On January 11, 2011, the People’s Liberation Army tested a new J-20 stealth fighter jet. The move surprised the visiting American Secretary of Defense, Robert Gates, who had been given no warning of the test. But far more troubling than the jet itself was the fact that Chinese President Hu Jintao was as surprised as Gates. The head of the People’s Republic of China (PRC) had been blindsided by his own military.

It was a telling moment that belied the conventional story of civil-military relations in China. That story begins in 1929, when the Chinese Communist Party (CCP) convened a meeting in Gutian, in Fujian province, that established as an inviolable principle the Party’s authority over the military. The Red Army, the CCP declared, was to implement the revolution under the command of the Party. Power grew from the barrel of the gun, as Mao Zedong put it, but the Party held the gun. In the official history, that was that: The military was subservient to the Party, and from then on the ideological work of enforcing that subservience fell to all Chinese citizens.

That is how the story is told, but the reality is less tidy. There were no guarantees in 1929 that the CCP would ever come to power. When it did, its establishment benefitted from individuals and groups who were often not directly answerable to anybody. Their ties to the CCP were frequently a matter of shared interest or passion more than any codified doctrine of civil-military relations.

None of this, however, stopped Xi Jinping from invoking the spirit of 1929 when he convened a new Gutian conference in September 2014. The traditions established at the 1929 gathering, Xi explained, needed to be carried forward. Guojiahua, the idea that the People’s Liberation Army should serve the country, not the Party, was erroneous and unacceptable. The military could not have an agenda independent of the Party.

Xi delivered another warning at the 2014 Gutian conference: on the perils of corruption. Corruption had to be rooted out, he announced, and the armed forces would do well to reflect on the Xu Caihou case. Xu, a former Vice Chairman of the Central Military Commission, had been jailed for bribery in June. There was much work to be done to keep the military pure, and Xi, so he declared to the assembled officials, was not going to slacken in that work.

In Xi’s decision to emphasize Party authority alongside the scourge of corruption, there is evidence of how precariously Chinese governance operates today. Had Party authority indeed been unquestioned, there would have been no need to assert it so insistently. If there were no resentment of the civilian government within the People’s Liberation Army (PLA), and perhaps even examples of insubordination, there would have been no reason to invoke the 1929 meeting in the first place. Many observers see Xi Jinping’s anti-corruption campaign as a pretext for just another Party purge by a power-hungry would-be dictator. That a purge is going on there can be no doubt, but Xi is also driven by genuine concern about both the insidious damage being done to China by corruption and the involvement of the military in that corruption.

Xi’s dilemma is unenviable. He has to continue the drive, and the purge that is part of it, because he has staked his credibility on cleaning up the mess. More important, he needs to alleviate growing popular discontent and recover lost assets at a time when the economy is slowing somewhat and people are increasingly angry about entrenched inequality and anxious about the future.¹ But he has to know that his anti-corruption drive threatens PLA members’ interests in unpredictable ways. To incur the wrath of men with guns is not something to be undertaken lightly. A military coup, once unthinkable in the PRC, is now conceivable.

To understand why, it is worth remembering that civil and military spheres have never been as neatly separated in China as in the West. “The sky is high and the emperor is far away” goes an ancient Chinese proverb: Central authorities were distant from the day-to-day lives of their subjects in such a vast state, so civilians had to assume functions that in the West would be considered best left to the military. When steppe nomads raided settlements, there was no time to wait for the emperor to send troops; people had to organize their own defense. Hence the stories of heroes fighting in China’s wilderness: men who with mighty staffs and sharpened spears, righteous fists and brave hearts, could defy and defeat anyone from marauding robbers to corrupt officials.

Myth and history came together to weave such romances deep into Chinese culture. The story is still told of the village of Sanyuanli, where, during the Opium War, local farmers surrounded and drove away British troops while the cowardly Qing court did nothing effective. The villagers prevailed principally because rain had soaked British muskets and because the British themselves were keen on a negotiated settlement, but that did not prevent the incident from being remembered as an example of Chinese martial valor.² The Taiping Rebellion, which claimed some twenty million lives, was started off by Hong Xiuquan, a civilian who thought he was God’s younger son and who gathered civilians to the ranks of his heavenly army. The militias that sprung up to deal with the Rebellion were led mainly by civil servants independent of the Qing state’s armies; that these militias chose to offer their loyalty to the state was a matter of serendipity, not doctrine. The warlords who carved China up in the aftermath of the Qing’s fall were people to whom civil-military distinctions mattered little. They were armed men who governed bits of territory through shifting combinations of fear and the provision of social services. Mao Zedong’s concept of a people’s war, too, rested on the notion that hungry and dispossessed civilians could wage war. The Communists’ rise to power in fact owed much to its capacity to erase lines separating the civilian world from the military one.

After the Party took control, civilians continued to undertake jobs that in the West would have been left to a professional army. The PLA saw action in the Korean War, but so did young peasants from across the country, who dropped their ploughs and headed for the front to “resist America and help Korea.” In the late 1960s China launched a military intervention to support the Communist Party of Burma, and it too relied heavily on civilians from the southern provinces streaming down to do battle. The Red Guards as well, whether fighting each other, their elders, or foreign missions, saw themselves as troops. Theirs was the language of siege, command, battle, and conquest. Even today, the conflict with other Asian states over the Spratlys and Paracels involves civilian fishermen willing to fight for land they see as theirs. They are encouraged by the state, certainly, but the encouragement works because it builds on a tradition in which military action is dispersed, something anyone can conduct if the circumstances are right. Video games that allow Chinese to kill Japanese war criminals, or nationalist protests running well ahead of where the government wants them to go and spilling into anti-Japanese violence, show how deep this tradition still runs. Militias that are not terribly well regulated have become a hallmark of Chinese life. To call China a militarized culture implies discipline. A “militia-rized culture”, where anyone can play at being a hero of the marsh, is closer to the mark.

There are, to be sure, other places where such cultures exist. Afghanistan and the Kurdish areas of the Middle East come to mind, as do lands inhabited by the Tuareg and Chechens. And as in other places, a militia-rized culture, while it has its uses, can be dangerous to a sitting government. If armed force is divorced from state authority, there is no reason it cannot turn against said authority. When the Mandate of Heaven passes from a government, it becomes a fair target. Whether it is the Sanyuanli villagers defying the British, Hong Xiuquan fighting the Qing, or Mao defeating the Kuomintang and the imperialists, stories abound of the Chinese people taking on authorities they disliked. And in Xi’s China there is plenty for a disgruntled marsh hero to fight.

Moreover, angry, unmarried, and underemployed young men—always a potential source of civil strife—loom everywhere in China, a legacy of the one-child policy and rampant female infanticide. Some who are wealthy and well connected can simply leave instead of voicing discontent, parking their assets in Boston or Manhattan real estate. But for the vast numbers of the discontent who remain, sustained uprisings directed against a disengaged, oppressive regime are a real and growing possibility. So the issue is not just civil-military relations; it is societal-civil-military relations, a triad rather than a dyad of potential trouble.

Xi Jinping knows all this. He is painfully aware that addressing the sources of discontent requires showing that the state actually cares about the well-being of its citizens. This is where the anti-corruption drive and the emphasis on rule of law come in. Corrupt functionaries, Xi wants his subjects to know, will no longer be able to get rich at the expense of the people. That this campaign has to apply to tigers as well as flies shows how deep Xi must believe the discontent runs. Taking on officials like Zhou Yongkang, Bo Xilai, and, most recently, the spymaster Ma Jian was not something to be done lightly, even if they were political foes. The risk of damaging Party unity by attacking them was considerable, and presumably would not have been taken had Xi not considered the risk of inaction even greater. For the Chinese people to truly believe that their Chairman was bent on ridding the country of corruption, high functionaries, not just petty ones, had to fall. And the campaign had to go beyond the civilian sphere, for the military, too, is seen as being in need of subjugation to the law.

It remains hard to get concrete information on China’s military-industrial complex, but PLA personnel are clearly perceived to have profited from the system in ways that ordinary Chinese cannot. With the PLA’s involvement in a range of companies—from defense to petroleum, aerospace to infrastructure—the opportunities for skimming money abound. The evidence, as prosecutors frequently point out, can be found in the cars driven, in the favors done and received, in the red envelopes passed silently as gifts. Gu Junshan, connected to Xu Caihou, stands accused of siphoning off 30 billion yuan, bribing people by offering them the keys to a Mercedes filled with gold. Xu himself allegedly had received sufficient ill-gotten gains to fill ten trucks. Gao Xiaoyan, one of the few women to have reached the rank of PLA Major General, is now under investigation on bribery charges stemming from her work for a PLA hospital. The very opacity of the system makes it harder to pass off military dealings as legitimate. There is always room to imagine corrupt activity. If corruption is to be truly rooted out in China, the PLA will have to be subject to relentless investigation, too.

Evidence of wrongdoing has not been presented in ways likely to satisfy an American or British court, but corruption is nonetheless a real problem in China. All else equal, it is easier to be corrupt in a murky economy, with a legal system far from explicit, consistent, and impartial. The opportunities for truly damaging corruption abound, which is why there is little reason to doubt Xi’s sincerity and determination. The current chairman was a youth doing hard labor during the Cultural Revolution. Just what lessons he drew from this experience are unclear, but he might well have perceived Mao as rejuvenating China’s national strength by tearing into the cadres at the very top of the political apparatus. Officials were a threat to national security then; they are a threat to national security now, and it is Xi’s turn to deal with them. Xi, one suspects, is that most unpredictable and misunderstood of leaders: a true believer with a mission.

There are two main problems bedeviling Xi’s approach toward the PLA: its lack of credibility and its impact on troop morale. As to the former, in the absence of an independent ombudsman that could investigate Xi as ruthlessly as it does his political enemies, the anti-corruption campaign can never seem wholly honest and just. As already noted, it is widely perceived as an old-fashioned settling of political scores, and this simply comes with the political real estate. After all, Xi’s predecessors used similar campaigns and similarly high-sounding rhetoric to eliminate political opponents. Again, the weakness of Chinese rule of law comes into play. While optimists have made much of Xi’s emphasis on the “rule of law” and the “constitution”, the fact is that Article 51 of the constitution specifies that citizens may not infringe upon the interests of the state—a clause vague enough to mean that Xi, as the paramount representative of the state, will decide when citizens start infringing. The credibility problem is exacerbated by the fact that none of Xi’s political cronies has been prosecuted.

The second problem is that Xi’s campaign damages China’s national security planning. Even the non-corrupt in the PLA have little way of knowing when they have crossed the line, for the line can shift at the whims of their chairman. The virtues of unpredictability have been famously espoused by Sun Tzu (which may be part of Xi’s rationale as he takes on vested military interests), but under current circumstances it hurts more than it helps. The reason is simple: The national security calculus in China right now demands massive military modernization, and that costs money. It is risky—indeed, potentially fatal—to ask for money for weapons, if the chairman suspects that the money will be stolen. The PRC’s military budget has already grown to $132 billion (it is probably much higher). As military planners attempt to counter American and others’ capabilities, it will only swell further. But if Xi’s anticorruption drive makes senior PLA officers wary of proposing budget hikes, it may anger patriotic and nationalistic professional soldiers who might feel that they are betraying the national trust.

In an unpredictable situation, where an attempt to do their jobs by asking for investment in better military systems might lead to an investigation, PLA members have, broadly speaking, three options. The first is to keep a low profile and cooperate with Xi; the odds are reasonable that cooperation will ensure safety. Most seem to be following this particular course. But if one is already under investigation or if one is connected, however loosely, with someone Xi happens not to like, keeping a low profile might not work. A second option is therefore suicide. Two admirals, Ma Faxiang and Jiang Zhonghua, leapt from tall buildings; a general under investigation for corruption, Song Yuwen, is said to have committed suicide by hanging himself.

The third option is resistance. Men with guns always have the option of using them. One or more groups of officers who feel their well-deserved economic gains are being threatened might well opt to take Xi on militarily. As far as we know, such things have not happened under the reign of the CCP (not least because under Mao and Deng, the line between civilians and the military was blurred indeed), but China has historically been prone to the military taking a hand in political affairs. It was a Ming general, Wu Sangui, who first opened China to the invading Manchus. It was Yuan Shikai, a Qing general, who usurped the leadership of the Republic of China at its birth. It was a benevolent general, Zhang Xueliang, who kidnapped Chiang Kai-shek, his civilian leader, and forced him to come to terms with the CCP. There is no guarantee that such military interference will never happen again. A combination of motives—an eagerness to protect a privileged economic place, the fear that if one does not strike, one will be struck down, the idea that one must protect one’s country from a dictator bent on tearing it apart—could impel soldiers to try to seize political power. These are the circumstances in which military coups happen elsewhere; there is no reason to suppose that China is exceptional in this regard. There, as anywhere else, it is impossible to know what will happen when one threatens heavily armed men.

A successful military seizure of the Chinese state could take several forms. The simplest would involve a senior officer or group of officers cooperating silently, gathering enough power, and then suddenly placing Xi under house arrest. Martial law would be declared—for the purest of patriotic motives, of course—and life would go on as normally as possible under such circumstances. Bloodshed would be minimal; our imaginary cabal would have gathered enough support to make resistance futile. They might then be either a transition back to Party rule, or the declaration of a new government that might or might not keep the policies of its predecessor largely unchanged. As military coups go, this would be the cleanest, neatest option.

It is an unlikely outcome, however, because the PLA is divided. Many within it have a stake in Xi’s survival. Far more probable is a scenario in which officers in charge of one of China’s military districts—in Sichuan, say, or perhaps in Jilin—decide that they have tolerated more than enough interference from the central government and declare war. There might be considerable local support for such a move; regional identities remain strong within China and resentment of a rapacious central government is easy to foster. Bo Xilai fell in part because of his popularity in Chongqing. Affordable housing and the idea that he would not let his Chongqingers down made Bo a hero to many locals. Beijing’s arresting him was for many just another example of the central government interfering with Chongqing’s well-being. Capitalizing on local discontent and China’s militia-rized culture, an enterprising military commander could well gather enough strength to challenge Beijing.

Were such a thing to happen, China’s fate could go in one of several different directions. If our imaginary commander were strong enough, an outright seizure of the capital after long, bloody warfare would be one outcome. Mao Zedong, after all, managed to seize power and unify the country after battling a series of foes. But given Xi’s strength, outright victory would be unlikely. Instead, one can expect a bloody stalemate, with the country dividing along north-south lines as old as China itself. “Two Chinas”, to use that dreaded phrase, could emerge. Balkanization might not stop there either. Once other military commands see the possibility of successful defiance, they too might act. Xi might find that quashing secessionists costs more blood and money than he can get his hands on. China might fall back into a new Warring States or warlord era, in which little fiefdoms spar, subside into coexistence, and then start sparring again.

All of this is purely speculative, of course. Fear of chaos and the patriotic education system provide a strong deterrent to such action. But unlikely things happen all the time. The survival of the CCP in the years following Gutian is one of them: The odds were stacked against the small band of peasants and dreamers who survived purges, a very long march, and Muslim warlords. It is worth remembering, too, that a unified China is far less of a norm in the five millennia of its history than what the official record claims. The country has fallen apart suddenly and violently many times in its past, often precisely because of the sorts of conflicts one sees unfolding today.

The rupture between civil and military spheres has often been heralded by academics as propitious for democratization. Transitions to democracy in Taiwan and Southeast Asia are cited as hopeful exemplars.³ However true the argument might be for Taiwan (though it fails to take account of the particular nature of civil society there, as well as of American pressure for democratization), there is little reason to hope for a peaceful, let alone democratic, outcome from civil-military conflict in China. Xi’s replacement would probably be another strong authoritarian leader, perhaps one more nationalistic and belligerent in his conduct of foreign affairs. In China, revolutions from below have tended toward violence, ending with the displacement of one despotism by another.

Even that might be better than a China falling to pieces. Disunity in China has meant immense violence in the past, with the effects often spilling over into neighboring countries. Given China’s integration into the global economy, internal chaos could spell trouble the world over.

The possibilities of military interference and subsequent trouble are not lost on Xi. He is too steeped in China’s past, and too savvy a political operator, not to realize that enemies can strike suddenly and without warning. Preventing military dissent was part of the rationale for forming his new national security council at the Third Plenum in 2013. Modeled in part on its American counterpart, the new body is tasked with both domestic and international security affairs; its mandate includes addressing terrorist threats from restive minorities as well as planning to neutralize American military power. It is also meant to be a “unified” structure for dealing with national security, with Xi himself at its head.

This means that the PLA’s perspectives on national security must now pass through a body that Xi chairs. The council is thus a way of reinforcing Party control over the military. It is a way of asserting authority, of making sure that marsh heroes do not go too far. (Xi might also have studied how Mao dealt with recalcitrant generals; Mao purged Peng Dehuai for an alleged anti-Party conspiracy. The new security council could allow Xi to keep an eye out for such conspirators). At one level, this is comforting: One does not want situations in which the Party Chairman is caught wrong-footed, as Hu Jintao was, before visiting senior American officials. But from the perspective of disgruntled PLA members, it constitutes another check on their capacity to do what they deem necessary.

To address resistance from within the military, Xi has also tried to reform the culture within the PLA. One initiative is to improve the auditing of military expenditures. The absence of external supervision has been cited as a reason for graft. Tighter regulations will keep power controlled and directed toward the goals of national security. Military modernization is hurt, after all, when funds that should go to purchase long-range missiles are purloined to buy bars of gold.

But much more important is the attempt to reorient troop loyalties. The risk of military dissent arises from the fact that soldiers might feel more loyal to their commanders than to Xi. A move is underway to educate them on the egregiousness of such thinking. Instead of answering to the “rule of man”, senior commanders and military academy leaders are making clear, troops must answer to the “rule of law”, which means obedience to the Party. Commanders, Major General Pan Liangshi recently declared, should have a “legal mind.” More than any Chinese leader since Mao, Xi seems bent on forging a cult of personality. The public appearances, the charismatic speeches, and the trinkets bearing his image help to portray him as the Great Helmsman, the embodiment, as with emperors of old, of the “rule of law” to whom the PLA’s soldiers owe ultimate allegiance as Commander-in-Chief. All this is being touted as the discipline crucial to combat readiness. It is also the discipline crucial to preventing the military from spiraling out of Party control.

It is difficult to take issue with Xi’s drive to control the military. Any civilian government wants to know that troops are responsive to the government, not to whoever happens to be in charge of their regiment. Anyone familiar with what revolutionary upheavals have looked like in China’s past can understand and even empathize with Xi’s concern.

Just how successful he will be, though, remains to be seen. That he must do something about corruption is undeniable, but can he subdue it without creating a new legal order that threatens Party rule too? True transparency is dangerous; it ultimately means setting up an authority not answerable to Xi or the Party. To have such an authority would be to place Xi and his allies at considerable risk. But absent such an authority, the anti-corruption drive will never be complete, never beyond cynical reproach; it will remain little more than a source of uncertainty and fear. And uncertain, fearful people can do shocking things.