The military suppression of demonstrators in Beijing in the summer of 1989 was a stark reminder that China's Communist Party leaders had never lost sight of Mao Zedong's famous observation, "Political power grows out of the barrel of a gun." Strangely enough, scholars of China's modern history have often been more reluctant to acknowledge the political importance of military power in the founding and survival of China's Communist government. Explanations of the Communist Party's political strength have usually paid more attention to ideological or organizational factors than to the control and application of armed force. It would be equally wrong to interpret the party's political success only in terms of military force, but the recurring and crucial political interventions of the People's Liberation Army suggest the need for a more balanced appraisal of the importance of military power in modern China's political development.
The role of the military in modern Chinese politics can only be properly understood when placed in its historical context. In 1921, the year in which the Chinese Communist Party was founded, China was suffering endemic civil warfare, which had begun a decade earlier with the 1911 overthrow of the last imperial dynasty and was fed by continuing political struggles for the control of the Republic established at that time. By the end of this decade of civil war, the possession of armed force had become an essential determinant of political power. The main benefactors of this condition were a large number of competing military commanders, usually referred to as "warlords." Relying on the personal command of their armies, these warlords extended their control over government administration and fought wars with each other to expand their political influence. Mao's recognition
of the linkage between military force and political power was not simply the result of abstract theorizing. Rather, it was a practical acknowledgment that under such conditions the Communist Party's own political struggle would depend not only on the battle for the hearts and minds of the Chinese people but also on the control of military power.
The study of the emergence of warlordism in the early twentieth century thus provides a crucial starting point for an understanding of military-civil relations in contemporary China. Unfortunately, scholarly research on this subject has been quite uneven. On one hand, many Chinese historians have been reluctant to focus attention on this painful chapter in China's modern history. Western historians, on the other hand, often seem disconcerted by the difficulty in finding meaning in the constant civil wars and complicated political maneuvers that characterized the warlord period. In the mid 1960s, the publication of two ground-breaking warlord biographies by James Sheridan and Donald Gillin attracted more attention to this subject from Western scholars. In the decade from 1970 to 1980, a number of important studies appeared in the West that took a closer look at various individual warlords, or warlord cliques, and their political struggles. In the late 1970s, historians in the People's Republic of China also began to show new scholarly interest in this topic. It was at this point, though, that Chinese and Western research parted company. In the 1980s, just as the field of warlord studies was gaining legitimacy among Chinese historians, Western scholarship on the topic went into a decline, from which it has yet to recover.
The lack of sustained Western scholarly attention to Chinese warlordism suggests that the historical relevance of this topic is still not fully appreciated. Most textbooks continue to give the warlords and their wars only a brief treatment before moving on to more detailed examinations of other, by implication more important, developments. Warlordism thus tends to appear as an aberration with little lasting impact on Chinese society or politics. The neglect of this subject in Chinese history is curious in contrast to the work of historians in other fields who have found war to be a catalytic force for far-reaching social, political, and economic changes. Some approaches used in past Western studies of Chinese warlordism may have inadvertently contributed to the perception of its comparative unimportance. Most studies have focused narrowly on the warlords themselves, through biographies of individual warlords, histories of warlord factions, or studies of warlord politics. This may have made it easier for
others to dismiss warlord studies as a self-contained and rather esoteric field. A shift to more topical and contextual approaches may encourage greater appreciation of the connections between warlordism and the broader transformation of modern Chinese society. This study hopes to make some small contribution toward meeting this challenge through a close examination of the emergence of modern Chinese warlordism.
Its identification as the subject of this study necessitates some initial consideration of how warlordism is to be defined. One question that immediately arises is whether to use this particular term at all. In Chinese, the word warlord (junfa ) has a particularly pejorative connotation. Historically, this has resulted in attempts to distinguish, often on the basis of political criteria, between bad "warlords" and good "military commanders." To avoid such value judgments, some Western scholars have substituted the term militarism for warlordism , thus identifying all the military strongmen of the Republican period by the supposedly more neutral label militarist . One argument against this usage is the considerable baggage of historical and theoretical meaning acquired by the term militarism since its initial appearance in the mid nineteenth century. For example, militarism is frequently used to refer to martial attitudes, ideologies, or policy orientations. Certainly modern Chinese society produced advocates of military strength and martial virtues who might be labeled militarists in this sense. However, not all militarists of this type were military men, and not all powerful Chinese military commanders were militarists in their ideological orientations. The term militarism therefore does little to clarify the particular features that characterized the military's role in early Republican society and politics.
One concept employed by social scientists that comes closer to describing the more restricted sense of militarism seen in the Chinese case is praetorianism . With the political role of the Praetorian Guard in ancient Rome as its referent, praetorianism has been defined as "a situation where the military class of a given society exercises independent political power within it by virtue of an actual or threatened use of military force." However, this term has also undergone a certain degree of elaboration by social scientists, even to the point of expanding its application to groups other than the military. Although perhaps remaining useful as an encompassing term, praetorianism is still too broadly defined to serve as a precise description of the specific form of military rule found in Republican China. Unlike the classic cases of military intervention usually referred to as examples of
praetorianism, the military in Republican China did not exert its political influence as a single class or as a corporate body. Rather, a multitude of relatively autonomous military commanders competed with each other for political power. One way to describe this situation might be to use a qualified term such as fragmented praetorianism or, as suggested by Donald Sutton, fragmented militarism . Definitional clarity, however, seems best served by a retention of the label warlordism for this distinctive fragmented form of military rule. The word warlord itself also conveys an image of the personal power of individual commanders that was another essential feature of the Chinese phenomenon. No abstract term describes the characteristics of the military power-holders of the early Republican period more clearly than the label applied to them by people who lived under their rule.
In reaching a conclusion on the terminology to be used, the above discussion also highlights the main points that should be included in a definition of warlordism . Adapting the definition of praetorianism , warlordism may be described as a situation where a number of individual military commanders exercise autonomous political power by virtue of the actual or threatened use of the military force under their personal control. This definition accents the importance of understanding not only why military intervention in Chinese politics occurred, but why it took the particular form it did.
The frequency of military coups in Africa, Asia, and Latin America after World War II spurred a lively scholarly debate among social scientists seeking to explain the reasons for military interventions in the political arena. While numerous models have been proposed, many scholars recognized early on that no monocausal model could explain the variety of experience represented by this phenomenon in different nations and at different times. Indeed, the main effect of empirical studies seeking to prove or disprove various models has been to accentuate the diversity of factors that have contributed to specific military interventions. One response to this situation might be to build increasingly complex typologies of military-civil relations to account for the variety of individual cases. The closer models derived from such typologies come to approximating particular cases, however, the more they become merely descriptive rather than explanatory. Historians, concerned more with the explanation of specific historical events or phenomena than with the construction of universal models, may be less dismayed by this situation than other social scientists. For a historian, the general literature on military interventions is useful, not in providing a definitive model into which to fit Chinese warlordism, but
in supplying different analytical approaches that may all contribute to a fuller understanding of this particular phenomenon.
Most attempts to explain why the military intervenes in politics can be subsumed under one of three general approaches. The first finds the cause of military interventions within the military itself. Proponents of this approach emphasize features of military organizations that contribute to their capacity for political intervention, such as technical expertise and training, cohesive and hierarchical organizations, or nationalist orientations. The second school of interpretation looks for the causes of military interventions not within the military but in the structural or societal conditions of the countries where they occurred. Samuel Huntington exemplifies this approach in his observation that "the most important causes of military intervention in politics are not military but political and reflect not the social and organizational characteristics of the military establishment but the political and institutional structure of society." Huntington theorizes that "praetorianism" occurs when the development of a society's political institutions fails to keep pace with expanding political mobilization. Other researchers have focused on similar cases of systemic disequilibrium—evidenced in economic dislocation, political disorder, or social tension—as providing the main context for military interventions. In this approach, then, the military is largely drawn into politics by the failure of civil government or civil institutions to deal effectively with these systemic problems. The third approach represents an attempt to bring idiosyncratic human motivation back into consideration. Samuel Decalo largely defines this approach in a study of military coups in Africa, where he argues that personal factors, particularly the self-interested motives of coup leaders, play a primary role in military interventions.
On a theoretical level, the above approaches are sometimes posed as mutually exclusive interpretations. Because they deal with distinctively different levels of analysis, though, these approaches in application often turn out to be more complementary than antagonistic. Indeed, the evidence supporting each of these approaches suggests that the motives of military commanders, the military's organizational characteristics, and general societal conditions all play some role in influencing military interventions. The exact mixture or comparative importance of these different factors depends on the specific conditions of particular cases. The goal of this study is not to advance any one of these general approaches but to provide a synthetic explanation for the rise of Chinese warlordism that is informed by them all. At the
same time, the categories of these three approaches provide a useful framework in which to analyze past explanations of the origins of Chinese warlordism and to suggest the directions this study should take.
At first glance, the pattern of personal power represented in warlordism seems to support the approach that finds the source of military interventions in the personal motives and ambitions of military men. Although this approach is rarely represented in more recent scholarly studies, contemporary observers in the warlord era often blamed the conditions of warlordism on the self-interested actions of the warlords themselves. A 1924 article tops a list of answers to the question, "Why are there so many civil wars?" with the simple statement, "Because there are warlords seeking personal profit." A simple consideration of the particular nature of Chinese warlordism, though, raises some problems with this approach. The importance of idiosyncratic factors seems most plausible in explaining the actions of a single charismatic military leader in the context of a specific coup situation. However, a general trend toward political intervention by numerous military commanders, such as seen in the case of warlordism, requires a broader explanation. Assuming that military officers in early twentieth-century China were not radically more ambitious or self-interested than their counterparts in earlier or later periods, personal factors do not explain why warlordism occurred at the particular period when it did. Furthermore, whereas the behavior of many warlords in specific events can be understood in terms of their personal motivations, it does not explain how they were able to bend armies to their own political wills in the first place. The conditions that gave rise to warlordism must therefore be sought within a context broader than the individual interests of the warlords themselves.
The next level of analysis, military organization, has provided the basis for several important interpretations of the rise of warlordism. One theory, mainly associated with Luo Ergang and Franz Michael, traces the roots of China's twentieth-century warlords back to the development of "personal armies" in the mid nineteenth century. In this view, the independence of these armies caused an increasing devolution of state power into the hands of regional military commanders, who eventually, after the collapse of the last imperial dynasty in 1911, became warlords. Thus, warlordism is traced back to general changes in Chinese military organization and to the specific structure of personal and regional military power that it created. A contrasting organizational theory can be found in Donald Sutton's case study of
the Yunnan provincial army, where he proposes to reveal the origins of warlordism "by studying organizational behavior within a single army." Unlike the previous theory, Sutton determines that the Yunnan army at the end of the Qing dynasty was a relatively modern, impersonal force. In his view, the relative strength of the Yunnan army as a cohesive military organization contributed to the establishment of regional "militarism" in Yunnan during the 1911 Revolution. Likewise, the disintegration of the Yunnan army's cohesion under the early Republic eventually gave rise to full-blown warlordism or "fragmented militarism."
The stark contrast between these two theories immediately suggests a weakness in purely organizational explanations of Chinese warlordism. Both theories fail to account fully for the organizational diversity of the late imperial and early Republican Chinese army. The first theory assumes a continuity in personalist military organization, which Sutton's study does much to disprove. At the same time, while proposing that the experience of the Yunnan army "could serve as a microcosm of an almost universal two-stage process," Sutton also acknowledges that "the Yunnan Army was not typical for its time." Indeed, he admits that "provincial armies differed greatly in their size and power, their military effectiveness, their internal structure . . ., and in their cohesion." The variety of forces that ultimately provided the power bases for different warlords suggests the difficulty in attempting to trace warlordism back to a single form of military organization. This situation also suggests that an explanation of the rise of warlordism must ultimately consider conditions extraneous to the military itself.
This finally brings us to the approach that seeks the sources of the military's political role in structural or societal conditions. One example of this approach is a formulation that has had a major influence on how historians in the People's Republic of China interpret the origins of warlordism. In a 1928 analysis of China's recent history, Mao Zedong linked warlord conflicts to China's "localized agricultural economy (instead of unified capitalist economy) and the imperialist policy of division and exploitation by marking off spheres of influence." Warlordism is thus seen as having its roots in the dual context of China's "semi-feudal economy" and foreign imperialism. Mao was not the first Chinese political thinker to suggest these connections, but his pronouncement had the effect of canonizing them for Chinese Communist historiography. Unfortunately, as a result of Mao's imprimatur, Chinese scholars hesitated to analyze this theory in any
detail, let alone consider alternate interpretations. Only recently have Chinese scholars begun to look beyond these broad generalizations and to acknowledge some of the problems in applying them. Even so, this interpretation has largely continued to define the parameters of inquiry into warlord origins among Chinese historians.
The suggestion of a strong linkage between warlordism and imperialism originally gained popularity in China during the anti-warlord, anti-imperialist movement of 1919 (usually referred to as the May Fourth Movement). In essence, this theory suggested that the imperialist powers replicated their rivalries in China by supporting specific warlords in their spheres of influence, thus keeping China weak and divided for easy exploitation. This view primarily appears in stock descriptions of warlordism by Chinese scholars, but some Westerners have also incorporated it into their discussions of the factors giving rise to warlordism. In practice, though, hard evidence of a close causal relationship between imperialism and warlordism has been fairly difficult to come by. For example, a detailed study by Odoric Wou has showed that one major warlord, Wu Peifu, usually failed in his efforts to gain political or material support from the foreign powers usually identified as his backers, Great Britain and the United States. In recent studies, some Chinese scholars have also acknowledged that most minor warlords, particularly those in remote locations, had minimal, if any, contact with foreign powers. They have also noted that even the larger warlords, rather than being simply beholden to the foreign power whose sphere they inhabited, had more complicated, and sometimes antagonistic, relations with various foreign powers. Such circumstances show that the policies of imperialist nations did not directly foster warlord rule in the sense implied, for example, in charges that the United States and the USSR instigated military coups in their client states in the Cold War era.
Although the assumed linkage between imperialism and warlordism in its original application seems insupportable, the influence of imperialism should still be recognized as part of the general political context in which warlordism arose. The threat of imperialism was after all a major catalyst in the nationalism that inspired Chinese politics in this period. Attention should also be paid to specific circumstances created by imperialism that may have influenced warlordism's course. Even so, the effects were often contradictory. For example, the foreign power consensus recognizing only Beijing as the seat of China's national government after the fall of the Qing dynasty favored the reunification of the country, not the political independence of multiple
warlords. On the other hand, the provision of foreign loans and the foreign guarantee of customs revenues to the Beijing government may have encouraged its proclivity to use military force to resolve political conflicts, thus strengthening the military's political role. In the end, such influences taken together are still insufficient to find imperialism to be a main causal agent in the rise of warlordism, except in the broad sense that the whole political development of modern China would have been different in the absence of foreign pressure. The most important conditions specifically shaping the emergence of Chinese warlordism were domestic, not foreign.
The other half of Mao's pronouncement does posit a general domestic condition, the existence of a semifeudal economy, as providing a foundation for warlordism. Although this part of Mao's statement is again more often quoted than analyzed, its main application stresses the manner in which China's low level of economic integration contributed to broader political disintegration. As such, this condition could be included among other "latent factors"—such as China's geographical expanse, poor communications, ethnic diversity, and particularistic loyalties—identified by other scholars as contributing to China's susceptibility to political fragmentation in the warlord period. The problem with all these factors, though, is that they were applicable for a much longer period than just the warlord era. What needs explanation is why the centripetal forces that generally kept China united were overcome in this particular period by centrifugal forces. Equally important is the need to explain why political fragmentation in this period took a military form. Broad generalizations about the nature of Chinese society or its economy do little to answer such specific questions.
An effective structural explanation of warlordism must show a more direct connection between particular societal conditions and the emergence of military men on the political scene. One explanation that comes closest to meeting this requirement postulates the importance of a political vacuum created by the 1911 overthrow of China's last imperial dynasty to the rise of warlordism. Scholars who express this view generally suggest that the lack of strong civil forces and institutions enabled the military to seize political power following the 1911 Revolution. Because this view presupposes the organizational strength of the military in making this move into politics, it is often found in combination with one of the theories of warlord origins that stress organizational developments inside the military. Thus, the military forces waiting in the wings in 1911 are either the personally oriented
regional armies described by Luo and Michael or the cohesive modern military organizations depicted by Sutton. In this case, then, structural and organizational theories neatly combine together to explain the emergence of warlordism.
Although many works cite some form of this political vacuum theory, the exact nature of the civil collapse that is assumed to have resulted from the 1911 Revolution has never been the subject of rigorous analysis. Indeed, the theory appears to make a questionable equation of civil government with the imperial system, and so sees their collapse in tandem. This seems to underestimate the adaptability of China's civil bureaucracy to changing political circumstances. More important, it fails to take into account the substantial expansion of civil politics that took place outside the imperial government before the 1911 Revolution and its potential to provide a new foundation for civil rule. Thus, it seems likely that a failure of civil politics, rather than a true civil political vacuum, would better describe the conditions that opened the way for warlordism.
This brief survey of the various approaches and theories that have been applied to the origins of warlordism shows that much work still needs to be done on this topic. Many of the generalizations that have been made about the rise of the warlords do not stand up to more careful historical analysis. Other theories have been flawed by their in-attention to the precise social and political conditions under which warlordism appeared. The goal of this study is to take a fresh look at the combination of factors that created warlordism in China, with a particular effort to stay close to the social and political "ground" on which this development occurred. This is done through a historical approach that relies primarily on a chronological narrative. Within this narrative, however, an attempt is made to focus on a number of overlapping conditions that ultimately transformed military commanders into politically dominant warlords.
The first chapter of this study begins with an examination of Chinese military organization in the late Qing period. Although the rise of warlordism does not seem tied to any one form of military organization, the organizational diversity and fragmentation of the Chinese army had a shaping influence on the particular fragmented form of military rule that eventually emerged. Ironically, while fragmented military organization originally served dynastic interests by weakening the military's potential for political intervention, once the military began to play a political role under the Republic, this fragmented organization contributed to the growth of political disunity.
The second important condition influencing the rise of warlordism deals with what S. E. Finer has called the military's "disposition" to intervene in politics. This issue is introduced in Chapter 2 through an examination of the politicization of military men in the late Qing period. This chapter shows that a change in the social composition of new modern-style forces raised in the Qing dynasty's last decade and a half helped to link the military to burgeoning reform and revolutionary movements. The resulting politicization of the military led to its first open political intervention—participation in the 1911 Revolution.
Although military men may have whetted their political appetites in 1911, the revolution itself did not lead immediately to the appearance of warlordism. Rather, as seen in Chapters 3 through 5, different civil political alternatives arose at provincial and central levels in the Republic's first years that were to some extent able to forestall the establishment of military rule. A failure to reach a consensus on these political alternatives, however, ultimately led to a final condition that would ensure the emergence of warlordism—the militarization of politics. The concluding chapters of this study show how politics was militarized as competing political authorities looked to military men to resolve their fundamental political conflicts. To borrow Finer's terminology again, the civil wars provoked by these political conflicts provided continuing "opportunities" for military interventions in the political arena. Under these conditions, the gun soon became the most important political determinant. Political reliance on military force continued the politicization of the military and in turn politically empowered military commanders. Military men were thus given many opportunities to bend political conflicts to serve their own political ambitions. Military commanders who took advantage of these opportunities thus became "warlords."
The historical approach of this study emphasizes the unfolding of warlordism as a process. For the most part, military rule in Republican China did not begin with abrupt military seizures of power or military coups directed against civil authority. Rather, it began as a process of political aggrandizement by various military commanders acting through posts or offices that were for the most part legally obtained. Equally important, the process of warlordism was uneven. Warlordism did not appear at a single moment throughout China, but at different times from one area to the next. Even within a single area, the path to warlordism might vary from one military commander to the next. The process depended on the particular circumstances, or "opportunities," in different areas, and the individual response, or
"disposition," of military commanders to circumstances presented to them. It is not surprising, then, that scholars sometimes differ over the exact dating of the beginning of the warlord era. Different dates are possible depending on what level of politics, what region of the country, or what individual military commander is being examined.
Recognition of the uneven development of warlordism created some problems in determining the proper context for this study. Simply given China's size, not to mention the multitude of warlords and the complexity of their experiences, any attempt at an all-inclusive national study would make it difficult to carry out a detailed empirical examination of the rise of warlordism. The solution to this problem was a narrower case study. Nonetheless, a case history of one warlord or of one warlord clique, the approach followed by most past warlord studies, would increase the difficulty of differentiating between the general process of warlordism and potentially unique experiences. The obvious solution to this problem was to focus on a particular region that would include more than one warlord or warlord clique, and that would allow for closer attention to the social and political context in which warlordism arose. Although provincial studies have become common in the China field, a study of one province carries some of the same problems of exceptionality as the study of one warlord. The compromise solution adopted here was a two-province study that would not be so broad in scope as to prevent detailed historical exposition, but broad enough to allow the inclusion of more diverse experiences as a basis for generalization.
The two provinces chosen for this study are the neighboring provinces of Hubei and Hunan in central China (see map). In area, Hunan is somewhat larger than Hubei (approximately 84,000 square miles to 73,000 square miles respectively). At the beginning of the twentieth century, each of these provinces was estimated to have a population of around 25 million people. The two provinces are geographically joined by the central Yangzi plain, which spreads westward from Wuhan in Hubei along the triangle formed by the Yangzi and Han rivers and then southward beyond Dongting Lake into north-central Hunan. The geographical linkage of the two provinces is represented in their names, which simply designate their relative positions north (Hubei) and south (Hunan) of the plain's largest body of water, Dongting Lake (five to six thousand square miles at its highest level). Enriched by alluvial soil and watered by numerous rivers and lakes, the central Yangzi plain shared by these two provinces is one of China's most fertile agricultural regions.
Hubei and Hunan have traditionally had strong economic and political ties. Located at the juncture of the Yangzi and Han rivers, Wuhan (actually composed of three cities: the Hubei provincial capital, Wuchang; the commercial port of Hankou; and the industrial center, Hanyang) was historically the trade center for the entire central Yangzi region. Hunan was particularly tied into this regional trading network by geographical factors. The waters of all four of Hunan's
main rivers (the Xiang, the Zi, the Yuan, and the Li) converge around Dongting Lake and then flow via the Xiang River into the Yangzi River at the Hubei-Hunan border upstream from Wuhan. Trade thus flowed naturally along these waterways and helped forge strong commercial links between Hunan and Hubei. Historically, the close connections between these two provinces were reflected in political administration. During the Ming dynasty (1366–1644), Hubei and Hunan actually formed one province, Huguang. The separation of the two provinces occurred during the Qing dynasty (1644–1911), but even then they remained joined under the supervision of the Huguang governor-general, whose capital was at Wuchang. The existence of close geographical, economic, and political ties between Hubei and Hunan helps give this study more cohesion than would have been possible with a more arbitrary selection of two provinces. Study of these two provinces also benefits from the excellent foundation provided by Joseph Esherick's work on the 1911 Revolution in Hunan and Hubei, as well as other previous case studies of these provinces.
Beyond the general case that might be made for the advantages of a combined study of Hubei and Hunan, my selection of these two provinces for a study of the emergence of warlordism was ultimately based on more specific methodological considerations. Provinces that have been the subject of past warlord studies have usually been located on China's periphery, heightening the problem of exceptionality. The choice of Hubei and Hunan was therefore a deliberate effort to redirect attention to central China in the hope of both increasing the validity of generalizations and aiding contrasts with previous provincial studies. These two provinces also provide instructive cases because they were frequently at the center of the successive political and military conflicts, beginning with the 1911 Revolution, that provided an important context for the rise of warlordism. In particular, the two provinces lay along a shifting boundary between centralizing efforts from Beijing and provincial attempts to defend local self-government that was at the root of many of these conflicts. The political tensions and divisions within this two-province arena in this era thus provide a greater comparative framework than might be assumed from the historical closeness of these two provinces. The comparative value of a study of these two provinces is also enriched by the diversity of military units, from central armies to local and provincial forces, found in Hunan and Hubei during this period. The frequency of military conflicts within these two provinces, linked to political struggles that often went beyond their borders, again played a role in the rise and fall of
various forces and the movement of different armies in and out of the region. Although more confusing than the cases of provinces where single warlords or warlord cliques maintained dominant positions for longer periods of time, the complexity of military forces in Hubei and Hunan presents a clearer picture of the dynamics that made military force the main determinant of political power.
Generalizations drawn from any case study, no matter how carefully chosen, can only be tentative. In the course of this study, it will become apparent that the experiences of Hunan and Hubei were in themselves unique in a number of respects. Neither province can be presented as totally representative of other provinces or of China as a whole. That caveat does not, however, reduce the value of this case study. Indeed, acknowledging the diversity found in these two provinces actually serves the purposes of this study by showing the manner in which the general process of warlordism unfolded in the context of varying local conditions.