Even so, some spatially extensive imperial polities did eventually appear under the Fatimids, Almoravids, and Almohads. The Fatimids depended on the military muscle of Berber tribes. The Almoravids received at least initial support from local nomads, and the Almohads were backed by Berber tribes from south of the Atlas. All of these nascent empires formed and matured in predatory interaction with more substantial agrarian hinterlands beyond their original heartlands. The Fatimid state’s center of gravity first shifted to Tunisia and then on to Egypt. For the other two, the Iberian peninsula provided vital material resources and depth. Overall, outcomes were consistent with basic expectations related to the “steppe effect”: while the often smaller scale of interactions between settled and steppe zones helps account for the relative paucity of indigenous imperial formations in North Africa, significant scaling-up of these interactions was essential for those few that did in fact come into existence.
SOUTHEAST ASIA
Were there any regions that were as well protected from the steppe as Latin Europe? Southeast Asia is the most promising test case: endowed with enough natural resources to support complex societies and large-scale state-building, yet far removed from the Inner Asian grasslands or any other steppe frontier, it was never dominated by hegemonic empires.
On the continent, a first phase of scaling-up commenced in the ninth century CE: Pagan expanded in Burma, Angkor in Cambodia and Thailand, and Dai Viet in Vietnam. Largely confined to the lowlands, these sizable but brittle polities were propped up by close cooperation with religious institutions—comparable perhaps to what had happened much earlier in ancient Egypt and Mesopotamia—and by favorable environmental conditions fostered by the Medieval Climate Anomaly that ensured strong monsoon rains. In the thirteenth and fourteenth centuries, all of them succumbed to escalating problems associated with population growth and climate change.
The period from the late thirteenth to the mid-fifteenth century was thus defined by political decentralization, followed by a phase of renewed concentration that lasted until the mid-sixteenth century. Even so, the latter did not result in hegemonic unification: in Lieberman’s count, the number of polities merely fell from twenty-three in 1340 to nine or ten in 1540. In a region of a little more than 2 million square kilometers, this reflected a considerable degree of fragmentation.
Even this modest consolidation ended in a shorter and less dramatic phase of collapse in the second half of the sixteenth century, which was reversed by the expansion of Burma and Siam across western and central continental Southeast Asia while devolution continued farther east until Western colonial powers began to intervene. Over the long run, in terms of internal structure, the political landscape gradually transitioned from “solar polities” composed of quasi-sovereign satellite regimes that were loosely attached to small cores to more cohesive states with stronger centers—a shift not wholly dissimilar to what we observe in post-Roman Europe.99
Maritime Southeast Asia, split into many islands and even more distant from any steppe, was completely immune to steppe inputs: the Mongol attack on Java in 1293 remained a freakish outlier. Local horses were small and imports rare. In keeping with the “steppe effect” model, the entire region was characterized by enduring polycentrism. Segmented vassalage empires heavily relied on naval assets.100
I confine myself to the Old World. Prior to the reintroduction of the horse by European settlers, the New World likewise lacked steppe inputs, an absence that makes the imposing scale of the Inca empire even more remarkable than it would otherwise be: it joins the Roman empire as the only other genuine example of hegemonic imperiogenesis detached from the steppe.101
“Follow the Horses?”
Despite its brevity and inevitable simplifications, this survey has revealed a strong correlation between military inputs from the steppe and the scale of state formation across the premodern world. If taxes, highlighted in my discussion of fiscal institutions earlier in this chapter, were indispensable for empire-building, so were horses. The conspicuous scarcity of large empires in regions that were ecologically well equipped to support them but were sheltered from major grassland zones highlights the causal dimension of this association. My series of vignettes of the “steppe effect” from Western Europe to China focused on the mechanisms that rendered proximity to steppe frontier zones conducive to imperiogenesis on both sides of the ecological divide: the pooling of military assets for the purposes of predation, preemption, and defense; the dissemination of steppe-sourced military techniques; the infiltration and repeated takeover of exposed agricultural regions by steppe warriors, as well as responses to these intrusions.
The agrarian empires that were forged in these complex interactions were often large because they were close to the steppe, rather than close to the steppe because they were large. Hoffman, in his dismissal of geographic factors, reverses causation when he claims that
large states like China were more likely to abut thinly populated regions where low rainfall would rule out sedentary agriculture and where herders, hunters, and armed raiders could thrive.… The large neighboring state would then face the risk of attacks by these nomadic groups, but the ultimate cause behind that threat would not be low rainfall in a nearby region but rather the size of the state itself, which was the result of politics.
Alongside an exceedingly rich historical record of which I have merely scratched the surface, statistical analysis of the correlation between the odds of China being unified and the frequency of nomadic attacks on it contradicts Hoffman’s reading. Behind all this lies a more fundamental problem—his notion that political history created “different political geographies,” and that “physical geography” can be juxtaposed with “political history” as if they were autonomous features. They were not. Political outcomes, such as patterns of state formation, were in no small measure contingent on geographic factors.102
Wherever it was ecologically feasible, Inner Asian interventions intensified over time. Although we ought not to underestimate their role even in some instances of ancient Near Eastern and ancient Chinese state formation, they did not generally make a decisive contribution to the emergence of the earliest empires. Inputs were mostly indirect, such as the diffusion of the chariot. As steppe cultures developed, their influence steadily grew. From the beginning of the fourth century CE, China’s macro-political evolution was increasingly shaped by it, and from the seventh century CE onward, conquest regimes from marginal zones became the principal drivers of state formation in the Middle East. In South Asia, irruptions commenced near the end of the first millennium BCE and gave rise to the most dominant empires of the second millennium CE. In the High Middle Ages, Eastern Europe was likewise reshaped by interactions with the Inner Asian grasslands.
The impact of the steppe frontier peaked in the early modern period: whereas the Mughals ruled 5 million square kilometers and 175 million people in 1700 and the Qing 12 million square kilometers and more than 300 million people in 1800, the largest states of the “protected zone”—France, Burma, Siam, Vietnam, and Japan—in 1820 covered between 300,000 and 900,000 square kilometers and housed between 4 million and 32 million people.103
One final observation. Proximity to the steppe mattered not only because of the competitive dynamics it engendered among agriculturalists and pastoralists alike. It also skewed the deployment of resources in ways that could have a profound impact on polity size. Chiu Yu Ko, Mark Koyama, and Tuan-Hwee Sng have developed a simple model that tests for the effects of threat direction and strength on state formation. They compare a scenario in which a large region faces a severe threat but only from one side—the Chinese case of relative isolation combined with exposure to the steppe to the north—with one in which it faces two-sided but weaker threats, a stylized analogue of Europe with its robust linkages to Asia and North Africa but greater detachment from the steppe. In their model, political centralization—a single empire—turns out to be more resilient to a one-sided threat. On the other
hand, fragmentation—a state system—is more resilient to a two-sided threat as long as military power projection by these external opponents is not efficient. The absence of strong steppe inputs satisfies this last condition.
These predictions mesh well with what we observe at opposite ends of Eurasia. In East Asia, the concentration of challengers to the north favored hegemonic empire among the agriculturalists, as well as the decentering of the capital cities toward the threat zone, a feature well documented for most Chinese dynasties. In Europe, the absence of a severe one-sided threat facilitated decentralization. Otherwise, overwhelming challenges might have destroyed members of a state system that were concurrently engaged in competition with their peers. The lack of such challenges thus both encouraged and enabled European polycentrism.104
CHAPTER 9
Culture
BUT WHAT ABOUT CULTURE? Didn’t the First Great Divergence also grow out of beliefs and ideologies that influenced state formation? It is easy enough to enumerate contrasting ideational traits that distinguished Europe from China and may plausibly be linked to observed outcomes—and I will indeed do so in the course of this chapter, just as I already identified a series of putatively significant environmental features. Yet the chief question is that of priority. Within historical time frames, the physical environment was a given; culture was not.
Cultural evolution unfolded within the geographic and ecological constraints of that physical environment. As agriculturalists and pastoralists created niches that allowed them to elaborate their respective modes of subsistence, they were in many cases unable to overcome the most fundamental obstacles to transformative change, such as temperature and precipitation regimes. Dramatic interventions in the environment that were driven by political preference were rare, and with the exception of China’s Great Canal were bound to remain rather limited in their overall impact on state formation.
It is not enough to show that culture and ideas mattered. That is banal. We also need to assess the extent to which they had been conditioned by the often intractable ground rules imposed by the physical environment. As we will see, while geography and ecology did not determine the cultural preconditions for macro-social scaling-up, they shaped and reinforced them through the outcomes they helped bring about. Over time, this caused different trajectories of state-building to become associated with different sets of cultural traits and ideological commitments, especially in elite circles.
LANGUAGE AND WRITING
China
Chinese logographic script can be traced back to the Oracle Bones of the Shang period in the late second millennium BCE, when it already appeared to be well developed. It underwent considerable evolution in the Western Zhou and Warring States periods. Two versions—now dubbed “seal” and “early clerical”—that had been systematized in the state of Qin spread across China proper upon conquest. Although cursive versions developed and were disseminated alongside them, different writing styles coexisted without resulting in meaningful regionalization. The writing system eventually became more fixed in late antiquity (as the so-called regular script).1
The Chinese language is likewise documented from the Shang period onward. Rime dictionaries reflect attempts to impose uniform pronunciation of characters. The creation of the most famous specimen, the Qieyun of 601 CE, coincided with the imperial restoration undertaken by the Sui and arguably sought to reconcile differences in usage between northern and southern China. Even so, regional deviations from Classical Chinese, which had crystallized during the Warring States and especially the Han periods, persisted.
At the same time, two factors contributed to linguistic unity. First, the fact that Classical Chinese dominated formal writing throughout the imperial period ensured a great degree of cultural homogeneity among the literate elite. From the twelfth century CE, vernacular Chinese came to be dominated by northern China dialects (Old Mandarin). “Official speech,” rooted in this tradition, was employed as a uniform spoken language for late imperial—Ming and Qing—officials to guarantee mutual intelligibility. It was only in the twentieth century that this version was turned into modern Standard Chinese.2
Second, linguistic diversity within the core regions of Chinese had long been relatively moderate (figure 9.1). The long reach of Mandarin dialects was in no small part a function of the considerable ease of communication across the Central Plain—their original and much smaller core zone—as well as their more recent dissemination to the colonized frontier regions of Sichuan and Yunnan and later on Manchuria and Xinjiang. Conversely, much greater linguistic diversity survived in the southeast and south of China.
This pattern, not coincidentally, maps onto genetic variation, which is low in northern China and much higher not only between north and south but also among southern groups. The unusually high degree of homogeneity in the north was facilitated by gene flows across flat and politically united terrain.3
All this suggests that for all its diversity, China was relatively unified in ways that mattered to state formation. Elite communication was eased by a single writing system that allowed exchange across regional dialects, by an elite language that, unlike Latin, was never allowed to fade, and by uniform pronunciation among state agents, imposed from above. Moreover, the physical environment of northern China was conducive to demotic integration.4
Western Eurasia
No comparable unity existed at any time in western Eurasia. At the elite level, the Roman empire was resolutely bilingual and bi-alphabetic. Thanks to antecedent Greek migration and Macedonian imperialism, Greek was dominant in the empire’s eastern half, whereas Latin became dominant in the west. While the two alphabets shared a common root and did not differ greatly, the languages were mutually unintelligible. The state class helped preserve this dualism: although Latin was obligatory in the military (and in Roman law), the elite of the politically dominant west embraced bilingualism as a mark of high culture. Greek received an additional boost from the spread of Christianity, which was long dominated by Hellenophones from the east.
FIGURE 9.1 Modern distribution of Chinese dialect groups.
In addition, parochial languages continued to be spoken by large segments of the empire’s population. Aramaic was common in the east, in Syria and Palestine, and a substantial corpus of literature was produced in Syriac. The Egyptian language flourished in its homeland and even acquired a new alphabet, Coptic, that was likewise employed in literature, even as Greek retained its leading position in this domain. Punic survived in North Africa until late antiquity, with Latin script being used to record it in inscriptions. In Gaul, spoken Celtic likewise persisted throughout the Roman period, quite possibly on a large scale. Asia Minor boasted a whole range of local languages: Phrygian even made a comeback in Christian inscriptions. Thracian languages are likewise known to have survived into late antiquity, and in Britain, Celtic endured and receded only after the end of Roman rule. By analogy, there is no compelling reason to assume that local languages disappeared from other regions such as (Iberian and Celtiberian) Spain, the (Celtic) Alps, and the (Illyrian) northern Balkans.5
In these regards, the sphere of Roman power was substantively more diverse than that of the contemporaneous Han dynasty: there was no single elite language and writing system, and parochial languages were much more prominent. The east–west split in elite circles was arguably the most consequential: while logistical constraints—foreshadowed by limited economic integration beyond often state-sponsored transfers and conduits of exchange—would have greatly contributed to the empire’s eventual partition, it was hardly a coincidence that it fractured almost precisely along this linguistic fault line.6
Post-Roman developments further added to the mix. Arabic was introduced to the Levant and the Maghreb as well as, albeit less perduringly, to the Iberian peninsula. In Europe, the impact of the Germanic and Slavic diffusions varied a lot across regions. Gothic never became dominant in the Iberian peninsula, and Frankish (Old Franconian) made only limited
inroads into Gaul, even though its northern and eastern reaches remained bilingual for centuries. Italy, where Latin was more deeply entrenched than elsewhere, was even less deeply touched by the presence of Goths and Lombards.
Other areas were more strongly affected. In England, Celtic (Common Brittonic) languages mingled with and were gradually pushed back by a complex array of Old English (introduced by Anglo-Saxon settlers), Old Norse (from Danes and Normans), and Old French (again via the Normans). On the continent, Germanic languages came to dominate Bavaria and Austria. Slavic infiltration took over much of the Balkans and the western Carpathian region.
Regional versions of Latin slowly but surely diverged into separate Romance languages. From the ninth century onward, in a process I briefly summarize in the Epilogue, literate residents first of France and then of Italy and the Iberian peninsula began to write down their own vernacular forms of Latin. As a result, even as the church maintained a single Latin standard, the language gradually lost its monopoly on literate and elite discourse. Instead of being blocked, linguistic regionalization was readily acknowledged and facilitated by political divisions. In the High Middle Ages, intensifying state formation actively promoted this process: at that point, most writing was no longer done in Latin at all.7
In all of this, it is important to distinguish cause and effect. China’s single writing system existed because it had been sponsored by a single state. Mandarin was both imposed by the central state and spread under the umbrella of imperial unity and expanding settlement in the peripheries. Empire sustained and on occasion impelled cultural harmonization, and was in turn strengthened by it.
Despite this mutual reinforcement, basic counterfactuals suggest that the arrow of causation from state power to cultural unity was stronger than the one pointing in the opposite direction. Classical Chinese had been cemented by four centuries of Han rule, and its subsequent susceptibility to splintering leaves no doubt that political fragmentation mattered: had it run unchecked, linguistic divergences might well have continued, ending mutual intelligibility even in elite circles. Multiple writing systems might have emerged alongside these vernaculars. Nothing in the Chinese language or writing system per se favored unity and consolidation: it was the exercise of political power that provided the necessary impetus. As we will see shortly, the same principle applied to content, to the ideas that were being maintained and spread by means of a unified elite script and language.
Escape From Rome Page 32