structures in terms of how it affects the trajectory of religious change. According to the data that Turchin and Stark accept, the Empire was over half Christian by around 360. By the 400s it was overwhelmingly Christian. Nevertheless, in 529 Justinian closedthe Academy in Athens which was still the locus of pagan philosophical thought. The Diaspora of Neoplatonic pagans remained active until the Islamic period in Alexandria, and likely influenced the Sabians of Haran. By the time of Justinian these pagans could only draw from a small subset of the Empire’s population, those whose families remained loyal to the old religion, or, those of other minority religions such as Judaism or Samaritanism. The similiarities to dhimmis under Islam is again rather clear. But the point I want to make here is that despite the presumed autocatalytic dynamics operative through the Christian Empire, philosophers still remained pagan! There were particular incentives within the philosophical culture which fostered adherence to a pagan religious outlook. The autocatalytic process does not operate across the full sample space to the same extent. While most of the Empire was being immersed in a religion which was a synthesis of Roman institutions, Greek philosophy and Hebrew theism, a subset of the population of philosophical inclination was being drawn into a religious system descended from Hellenistic paganism. This quasi-philosophical world-view was the one that drew the pagan convert Julian to the Apostate. It is notable that Julian, a self-conscious Hellenist in his fashions, was relatively well-educated and manor-born in comparison to the military populists who were dominant between 280 and 400. Though the first illiterate Roman Emperor did not come onto the scene until the early 6th century, there was a wide range of cultural sensibilities, from philosopher-kings and scholars such as Marcus Aurelius and Claudius, to military tryants and autocrats such as Decius and Diocletian.

This “different world” was not operative only among philosophers. The Frankish general Arbogast was the son of a Romanized German, and yet he is known to have been a pagan of classical Roman sensibilities. Arbogast led a pagan senetorial rebellion against Theodosius the Great, and was defeated. Because history has minimal interest in losers we do not truly understand with any clarity how it was that a barbarian by ancestry was acculturated to the world-view of the pagan Roman elite at this late date (Roman society had become far more xenophobic and prejudiced against barbarians than it had been earlier by the 4th century). But books such as The Making of a Christian Aristocracy: Social and Religious Change in the Western Roman Empire draw upon a wide range of textual evidence on the late Roman senatorial elite of the West to imply that they did not make the final turn to Christianity until after 400. Additionally, a deeper analysis of the shape of religious variation smokes out intriguing patterns. Roman senators of the 4th century who were Christian were much more likely to be new men, parvenus dependent upon imperial patronage. They were more likely to have risen through the military or civil service, as opposed to having inherited their status. Additionally, Christian senators were also more likely to come from Gaul and other provinces on the frontier, while the pagans were more likely from the old imperial core, Italy and North Africa, two regions relatively insulated from the disturbances of the 3rd and 4th centuries.

Within the status hierarchies of the old senatorial elite paganism and its attendant familial cults had a strong attraction. Even in the mid-5th century, high status nobles such as the general Marcellinus, were devout pagans. Another locus of pagan power in the army is rather clearly illustrated by a revocation of the expulsion of non-Christians from the officer corps by Theodosius II early in the 5th century; so many officers, including generals, protested and offered to resign that it was judged to be impractical in implementation. There were enough crypto-pagans on the ground that as late as the reign of the Emperor Zeno in 474 there were hopes that paganism would be restored as the official religion. In any case, that wasn’t to be, at some point the cause of paganism during late antiquity was as futile as that of Roman Catholicism in England by the 17th century. But, with hindsight I think we need to not forget that inevitable dynamics didn’t seem so inevitable back then, and different incentives and social networks intersected across the same time and space. Similarly, an autocatalytic process might have been operative in terms of conversion to Islam in the Levant, but even in 1900 around 10-20% of Palestinian Arabs were Christian, and across the coastal mountain ranges of Syria-Lebanon Christians and heretical Muslims (Druze, Alawites) were more numerous than Sunni Muslims (emigration in this case was so strongly biased toward Christian Arabs that the proportions would have changed a great deal even without the differential birthrates which came to the fore in the 20th century).

Despite the specific twists and conditonalities, I do think that the null model of the autocatalytic expansion of particular religious groups is useful. In Persia I believe we have an excellent case study in Mesopotamia which suggests that ethnic polytheism naturally tends to cede ground over time to universal monotheism. As I have outlined I think the likelihood that there was an exogenous confound is sharply dampened in this one scenario. Obviously there isn’t the issue of Christian blackmail (i.e., monotheistic states after the fall of Roman had a cheery habit of threatening to invade unbelievers because of the fact that they were unbelievers), nor sponsorship by Christian elites. Granted, like Ireland the Christianization of Mesopatamia might have been facilitated by the mediating role of local notables wishing to integrate themselves into the transnational luxury trade. It seems that Semitic ethnicity was a bar to conversion to Zoroastrianism, and the Arab federates of the Sassanids, the Lakhimids, were not surprisingly Christians (one could argue that the incentives of the pagan piligrimage trade were one of the reasons that the nobles of Mecca did not align themselves with a world religion). But in a pre-modern society there simply wasn’t as much individual choice, and patrons followed their clients, whether those patrons were an individual or a corporate entity like a guild or village council.

But the Sassanid Empire also expanded east, into Central Asia and the Punjab. These were regions where Zoroastrianism was simply not much of an option for those non-assimilated to Persian ethnicity or identity. And not surprisingly, Nestorian Christianity was influential along the trade routes. Arnold Toynbee alluded to a stillborn Nestorian civilization, and it was thanks to the reach of the Zoroastrian Sassinid Empire that Nestorianism spread so far and wide. In the 8th century Nestorians were a prominent power not just in Central Asia, but also in China. It seems that the Christians of Kerala were originally affiliated with the Nestorian Church of Mesopotamia. And, it is well known that Nestorians were still extant among the Turco-Mongol peoples swept up in the expansion of the armies of Genghis Khan; the mother of Kubilai Khan was a Nestorian Christian.

And yet, what happened here? Shouldn’t the autocatalytic process have increased the frequency of Nestorianism so that it dominated all these regions? In Persia itself Nestorianism declined with the rise of Islam. There are attested conversions, but it seems pretty clear that preponderant ancestry of modern Iraqi Muslims are from Aramaic speaking Christian peasants. In China there was a major suppression of foreign religions in the mid-9th century. This seems to have nearly extirpated Nestorianism, and driven Manichaeanism (which also came from Persia) to such low numbers that it went extinct in a few centuries, and also set Islam back quite a bit (modern Chinese Islam probably owes more to the influx of Central Asians with the Mongol Empire in the 13th and 14th centuries than the original expansion of Islam into China in the 7th and 8th). In India Nestorianism flourished in Kerala, but did not spread to any other region. One would assume that 1,500 years was long enough for autocatalytic dynamics to kick in….but it seems that Kerala’s Christians (who are by and large no longer identified as Nestorian, though they retain Syrian affinities) turned into another caste. The modern spread of Christianity in India was spurred by British raj and Western missionaries, though Syrian Christians were often critical conduits.

The case of India is important enough to inspect with greater detail. India is the only civilization which has produced a world religion besides the Middle East. Indians will generally assert that Hinduism, Buddhism, Sikhism and Jainism are Dharmic religions set against the Abrahamic religions of of the Middle East. The gap is a real one obviously. Hinduism and Buddhism are very different, but ultimately they deal in the same semantic currency and there are institutional resemblences. In Myanmar the Indians who have remained are by origin either Muslim or Hindu. The latter have been consistently changing their religious identity to that of Buddhism. Muslims have been to a far lesser extent. Though there are tensions between the Chinese and Thai in Thailand, and a religious gap between Chinese Mahayana sects and the Therevada Buddhism of Thailand, the mixing between the communities is rather fluid when compared to the situation in Malaysia, as the relations between Chinese and the Muslim peoples of the Malay archipelago is fraught with more tension. There are orthopraxic gaps which make this comprehensible; the food taboos of Buddhist priests and monks, whether Mahayana or Therevada, are rather intelligible to each other (generally derivations from Indian vegetarianism). In contrast, the Muslim aversion to pork does not generally allow for easy communal meals with Chinese, for whom pork is nearly the obligate meat. I recall that when there were riots in Java in the 1990s against the Chinese many fled to Hindu Bali. Here the proximate dynamic isn’t simply reducible to civilizational gaps, after all, both the Balinese and Chinese are outsiders in the mix of the Muslim majority and so a natural empathy might arise (and a substantial number of Chinese Indonesians are Christians, even if only nominally). But on a coarser scale increasing the N I think Turchin’s model of a metaethnic border is probably viable and useful, even if it is not likely the avowed rationale given for conflicts, it may lurk in the background as a necessary framing condition, or at least one which increases likelihood.

The East broadly, the Indian and Chinese cultural orbits, are interesting cases when it comes resistence against the expanding orbit of the One True God. On a whole, it’s taken some hits. Around 1/3 of South Asians now subscribe to an Abrahamic religion. Island southeast Asia was lost to Islam relatively recently from the Hindu-Buddhist bloc. The Dutch helped along the process in Java because of their rivalry with the Hindu kingdom of Bali (eastern Java was the center of a Hindu kingdom allied with Bali until the 18th century). China has a non-trivial Muslim minority. Myanmar, Thailand and Indochina all non-trivial large Abrahamic minorities. Hong Kong, Taiwan, Singapore and South Korea have non-trivial and powerful Christian populations. Japan has a small, but influential, Christian minority. On the other end of the balance sheet, in the secularizing West ideas from Dharmic religions are very popular among the elites, and some, such as reinarnation have penetration rates as high as 25%. But the influence is less institutional and organizational than it is a percolation of ideas and assumptions.

Let’s look at India first. By India, I’ll include the states not currently in the Republic of India, since before 1947 India meant the whole subcontinent, though Bhutan, Nepal and Sri Lanka were often bracketed off because of their Buddhist leanings. About 1/3 of the Indian population, broadly, is Muslim. Islam was present in India from the 8th century. Sindh was conquered by the Ummayads. In the next few centuries military incursions were minimal, but Arab mercenaries and merchants were a prominent force. The large number of Muslims in Kerala is a function of trade much more than the later rise of foreign Islamic dynasties since most of the time Hindu rulers were preeminent in this region (see Vasco da Gama’s reports). In the north of India Muslim warlords were dominant after 1000, and exclusive at the top of the totem pole by 1200. This situation persisted for 500 years until 1700, at which point political fragmentation was the dominant dynamic. After 1700 some non-Muslim groups rose to parity, but only after 1800 and British rule was the political supremacy of Muslim elites off the table.

So a 500 hundred year window of near domination, and quite a bit of power (1000-1200 was a long rearguard action on the journey to extinction on the part of Hindu kingdoms in northern India, not one of parity). And yet only 1/3 of the population is Muslim? First, autocatalytic dynamics assume a level of connectedness perhaps inappropriate in South Asia. Though there were no north Indian Hindu kings, many of the great vassals, rajputs, remained Hindu. So there were mediators who continued to foster the production of Hindu religious ritual through their patronage. There were many instances of conversion, but it seems clear from the extant biographical data Hindu warlords did not want to turn their back on their own cultural heritage, as would be an implication by conversion to Islam (they would also remain inferior in status to Muslims from Persia or Central Asia). There is an element of irony in this because it seems likely that some of the rajputs of northern India were themselves immigrants from Central Asia who filled the power vacuum after the collapse of the Gupta dynasty in the 6th century. But like the Tibeto-Burman Ahoms of Assam later, they became defenders of Hindu Indian cultural traditions on the metaethnic frontier. Additionally, Muslim power projected rather raggedly into southern India for much of this period, where the Empire of Vijayanagar flourished. Though Vijayanagar was contested, and eventually conquered, by south Indian Muslim dynasties, it remained a separate locus of patronage for Hindu cultural production during the period of Islamic domination. Finally, it must be remembered that India is a highly segmented society, and many villages were run by Hindu landlords (patels, thakurs, etc.) who served as mediators between the new Muslim overlords and the masses.

With the fall of the Mughul raj and the rise of the British Hindu notables quickly rose up to fill the void and stepped into the shoes of the Muslim ruling classes to administer India. This shows that a reservoir of non-Muslim elite talent always remained extant. Some of these were no doubt patronized by Hindu dynasts such as the Marathas and those of Vijayanagar. Others were patronized by Hindu vassals of the Muslim dynasties, such as the rajputs. And some of them were patronized by the Muslims themselves (e.g., theKayasthas served the Muslims more than other high caste groups which had a tradition of literacy). The Sunni Muslim elite seems to have taken a role as a rentier caste, opening up niches for enterprising non-Muslims. It is interesting that some of the most economically successful Muslims in the Indian subcontinent are the marginal Ismailis, who were persecuted by the Mughals and forced to convert to Sunni Islam.

Not only are there complex patterns vertically up and down the class ladder, but one must look at the conversion patterns as a function of geography. In modern India it is no surprise that aside from Kerala and centers of Muslim dynasties (e.g., Hyderabad) that Islam is relatively thin on the ground in the south compared to the north. Additionally, in Orrisa there are very few Muslims, and this is an isolated and frankly backward region which was less exposed to outside currents. But, it is important to note that the Muslim heartland around Dehli remained predominantly Hindu across all those centuries. It is no surprise that Muslims are the majority along the western fringe, not only are these regions closer to the demographic sources of Turkic and Persian immigration which buttressed the Islamic dynasties as soldiers and bureaucrats, but the Sindh was under direct Muslim rule far longer than any other region. Yet in Pakistan it is in Sindh which has the largest Hindu minority (likely due to the relative easy of population exchange along the Punjab border as opposed of the Thar boundary to the east of Sindh). Additionally, of course the other locus of Muslim majority in the Indian subcontinent is far to the east, in Bengal. Not only is it in Bengal, but there is a consistent pattern that the further east you go in Bengal the more Muslim the population gets, with the most pious region the southeastern district of Noakhali. When the British census revealed that there were more Muslims than Hindus in Bengal in the late 19th century they were somewhat shocked.

In fact, Bengal has been under Muslim rule only a century or two less than the Punjab, so the difference of duration isn’t that great. But, it is notable that prior to the Muslim conquest these two regions were relatively weak in terms of institutional Hinduism, and Bengal was the last region of India to host a flowering of Buddhism. The social and institutional robusticity of Indian religion, the set of beliefs and rituals which became Hinduism, did not characterize the Punjab or Bengal during this period. Polities in these areas were more often aligned with the “losing” cultural faction, and divided within themselves. The Rise of Islam and the Bengal Frontier, 1204-1760 documents the confluence of social and economic conditions which allowed eastern Bengal, what became Bangladesh, to become a mostly Muslim domain while western Bengal, now part of India, remained mostly Hindu. It is important to remember that eastern Bengal was to a large extent not characterized by the Malthusian trap which we see today, rather, for most of the past 1,000 years its the frontier served as a demographic release valve as peasants cleared the forest under the supervision of expanding elites. Those elites were of course mostly Muslim (though the capital input might have been from Hindu moneylenders). A combination of the relative weakness of extant Hindu institutions in eastern Bengal, combined with the emergence of a new non-Hindu elite, and, an expansion into a frontier so that a small number of pioneers might serve as genetic and cultural “founders” makes the fact that Bengal was much more fruitful for Islam than the central Gangetic plain much more comprehensible. Recall that I observed that there is data which suggests that elites on the geographic margins, the frontiers, were more open switching to the Christian religion and abandoning their older customs and traditions during late antiquity. In contrast, the old civilized cores, such as Italy and Greece, were notable for remaining pagan longer than new frontier metropoles such as Constantinople or Antioch. A similar difference might have applied to Iran and Central Asia, where the latter was Islamicized earlier than the former.

Of course, the analogy between paganism and Hinduism is not very strong. The robusticity of Indian socio-religious structures in the face of domination by another socio-religious framework is impressive and makes it very different from Babylonian paganism. Just as elite Roman senators were resistent to attractions of Christianity for a relatively long period, attempts by Christian missionaries to convert Indians have had to focus on the lower castes and maringally Indianized (e.g., the Tibeto-Burman tribal peoples of the northeast). Once the ball starts rolling though that is a sign that there as an institutional vacuum which Christianity can fill; the instances of forced conversions of pagan and Hindu Nagas by Naga fundamentalist Christians illustrates the power of autocatalyic peer “pressure.” On the other hand, among higher caste Hindus it seems that the logistic growth curve tends to saturate at a lower level.

A shift to southeast Asia highlights the importance of conditional parameters in these autocatalytic dynamics. Christianity is a minority religion in most of mainland southeast Asia. But, it is an ethnic religion. Specifically, it is a religion which is popular among minorities who have traditionally been at the cultural and economic margins. Lowland southeast Asia has been dominated by powerful kingdoms with a very strong self-conscious identity as vessels for Therevada Buddhism. Buddhist monarchs in southeast Asia even sent aid to those attempting to kickstart the Buddhist revival in Sri Lanka during the period of British rule when the elites were converting to Protestantism (a dynamic which was halted, and reversed). Groups such as the Karens, Hmong and tribes of the Montangnard highlands resisted conversion to Therevada Buddhism for a very simple reason: those who converted became Bamar/Burman, Thai or Vietnamese. The tension across ethnicites was strongly diluted when the religious barrier disappeared. On the other hand, as we have seen above, non-institutional paganism tends to lack robusticity over the long term. Tibet and Japan both manifested the same dynamic which I asserted for pagan Europe during the process of the shift toward a universal world religion. In both these cases the correction was temporary, and the setbacks rolled back as Buddhism eventually embedded itself as the dominant institutional religion of the culture.

The arrival of Christianity changed the game. Groups like the Karens observed correspondences between Christian theology and their own indigenous religion, but, seeing at how similar and convergent supernatural concepts tend to be I don’t believe that similarities would be hard to observe (Christianity was repeatedly confused as a form of Buddhism in East Asia). Today a large proportion of Karens are Christian, but not all. A significant number are Therevada Buddhists, and not surprisingly these tend to be much less hostile to the central government, and even complain of persecution at the hands of Christians. About 10 years back I recalled reading about the conversion of the only non-Christian resistance leader among the Karens. It is obvious that the Karen resistance has religious overtones, and the Christian identification operates synergistically with their historical self-conception as a separate people. A similar process occurred in Indonesia after the suppression of the Communist Party. Many Chinese and secular Javanese became Christian because that was an open option available to (one had to affiliate with a religion in Indonesia during Suharto’s regime) them. At this point their children are no doubt sincere believers, but one can not ignore the contingent parameters which drove what is now an autocatalytic process within these networks (also, a minority of very nominal Javanese Muslims are making the switch to Hinduism, as they view it as a more authentic expression of their outlooks and identity as Javanese).

In China, South Korea, Singapore and Taiwan Christianity is a religion with some influence. Christianity is a firmly bourgeois sect in Taiwan and Singapore. Though growing in Singapore, recent data suggests that Buddhism is now rapidly rising in proportion faster by siphoning off Chinese religionists, and many of the college educated cohorts are switching to Buddhism instead of Christianity as would have been the norm one generation ago in the process of upward mobility. In Taiwan Christianity has been stagnant for the past generation, likely going through a contractionary phase. Again, the same dynamic of Taoists and Chinese folk religionists switching to Buddhism is noticeable here. What’s going on? I think the key here is that Buddhism is a familiar religion, it simultaneously can play the role of Chinese ethnic religion and world religion. In contrast, it seems that Christianity often opens up a fissure with the family because of its more stark rejection of “pagan” practices which older members might continue out of custom and habit (it seems in Singapore that Christianity drew disproportionately from secular educated segments where these familial concerns would likely loom less).

One important point to emphasize with East Asia is the relatively weak role that institutional religion has played within these set of societies. In China, Japan and South Korea though Buddhism was the most common religion before the arrival of Christianity, it was neither dominant or elite. The Tokugawa demand that Japanese families affiliate with a Buddhist temple in the process of extirpating Christianity was not due to Buddhist piety, in fact, the predecessors of the Tokugawa had gone to great lengths to break the back of institutional Buddhism as an alternative power center in the 16th century. Rather, religion was an instrument of social control and tool for sifting the loyalists from those prone to sedition; in this case, Catholics whom the regime assumed to be supportive of rival southern daimyos and foreign powers. In China and Korea Buddhism was the primary institutional religion above the level of local cults and shamans, but, it was also kept at remove from the centers of power because of the skepticism of the Confucian mandarins toward otherworldly religion. The rise of Christianity in Korea was not truly at the expense of a vibrant Buddhism, it was within a very secular context when it comes to rival institutional structures. South Korean Buddhism to some extent has been prodded by Christianity, going so far as create its own religious television channel. Many would argue that Christianity was the best thing that could have happened to South Korean Buddhism, which had become a moribund mountain cult.

South Korea is an interesting case beause it has the largest Christian population proportionately outside of the Philippines in Asia. But they are only 30% of the population, and many Americans are surprised that around 20% of Koreans are avowed Buddhists and 50% have no affiliation. Again, this is due to the expectation that societies transition from religious monopolies (as in the West) toward a more relaxed regime. In Korea religion was marginalized as a source of authority by the center, and its new power in coalescing individuals is a throwback to the period before 1300 when Buddhism was likely an important cog in the process of “Koreagenesis.” Again, I suspect the insight that metaethnic frontiers require ideological cement is important to keep in mind. With the period of Japanese colonialism many patriotic Koreans began to look to Christianity as a source of resistence, and Christians went from 5% to 25% of the population within 2 generations. But from the data I’ve seen over the last 15 years that rate of growth has slowed radically and the proportion has been only inching upward. Part of the issue might be the correlation between higher levels of education and Christianity, and the lower fertility of these groups. Additionally, it seems that many Korean Christians are now switching across churches, implying that the social networks are becoming tapped out of new low hanging consumers.

These data from various East Asian countries, and the example of Japan as a nation where autocatalytic Abrahamic dynamics seem to simply not operate, suggest that the logistic growth curve for Christianity has limits in penetrating all societies. South Korea was a best-case-scenario for a variety of reasons, but even here it seems that the trajectory has slowed down. The standard model in Christian or Muslim nations is for the logistic growth curve to move up toward an overwhelmingly majority, if not nominal uniformity. But East Asia has long been characterized by pluralism, and even when Buddhism was ascendent as during Tang China, Silla Korea or Fujiawara Japan, there tends to be a correction and institutional religion can never “swamp” the culture. The Indian case is separate because there you have a robust institutional religious system which weathered the rather large exogenous Abrahamic shock, and it is important to note that Indian religious systems were exported long ago into East Asia, and yet they have never attained the level of cultural monopoly that they exhibit within India, or, southeast Asia.

During the 16th century Japan was very open to Christianity. Daimyos in the south of the country converted to Christianity and brought their peasants with them. As much as 10% of the Japanese population might have become Christian at some point. The Tokugawa exterminated this population aside from a few crytpo-Christians whose orthopraxy had been Buddhaized to the point where they had evolved into a new religion. When Japan was reopened in the 19th century many Christians assumed that it was going to be easy pickings…in fact, it didn’t turn out like that. Though Christians were influential, including likely many women in the imperial family, they never achieved critical mass. Sociology is deterministic only with particular background conditions. No doubt some scholars assume that Japan would have become totally Catholic if the growth rate persisted and was extrapolated, but why make that assumption? Perhaps the Christians of Japan would have become the Moros of the Phillipines, or the Catholics of Sri Lanka, a substantial minority, but eventually sealed off into their own social networks.

This post obviously got a little out of hand. You can probably tell that I like both general deductive models, and an attention to contingent detail. There can be a general trend (e.g., we all die) with variations along the way of interest (e.g., what we do before we die). It seems that autocatalytic process will result in Africa becoming totally Muslim or Christian. On the other hand, if it takes 1,000 years for India and China to become totally Christian or Muslim…well, I’m not sure if that certain projection is really that useful seeing as how 1,000 years is a long enough time that a lot of the background parameters could change. Additionally, there are various frequency dependent dynamics and mixed morphs which are likely operative in these historical social trajectories that I think are being left out in this treatment.

  • omar

    The notion of state sponsorship undermining a religion may be getting a try0ut in Iran….

  • Thomas

    Well this post is one that I am starring in Google Reader and definitely going to come back to. Thank you very much for the work that you put into it.

  • Justin Giancola

    I was gonna say you guys are champs for getting through this so soon! :)

    by the way razib, have you ever thought of being a history teacher? mine is good, but dang, I think you could give him quite the run for his money.

  • Razib Khan

    #3, humanities people need to know foreign languages. that i can’t do.

  • Justin Giancola

    you can’t bluff any hindi? bangla? (had to look that one up btw, ha!)

    I think private schools can higher whomever they want. You could just show up a few days a week; rant off some stuff my memory; get that $. :)

  • Onur

    A similar difference might have applied to Iran and Central Asia, where the latter was Islamicized earlier than the former.

    Here you must have meant only Greater Khorasan (including Transoxiana) by “Central Asia”. An important point, Greater Khorasan was almost exclusively Iranic-speaking until the first few centuries of the 2nd millennium CE, when the tide started to turn in favor of the spread of Turkic in parts of it. So those who converted to Islam in Greater Khorasan were almost exclusively Iranic-speaking in the early centuries of Islam. Greater Khorasan was an Iranic-speaking realm when it had been effectively Islamized. Also keep in mind that Greater Khorasan is traditionally a part of Greater Iran and outside the realm of Turan.

  • TGGP

    The Political Economy of Beliefs is a sort of modern economic take on Stark’s hypothesis.

    Hill tribes affiliating with Christianity as a way of distinguishing themselves from the Dharmic majority/government is a significant topic in James Scott’s “The Art of Not Being Governed”. He also discusses how the proto-states/kings of southeast asia took inspiration from the Indian example.

  • Spike Gomes

    Just some minor notes:

    The Kakure’s orthopraxy more drifted towards the Shinto influenced traditions of the Japanese religious sphere (though for the time period we’re talking about, neither existed as we know them now), this is due to the fact that much Buddhist practice was institutional whereas “Shinto” practice was home and village based.

    Almost certainly Catholicism would have become a minority, strongly regional faith in Japan. Where it was strongest was a peripheral rather poor part of the nation that lacked much institutional penetration by Buddhism followed distantly behind by several urban centers.


    If I were to finish my Ph.D. in Japanese religion I would need to have classical Chinese and Japanese comprehension skills down, despite the fact my interest and research focus was solely on modern Japan. Language reqs are the one way many humanities programs can IQ filter people in the programs.

  • Razib Khan

    Language reqs are the one way many humanities programs can IQ filter people in the programs.

    hey, the correlation between language skills and IQ is imperfect :-)

    you can’t bluff any hindi? bangla? (had to look that one up btw, ha!)

    literacy is probably more important than verbal fluency. and no, hindi is not close enough to bangla. they’re probably as far as french as italian are from each other, if not more (i think hindi, punjabi, and gujarati, are like spanish, italian, and catalan, in that they’re a relatively tight clade).

  • syon

    Razib Khan:”#3, humanities people need to know foreign languages. that i can’t do.”

    Well, based on my experience as a humanities guy getting a phd, the language bar is pretty low; most graduate programs have “reading knowledge only” programs that are a lot easier to pass than courses with a spoken component.

    Spike Gomes:”Language reqs are the one way many humanities programs can IQ filter people in the programs.”

    I’m going with Razib on this. I’ve not noticed any real correlation between IQ and facility for languages in the graduate reading knowledge courses that I’ve taken. Success in such courses seems to be more a matter of a having a good work ethic ethic than intelligence.Indeed, in my French class, the star student was a girl who seemed to be at the low end of the graduate school IQ norm.

  • Spike Gomes

    syon: Trust me, Classical Chinese is more g-loaded than French. European languages have the advantage of being alphabetic and with somewhat familiar grammatical rules. Once you get into ideographic stuff with implied meaning between characters, the imaginary CPU fan on your brain goes into high speed cooling mode.

    Seriously though, it is not a perfect correlation, but there is *something* there, otherwise there wouldn’t be the huge body of evidence that shows the disparate rates of time it takes to get to fluency in various languages.

    From my anecdotal experience, it was nothing to “get” Spanish. I never had to study until well until the second year. Japanese was harder from the start and a slog from the third year on. Kanji memorization is highly G-loaded, IMHO. It’s why one of the common humor segments on Japanese talk shows is testing celebrities on their kanji reading and writing skills, and why a constant refrain in Japanese newspapers is on the decline of Kanji literacy (my gut feeling is that this is overblown, and has always been the case that the vast majority of Japanese adults did not have a full grasp of the Joyo Kanji they should know by High School graduation). I never took any of the Classical languages, but my office mate did, and they seemed to be on the next level of difficulty.

    For the folks in my general field, it didn’t stop there, however. With Japanese you could get by with those three. Those in Chinese and Indian religion had to learn German and/or French in order to read scholarly works, as well as multiple classical forms, and you don’t get *real* cred unless you do something like self-learn Soghdian or Tocharian like the Chinese History prof who cross-taught courses in our department.

    Trust me, it’s a bit more than just picking up some Italian or German on the side.

  • trajan23

    One thing I’ve noticed while completing my language reqs is that women, on the average, seem to do slightly better than men. Is there any data on this? Does the female brain “get” second languages more readily?

  • syon

    Spike Gomes:

    Well yes, I completely agree that some languages are much more difficult for an English speaker to learn. And, yes, I also agree that this greater level of difficulty imposes greater cognitive demands. However, once a certain IQ threshold is reached, I’m just not sure that higher IQ equals greater facility with languages. I’ve known many mediocre intellects (by graduate school standards;compared to Joe Average, they are, of course, really smart) who are demons at mastering languages. I think that there is some non-IQ related aspect of mental functioning ( talent, to use an old fashioned term) that is involved in the process.

  • Spike Gomes


    I don’t disagree in some respects. I don’t think it takes much smarts at all to get most *living* European languages, or even conversational non-Indo-European languages (save for highly inflectional stuff like Finnish or heavily tonal languages like Vietnamese). It just takes a knack for it.

    However, the written forms of some languages aren’t really completely mastered by the native speakers of the language (frex Japanese Kanji), and moreover, certain classical languages were designed to serve as “gatekeepers”. They were never really “living” in the sense that they were an accurate reflection of what was mostly less complex dialectical spoken language. You didn’t want people to pick up an esoteric Buddhist text and read it easily. Even a native Chinese would have to be somewhat brainy to study and understand Classical Chinese.

  • Justin Giancola

    I totally over-looked how this has been posted before. It seemed really familiar as I was reading it. I had already found this somehow, but I do really think the posts from the past is a great institution for those who have missed powerful classics.


Discover's Newsletter

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

Gene Expression

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

About Razib Khan

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


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar

Login to your Account

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

Not Registered Yet?

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