Imperial China, Chs 2-4: Liao hybrid government and ‘concavity vs convexity’

These chapters examine the Liao Empire (916-1125), a dynasty of warrior-nomads who were key players in Asian regional politics. There were lots of interesting details in these chapters, but it’s the Liao’s style of government that I’ll focus on here.

The core problem faced by steppe-warriors who conquered parts of China in this era was how to retain their martial, mobile ways. Once they conquered a sedentary people, the temptation was to settle and enjoy the spoils of their victory – but then they’d end up becoming ‘little Chinese’, losing cultural integrity, and being absorbed into China in the future. The steppe people in question here are the Khitans, and their empire stretched across the north of ‘Inner Asia’ (the peninsula in the bottom right is Korea).

Source: Wikimedia Commons

Abaoji was the Khitan Great Khan from 907-926 (later declared the founding emperor of the Liao dynasty) and he set out a dual administration model that the Liao would employ long after his death. They used a steppe system of government for the steppe part of their empire, and a Chinese system for the sedentary part that they were conquering. The Khitan elites had to function in both spheres, Mote writes: ‘many became culturally dual in order actually to function and not merely play figurehead roles.’ Abaoji and his successors were simultaneously the Great Khan at the top of one system and the Chinese emperor at the top of the other.

The model was ingenious. The ‘Northern Chancellery’ would look after military, Khitan and other tribal matters, including tribute from other steppe peoples in the empire. The ‘Southern Chancellery’ would govern sedentary society along Chinese lines, collecting taxes from agriculture and trade.

The traditional steppe society of the Khitans was not easily reconciled to their new role, however. Tribal chieftains were accustomed to ‘electing’ a Great Khan (essentially acclamation by his peers). This worked well on smaller scales, as the leader could be replaced without widespread conflict and you could select for the martial traits the Khitan wanted. But when you governed an empire, disputes in succession led to civil war. The old method wasn’t sustainable at this bigger scale.

The Chinese had handled this issue (as they would in future) by using a system of fixed succession and extreme subservience to the emperor. The Liao eventually established a succession protocol based on primogeniture some years following Abaoji’s death, modelled on the Chinese system. The actual Chinese system was much more complex than this, however, and here I will quote Mote at length:

At its best, of course, the Chinese government embodied many other elements: cultural and institutional limitations on the power of the throne that were not apparent to Inner Asian visitors. A responsible civil bureaucracy, advised by military associates, in possession of systematically assembled information and observing powerful precedents, provided the Chinese throne with judgement and guidance. What appeared to be servile bureaucrats at the Chinese court were in fact self-assured officials carrying out political actions largely of their own making. The Chinese Emperor was a despot limited by intricate cultural and bureaucratic restraints; in fact, he shared authority with his scholar-officialdom. It was difficult for the outsider to sense such complex interrelationships behind the pageantry. Yet he would have been correct in perceiving the strength of the Chinese Imperial institution, particularly the stability that was provided by normally unchallenged tenure of reign and by predetermined succession.

This made me think of Vitalik Buterin’s post on convex and concave worldviews. It seems that the Chinese system was an example of concavity:

(N.B. it’s a bureaucracy, but I chose ‘technocracy’ as there’s a civil service exam, and the impression given by Mote is that there was a degree of competence associated with the scholar-officials.)

As you can see in the graph, the most effective system is the hybrid one that combines both monarchical and technocratic aspects. As Vitalik Buterin shows in his post, it’s also possible that trying to combine two systems is actually worse than just pursuing one; this is the convex worldview, where the graph is a valley instead of a hill. The bottom of the valley is halfway between the two systems/policies – the ‘dime store ditch’ in the words of Ben Thompson, who used a convex graph to describe the West’s response to coronavirus.

(Obviously the concave trade-off above didn’t hold true forever, as I expect to explore in future posts on the collapse of various dynasties.)

The same principle of concavity might have undergirded Liao success in importing Chinese institutions. It is surprising, Mote points out, that the Khitan leadership retained decision-making power amidst a proliferation of titles and authorities in the new, hybrid government. Karl A. Wittfogel and Feng Chia-sheng’s theory of ‘creative misunderstanding’ seeks to answer this question. Organisational capacity was improved as Chinese political forms were borrowed and imposed on top of Khitan society, even if it was not a proper copy of the original structures:

The resulting forms retained some of the resilience of tribal society while integrating that with the exploitative government imposed over the sedentary populations. They created the continuity and stability of the ruling institution without feeling the need for resolving all the non-rational features of that juxtaposition. Never stopping to rationalise all these contradictions, they created an imperfect system but nonetheless served Khitan ruling interests very well: it strengthened their tribal empire without undermining their society’s structural integrity.

[N.B as summarised by Mote.]

Another happy medium, then:

The Khitans’ success in retaining both their steppe character and adopting the forms of Chinese rule was also based on a specific role for the elite. Some, at least a significant minority, were expected to become culturally dual and thus able to rule the southern, Chinese part of the empire, while the rest of the population would remain in the steppe. This appears to have been accepted sufficiently by both elite and populace that those culturally dual Khitans still regarded themselves as – and were regarded by the populace as – Khitan, despite their cultural acclimatisation. The ‘social base’ of the Khitan was thus preserved.

Another interesting aspect of these chapters was the way the Khitan amalgamated the various steppe peoples under their rule. These groups were mobilised for military purposes, and that provided both a motive and means for reducing the complexity of the population under Khitan control. Mote notes that the Mongols were later to amalgamate the populations under their control to a much greater extent.

This simplifying tendency also presented itself in the realm of law. Liao law ‘[moved] away from individual oriented “ethnic law” towards the realisation of statewide uniformity in “territorial law,” a considerable advance both in terms of legal concepts and as a practical measure for providing uniform conditions throughout the area under the Liao dynasty’s administration.’ Again, simplification makes life a lot easier for governments if they can pull it off.

Additionally, the Khitan had varying degrees of control over different populations within their empire. There were the core Khitan and those closely associated with them, then six ‘large tribes’ who were not directly administered by the state but instead had ‘kings’ appointed by the Liao. Then there were ‘various tribes’ who were supervised by Khitan officials, some of which appear to have lived partly within and partly outside Khitan control. Mote writes: ‘Those borders appear to have been fluid, changing according to the varying abilities of the Liao armies to control them.’ Beyond that were seventy-eight ‘subordinate states’, who either had requested diplomatic recognition from the Liao or paid them tribute.

Political control, then, was not a hard boundary but more of a gradient between different groups and polities. This reminded me of the contested boundary between the Eastern Roman empire and the Sassanids, and the role of Arab proxies there (see Tom Holland’s In the Shadow of the Sword for discussion of this). Control of territory wasn’t necessarily total, and there was no hard line separating one state from another.

This hybrid government combined Chinese and steppe features to great effect, allowing them to rule over both nomadic and sedentary peoples for two centuries.