Civil War in Guangxi
The Cultural Revolution on China's Southern Periphery
Andrew G. Walder




Why was Guangxi’s violence so much more intense than that of other regions of China? This straightforward question breaks down into several separate ones. The first concerns the origins and nature of the conflict between the Allied Command and April 22 factions, and the issues that motivated their mutual animosity. These groups did not previously exist, and began to form only in February 1967, in the wake of a mishandled rebel power seizure over the provincial government in the capital city of Nanning. The factions took full form in Nanning in April of that year and spread steadily across Guangxi’s other cities and counties in the months to come. Who were these groups, how and why did they form, and what set them on an opposed course? How far did they spread across Guangxi, why did they spread, and how deeply did their conflicts extend into rural communities? It would seem essential, as a start, to closely examine the political divisions that were the ostensible motivations for the violence.

A second question is why so much of the violence was focused on politically stigmatized households, whose family status resembled caste distinctions handed down over generations. Why would violence against these households, largely uninvolved in factional conflicts, become so widespread? Was this development inadvertently spurred by the factional warfare carried out in the name of the Allied and April factions? These political categories, and the associated stigma, existed all across China. Why did these groups become such a focus of political violence in Guangxi?

A third question concerns the scope and intensity of the violence. Factional warfare was common across China during this period, and the restoration of political order exacted the highest death tolls of this period nationwide.1 Politically stigmatized “enemy” households existed everywhere, and they were routinely persecuted. Yet the death tolls from factional warfare, its suppression, and the persecution of stigmatized households were far higher in Guangxi than elsewhere. What caused political violence to be more widespread and intense once it had gotten under way?

There are two broad approaches to these questions. The first is to characterize the political processes that generated high death rates—a focus on what happened, on how events unfolded into a deadly climax in 1968. What actors were responsible for killings, and why did events unfold the way they did? The second approach is to set aside narrative accounts and identify distinctive regional characteristics, defined by history, ethnic diversity, or economic and political structures, that may have intensified political violence to levels that surpassed those of other regions. To simplify, the first approach is to focus on political activities as a process, tracing them as a narrative flow of events; the second is to identify static structures or variable characteristics of the region that set it apart. In evaluating potential explanations, we need to keep another distinction in mind. Our problem is not to explain political violence in Guangxi, but to ask why violence in Guangxi was so much more severe than elsewhere. These two questions are not the same, and I will return to this distinction later in this book.

Political Processes

There are three different claims about Guangxi’s politics in the scholarship about this period, and all are relevant to one or more of the questions that I have posed. The first places a heavy emphasis on orders issued from Beijing on July 3, 1968, which called for the ruthless suppression of continuing factional resistance to the long-delayed reimposition of political order in Guangxi. The second characterizes factional conflict in China during this period as a struggle between conservatives with vested interests in the status quo and radicals who sought to challenge the existing system of power and privilege. From this perspective, Guangxi was a particularly violent example of how far the powerful were willing to go to defend their privileged positions. The third draws on sociological theories about intergroup violence as collective behavior, and it attributes the unusually high death tolls to the mobilization of community-level antagonisms toward politically stigmatized households in “reactionary” class categories.

Beijing’s “July 3 Orders”

Official histories published in China provide detailed chronological narratives with a bare minimum of interpretive structure. They recount major events and describe in detail the formation of factions and subsequent conflicts, while providing little or no interpretation or analysis, never framing puzzles to which alternative explanations might apply.2 The narratives implicitly condemn the violence—the primary reason for describing it in such detail. The explanation for the violent outcome places overwhelming emphasis on the impact of the “July 3 Orders” (qi.san bugao).3 This was a harshly worded directive from Beijing in 1968 that mandated the restoration of order in Guangxi by force, designating rebel holdouts as class enemies and agents of foreign powers. In these accounts, the orders are condemned as extreme leftism, and the wave of killings that followed is framed as the consequence of an erroneous political line. From this perspective, the July 3 Orders are the primary feature that set Guangxi apart from other regions.

The July 3 Orders indeed triggered subsequent events and provided a political opportunity for Guangxi’s military and security forces to apply overwhelming force against a stubborn insurgency. But an interpretation that places such singular emphasis on a directive from the nation’s capital portrays the outcome as a foregone conclusion, the following of administrative orders by civilian and military authorities. The restoration of political order elsewhere in China during this same period also generated the highest observed death tolls.4 This raises the question of why the July 3 Orders resulted in a much more deadly outcome in Guangxi. As we shall see in later chapters, mass killings in rural Guangxi began well before July, so the orders appear only to have accelerated processes that were already under way.

Political Struggle between Conservatives and Radicals

What political processes were unleashed by the July 3 Orders? Two very different interpretations have been applied to this question. The first is that the Allied Command faction represented forces that were aligned with the existing power structure, against a more radical April faction that sought to challenge it. This interpretation is supported by the observation that the Allied faction was in a mutually supportive relationship with the Guangxi Military District and local PADs. Moreover, the stance that defined their political orientation was their declaration of support for Wei Guoqing, Guangxi’s longstanding top official, who in early 1967 was appointed to lead military control forces. Wei’s highly unusual retention as head of military control forces cut directly against the grain of the radical Maoist offensive against “revisionist” power holders in late 1966. The April faction, by contrast, was adamantly opposed to Wei’s appointment as head of the military forces that served as an interim substitute for a collapsed civilian government, and they fought long and hard to resist the imposition of order by units aligned with the Guangxi Military District. From this perspective, violent conflict was a struggle over the reimposition of the existing order, with one side fighting to defend it, while the other sought to overturn it.

This interpretation is a variant of the interest group interpretation of Cultural Revolution factionalism known as the “radical-conservative hypothesis.”5 At the core of this notion is the idea that individuals who enjoyed advantages in existing structures of power and privilege—in particular party members, households headed by revolutionary veterans, party cadres, and the military—had an inherently conservative orientation and would gravitate toward factions that sought to defend these structures and reimpose order through repression. This interpretation is popular with émigré authors, often former participants in these conflicts, who fought against the imposition of military control and saw themselves as the true rebels against injustice and oppression.6 It also appears as a central organizing theme in critical histories of the Cultural Revolution penned by Chinese authors who dissent from official interpretations. These accounts posit a struggle between a “bureaucratic clique” and rebel forces, manipulated by Mao, which ultimately resulted in the victory of the former.7 One account of the Cultural Revolution in Guilin portrays the struggles as a rebellion by a radical April faction against the provincial power structure and their conservative defenders in the Allied faction.8

This interpretation is lent further plausibility by the rhetoric of the two sides. The Allied Command eventually portrayed their opponents as counterrevolutionary forces seeking to overturn the Communist Party and the political order. The April faction portrayed themselves as a progressive force seeking to overthrow reactionary elements defending the old regime. From this perspective, the violence was an expression of the deeply rooted nature of the political forces in conflict over existing structures of power and privilege. The implied explanation for Guangxi’s unusually high death tolls is that they illustrated the extreme violence to which the powerful were willing to resort in defense of their privileges, once authorized to do so by Beijing. This analysis views the political processes in Guangxi as identical to those elsewhere in China, with the July 3 Orders as a catalyst that removed constraints on extreme repression.

This line of analysis offers solutions to two of our puzzles—what political forces the two factions represented and why their conflict became unusually violent. It views the two factions as a reflection of latent political divisions that existed prior to this period, and that found organized expression in a period of disorder. And it attributes the violence to the enormity of the stakes involved and the willingness of the powerful to defend their privileged positions at any cost. This characterization of the conflict, however, does not address the widespread persecution of politically stigmatized households, or the ferocity of the violence unleashed against them. To approach these questions, we need to consider other ideas.

Community-level Collective Behavior

A very different portrayal of political processes in Guangxi directly addresses these shortcomings. It views the struggle between factions so often portrayed as radicals and conservatives as largely irrelevant to the generation of the unusual death tolls. In this account, factional struggle was restricted almost entirely to cities and towns, while all but a small percentage of deaths were in rural villages, where conflicts between the two factions did not penetrate. This analysis attributes the high death tolls to collective behavior in villages inadvertently touched off by claims in the July 3 Orders that the April faction was in collusion with historical class enemies. This framed stigmatized “class enemy” households as targets of retribution, and killings spun out of control in remote rural communities as an unintended consequence of official rhetoric during a campaign to suppress an urban insurgency.9

This explanation draws on theories about social processes that generate intergroup violence and genocide. In this case, the social categories that created group antagonisms were defined by political labels affixed to households at the time of the Communist seizure of power. The families headed by members of former propertied social classes (landlords, rich peasants), those affiliated with the Nationalist regime, or those who fell afoul of early campaigns against real or suspected opponents of the Communist Party were collectively labeled as political enemies, and frequently referred to as “four type elements” (silei fenzi). The “four types” were “landlords, rich peasants, counter-revolutionaries, and bad elements”—labels given to individuals and households when the new regime was established. With the addition of “rightists” (youpai fenzi), individuals sanctioned in later political campaigns in the 1950s, the stigmatized categories expanded to “five type elements” (wulei fenzi)—di fu fan huai you. The labels had the caste-like feature of being inherited across generations. Individuals in these families were systematically discriminated against in job assignments on collective farms. During political campaigns they were the first to suffer suspicion and punishment.10

According to this explanation, Guangxi’s military authorities sought to discredit the April faction by charging that they were conspiring with former class enemies who sought to overthrow Communist Party rule. The July 3 Orders explicitly embraced this portrayal of stubborn April faction resistance, implicating politically stigmatized households, which were labeled and stigmatized precisely for their historical opposition to and presumably enduring animosity toward the new regime. The campaign to suppress the April faction, in this account, inadvertently unleashed collective violence in villages that focused on these households, particularly in rural regions where there were few if any members of the April faction.

High death tolls were therefore not due to actions by the Allied faction or agents of the state, but instead were the product of mass killings in villages, where popular antagonism toward these stigmatized groups was activated by the political rhetoric of the military authorities. The campaign to suppress the April faction incited residents of rural communities to engage in campaigns of extermination against their own neighbors, who were otherwise uninvolved in the political battles of the towns. This explanation invites parallels with genocidal intergroup violence in Rwanda in 1994; communal warfare among Hindus, Muslims, and Sikhs in the Punjab, Bengal, and other regions spurred by the partition of India in 1947; and ethnic cleansing by local militias in Bosnia-Herzegovina during the Yugoslav wars of the 1990s.11

While state actors often facilitated the violence in these other settings, ordinary citizens participated actively in killing other members of their communities. In Rwanda: “the violence was low-tech: many perpetrators used ordinary farm tools, such as machetes, clubs, and hoes, to kill. The violence was public, face-to-face, crowd-enforced, and neighbor sometimes killed neighbor.” “Perpetrators in rural areas were ordinary men. They were fathers, husbands, and farmers who had average levels of education and who had no prior history of violence . . . for the most part, Rwanda’s genocide perpetrators . . . were regular citizens.”12 In Bosnia, “much of the Serb population took part in the genocide against Bosniaks [Muslim Bosnians], either directly or as bystanders.” “Witnesses knew their attackers—their Serb neighbors—who participated in the looting, raping, torturing, and killing of Bosniak men and women.”13

This proposed explanation, unlike the others, directly addresses the puzzling victimization of stigmatized households, placing it at the center of attention. It seeks to explain how a conflict between urban factions could have spilled over into rural regions where these conflicts seemed to be largely absent, mobilizing ordinary residents into acts of widespread violence against families that were apparently uninvolved in the larger struggles. It also implies that Guangxi’s high death tolls were a product primarily of community-level collective behavior in rural regions, which generated violence that went far beyond what the authorities anticipated.

This line of analysis sets aside questions about the nature of factional conflicts, which are viewed as tangential to the rural killings. Instead, it directly addresses violence against stigmatized households. But it does not provide an equally compelling explanation of why the persecution of these households was so much more severe in Guangxi than elsewhere. There were stigmatized households everywhere in China, and there were similar outbreaks of mass killings of “enemy” households in other regions, although they were rare and isolated events.14 Why would mass killings of this type elsewhere in China be limited to a small number of localities, whereas they spread like a tidal wave across Guangxi? One possibility is that Guangxi had distinctive features that may have served to escalate community-level conflicts to an unusual degree.


1. Walder and Su (2003); Walder (2014; 2019, 173–87).

2. For example, Guangxi Cultural Revolution Chronology (1990).

3. Central Committee, Chinese Communist Party (1968); and the account in Bu (2008, 707–15).

4. Walder and Su (2003); Walder (2014; 2019).

5. The term, and the interpretation, was first developed by Lee (1978). Variants of the argument are Chan, Rosen, and Unger (1980) and Rosen (1982). It has been widely influential, although it has often appeared in modified form (e.g., Andreas 2002; 2009).

6. For example, Liu Guokai (1987), who was aligned with the defeated Red Flag faction in neighboring Guangdong.

7. For example, Yang (2021).

8. Hua Linshan (1987; 1996). Hua was a native of Guilin and sympathized with the April 22 faction.

9. This is the highly original approach offered by Su (2011).

10. Unger (1984); Walder (2015, 108–12).

11. For example, Strauss (2006, 65–121) on Rwanda; Brass (2003), Copland (1998; 2002) and Khan (2007) on the Indian subcontinent; Bećirević (2014, 54–143) and Naimark (2001, 139–84) on Yugoslavia.

12. Strauss (2006, 1 and 96).

13. Bećirević (2014, 85 and 88). Similar cases of extreme intergroup violence come from 1941, in war-torn Europe: Bosnia (Bergholz 2016) and Poland (Gross 2002).

14. The widely cited cases were Dao County, Hunan, in late 1967, and neighboring counties in Lingling Prefecture, and two rural counties in the Beijing suburbs in late 1966 (Song 2002; Tan 2010; 2017; Yang 2021, 346–54).