The Culture of National Security: Norms and Identity in World Politics, by Peter J. Katzenstein, editor

Cultural Realism and Strategy in Maoist China

Alastair Iain Johnston

Where does realpolitik behavior come from? Classical realists might attribute states' preferences for unilateral, competitive, coercive strategies to human nature, greed, a hardwired desire to maximize power. Neorealists have attributed it to the uncertainty generated by anarchy, mediated by different distributions of material capabilities. Both explanations are unsatisfactory because both have difficulty accounting for the considerable volume and consequentiality of non-realpolitik behavior. Some scholars have recognized this and have conceded that non-realpolitik, "deviant" behavior is more likely a product of cooperative ideas or institutions, but at the same time they have reaffirmed that "nondeviant," realpolitik behavior remains best explained by the key causal variables in their respective versions of realism. To some extent, those with a constructivist and ideational theoretical bent have unwittingly reinforced the hegemony of realist theory in the realm of realpolitik behavior, precisely by focusing on non-realpolitik behavior. It is critical, then, that construc tivists and their fellow travelers take up the challenge and see how far ideational arguments can go in accounting for realpolitik behavior. ₁

This essay is an initial attempt to do just that through an analysis of the China case. Essentially I argue that China has historically exhibited a relatively consistent hard realpolitik or parabellum strategic culture that has persisted across different structural contexts into the Maoist period (and beyond). ₂     Chinese decision makers have internalized this strategic culture such that China's strategic behavior exhibits a preference for offensive uses of force, mediated by a keen sensitivity to relative capabilities. These preferences are often a reasonably accurate guide to strategic behavior. The persistence of an ideationally based hard realpolitik, however, suggests that structural accounts of realpolitik behavior are incomplete, precisely because this empirically observable cultural realpolitik has persisted across vastly different interstate systems, regime types, levels of technology, and types of threat. And it persists into the post-Mao period at a time when objectively and subjectively China's threat environment is the most benign in several decades. Since I have already looked at the presence and effects of this realpolitik strategic culture in traditional China, ₃     in this essay I examine Maoist China, with some reference to the ancient past, to determine the degree of continuity in Chinese realpolitik.

Mainstream realist theorists react to these claims by arguing that cultural realpolitik is epiphenomenal, a product, say, of the logic of anarchy. The realist retort, however, rests at different times on three problematic claims: (1) only deviant, non-realpolitik behavior may be ideational in origin, but, by implication, nondeviant behavior is not explained by ideational variables; (2) realist theory, whether classical or structural, makes determinate predictions about state strategies in the absence of some ideational interpretation of the meaning of material capabilities; (3) realpolitik ideology is epiphenomenal. An analysis of Chinese realpolitik, however, challenges these claims, or at least shows that they are too weak to dismiss cultural realism as a third, ideationally rooted, explanation for realpolitik behavior.

The essay begins with a quick summary of the content of traditional Chinese strategic culture and then moves on to a discussion of the conceptual and methodological issues involved in rigorously analyzing ideational sources of strategic choice. It then applies this discussion to the analysis of Maoist strategic culture and Chinese conflict management behavior in the post-1949 period. Since traditional Chinese and Maoist strategic cultures make predictions about behavior similar to those made by a determinate structural realpolitik model, the Chinese case raises critical questions about ideational versus structural explanations of strategic choice. In the last section I examine these questions, arguing that a structural realpolitik model can in fact be subsumed within an ideational realpolitik strategic culture, that the latter is not epiphenomenal. At the very least, structure cannot account for Chinese realpolitik.

Why China?

China poses interesting problems for the analysis of ideational influences on strategic choice. Many have assumed that the China case should turn up evidence for cross-cultural differences in strategic thought and practice. In other words, China could be a hard case for hypotheses derived from the Western strategic experience or a relatively easy case for culturally and historically contingent explanations. Practically all the scholars, Chinese or Western, who have studied Chinese strategic thought have argued that the Chinese have persistently exhibited what are essentially nonrealist predispositions. According to standard interpretations, from the core notions of this strategic thought we should expect that as a political actor becomes stronger in relative terms it becomes less, not more, coercive, seeking to induce potential adversaries to submit by magnanimously offering them legitimacy or material wealth. Those who have not studied the Chinese case, but who have tended to rely on this literature to make the claim that ideational and cultural sources of strategic choice do matter, consequently have tended to reinforce this perception of the "China difference." ₄     China, then, could be a crucial case both for those who would privilege anarchi cal structures and self-help behavior and for those who would privilege norms, cultures, ideas, and other shared, socially constructed, ideational influences on strategic behavior.

A closer look at the Chinese case, however, suggests a more complex picture. ₅     The short of it is, the Chinese strategic tradition does not embody only one set of clear strategic preferences. Rather, there are at least two different strategic cultures. One, derived from what can be called a Confucian-Mencian paradigm, places nonviolent, accommodationist grand strategies before violent defensive or offensive ones in a ranking of strategic choices. This preference ranking is associated in the core texts in Chinese strategy with language that reflects the Confucian-Mencian stress on "benevolent," "righteous," and "virtuous" government as a basis of security. This language casts military force as "inauspicious," to be used only under "unavoidable circumstances," and stresses the submission of the enemy without the resort to force. ₆

The other set of strategic preferences comes from what could be called a parabellum, or hard realpolitik, paradigm, which generally places offensive strategies before static defense and accommodationist strategies; at the violent end of the spectrum, the preference for offensive actions depends on the "softness" or the "hardness" of the realpolitik axioms in a particular text. This paradigm reflects a set of characterizations of the external environment as dangerous, adversaries as dispositionally threatening, and conflict as zero-sum, in which the application of violence is ultimately required to deal with threats. Moreover, this paradigm explicitly embodies a key decision axiom--the notion of quan bian--which stresses absolute flexibility and a conscious sensitivity to changing relative capabilities. The more this balance is favorable, the more advantageous it is to adopted offensive coercive strategies; the less favorable, the more advantageous it is to adopted defensive or accommodationist strategies to buy time until the balance shifts again.

These two sets of preferences, however, do not stand as two separate but equal strategic cultures. Rather, the Confucian-Mencian language represents an idealized discourse. None of the ancient texts on strategy that I examined (with one exception) devotes very much space to any detailed, concrete application of Confucian-Mencian concepts of security. Moreover, a number of the texts, along with some historical commentaries and annotations, implicitly or explicitly relegate these vague strategic axioms to indistinct golden ages of sage kings and legendary rulers, thus suggesting the historical and strategic irrelevance of these axioms. Finally, critical security concepts in the Confucian-Mencian paradigm are causally disconnected from, or only indirectly related to, the defeat of the adversary or the security of the state. Contrary, then, to most of the Chinese and Western literature on Chinese strategic thought, traditional China's operational strategic culture exhibited marked hard realpolitik tendencies.

The predominance of the hard realpolitik, parabellum strategic culture seems confirmed by the fact that much of the strategic practice in imperial China reflected these realpolitik preferences. My analysis of the Ming dynasty (1368-1644), for instance, indicates that realpolitik decision axioms showed up in the cognitive maps of key strategists during the Ming dynasty as they debated how to deal with the Mongol threat from north and west of China. Empirically, Chinese strategic choices tended to reflect this decision calculus to the extent that in periods of clear military advantage the Ming tended to act more aggressively toward Mongols than in periods when the relative capabilities shifted out of Ming favor. Indeed, Ming strategists tended to argue explicitly that static defensive and accommodationist strategies (e.g., peace treaties) were temporary fixes, rather than culturally preferred ways of dealing with external threats. This calculus is similar to that discovered by Paul Forage in his detailed historical analysis of the Northern Song dynasty's (a.d. 960-1127) strategy toward the Xi Xia "barbarians" along the northern border in the eleventh century. ₇

These conclusions are based only on an analysis of Chinese strategic culture before China's integration into the European/global state system in the twentieth century and before the importation into China of "Western" liberal democratic and Marxist-Leninist ideologies. Does this parabellum strategic culture persist across this transition? To telegraph the findings of this essay, Maoist strategic culture does indeed represent continuity with the past, reinforced by modern Chinese nationalist and Marxist-Leninist influences on strategic preferences. The evidence suggests, as well, that China's conflict management behavior after 1949 has been generally consistent with hard realpolitik strategic axioms. The fact that these axioms have persisted into the 1980s and 1990s when China has become increasingly integrated into international economic institutions and when its threat environment is the most benign since 1949 (a phenomenon that neither neoliberal nor neorealist approaches can comfortably account for) suggests that China's realpolitik behavior is ideationally rooted. As I will suggest at the end of the essay, this possibility raises complicated conceptual and methodological questions in explaining where realpolitik behavior comes from.

Some Conceptual and Methodological Issues

To date, many of those who have explicitly used the term strategic culture have tended to define it in ways that make it unfalsifiable and untestable. Especially egregious in this regard is what could be called the first (and most influential) generation of studies in strategic culture. ₈   Definitionally, this literature subsumed both thought and action within the concept of strategic culture, leaving the mechanically deterministic implication that strategic thought led consistently to one type of behavior. ₉   The literature also tended to include everything from technology to geography to ideology to past patterns of behavior in an amorphous concept of strategic culture, even though those variables could stand as separate, even conflicting, explanations for strategic choice. This left little conceptual space for nonstrategic culture explanations of behavior. As a result, the work took on a mechanistically deterministic hue and concluded that there were obvious and easy differences in the strategic cultures of different states. Methodologically, there was little explication of, let alone agreement on, the process of observing a strategic culture. The literature is unclear about the sources one should look to for representations of strategic culture, the analytical methods one should use to sort out deep structures in strategic thought from symbolic or instrumental elements, how strategic culture is transmitted through time, and how it affects behavior.

We need, then, to construct a more rigorous concept of strategic culture that specifies its scope and content, the objects of analysis, the historical periods from which these are drawn, and the methods for deriving strategic culture from these objects. Then it is necessary to explicate a research strategy that can credibly measure the effects of strategic culture on the process of making strategic choices. I have explored these issues elsewhere, ₁₀   but briefly put, the research strategy should involve three steps. The first is to come up with a definition of strategic culture that is falsifiable. The second is to test for the presence of strategic culture in the formative "texts" of a particular society's strategic traditions. The third is to test for the effect of strategic culture on behavior.

As for the first step, paraphrasing heavily from Clifford Geertz's definition of religion, ₁₁   I define strategic culture as an integrated system of symbols (i.e., causal axioms, languages, analogies, metaphors, etc.) that acts to establish pervasive and long-lasting strategic preferences by formulating concepts of the role and efficacy of military force in interstate political affairs, and by clothing these conceptions with such an aura of factuality that the strategic preferences seem uniquely realistic and efficacious. Specifically, strategic culture as a "system of symbols" comprises two parts. The first consists of basic assumptions about the orderliness of the strategic environment, that is, about the role of war in human affairs (whether it is aberrant or inevitable), about the nature of the adversary and the threat it poses (zero-sum or variable sum), and about the efficacy of the use of force (the ability to control outcomes and eliminate threats and the conditions under which it is useful to employ force). Together these make up the central paradigm of a strategic culture (figure 7.1). ₁₂

The second part of strategic culture consists of assumptions at a more operational level, about what strategic options are the most efficacious for dealing with the threat environment as defined by the central paradigm. It is at this second level that strategic culture begins to affect behavior directly. Thus the essential components or empirical referents of a strategic culture will appear in the form of a limited, ranked set of grand strategic preferences that are consistent across the objects of analysis and persistent across time. ₁₃   They are not, therefore, necessarily responsive to changes in noncultural variables such as technology, threat, or organization. At the high end of the three dimensions we should expect strategic preferences to reflect a hard realpolitik central paradigm, that is, to show a preference for offensive over defensive over accommodationist strategies. ₁₄   At the low end we should expect the opposite preference ranking, consistent with an idealpolitik central paradigm. ₁₅

I use ranked preferences instead a simple menu of strategic options because if different societies have different strategic cultures they ought to put different weights on these choices--that is, rank them differently. Ranked preferences allow for testing for consistency in strategic culture within systems and thus for differences between systems. ₁₆   This approach also provides a concept of strategic culture that is falsifiable. If preference rankings are not consistent across objects of analysis across time, then a single strategic culture can be said not to exist. Additionally, I use strategic preferences that are ranked because that approach should yield, ceteris paribus, explicit predictions about behavior, thus making it more possible to distinguish a strategic culture model of choice from other models.

The next step in this research design is to test for the presence of a strategic culture by testing for the congruence between the strategic pref erence rankings across the cultural objects of analysis (or "texts"). This requires first determining what are the "artifacts" in which one expects to find a culturally based set of ranked strategic preferences. In principle, there could be a large variety, including the writings and debates of strategists, military leaders, and "national security elites," or war plans, or even images of war and peace in various media. ₁₇

Fortunately, in the study of China one does not have to be too arbitrary in sampling these objects of analysis. In my work on traditional China, I used a set of texts--the Seven Military Classics--which together formed the core of Chinese strategic thought and military education from the eleventh century on, though some of the texts had existed individually since 500 b.c. ₁₈   In the study of Maoist strategic culture, Mao's own writings form the obvious sampling base. For this study I have chosen a handful of texts from different periods in Mao's life when he faced different strategic contexts. The assumption here is that if there is indeed a persistent Maoist strategic culture, the central paradigm and related strategic preferences should be consistent across these different texts, across time and different strategic contexts. ₁₉

Having chosen the objects of analysis, the next question is, How should one extract the central elements of a strategic culture, if indeed one exists? Here I rely primarily on a modified form of cognitive mapping supplemented by symbolic analysis. Cognitive mapping seems appropriate because the researcher is interested in what the texts appear to be telling a strategist about what to do, how to rank options, and thus how to make choices. Causal judgments are a key step as decision makers reason about how certain types of behavior will affect their environment in such a way as to secure basic foreign policy goals. ₂₀   Cognitive mapping is precisely a technique for uncovering causal linkages between certain behavioral axioms and their estimated behavioral effect. ₂₁

As for symbolic analysis, the literature on cultural analysis in anthropology and organizational studies suggests that symbols (e.g., analogies, metaphors, key words, idioms) are the vehicles through which cultural forms (e.g., shared rules, axioms, preferences) are manifested empirically, such that culture can be communicated, learned, or debated. ₂₂   From a symbolic perspective, then, strategic culture may be reflected by symbols about the role of force in human affairs, about the efficacy of certain strategies, and hence about what sorts of strategies are better than others.

The third stage in research involves analyzing the relationship between strategic culture and behavior. Here one has to trace the presence of a strategic culture from the objects of analysis up to the strategic assumptions of key decision makers in the historical period of interest. Since Mao was the key decision maker in post-1949 China until his death in 1976, I look at the possible influences and parallels between traditional strategic culture and Mao's strategic decision axioms.

The next substep, then, is to test for the effects of decision makers' preference rankings on politico-military behavior. Here the primary methodological issue is how to conceptualize the relationship between culture and behavior. In essence, the research problem at this stage is to control for the effects of culturally exogenous variables. This is not a clear-cut process. There are a number of ways of conceptualizing the relationships between strategic culture and other exogenous independent variables. ₂₃   The key issue is how to measure the effects of a constant or slow-to-change variable on outcomes that are supposed to vary (strategic choice). My preference is to treat strategic culture as a consistent set of ranked preferences that persists across time and across strategic contexts. Decision makers are sensitive to structural or exogenous conditions (i.e., relative capabilities) in a culturally unique way, such that interaction may (though it need not) yield unique predictions. This conceptualization allows one to consider strategic culture as a constant that, in interaction with a structural intervening variable, creates variation in the overall independent inputs into strategic choice. ₂₄   One can then test the influence of strategic culture against noncultural variables.

But this then raises the obvious question: What alternative models are "out there" against which a strategic culture model can be tested? At first glance, the most obvious competitive model would be a structural realpolitik model. One version of this model posits that states' decision makers generally share an undifferentiated interest in expanding the influence of the state. Given that interest, states will expand as long as their resources allow, since greater relative capabilities increase the probability of success of expansion. ₂₅   This structural realpolitik model would therefore predict that as a state consolidates and mobilizes resources it will adopt increasingly coercive strategies. The state will become more, not less, belligerent. ₂₆   As the relative power capabilities decline, the state turns to less offensively coercive, more static defensive strategies, and from there to more accommodating strategies as temporary fixes for disadvantageous conditions. One could pit this model against a strategic culture model as long as the latter made distinctive predictions about strategic choice, or as long as some form of critical experiment could be set up to test for additional sets of predictions if the initial sets were similar.

As I have suggested, however, China's hard realpolitik strategic culture does not make predictions that are unambiguously different from this more determinate version of structural realpolitik. Yet the ideational explanation should not be dropped simply on the grounds of parsimony; to do so in the absence of some sort of critical test would be to make an important theoretical choice for aesthetic reasons. Moreover, to do so would be to accept the logical fallacy that ideational models can explain only non-realpolitik "deviant" behavior and that "nondeviant" realpolitik behavior is not ideationally rooted. As I will argue, because of its inadequacy in accounting for the persistence of cultural realpolitik axioms, structural realpolitik should be abandoned as the "null" hypothesis. This choice does not mean abandoning competitive hypothesis testing, but it does suggest that the most logical alternative models to cultural realpolitik are themselves ideational in nature. More of this later.

The Maoist Central Paradigm

As I suggested above, strategic culture provides answers to the three broad interrelated questions in the central paradigm, and from this paradigm should flow logically connected strategic preferences about how to deal with threats to security. Traditional Chinese strategic thought tended to provide answers toward the hard parabellum end of these three dimensions. Operationally, therefore, the strategic preference ranking tended to place offensive strategies above defensive and defensive above accommodationist. Where does Maoist strategic thought fit along these three dimensions?

The Nature of Conflict

Mao's writings are nothing if not paeans to the constancy of conflict and struggle in human affairs. The starting point is Mao's theory of contradictions. ₂₇   While scholars debate whether Mao was simply inheriting traditional Chinese concepts of the dualism of existence or Hegelian dialectics through Marx, Engels, and Lenin, or was fusing the two, most agree that for Mao contradictions were the driving force of all natural and human activity. ₂₈   The resolution of contradictions within a thing was the funda mental source of its transformation and development. Conflict between contradictory elements drove nature and history. The resolution of one contradiction led to the creation of or superordination of another contradiction. Balance or the stability of equilibria in natural and social phenomena was relative, imbalance was absolute. For Mao, conflict didn't require a solution, it was the solution to political problems. Hence politics and international affairs were processes by which contradictions--whether between classes or states--were resolved through conflict, leading to the emergence of new contradictions.

In Mao's view, conflict in human affairs was not only inevitable but also desirable. Harmony was transitory and undesirable. He came to this conclusion well before he had accepted Marxist-Leninist normative arguments that class conflict was central to historical progress. In his late teens he--like many from his nationalistic generation--had come to believe that China's political weakness in the face of imperialism was not just analogous to but was also a product of the physical weakness of the Chinese people. His class notes, written when he was a student in Changsha, indicate that he accepted the view that people who were lazy, indolent, and weak could not progress; they would fall behind, decline, and die. Likewise, states that were lazy, indolent, and weak would fall behind other states, decline, and be exterminated by others. People and states needed to struggle; they required a spirit of vitality and vigor. ₂₉   Here one finds parallels to the social Darwinism arguments, made by nineteenth-century European nationalists, holding that extended periods of peace atrophy the physical capabilities and will of the state. ₃₀   But there are parallels as well in traditional Chinese strategic thought and practice. Ming dynasty strategists, for instance, also lamented that prolonged peace led to the decline in military preparedness and will that in turn encouraged Mongol enemies to exploit this weakness. ₃₁

This notion that conflict was ubiquitous and inevitable dovetailed with Mao's embrace of Marxism-Leninism in his late twenties. Class contradictions compelled history forward, and the resolution of class contradictions was an inherently conflictual process, aimed fundamentally at the elimination of the adversary. His analogy for politics seemed to summarize these axioms: "Politics is bloodless warfare. Warfare is bloody politics." ₃₂   

The Nature of the Enemy

In ancient Chinese strategic thought, the nature of the enemy was defined by the concept of righteous war (yi zhan). Generally the concept meant "sending forth armor and weapons in order to punish the unrighteous," namely, those who bullied weaker states, killed their own people, insulted other states, and otherwise rebelled against the established political and social order. In the face of unrighteous behavior, the violent destruction of the enemy was both necessary and desirable. ₃₃   Within this context the ends justified the means: once the ends of war were deemed righteous, then actions that in another context could be unrighteous (i.e., invasion and killing) were infused with moral intent. ₃₄   Since the adversary was a threat to the moral political order, the contest was explicitly zero-sum: the enemy could not be won over but had to be destroyed. ₃₅

Like the ancient Chinese, Mao also developed a concept of righteous or just war in which the enemy was defined as dispositionally apt to threaten one's own fundamental values as a class or state. For Mao, just wars were wars conducted by oppressed classes or nations. Unjust wars were those undertaken by oppressors, whether classes or nations. Hence, conflicts between oppressed and oppressor were zero-sum in nature. Under these circumstances, any and all strategies and tactics were acceptable. Well before he embraced Marx, Mao criticized Song Rang Gong of the Spring and Autumn period for his moral chivalry. Later, in a major work on strategy titled On Protracted War, Mao repeated this critique that in a just war strategy and tactics are questions of methods, not morality. ₃₆

Applied to politics, Mao's theory of contradictions complemented the zero-sum conceptualization of the enemy inherent in his vision of just war. Mao divided contradictions into two sets of categories: principal and secondary, antagonistic and nonantagonistic. He saw the principal contradiction in any particular phenomenon as playing the "leading and decisive role." ₃₇   The secondary contradiction played a subordinate role and did not immediately drive the development of a phenomenon. Once the principal contradiction was resolved, however, the secondary contradiction could turn into the new principal one. He developed the notion of antagonistic and nonantagonistic contradictions to resolve the problem of conflict within socialist parties and societies. If contradictions embodied all things, then even after socialist societies had abolished class society, contradictions had to exist. These contradictions, Mao argued, were nonantagonistic, in the sense that their resolution did not necessitate the violent eradication of the adversary required by antagonistic contradictions. Nonantagonistic contradictions existed "among the people," that is, within and between progressive classes and social groups remaining inside a socialist movement or state.

In combination, these types of contradictions created a two-by-two matrix (figure 7.2). Contradictions in the first cell were those in which conflict was inherently zero-sum and thus could be resolved only by eliminating the adversary. There could be no compromises with actors in this cell, except for very strict tactical purposes designed to weaken the enemy. Contradictions in the second cell were temporarily subordinate to the resolution of the principal contradiction. Actors in this cell could be temporary allies, with whom one could form a united front but against whom one also prepared for inevitable conflict. ₃₈   Actors in the third cell were those with whom the resolution of conflict required not their elimination but rather their transformation (e.g., through political education). Contradictions within this cell were primarily within one's own in-group, not between an in-group and an out-group. Actors in the fourth cell were the least threatening to one's security or well-being and constituted the most credible political allies.

The question is, How did an actor get put in the first cell in Maoist strategic thought? What was the threshold of threatening behavior beyond which the conflict was defined as antagonistic and principal? It is hard to answer this question with much precision. In effect, it asks, Where on the first two dimensions of the central paradigm does one put Mao in relation to other texts in China's strategic tradition? In the traditional texts on strategy, the threshold beyond which a conflict is defined as zero-sum varied somewhat from text to text.

As a Marxist-Leninist, Mao drew upon his class analysis to provide this threshold, and it appears to have been quite low, perhaps lower than that of the traditional texts. In principle, all oppressive classes belonged at some point in cell 1, depending on the immediate political problem (e.g., the overthrow of the nationalists or the consolidation of political control in the new China). As a Chinese nationalist, Mao could easily put all major powers (whether socialist or capitalist) in cell 1, depending again on the strategic problem at hand (e.g., facing American imperialism in the 1960s or Soviet social imperialism in the 1970s).

In sum, according to Mao's strategic thought, at some point in the development of contradictions between two adversaries this relationship would enter the realm of the principal antagonistic contradiction. At that point the nature of the conflict with the enemy became zero-sum, and negotiation, compromise, logrolling, and suasion were essentially ruled out. One could venture, then, that Maoist thought defined most class-based disputes and all threats to Chinese territorial and political integrity as inherently zero-sum conflicts.

The Role of Violence

Given the ubiquitousness of contradictions and conflict in human affairs and given also a zero-sum conceptualization of the adversary inherent in the process of resolving antagonistic contradictions, it is not surprising that Mao placed a great deal of stock in violence or in the employment of overwhelming force to eliminate adversaries. Thus superior force was a key ingredient in the forward march of history. This principle held between oppressed and oppressor states as well. For Mao, there was both a strong ideological and pragmatic acceptance that violence inhered in human social processes and that preparations for and the use of violence were essential for self-preservation and self-development. In Mao's view war was the highest form of struggle, the most efficacious means of resolving antagonistic contradictions among classes and states. ₃₉   Mao accepted Clausewitz's notion that war was a continuation of politics by other means, but he did not accept the potential restraints that this axiom placed on the scale or conduct of war. If the nature of politics was the struggle between the just and the unjust, then just war was a valid continuation of politics. And given that the unjust enemy was unlikely to be satisfied with the mere political resolution of conflict, then just war was not a last resort. ₄₀   Force did not come at the end of a process of political give-and-take that had stalemated, leaving no other choice but coercion. The use of force depended wholly on whether the adversary capitulated. Since class and national enemies were not predisposed to capitulate short of the use of force, violence was in all likelihood a necessary part of the process of dealing with threats. ₄₁

Mao came to this conclusion at an early age, though somewhat tentatively. In his notes from his school days in Changsha, for example, he recorded that those without strong power could not complete their tasks or be successful, whatever their endeavors. ₄₂   Despite this, as late as mid-1919 he was reluctant to endorse fully violent revolution in China, arguing in one essay that radical change that used "power politics" to overthrow power politics would only result in more of the same. ₄₃   The radicalization of Chinese intellectuals during the May Fourth movement in 1919, however, pushed Mao to embrace Marxism-Leninism in 1920 and 1921 and to embrace, intellectually at least, the role of violence in social transformation. In 1926, after a lengthy investigation of the radicalizing peasant movement in Hunan, he came to believe that armed uprising was the only way to resolve principal antagonistic contradictions. Subsequently he stressed that when carrying out a revolution one had to use "blade against blade, and rifle against rifle," because class enemies were not going to submit voluntarily. ₄₄

For Mao, then, the causal relationship between military power and security was straightforward. As he remarked in reference to a perceived threat that the U.S. might intervene in the Chinese civil war to defend the collapsing Guomintang (gmd) regime, "The stronger and more resolute the power of the people's revolution the lower the possibilities that the United States will directly interfere militarily." ₄₅   Mao was not referring here abstractly to the revolutionary enthusiasm, cohesion, or organization of Chinese people; he was referring to the military power of the Chinese Communist Party (ccp).

To get a more systematic handle on the role of military force in the achievement of security, it would be instructive to take a closer look at the cause-effect relationships and symbolic content in Mao's writings on strategy and statecraft.

Problems of Strategy in China's Revolutionary War (1936)

According to the cause-effect relationships in this text (figure 7.3), victory over the adversary depended on variety of inputs or "causes." Among these were the effective use of intelligence to spot the enemy's weak points (211.5.1) ₄₆   and then the use of military force to attack those points (e.g., 184.2.4; 201.1.1; 224.2.1); immediate retreat after achieving tactical victory, followed by a second attack (232.1.2); the concentration of superior military force (e.g., 199.1.4; 233.2.1; 238.1.1); and the annihilation, as opposed to the mere routing, of enemy forces (e.g., 202.2.2; 224.2.1; 248.3.2). Victory was also a function of a commander's abilities, in particular his understanding that war was the highest form of resolving contradictions (180.1.3), and his subjective ability to understand the objective limitations on the use of force (e.g., 188.1.2; 190.2.4; 232.1.3). Absolute flexibility was also critical for victory. A strategist must be able to act according to circumstances and demonstrate tactical flexibility (e.g., 184.1.2; 232.1.1). Other causes of victory included flexible borrowing from the military experiences of other states (180:5.2; 181.2.10) and the correct military and political leadership of the ccp (e.g., 192.1.2; 194.1.3).

Defensive operations were also sources of victory, but only as temporary stages in the offensive application of force. In one cause-effect relationship, Mao linked the loss of territory from retreat to victory over the enemy, but it is clear from the context that this strategy involved trading space for time and then using that time to create conditions advantageous for the attack by, say, tiring the enemy or forcing it to make mistakes that then could be exploited (e.g., 217.3.9; 221.1.2; 232.1.2; 234.4.3). Indeed, defensive operations were the first operational step in defeating the enemy. A defensive retreat in the face of an enemy offensive, or the use of positional defense of key points alongside mobile warfare, led to the defeat of enemies in a civil war (207.4.4; 217.3.8) and more generally to military utility (e.g., 199.4.2; 215.3.2; 242.2.1). Defense also had instrumental political value. It lulled the enemy politically (207.4.5), and in just wars it could rally conservative or "backward" elements in society around one's cause (207.4.6). But a defensive retreat was only a prelude to the shift to the strategic offensive. Passive defense or pure defense reduced the scope for initiative (234.4.1). "Active defense"--defense for the purposes of counterattacking and taking the offense--had positive utility and was a source of victory (e.g., 207.3.1; 233.1.6; 234.2.1).

As for the ability to attack the enemy with overwhelming superiority, this was rooted in the fighting power of the military, which in turn was dependent on popular support, advantageous terrain and weather, military organization, and the confidence and experience of cadres and commanders (e.g., 206.3.1; 206.5.1; 212.2.2; 222.4.3; 223.2.4). Favorable changes in relative power, then, led to an increasing ability to concentrate superior military forces for offensive operations (e.g., 199.1.6; 202.4.1; 208.4.1; 215.5.1; 248.4.1). In other words, the relationship between the human element (e.g., morale, popular political support, etc.) and military victory was indirect, mediated by the ability to apply superior armed force against a weakening enemy.

A couple of points are worth mentioning here. First, it is clear that this text does not advocate passive defense (e.g., retreat or static defense) as a permanent, effective means of dealing with an external threat. Rather, it develops the notion of active defense, whereby in an initial period of strategic weakness one relies on retreats, hoping to "lure the enemy in deep" or wear it down through small counteroffensives within the context of strategic defense. Once the relative balance of power--here an amalgam of human and material variables--begins to shift in one's favor, one should go on the attack. The end point of this attack is the strategic annihilation of the enemy's ability to wage war. The entire process--from strategic defense to strategic offense--is called active defense. This terminology is dictated by an instrumental need to frame one's own actions as entirely defensive and just, a position that is important for winning popular support and sympathy. ₄₇   The term active defense is probably important at a deeper symbolic level as well: the concept of just war requires that the enemy's threatening actions be defined as dispositional rather than situational, hence morally beyond the pale. Thus the enemy's nature is to threaten; one's own is to defend.

Second, the text stresses that strategists must exercise patience and absolute flexibility at the tactical and strategic levels, what Mao called the "flexible application of principles according to circumstances" (an zhao qing Kuang huo yong yuan ze). ₄₈   The shift from the strategic defensive to the strategic offensive should come only when conditions are ripe--namely, when the relative balance of composite capabilities shifts in one's favor. A strategist has varying degrees of control over when and how this shift occurs. Part of the process involves actively wearing the enemy down militarily. It also involves "conserving one's own strength," avoiding decisive engagements, building one's own military capabilities, and seeking out political and military allies. In both cases, however, the key is to remain flexible enough to exploit opportunity when shifting from the strategic defensive to the strategic offensive. This flexibility axiom parallels the notion of quan bian found in ancient texts on strategy and statecraft. Absolute flexibility was at the core of the strategic advice in traditional China. As a Warring States text, the Si Ma Fa summarized it nicely: "As for war, it is [a question of] expedient assessment" (zhan zhe quan bian ye). ₄₉   A Ming dynasty (1368-1644) text put it best: "As for the way of employing the military instrument there should be no constant form in either attacking or defending; there should be no constant rules for either dispersing or uniting forces; there should be no constant time period when one is in motion or at rest; there should be no constant directional momentum, either when extending or retracting [one's forces]." ₅₀

Mao was quite at home with this type of thinking. One could not be bound by one set of methods (e.g., strategic defensive) when the annihilation of the enemy required the adoption of another set (e.g., strategic offensive). And the strategic offensive required superior capabilities. It is not entirely accurate, then, to characterize Mao's strategic thought as only stressing "using weakness to overcome strength" (yi ruo sheng qiang) by dint of superior morale or political mobilization. ₅₁   Mao outlined a process--active defense--by which an initially weak military entity could acquire the human and material resources for eventually defeating an initially stronger military entity. But the defeat, or annihilation, of that enemy entity at the tactical and finally the strategic level required military superiority. In other words, a humanly richer but materially impoverished military force could not defeat a materially superior enemy. At the moment of defeat the "just" side would have to have superior human and material capabilities. I will come back to this point shortly in the discussion of Mao's Moscow speech in 1957.

On the New Stage (1938)

This short essay (figure 7.4) on the creation of an anti-Japanese united front with the gmd is essentially a discussion of how to handle secondary antagonistic contradictions, the second cell in figure 7.2. Mao readily moved the conflict with the gmd from cell 1 to cell 2, as Japanese imperialism became the primary threat to Chinese national development in the 1930s. The causal arguments in this essay were straightforward: Japanese imperialism threatened the survival of all classes in the Chinese nation. Hence the security of the state was an "effect" of the defeat of the enemy's "savage and protracted" warfare (191.1.3). Cooperation with the gmd was causally connected to the defeat of Japanese imperialism (e.g., 179.2.7), as was the application of violence ("the barrel of the gun") (190.2.1), though from context it seems clear that the former runs through the latter to lead to victory. ₅₂   Indeed, there was a feedback relationship between the united front and the Japanese invasion, such that the latter compelled the former to develop, while the former would lead to the defeat of the latter.

The development of the united front, in turn, required that the CCP make political concessions to the gmd, such as refraining from expanding ccp-controlled territory and from organizing secret cells within the gmd (e.g., 185.1.3; 187.3.1; 188.3.1). United front activities were not to be limited to the gmd, however. Mao also argued that alliances with other states opposed to Japanese imperialism were, in a general sense, of great utility (191.1.1; 191.1.2). Concessions with the secondary contradiction were, of course, permissible. Concessions in the face of the primary contradiction, however, were politically and militarily disastrous and would assist the enemy's aggression (e.g., 193.1.1; 193.2.9). Mao used the initial Allied reaction to Hitler as an analogy. The unwillingness of the democracies to sanction aggressor states, and their endorsement of the policy of appeasement, were equivalent to assisting aggression.

Despite the focus on the united front, in this essay the defeat of the principal adversary also ultimately required the use of force. The united front within China, and between China and other anti-fascist states, would create superior capabilities to apply in the contest with fascism. Concessions toward or bargaining with the principal enemy were anathema and would only encourage aggression. These cause-effect relationships do not differ much from those set forth in the essay on strategy in China's revolutionary war. Rather this text focuses on one part of the process of creating superior power capabilities to deal with a zero-sum conflict.

The Present Situation and Our Tasks (1947)

This text (figure 7.5) outlines the strategy for the final overthrow of the gmd and the establishment of a new state. As in the other texts, here also victory over the adversary is the result of varied causes, but they boil down to two: political mobilization and the offensive use of superior military power. As for the former, popular political support and rigorous political work inside and outside the People's Liberation Army (pla) are both "causes" of victory over the gmd (159.1.4; 161.2.6). While the text links these two factors directly to victory in China's revolutionary war, comparatively little time is spent in the text expanding on this causal relationship. It is not unreasonable to view it as in fact indirect, with political mobilization being an input into the ability of the pla to defeat the adversary militarily.

By far the most attention is paid to the military process of defeating the adversary. Here victory is a function of putting the gmd on the defensive militarily through the application of people's war (157.1.6; 161.2.3) and through being absolutely superior in every specific campaign even though strategically inferior (160.1.15). In other words, relative capabilities matter. Advantages in relative capabilities allow the pla to attack, to shift from the defensive to the offensive against the enemy, attacking dispersed, isolated enemy forces first, crushing the enemy with a combination of frontal and flank attacks, and then attacking concentrated and strong forces second (e.g., 160.1.2-13). The causal paths lead directly from these actions to the defeat of the enemy, as well as to the elimination of imperialist influence in China (157.1.5; 160.1.14). The failure of the ccp/pla to oppose the gmd with military means leads directly to the forfeiting of China's future (158.1.1; 158.1.2).

In addition to being direct causes of the enemy's defeat, offensive operations lead to the diminution of the enemy's ability to fight. Specifically, offensive operations are causally linked to the destruction of the enemy's plans to take the war into ccp territory (157.1.2), pushing the enemy onto the defensive (157.1.1; 157.1.3), liberating territory (159.2.1), and improving the pla's relative ability to operate successfully against other enemy forces (160.1.12). Strategic flexibility--namely, the ability to abandon strategically disadvantageous points--fuels the enemy's underestimation of pla strength (161.2.6), which in turn leads to the defeat of the enemy (162.1.1).

Mao's Moscow Conference Speech (1957)

This speech (figure 7.6) to the Moscow Conference of Communist Parties in November 1957 was not a work on strategy proper. Rather it was a broader discussion of statecraft and the role of power capabilities in general in the achievement of strategic goals. In particular, Mao focused on the sources of the socialist bloc's strength, arguing that it was rooted in three things. One was the foreign policy crises facing imperialism--such as the Suez crisis, setbacks for U.S. influence in Africa and Asia, China's victory in the Korean War--all of which were indicative of the weakening strategic influence of Western imperialism. These crises were a direct cause of the East Wind's (socialism's) prevailing over the West Wind (imperialism) (e.g., 117.3.1;119.3.1).

A second set of causes for the socialist bloc's strength had to do with industrial and military capabilities. Here Mao argued that the bloc's emerging technological superiority (e.g., the Soviet Union's Sputnik launch, increases in steel production) would lead to the East Wind's prevailing over the West Wind (117.6.1; 118.1.2; 118.1.3; 118.2.5). This superiority was a cause of the West's political-psychological weakening (116.2.1) and contributed to socialism's global surge. In other words, socialism's material superiority, translated into political and military strength, would lead to socialism's prevailing over imperialism (119.6.2). ₅₃   Finally, Mao hinted that the military destruction of imperialism was also causally related to socialism's victory. Even in nuclear war, he argued, an atomic attack by imperialism would lead to retaliation by socialism (118.4.1; 118.4.2), which in turn would lead to the destruction of imperialism and the triumph of socialism (119.1.1; 119.1.4; 119.1.5). Short of nuclear war launched by "maniacs" in the imperialist camp, the invincibility of socialism would deter imperialism from fighting and would lead to everlasting peace (118.3.2; 118.3.3; 119.1.7).

As for China, Mao was also quite clear about the basis of its security. One factor was increased steel production (118.2.3; 118.2.6), which in turn was a product of Soviet assistance and the Chinese people's willingness to exert themselves for this goal (118.2.1). Another was its military prowess, as demonstrated in the defeat of the gmd (116.4.1; 116.4.2), the victory in the Korean War (117.2.1), and the American unwillingness to send forces to aid the French in Vietnam. As for the cause of military victory, the strategy was a piecemeal one, to destroy the enemy one by one (120.2.8). Offensive military actions were also causally related to compelling the enemy to compromise or capitulate (117.2.4). In other words, a military hard line against the Americans and/or their allies appeared to have desirable political effects.

The speech also provided evidence of Mao's consistent zero-sum conceptualization of adversaries. For him, the maintenance of an uncompromising strategic objective while flexibly blending compromises and struggle at the tactical level is a correct strategy when dealing with a zero-sum conflict (e.g., 122.3.1; 122.2.5). When two sides are unable to accommodate each other, the struggle or contradiction is inherently antagonistic, and those conditions must lead to the overthrow of the adversary. Socialism and imperialism exist absolutely exclusively of each other, and resolu tion of this contradiction leads invariably to the collapse of imperialism (122.2.2; 122.2.4).

A couple of observations about the patterns in these cognitive maps are in order here. First, in all the maps the management of the security problematique requires most directly the military defeat of the adversary. In one of the texts--Mao's Moscow speech--there is an alternative route, namely, socialism's technological and material (hence military) superiority, but it is not clear why that alone might lead to the destruction of imperialism. Mao hints that this superiority will deter imperialism from attacking socialism, giving time for the internal contradictions within the imperialist camp and its spheres of influence to cause its implosion. In any case, superior military power is clearly a critical causal element in the defeat of an adversary and the achievement of key political goals. When circumstances require the application of this superior military power, offensive uses of force are causally linked to military success. Defensive operations are only a temporary and early stage in the application of violence, and by themselves they are inadequate for achieving desirable political ends. The so-called human element is only one input into this process of applying superior military force offensively. Relative material power counts, and in fact determines military success. ₅₄   Most of the texts also highlight a concept of absolute strategic flexibility that along with the concept of just war lifts any a priori moral or political boundaries on the means by which the enemy is defeated.

A second general observation is that none of these causal relationships is inconsistent with the cognitive maps in traditional Chinese texts on strategy. Like these ancient texts, Mao's texts are essentially deeply structured around a parabellum or hard realpolitik central paradigm. The key question is, then, to what extent was Mao directly socialized in this traditional strategic culture? Unfortunately, on this point the evidence is rather spotty. The issue is also complicated by the fact that by the 1920s Mao had been exposed to Leninist and Clausewitzian ideas about statecraft and strategy. A case can be made that these traditions also embodied a parabellum central paradigm, hence the difficulty in separating out the effects of traditional Chinese influences.

We do have evidence, however, that Mao had read at least some of ancient texts on strategy as early as 1913, well before he had any contact with Clausewitz's or Lenin's works. ₅₅   One of his school notes from 1913 seems to paraphrase a key passage in the Si Ma Fa to the effect that it is legitimate to "kill people in order to give life to people" (sha ren yi sheng ren)--that is, destroy an enemy in order to achieve a greater good. ₅₆   This position is consistent with the absolute flexibility axiom at the core of Chinese conceptions of just war, and it embodies an obvious instrumentalism in conceptualizing the role of force. Later on, in his more mature works on strategy, Mao echoes again this axiom: "Whoever wants to seize state power and intends to preserve it must have a strong military. . . . We are for the abolition of war, we do not want war. But only through war can we abolish war, and if we want to get rid of the gun we must take up the gun." ₅₇

The occasional references to Sun Zi in Mao's later works all focus on the notion of absolute flexibility. Two of the four references in "Problems of Strategy in China's Revolutionary War" and in "On Protracted War" are to Sun Zi's axiom "Know the enemy and know yourself and in one hundred battles you will not be in danger," one refers to avoiding the enemy when it is stronger and attacking it when it weakens, and one refers to deception, or displaying a false "form" to the enemy. Arguably, all four references relate to gauging the nature of changing circumstances and exploiting these strategic opportunities.

As for other influences from works on traditional Chinese military history, we also know Mao was an avid reader of popular histories and novels set in the Warring States (475-221 b.c.) and Three Kingdoms period (220 b.c.-a.d. 280). These stories tended to stress the righteous use of force, often by militarily inferior groups. Operationally many of the accounts stressed ingenious political and military stratagems and hence were instrumental in creating the myth that Chinese strategic tradition stressed minimally violent solutions to security problems--but these were usually preludes to the application of offensive violence.

What Mao seems to have most clearly rejected was the minimal violence notion in Sun Zi that one could "not fight and subdue the enemy." After he emerged as the chief strategist for the ccp in the 1930s, Mao was apparently tutored on traditional Chinese strategic thought by Guo Huaruo, who until the mid-1980s was the ccp's most authoritative interpreter and annotator of Sun Zi's Art of War. Guo stressed that from a Marxist-Leninist perspective the notion of "not fighting and subduing the enemy"--the core of the conventional interpretation of Sun Zi--was un-Marxist, since class enemies could not be credibly defeated without the application of violence. ₅₈   This axiom was dismissed as "idealist" (wei xin zhu yi). In this instance, traditional Chinese texts were mediated by Marxist-Leninist arguments, but it is possible that Mao was receptive to this interpretation, given his socialization in the popular novels and histories.

In sum, while the evidence is scattered, it does seem that Mao was exposed to elements of the parabellum tradition in Chinese strategy thought before he was introduced to Clausewitzian and Leninist variations of parabellum. The net result of Mao's socialization in these three traditions was, arguably, a mutual reinforcement of the hard realpolitik tendencies in his strategic thought.

Strategic Preference Rankings

Since the Maoist and traditional Chinese texts share the parabellum central paradigm, we should also expect them to share the grand strategic preference rankings derived from that paradigm. That is, the Mao texts should embody a preference for offensive operations over static defense, and static defense should be preferred to accommodationist strategies. ₅₉   It is clear from the cognitive maps that Mao believed that final victory over an adversary in a zero-sum conflict required the offensive application of superior military force to annihilate rather than merely deter the enemy. The process of getting to this point, however, is a little complicated and obscured by the political language that Mao used to clothe his offensive preferences.

Starting from the least-preferred strategy, Mao's concept of principal antagonistic contradictions, and his notion of just war, ruled out the possibility of long-term cooperation or accommodation with an enemy. The goal of political struggles was to "preserve oneself and annihilate the enemy." It was unlikely, in Mao's view, that the other side in a "principal antagonistic contradiction" would willingly submit, or bargain away its existence.

As for static defensive strategies, Mao labeled these passive defense (xiao ji fang yu) or pure defense (dan chun fang yu), to contrast them with his preferred strategy of "active defense." Passive defense, he argued, involved methods essentially for blocking or obstructing an enemy who was on the offensive. These strategies were of only limited value for holding territory temporarily while other forces engaged in offensive operations within the context of the strategic defense. Passive defense put one in a reactive position and tended to force the dispersal and weakening of one's military power. ₆₀

Mobile defense--the tactical offense within strategic defense--was designed to create the conditions ripe for a shift to the strategic offensive.

Mao argued, then, that superior military forces applied offensively were decisive in defeating an adversary. In one specific discussion of the final push to defeat the gmd in the late 1940s, he wrote that ideally the annihilation of the enemy required accumulating 3:1 or 4:1 numerical superiority. ₆₁   Thus the shift to the offensive depended on relative capabilities: when these were advantageous, the just side should apply offensive violence to annihilate the enemy's war-making ability. Any other strategy, then, whether defensive or accommodationist, was contingent and should be adopted only when relative capabilities could not guarantee the successful offensive use of force. ₆₂

This preference for the offense was qualified, however, in two ways. Politically Mao stressed that offensive uses of force, whether at the level of operational strategy or grand strategy, should be named active defense. While just wars were not defensive ones, in the sense that the just side could declare war or initiate violence, they should nonetheless be called defensive. For one thing the term active defense was more politically palatable; it could be used in arousing righteous indignation among masses and soldiers or to attract sympathetic support from external sources. ₆₃   Just wars were also a legitimate response to what we would now call "structural violence." The source of conflict was the existence of an oppressor class or an oppressor nation(s). The oppressor was acting offensively, in that ultimately if it were not for the existence of this oppressor there would be no fundamental contradiction requiring the use of force for its resolution. ₆₄

Operationally, the offensive use of force was qualified by a preference against a first strike out of the blue. Instead Mao evidently preferred an offensive "second strike" (hou fa zhi ren). ₆₅   Again, the reasoning was both political and military. To strike the enemy, particularly its territory, first, without specific provocation would be to give it the sympathy of world opinion and would tar the just side with the politically damaging label of aggressor. ₆₆

Militarily, Mao's version of a second strike offensive was designed to compel the enemy to move first, thus allowing an opportunity to gauge its intentions and capabilities. One could thereby ascertain the enemy's weak points, and attacks on these points, not first strikes per se, were decisive in conflict. ₆₇   There are, in fact, parallels here to the strategic calculus in the traditional Chinese texts. The Sun Zi text speaks of "first putting oneself in an undefeatable position" and waiting for the enemy to put itself in a defeatable position. In other words, whether at the grand strategic or the operational level of strategy, defense and offense were linked. One waited to see how the enemy moved, while one's own strategic posture of apparent immobility and obscurity concealed one's capabilities and intentions. Once the enemy revealed its disposition, its weaknesses and strengths, then one shifted to the offensive, striking at the enemy's "empty" (xu) points and disarming it. One could use limited amounts of force to provoke the adversary into a definitive move. Defensive strategies established the parameters within which the enemy had to operate and thus allowed oneself to retain the initiative.

If the above characterization of the central paradigm and related strategic preferences in Mao's thought is accurate, then what sorts of expectations should we have about Chinese strategic behavior under Mao? First, we would expect the use of force to be framed politically as defensive and just, whether or not China initiated violence. This is not an unusual expectation; most states develop some symbolic language to frame behavior in culturally acceptable terms. We would also expect the Chinese to initiate force after establishing that conflict is imminent, given the disposition of the enemy and the zero-sum nature of the conflict. In other words, initiation would come after Chinese decision makers have concluded that force is "unavoidable." There ought to be a low threshold establishing when an adversary's moves indicate that conflict is inevitable. That is, once it has been determined that conflict is zero-sum, we should expect the Chinese leadership to initiate larger-scale conflicts, and given the premium placed on military initiative, we should also expect China to resort to force in these types of conflicts. In other words, there should be an observable tendency to link more isolated threats to a broader challenge to the fundamental values of the state. To the extent that there are gradations in the severity of this threat to the fundamental values of the state, severe threats will require the military annihilation of the enemy's capacity to continue to challenge China. We should expect too that in a general conflict situation, Chinese coercive behavior should be positively related to advantageous changes in relative military capabilities. In other words, once in a conflict situation, advantageous shifts in relative capabilities should be a necessary cause of the initiation of force.

Chinese Conflict Behavior

How well does Chinese conflict behavior after 1949 fit the expectations that follow from the Maoist variant of the parabellum strategic culture?

The first observation is that, of all the major powers, the prc has been quite prone to use force in foreign policy crises. According to the data set on foreign policy crises generated by Jonathan Brecher, Michael Wilkenfeld, and Sheila Rosen, the prc has been involved in eleven foreign policy crises through 1985 and has resorted to violence in eight (72 percent), proportionally more than the other major powers in the twentieth century. Comparable figures for the U.S., the ussr, and the uk from 1927 to 1985 are 18 percent, 27 percent, and 12 percent, respectively. And according to these researchers, the Chinese use of violence has been what they label "high intensity," involving "serious clashes" or "full-scale war." ₆₈

Second, these conflicts were all located along China's borders. Territorial disputes were thus crucial drivers in most of these crises. According to the Brecher and Wilkenfeld data set, in comparison with other major powers China was more likely to use violence in a dispute over territorial issues, employing it as a key conflict management technique in 80 percent of such crises (N = 5). For the U.S., the figure was 0 percent (N = 11); the ussr, 20 percent (N = 10); Britain, 8.3 percent (N = 12); France, 27.3 percent (N = 11); India, 33.3 percent (N = 3); and for all actors in the data set, 23.5 percent (N = 281). Chinese decision makers tended to see territorial disputes as high-value conflicts, partly because of a historical sensitivity to threats to the territorial integrity of the state. ₆₉

The definition of high-value, zero-sum conflicts was not limited to territorial issues, however. There seems to have been a tendency for Chinese leaders to define even political/diplomatic crises as high threat and to view force as a legitimate response. China used violence as the major response in 66.7 percent of the crises involving political/diplomatic issues (N = 3); the comparable statistic for all actors is 18.1 percent (N = 127). This would suggest that Chinese leaders were more apt to view a wide range of disputes in zero-sum terms, thus establishing a low threshold for deter mining what conflicts constituted a clear threat to the security of the state. ₇₀   Obviously these comparisons need to be handled with care. Given the small N of Chinese cases, a relatively small number of new cases in which Chinese behavior deviated from past trends could very quickly change the percentages. The data are, nonetheless, suggestive, particularly since the behavior is consistent with the hard realpolitik ideational structure behind Maoist understandings of conflict. ₇₁

The sparse anecdotal evidence we have of the Chinese decision-making process is consistent with the aggregate data. In most cases when China used force, the "threat" was considered high and the issue at hand tended to be a zero-sum one. In the Korean War case, for instance, before resorting to force Mao argued that not only did American military actions in the peninsula threaten China's industrial base in Manchuria but also domestic counterrevolutionaries could take advantage of an extended war with the U.S. to undermine the CCP's tenuous political control of the mainland. ₇₂   In the first Quemoy-Matsu crisis of 1954-1955, Mao viewed the imminent conclusion of the U.S.-Taiwan Mutual Defense Treaty as a two-pronged threat: one prong was the security threat posed by a formalized American military presence in a number of bilateral security arrangements around China's periphery; the other was the possibility that formal bilateral security relations between Taiwan and the U.S. might encourage the U.S. and other states to make a de facto endorsement of the concept of two Chinas, an outcome that Mao considered a threat to Chinese territorial integrity. ₇₃   The Sino-Indian border crisis of 1962 was a "conventional" zero-sum territorial dispute, but it came at a time of severe economic dislocation, in the wake of the Great Leap Forward and shortly after another "invasion" scare from Taiwan. Thus the urgency of preserving China's territorial integrity was accentuated by the sense of domestic economic and political crisis in China. There is some controversy about who initiated the Ussuri clashes with the ussr in 1969 and even less certainty about the decision process on the Chinese side. But one plausible argument is that in the face of a very real concern that China might be the next on the Brezhnev Doctrine list, Mao initiated conflict to signal to the Soviets that China was a risk-acceptant player and that any subsequent Soviet use of force would carry high costs for Moscow as well. To the extent that the Chinese had the Czechoslovakian case in mind, the issues at stake in this conflict--China's territorial integrity and political survival--were high-value ones as well.

Third, the Chinese have also been quite willing to initiate violence in disputes. In the Brecher and Wilkenfeld data, China initiated violence in 62.5 percent of those crises in which it ended up using violence (N = 8). To be sure, coding initiators in conflicts is an exceedingly difficult task, since it is sometimes hard to tell when a crisis or dispute began and which side was defending the status quo. Nonetheless, the Brecher and Wilkenfeld data are consistent with a rough estimate of Chinese initiation based on a preliminary version of the Correlates of War militarized dispute data. These data suggest that of the seventeen cases of militarized disputes in which the prc has been involved with the ussr through 1985, China probably initiated the dispute in eleven cases (65 percent). In seven of these eleven cases, the issue for China was either a perceived threat to Chinese territorial claims or an attempt by Beijing to modify territorial boundaries, a finding that is again consistent with the Brecher and Wilkenfeld data. ₇₄

Fourth, Chinese conflict behavior has tended to be sensitive to chang ing relative capabilities in ways that are consistent with a hard realpolitik strategic calculus. In other words, as in the Ming period, once in a conflictual situation, there seems to have been a correlation between an advantageous shift in relative capabilities and the PRC's initiation of hostile actions along a scale of violence. The question is, How did the Chinese or Mao determine relative capabilities? On the one hand, Mao often incorporated quite subjective measures about who was on the "defensive" and who was on the "offensive" in his assessment of the relative strengths of the U.S., the ussr, and China. On the other hand, Mao was also quite sensitive to relative material capabilities. A very strong theme running through his Moscow speeches in the fall of 1957, a year before the second Quemoy-Matsu crisis, was that the Soviet Union's technological breakthroughs, coupled with the rapid industrial growth rates in the socialist camp, translated into strategic power. ₇₅

It is not unreasonable to assume, then, that Mao used fairly rough estimates of relative industrial and military power to determine whether China had more or fewer capabilities with which to assert its interests in conflict situations. For the moment I will use the ratio of the percent shares of major power capabilities as a rough indicator of relative power. Very preliminary findings indicate, for instance, that in all three foreign policy crises with the U.S. (Korea, Quemoy-Matsu 1954-1955, Quemoy-Matsu 1958) the U.S.-prc power ratio had shifted in China's direction over the previous year. In the two foreign policy crises involving the ussr (the Ussuri River crisis of 1969 and the Chinese invasion of Vietnam in 1979) the ussr-prc power ratio shifted in China's direction. ₇₆   In the one China-U.S. crisis that did not lead to violence (Taiwan, summer 1962), China's relative power ratio with the U.S. declined. ₇₇   These are crises in the Brecher, Wilkenfeld, and Rosen data set in which China initiated direct military conflict with the adversary. This does not mean that an improvement in relative capabilities inexorably led to the initiation of conflict. Indeed, there are a number of dyad years in Chinese-U.S. and Chinese-Soviet relations, for instance, where such an improvement occurred and there was no military conflict. Rather, a favorable change in relative capabilities appears to have been a necessary but not sufficient condition. That is, without such a change, it appears that there would not have been any conflict. Given such a change, in the absence of a foreign policy crisis military conflict is not inevitable. But in the universe of cases of foreign policy crises, an advantageous shift in relative capabilities did accompany the initiation of violence.

Problems of Analysis

A couple of conclusions can be drawn at this point about traditional Chinese and Maoist strategic culture. First, the predominant Chinese strategic tradition does not differ radically from key elements in the Western realpolitik tradition. Indeed, the Chinese case might be classified as a hard realpolitik one, sharing many of the same basic tenets about the nature of the enemy and the efficacy of violence with advocates of nuclear war fighting on both sides in the Cold War or late-nineteenth-century social Darwinian nationalists. ₇₈   It is characterized by positions at the high end of the three dimensions that make up the central paradigm of a strategic culture. From these central assumptions flows a preference for offensive and/or preemptive strategies in dealing with threats.

Second, the Chinese case suggests that strategic culture is not a trivial variable in the analysis of strategic behavior. There is, at least in the Chinese case, a long-term, deeply rooted persistent and relatively consistent set of assumptions about the strategic environment and about the best means for dealing with it. Moreover, these assumptions appear to have a nontrivial influence on grand strategic choice in several different historical periods. These parabellum assumptions have persisted across different state systems in Chinese history--from the anarchical Warring States period, to the hierarchical imperial Chinese state system, to the increasingly interdependent post-Cold War period. In the 1980s and 1990s, even as China's economy has become increasingly integrated into the global economy, even as international economic institutions play an increasing role in directing China's development strategies, and even as China faces the most benign threat environment since 1949, hard realpolitik decision rules continue to dominate the Chinese leadership's approach to foreign policy and security affairs. Chinese approaches to global issues such as arms control, the environment, and human rights are still dominated by defection and freeriding decision rules. Chinese leaders openly admit that China's development goal is "a rich state and a strong army" (fu guo qiang bing).

These conclusions raise a host of complex implications and questions about the nature of strategic culture and its influence on behavior. The crux of these problems is, as should be apparent, that the predictions from a hard realpolitik model of Chinese strategic culture are similar to those from what could conventionally be called a structural realpolitik model. This situation poses obvious difficulties for any competitive hypothesis testing. Indeed, neorealists might immediately ask what the point is in constructing a strategic culture model of realpolitik when an "old-fashioned" realist model without reference to strategic culture might do just as well. ₇₉   Their argument would be that Chinese realpolitik strategic culture is epiphenomenal, not causal, a function of anarchical structures. I think this is a premature judgment. There are two broad kinds of responses to this charge.

The first is to admit that these two models make similar predictions and then to set up critical tests. A relatively useful test of realpolitik strategic culture is to look for periods in Chinese history in which, controlling for structural change, one could identify fluctuations in the strength of the strategic culture. If strategic choice were consistent with these variations (e.g., the absence or weakness of an identifiable realpolitik strategic culture correlated with accommodationist strategies even when conditions favored the use of force against an adversary), then the case for cultural realism would obviously be stronger.

There is suggestive evidence that this was the case in the Ming dynasty. Proportionately most "events" (e.g., years in which relative power favored the Ming and in which the Ming launched offensives) occurred in the first quarter of the dynasty, when overall structural conditions favored the Ming. Proportionately, however, during this period one was also more likely to see "nonevents" (e.g., years in which relative power favored the Ming and in which the Ming did not launch offensives). In general, the first quarter was a period in which Ming emperors were engaged in legitimating their dynastic rule. Part of that process involved issuing statements affirming the more benign Confucian approach to security. It is possible that in these microperiods Ming rulers, facing a relatively benign threat environment and concerned about establishing an image of a magnanimous rule, endorsed a degree of restraint in strategic choice. A more fine- tuned test, then, would be to look in more detail at each of these years of non-events and compare the strength of the parabellum calculus in the decision process for the years in which offensives were launched. To the extent that the strength of the parabellum axioms varied in the predicted direction across these two types of cases, one could make firmer conclusions about the effects of cultural versus structural realpolitik.

In the prc case, we would need evidence that in those foreign policy crises in which China did not resort to force hard realpolitik axioms were not as prominent. Unfortunately we are working with a very small number of cases. There is one instance in which China did not resort to force immediately in the face of threat to a high-value territorial security issue--namely, the U.S. decision in June 1950 to use the Seventh Fleet in the Taiwan straits to prevent a Communist invasion of Taiwan. Mao clearly considered this a major challenge to his plans to liberate the rest of China and another major threat, given the war in Korea. Yet the absence of an immediate Chinese violent response was the result not of a softening of Mao's hard realpolitik worldview but of his inability to do anything about the problem in the short run. This inaction is consistent with both a structural and an ideational realpolitik that is sensitive to relative capabilities. This case, then, does not offer a conclusive test of the two models. Overall, the kind of test outlined above may be relatively hard to set up, since there has been very little fluctuation in the hegemony of the hard realpolitik strategic culture, particularly in the post-1949 period.

A second test might be a cross-national one. Here one would look for cases in which one could plausibly argue that the hard realpolitik paradigm was replaced by a different strategic culture, cases in which changes in relative capabilities have not led to the coercive opportunism that one finds in Chinese history. My argument has been that the hard realpolitik strategic culture is a prism through which changes in relative capabilities are interpreted. Absent this paradigm, and changes in relative capabilities should mean something different. ₈₀   In this respect, liberal democratic zones of peace, based on shared identities, provide intriguing evidence. The argument is that shared identities reduce in-group exclusivity, hence the group values the "other" more. Since each side knows that it prefers accommodation and negotiation, and since each side believes that this is the case for the other, the central assumptions of the parabellum paradigm become hard to maintain. Thus the high measures on the three dimensions of the central paradigm should move dramatically toward low measures. Military conflict is not seen as imminent, conflict is not considered zero-sum, and violence is not considered efficacious. Accommodationist strategies are therefore more preferred than offensive ones, regardless of changes in relative capabilities. As a result, advantages in relative capabilities are not exploited, and disadvantages are not feared. Since the parabellum paradigm does not pervade decision makers' perceptions when they are dealing with other democracies, changes in relative capabilities should not have the same effect as on states with parabellum strategic cultures. The absence of war itself, and the virtual absence of militarized disputes, even as the relative capabilities of these states have changed or remained imbalanced, suggests, then, that there is no relationship between structural changes and violent conflict among liberal democracies. ₈₁   Such a finding would be consistent with absence of the parabellum paradigm.

Since I am not using the absence of war (the dependent variable) to posit the absence of parabellum (the independent variable), this is not a tautological argument. Rather, I am using a shared democratic identity to posit the absence of parabellum. If the absence of the parabellum strategic culture correlates with the absence of coercive exploitation of advantages in changing relative capabilities among democracies, then it seems reasonable to conclude that the presence of a parabellum strategic culture in China is correlated with the coercive exploitation of these advantages. Needless to say, a structural realpolitik model could not account for the relative absence of realpolitik interpretations of changing material capabilities in democratic security communities. Thus there is no logical reason to expect it to account for the presence of realpolitik interpretations in nondemocracies.

This brings me to a final test, one designed to see whether the content of China's strategic culture remained relatively stable over time as structure changed. If this were the case, then the charge of epiphenomenality would not stand. To be sure, the formative period for this realpolitik strategic culture was, arguably, the Spring and Autumn and Warring States periods (770-221 b.c.), a time when politics among the feudal states of central China were characterized by anarchical multipolar relations. But realpolitik axioms persisted across the rise of unipolar imperial states interacting with weaker nomadic tribal states (e.g., early Han dynasty from the mid-second to the mid-first century b.c., the Tang dynasty of the early seventh century b.c., the Ming dynasty in the mid- through late fifteenth century, when it was the largest empire in the world). And these persisted during periods of weak dynastic or imperial control when relations between Chinese and non-Chinese states, or among Chinese states, were bipolar or multipolar (e.g., the Three Kingdoms period (a.d. 220-280), the Northern and Southern dynasties period (a.d. 420-581), and the Northern and Southern Song period (a.d. 960-1279). ₈₂   These different systems cannot all be legitimately considered the same anarchical type even though, strictly speaking, in each there was no supreme authority regulating relations among the key actors. To argue that they were the same type of system would be to ignore the effects of the vast power differentials across these systems on actors' perceptions of their options. It would mean ascribing to the mere absence of supreme authority such a deterministic effect that relative power becomes unimportant. Structural realists can't have it both ways--they cannot sometimes argue that anarchy is the most important variable determining states' strategic choices and other times argue that power distributions are the most crucial. ₈₃   Indeed, one could argue that in lopsided unipolar imperial systems one ought to expect some dramatic variation from realpolitik behavior, since even though the system is anarchic in the strict definition of the term, the empire's survival is not threatened by weak, disunited tribes along the periphery. Massively asymmetric relative capabilities would suggest that the empire could afford to ignore or buy off these low-level threats. ₈₄   Yet, in the Chinese case at least, Chinese empires, especially at peak periods of power, often exhibited an offensive, coercive behavior rooted in a perception of adversaries as implacably hostile and threatening to the very survival of the system and in a distrust of the long-term efficacy of accommodationist strategies.

In short, that this hard realpolitik calculus undergirded the strategic decision making in the unipolar imperial state system during the peak periods of Ming power in the fifteenth century (periods when the Ming was most aggressive in its efforts to exterminate the Mongol threat, for instance) and in the multipolar and bipolar "anarchical" state systems in which twentieth-century China was situated suggests that its persistence is related not to particular distributions of power or to different kinds of anarchical state structures but to the transmission of a particular strategic culture. In Mao's case socialization in the parabellum strategic culture came both from exposure to some of its elements in traditional texts military strategy and history as well as from Leninist and Clausewitzian ideas. Hence there is at least some evidence that realpolitik axioms developed and persisted across different structural contexts.

But a second broad way to deal with the neorealists' charge of epiphenomenality is to argue that, if anything, it is the effects of structure that are epiphenomenal to realpolitik strategic culture. The argument that one has to set up a critical test between a material structural and an ideational model of realpolitik rests on the assumption that these two make predictions about behavior that are causally independent of each other: in one model, strategic culture interacts with changing relative capabilities to produce behavior; in the other, according to neorealists, changing relative capabilities alone explains behavior. But when neorealists make this latter link between capabilities and behavior, they in fact require an assumption about the meaning or implications of this change for state survival. The problem is, the basis of this assumption for neorealism is not clear. ₈₅   For some neorealists the answer is simple: anarchy is the primary cause of states' interests and preferences over actions because in conditions of uncertainty about the intentions of others, one's own security is assured ultimately only with sufficient military capabilities arrayed on one's own side (by either internal or external balancing). Thus under conditions of anarchy states will tend to interpret disadvantageous shifts in relative capabilities as threatening and dangerous. ₈₆

Yet for other neorealists, the relationship between anarchy and realpolitik is in fact not coterminus. This is because their argument contains a nor mative element that if states wish to survive in anarchy they have to think in realpolitik ways; otherwise they lose. Since there are losers in anarchy (and I don't think their proportion has diminished over time, as the selection argument would predict), then one has to assume there are states that don't act on these assumptions about interests and capabilities, and hence that there is at least some choice in the matter. As some neorealists will agree, the only way that changing relative capabilities--structure--can explain behavior is if decision makers think these changes matter for the security of the state. That is, states or state elites provide the realpolitik meaning of changes in structure. This meaning depends on how states conceive their interests.

So the question is, Where does this particular realpolitik interpretation of interests come from? Some neorealists are happy to acknowledge that these interests, and the process by which states interpret changes in relative capabilities in the light of these interests, are assumptions and are not endogenous to structures per se. At some level the empirical validity of these assumptions is important to neorealism. But of greater importance is that these assumptions are made in the first place. The way in which these assumptions are correct--in other words, where decision makers get these tendencies to interpret changes in relative capabilities in realpolitik ways--is of still lesser importance. Of course, this failure to problematize the interpretation of relative capabilities opens the door to the constructivist argument that, to the extent that neorealism has no well-defined theory of where state interests or preferences over actions come from, just as long as these interests are realpolitik ones and are present in anarchical structures, it is quite possible that empirically these interests and preferences come from realpolitik strategic cultures, independent of structure.

For other neorealists, the source of these realpolitik interpretations of changes in relative capabilities is anarchy. ₈₇   Anarchy produces uncertainty, and uncertainty produces fear of being exploited. It is fear that gives meaning to a change in power distributions. But empirically we know that there is considerable variation in the levels of fear. France has reacted very differently to the unification of Germany in the 1990s than it did to the same process in the 1860s. Western Europe of the 1990s, as a democratic security community of sovereign states, reacts to relative capabilities changes among them in very different ways than it has in the past. So anarchy, a constant, cannot account for dramatic variation in levels of fear. Thus anarchy cannot account for realpolitik interpretations of changes in relative capabilities.

If this is the case, then the initial puzzle above does not require a critical test, since all the Chinese case does is provide empirical evidence that structural realism's assumptions about state preferences are rooted in realpolitik strategic cultures. These two models, then, are not competing ones. Rather, structural realpolitik can be subsumed within the cultural realpolitik model. Indeed, cultural realpolitik is necessary to save structural realpolitik from the embarrassment that its assumptions may be wrong. If this is the case, then, the standard juxtaposition of norms versus interests or structural/rational versus ideational models is a false one. If what we have come to know as structural realism rests on the empirical presence of cultural or ideational realism, then realpolitik behavior is ideationally based, just as behavior that deviates from realpolitik behavior has ideational roots as well. This makes sense to those who consider interests and preferences to be socially constructed. There can be no interests that are not socially constructed, hence there can be no interests that are not rooted in ideational cues for ordering the environment. Thus one simply cannot conceive of interests' being rooted outside of ideas (e.g., in structural anarchy) or pitted against them.

This conclusion is not particularly good news for many realists. While it suggests that there is an empirical basis for the assumption in neorealism that states can tend to interpret the constraints of structures in realist ways, it also suggests that this empirical basis is independent of structure. In other words, if cultural realism provides the content for one of the key assumptions of structural realism, and if this content is independent of structure (to the extent that it inhabits decision makers' perceptions and persists across different structural contexts), then realpolitik behavior is independent of structure. Cultural realism saves structural realism the embarrassment of being empirically wrong about how states "think" but in doing so suggests that structure cannot cause realpolitik behavior.

This conclusion raises a problem for research methodology. It suggests, as Jackson rightly puts it, that ideas and interests should not be juxtaposed against each other because both are "concepts and therefore ideas" and because both can exist independent of anarchical structures. ₈₈   If this is the case, we then have to explain what exactly is the relationship between ideas, norms, cultures on the one hand, and structure on the other. While most scholars who take the "autonomous power of ideas" seriously would concur with the importance of this question, they come up with at least three different conceptualizations of the relationship between ideas and structure. One view is that material structure does in fact produce determinate predictions about behavior but that these are often wrong because ideas, norms, and culture sometimes mediate and thus skew the impact of structure on the decision-making processes. Scholars who start from this assumption thus explore cases that "deviate" from a neorealist prediction. Their conceptualization could be portrayed as in figure 7.7. ₈₉

A second view sees neorealist structures producing indeterminate predictions, providing limited possibilities and choices, which are then decided upon through the mediation of ideas, norms, and culture (see figure 7.8). Ideas finalize interests and preferences. ₉₀   In both conceptualiza tions, structure is implicitly or explicitly given enough content independent of ideas to produce behavioral predictions.

But a complete rejection of the ideas-interests dichotomy leads to a third conceptualization. Here ideas, norms, and culture generate structures--anarchical ones if the "strategic culture" is essentially a realpolitik one and institutionalized ones if the strategic culture is essentially an idealpolitik one. Behavior flows from structures, but these are given content and meaning by the ideational precursors. The behavior reinforces the ideational base upon which the super-"structure" rests (figure 7.9).

Of course, this still begs the question, raised in Wendt's important article, ₉₁   Where do these ideational bases come from and why do certain ideational and behavioral vectors appear at different times? In trying to explain where in the process of social interaction the disposition to identify other states as potential enemies against whom one should be prepared to use force developed, Wendt ventures that the emergence of a predatory state might teach other states to act in competitive power-political ways, though he fails to identify where a predatory state might come from. I am in no position to comment about the primordial origins of hard realpolitik strategic cultures, but I am willing to argue that their presence is a precondition for realpolitik behavior. ₉₂   The parabellum strategic culture--this learned resignation that disputes are settled through violence--may be one of the key ideational links in the chain leading from the appearance of interstate grievances to war. ₉₃ This does not necessarily mean that a cultural realpolitik model of strategic choice must be universally, cross-culturally valid or indeed constantly valid within one society across time. If this "style" of strategic choice is cultural in the sense that only those who "learn" it are likely to act along parabellum lines, then the Chinese case leaves open the theoretical and empirical possibility that other decision makers may not share hard realpolitik assumptions. The China case, then, underscores Vasquez's contention that realpolitik is "historically contingent and confined to certain issue areas" and Wendt's argument that structural anarchy is not the cause of self-help behavior. The test is to find a "system or issue area" characterized by non-parabellum or non-realpolitik behavior. ₉₄   At the moment, as I noted above, perhaps the best case for the argument that not all "successful" states act in realpolitik ways is the existence of the democratic security community or zone of peace, a community of sovereign states that in their interaction have abandoned the assumptions of the parabellum paradigm. ₉₅

The China case leads us to think harder about the ideational roots of realpolitik behavior. China is a critical case both for those who accept the independent causal effect of ideational variables and for those who don't. The former, constructivists and institutionalists alike, have tended to use ideational arguments to explain behavior that deviates from "standard" realist expectations. They will have to develop ideational explanations for "nondeviant" behavior in order to present a truly competitive challenge to dominant structural approaches. Chinese strategic behavior is a place to start. The latter will have to show why China's cultural realpolitik is not causal but epiphenomenal, despite its persistence across very different exogenous conditions. I don't think they can do it, for the reasons I've outlined here. In short, China is a hard case for both approaches, but it is the analysis of hard cases that drives theory forward.

---

I would like to thank the following people for their comments and criticisms of this essay: Peter katzenstein and other participants in the Social Science Research Council/MacArthur Project, as well as outside commentators at the three project workshops; members of the Olin Institute National Security seminar, and Thomas Christensen, Dale Copeland, Robert Keohane, Kenneth Lieberthal, and Stephen Walt. They are blameless for any shoddy analysis or errors of fact.

Note 1:  Alexander Wendt has recently articulated the importance for constructivist approaches of taking on this "hard case" of realpolitik behavior. See his "Constructing International Politics: A Response to Mearsheimer," International Security 20, no. 1 (Summer 1995). Back.

Note 2:  Parabellum comes from the realpolitician's axiom "si pacem, parabellum" (if you want peace, then prepare for war). This parallels a Chinese idiom, "ju an si wei, wu bei you huan" (while residing in peace, think about dangers; without military preparations there will calamity). Back.

Note 3:  Alastair Iain Johnston, Cultural Realism: Strategic Culture and Grand Strategy in Chinese History  (Princeton: Princeton University Press, 1995). Back.

Note 4:  Barry Gill, "The Hegemonic Tradition in East Asia: A Historical Perspective," in Stephen Gill, ed., Gramsci, Historical Materialism, and International Relations, p. 195 (Cambridge: Cambridge University Press, 1993); Justin Rosenberg, "Secret Origins of the State: The Structural Basis of Raison d'etat," Review of International Studies 18 (1992): 132; John A. Vasquez, The War Puzzle  (Cambridge: Cambridge University Press, 1993), p. 116. Back.

Note 5:  The following section draws from Johnston, Cultural Realism. Back.

Note 6:  Confucius (551-479 b.c.) was a philosopher of ancient China whose teachings (and their interpretations) became the basis of the predominant orthodoxy in political and moral thought in China. Mencius (390?-305? b.c.) was one of the more influential interpreters of Confucius' thought. At the risk of oversimplifying a complex body of thought, on questions of war and peace Confucianism stressed that external security rested on a ruler's ability to provide for the material and moral needs of his people through virtuous personal conduct and enlightened policies. This way, the people of the realm would be content with their lot, and potential enemies would willingly submit to partake of the ruler's magnanimity. Moral education was sufficient to transform potential enemies into willing and submissive allies. Confucius did not oppose military preparations, though he downplayed their role in the security of the state. Mencius, in particular, pushed Confucian ideas in a more extreme direction, arguing that a virtuous ruler had no need to use military force because he could have no enemies. It is highly questionable, however, whether this orthodoxy exercised much restraint on operational strategic thought and practice in Chinese history. See ibid. Back.

Note 7:  Paul Forage, "The Struggle for the Northwestern Frontier and the Consequences for the Northern Song" (paper presented at the annual conference of the Association of Asian Studies, Boston, March 1994). Back.

Note 8:  I concentrate on the first generation here because it is this work that has tended to dominate the literature. The first generation, which emerged in the early 1980s, focused for the most part on trying to explain why the Soviets and the Americans apparently thought differently about strategy in the nuclear age. Borrowing from Snyder's work on strategic culture and Soviet limited nuclear war doctrine, authors such as Gray, Lord, and Jones all argued in some form or another that these differences were caused by unique variations in deeply rooted historical experience, political culture, and geography among other macroenvironmental variables. See Jack L. Snyder, The Soviet Strategic Culture: Implications for Nuclear Options, Rand R-2154-AF (Santa Monica, Calif.: RAND Corporation, 1977); Colin Gray, "National Styles in Strategy: The American Example," International Security 6, no. 2 (1981): 21-47; Colin Gray, Nuclear Strategy and National Style  (Lanham, Md.: Hamilton Press, 1986); Carnes Lord, "American Strategic Culture," Comparative Strategy 5, no. 3 (1985): 269-93; David R. Jones, "Soviet Strategic Culture," in Carl G. Jacobsen, ed., Strategic Power: USA/USSR, pp. 35-49 (New York: St. Martin's, 1990). The second generation refers to a small number of studies that appeared mostly in the mid-1980s and focused on strategic culture as a kind of Gramscian discourse designed to reinforce the policy hegemony of strategists. See Bradley Klein, "Hegemony and Strategic Culture: American Power Projection and Alliance Defence Politics," Review of International Studies 14 (1988): 133-48; Robin Luckham, "Armament Culture," Alternatives 10, no. 1 (1984): 1-44. The third generation broadly includes work emerging in the 1990s that has focused on using culture, norms, and ideas as explanations for behavior. This literature is the most rigorous in conceptualization and methodology, but it is too new to have had much of an impact on the analysis of "strategic culture" in mainstream security studies. See Jeffrey W. Legro,Cooperation Under Fire: Anglo-German Restraint During World War II  (Ithaca: Cornell University Press, 1995); and the essays in this volume by Elizabeth Kier, and Nina Tannenwald and Richard Price. For a detailed discussion of the three generations of work on strategic culture, see Johnston, Cultural Realism. Back.

Note 9:  See Gray, "National Styles in Strategy," and Gray, Nuclear Strategy and National Style. Back.

Note 10:  Johnston, Cultural Realism, ch. 2. Back.

Note 11:  Clifford Geertz, The Interpretation of Cultures  (New York: Basic Books, 1973), p. 90. Back.

Note 12:  Here I rely on Taber's notion of a dominant paradigm as a collection of heuristics used "to guide the selection of problem-solving strategies for some specifiable period of time. Individual leaders may or may not be aware of the influence of a dominant paradigm on strategic choice" (Charles Taber, "Modern War Learning: A Markov Model" [paper presented at the Midwestern Political Science Association Conference, Chicago, 1987], p. 4). There are similarities here with operational codes and belief systems, except that strategic culture analysis focuses more squarely on historically rooted and collectively shared, rather than individual, belief systems. Back.

Note 13:  See also Wildavsky's "cultural theory of preference formation" in Aaron Wildavsky, "Choosing Preferences by Constructing Institutions: A Cultural Theory of Preference Formation," American Political Science Review 81, no. 1 (1987): 3-20. Back.

Note 14:  Lebow argues, for instance, that acceptance of the inevitability of war leads to preferences for preventive and/or preemptive strategies, since if war is inevitable, it makes sense to act before the enemy does. See Richard Ned Lebow, Between Peace and War: The Nature of International Crisis  (Baltimore: Johns Hopkins University Press, 1987), pp. 254-63. See also Snyder on the "cult of the offensive"; Jack L. Snyder, The Ideology of the Offensive: Military Decision Making and the Disasters of 1914  (Ithaca: Cornell University Press, 1984), pp. 17, 39, 199. Back.

Note 15:  A belief that war is an aberrant or at least a preventable event in human affairs ought to be associated with a non-zero-sum view of the adversary. If conflicts are in the main negotiable, then presumably the enemy has a price short of one's own capitulation. If this is the case, then ceteris paribus highly coercive, violent strategies--which entail severe economic and political costs--would seem to be strategies of last resort, since trade-offs and logrolling opportunities appear more cost-effective in managing  (as opposed to eliminating) the threat. Back.

Note 16:  David J. Elkins and Richard E. B. Simeon, "A Cause in Search of Its Effect, or What Does Political Culture Explain," Comparative Politics 11 (1979): 133. Back.

Note 17:  For a list, see Luckham, "Armament Culture," tables 2 and 3. See also Kier's list (essay 6 in this volume), which includes curriculum from military schools, training manuals, journals, languages, symbols, taboos, etc. Legro, in Cooperation Under Fire, examines planning documents, regulations, military exercises, and memoirs. Back.

Note 18:  These texts on statecraft, strategy, and tactics are still read and studied in the professional military education systems in Taiwan and the People's Republic of China. Sun Zi's The Art of War is one of the seven texts. For the only English translation of the complete set, see Ralph Sawyer, trans., The Seven Military Classics of Ancient China  (Boulder: Westview, 1993). Back.

Note 19:  Given the importance of sampling those texts that are most likely to capture the strategic culture(s) in a society, doesn't this sampling of one individual's works on strategy virtually guarantee that I will find congruence in preference rankings across texts? How can the ideas and axioms of an individual constitute a collectively shared, hence socially constituted, strategic culture? These are legitimate concerns. I have a couple of responses. First, to the extent that we are interested in explaining behavior, it makes sense to look for the presence of a strategic culture in the work of those who play a key role in decision making. As recent scholarship underscores, Mao was essentially in charge of Chinese foreign policy making at the strategic and even tactical levels from 1949 on. See Thomas J. Christensen, Useful Adversaries: Grand Strategy, Domestic Mobilization, and Sino-American Conflict, 1947-1958  (Princeton: Princeton University Press, in press); Robert Ross, The United States and China: Negotiating for Cooperation, 1969-1989  (Palo Alto: Stanford University Press, 1995); Chen Jian, China's Road to the Korean War  (New York: Columbia University Press, 1994). Second, given Mao's dominance of the policy process in China, coupled with his personality cult, Mao's "thought" formed the core of ideological socialization in China. To the extent that decision makers in China were exposed to antecedent strategic axioms on the basis of which they may or may not have made decisions, these axioms came largely from Mao. Given the special circumstances of strategic decision making in Maoist China, there is no easy way around the question of whether Maoist strategic culture was collectively shared. I would argue, however, that given the absence of any substantial contestation of Maoist strategic axioms during and after his life, the term Maoist strategic culture is still valid. That Maoist strategic axioms persisted after Mao died and after most of his economic, political, and cultural legacies were dismantled under Deng Xiaoping suggests that the strategic and military legacy was accepted and internalized by Chinese decision makers. There are literally scores of books on Mao's strategic and foreign policies that have been published in recent years. See Foundation for International Strategic Studies, ed., Huan qiu tong ci liang re: yi dai ling xiumen de guoji zhanlue sixiang [Global contradictions: A generation of leaders and their international strategic thought] (Beijing: Central Documents Publishing House, 1993); and Gu Yan, "Duli zizhu shi Mao Zedong waijiao sixiang de linghun" [Independence is the soul of Mao Zedong's foreign policy thinking], Shijie jingji yu zhengzhi [World economics and politics], no. 2 (1994): 30-33. Back.

Note 20:  David Dessler, "Notions of Rationality in Conflict Decision-Making" (paper presented at the annual conference of the International Studies Association, Anaheim, Calif., 1986), pp. 18-19; see also Judith Goldstein and Robert O. Keohane, "Ideas and Foreign Policy: An Analytical Framework," in Judith Goldstein and Robert O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions, and Political Change, p. 13 (Ithaca: Cornell University Press, 1993). Back.

Note 21:  Robert Axelrod, ed., Structure of Decision: The Cognitive Maps of Political Elites  (Princeton: Princeton University Press, 1976). Back.

Note 22:  See, for example, Charles D. Elder and Roger W. Cobb, The Political Uses of Symbols  (New York: Longman, 1983), p. 28; Lowell Dittmer, "Political Culture and Political Symbolism," World Politics 29 (1977): 577, 579; Lance Bennett, "Perception and Cognition: An Information-Processing Framework for Politics," in Handbook of Political Behavior  (New York: Plenum, 1981), 1:76; Earl R. MacCormac, A Cognitive Theory of Metaphor  (Cambridge: MIT Press, 1985), pp. 23-24; Yuen Foong Khong, Analogies at War: Korea, Munich, Dien Bien Phu, and the Vietnam Decisions of 1965  (Princeton: Princeton University Press, 1992), p. 10. Back.

Note 23:  These are explored in Johnston, Cultural Realism, ch. 2. Back.

Note 24:  One of the criticisms that structuralists sometimes level at those who work with cultural variables is that a constant, like culture, cannot explain change in state behavior. This charge fundamentally misunderstands the constructivist argument: ideas as independent variables are useful only because they interpret or give meaning to material facts. Thus changes in relative capabilities, for instance, mean something different to a realpolitician concerned about relative gains in a competitive world than they do to a liberal concerned about absolute gains. Variation in the predictions of cultural models comes from changes in the relevant material environment; the implications of these changes, however, depend on the content of the ideational constant. So constructivists should have no trouble using a structural intervening variable to make determinate predictions. Back.

Note 25:  Readers will note that this foreign policy offshoot of structural realism assumes that states prefer to maximize power, not simply to seek mere survival. This assumption is not uncontroversial, but without it realist models of strategic choice become indeterminate, just as economic expected utility approaches to decision making become indeterminate, or harder to model, without the use of money as the basis of the utility function. See Mancur Olson, "Toward a Unified View of Economics and the Other Social Sciences," in James E. Alt and Kenneth A. Shepsle, eds., Perspectives on Positive Political Economy, p. 218 (Cambridge: Cambridge University Press, 1992). If survival is the goal of states, one might expect a range of grand strategic preferences--from "defensive" expansionism in the name of the status quo, to static defense or deterrence strategies. On the other hand, if the expansion of influence and power is the primary goal of states, it makes less sense for them to rely solely on defensive or deterrence strategies. Thus the latter assumption about state interests leads to more determinant predictions about strategic behavior. In essence, to make structural realist models determinate, one has to plug in assumptions about state interests that are more commonly found in classical realist literature. This is essentially why Mearsheimer's version of neorealism is more determinate than most. See John J. Mearsheimer, "The False Promise of International Institutions," International Security 19, no. 3 (Winter 1994/95): 9-11. Back.

Note 26:  John A. Vasquez, "Capability, Types of War, Peace," Western Political Science Quarterly 39, no. 2 (1986): 321; John A. Vasquez, "Foreign Policy Learning and War," in Charles Hermann, Charles W. Kegley Jr., and James N. Rosenau, eds., New Directions in the Study of Foreign Policy  (Boston: Unwin Hyman, 1987), pp. 367-68. Back.

Note 27:  Mao Zedong, "On Contradiction," [1937], Selected Works of Mao Zedong  (Beijing: Foreign Languages Press, 1965), 1:311-47; Gao Tiejun, "Mao Zedong guofang xiandai hua sixiang chu tan" [A preliminary discussion of Mao Zedong's thinking on national defense modernization], Junshi lishi yanjiu [Studies in military history], no. 3 (1987): 61; Liu Chunjian, Shenqi de qi he: Mao Zedong Deng Xiaoping yu Zhongguo chuantong wenhua [A mystical unity: Mao Zedong, Deng Xiaoping, and China's traditional culture] (Shanxi: People's Publishing House, 1992); John Bryant Starr, Continuing the Revolution: The Political Thought of Mao  (Princeton: Princeton University Press, 1979). Back.

Note 28:  We know that from an early age, before he endorsed Marxism-Leninism, Mao accepted the notion that all phenomena comprise contradictory elements that create a process of constant change. In his marginal notes on a book about ethics that he read in 1917-1918 in his mid-twenties, Mao commented, "There is no life and death, only change. If this thing grows, that thing is eliminated. Life is not only life, and death is not only extermination. Merely a change" (Mao Zedong, " 'Lunli xue yuan li' pi zhu" [1917] [Critical comments on the Tenets of Ethics], in Mao Zedong zao qi wen gao [Mao Zedong's early writings] [Hunan: People's Publishing House, 1990], p. 200). Back.

Note 29:  Mao Zedong, "Jiang tang lu" [1913] [Class notes], in Mao Zedong zao qi wen gao, p. 585; Lucian Pye, Mao Tse-tung: The Man in the Leader  (New York: Basic Books, 1976). Back.

Note 30:  See, for instance, the writings of Heinriche von Treitschke, Selection from Treitschke's Lectures on Politics, trans. Adolf Hausrath (London: Gowans and Gray, 1914), and his Politics, trans. Blanche Dugdale and Torben de Bille (New York: Macmillan, 1916). As Frank Dikotter implies, however, it is more likely that Mao was influenced by Spencerian notions of organic intergroup competition as interpreted by Yan Fu. Darwin's work was not completely translated into Chinese until 1919. Instead, among early-twentieth-century Chinese intellectuals Herbert Spencer was one of the most influential European thinkers on the struggle between social groups; see Frank Dikotter, The Discourse of Race in Modern China  (Stanford: Stanford University Press, 1993), p. 104. Back.

Note 31:  See Johnston, Cultural Realism. Back.

Note 32:  He Taiyou and Zhang Zhongliang, Mao Zedong Zhanfa [Mao Zedong's Art of War] (Beijing: National Defense University Press, 1988), p. 11. See also Lan Shuchen, "Mao Zedong junshi sixiang yu zhongguo chuantong wenhua" [Mao Zedong's military thought and China's traditional culture], Junshi lishi yanjiu [Studies in military history], no. 3 (1987): 69. Back.

Note 33:  On righteous war, see Zeng Zhen, Tang Tai Zong, Li Wei Gong wen dui jin zhu jin yi [Contemporary translation and annotation of the "Dialogues of Tang Tai Zong and Li Wei Gong"] (Taibei: Commercial Press, 1986), p. 12; Lin Pinshi, Lu Shi Chun Qiu jin yi jin zhu [Contemporary translation and annotation of the "Lu shi chun qiu"] (Taibei: Commercial Press, 1986), p. 188; Shi Zimei, Shi shi qi shu jiang yi [Shi Zimei's teaching materials on the Seven Military Classics], Riben wen jiu san nian ed. (1222; Taibei: Taiwan National Museum), 31:9b and 34:17a; Niu Hongen and Qiu Shaohua, Xian Qin zhu zi junshi lun yi zhu [Translation and annotation of military essays by Pre-Qin scholars] (Beijing: Military Sciences Press, 1985), 1:397; Liu Yin, Si Ma Fa zhi jie [Commentary on the "Si Ma Fa"] (Taibei: Shijian Press, 1955), ch. 1. Back.

Note 34:  There are parallels with Machiavelli's view of morality and war: "Morality in Machiavelli's usage is entirely instrumental: it is part of a prince's arsenal to be used to greater or lesser effect. In no sense does it restrain state behavior--nor should one expect it to do so because the state in Machiavelli's treatment is beyond such restraint" (Michael Joseph Smith, Realist Thought from Weber to Kissinger [Baton Rouge: Louisiana State University Press, 1986], p. 10). For a fuller discussion, see Johnston, Cultural Realism, ch. 3. Back.

Note 35:  See Liu Yin [SKQS], Huang Shi Gong San Lue zhi jie [Commentary on Huang Shi Gong's Three Strategies] [Si Ku Quan Shu version] (Taibei: Commercial Press, 1983-1985), p. 726.103; see also Shi, Shi shi qi shu jiang yi, 36.7b, and Lin, Lu Shi Chun Qiu, pp. 211, 217. The demonization of an unrighteous adversary--and the corollary that any and all means of eliminating this enemy are legitimate--is described well by modern-day attribution theory. The adversary's behavior is seen as dispositional, while one's own behavior is circumstantial. This construct creates "an image of the situation for personal and public consumption that releases the subject from moral inhibitions and allows the subject to deal with the threat or opportunity without restraint" (Richard Herrmann, "The Empirical Challenge of the Cognitive Revolution: A Strategy for Drawing Inferences About Perceptions," International Studies Quarterly 32 [1988]: 183, 185). See also Thomas Hart, "Cognitive Paradigms in the Arms Race: Deterrence, Detente, and the 'Fundamental Error' of Attribution," Conflict and Cooperation 3 (1978): 150; John E. Mack, "The Enemy System," in Vamik D. Volkan, Demetrios A. Julius, and Joseph V. Montville, eds., The Psychodynamics of International Relations, vol. 1, Concepts and Theories  (Lexington, Mass.: Lexington Books, 1990), pp. 60-61; Howard F. Stein, "The Indispensable Enemy and American-Soviet Relations," in ibid., p. 82; J. Richard Eiser and D. van der Pligt, Attitudes and Decisions  (London: Routledge, 1988), pp. 45-66. Back.

Note 36:  Song Rang Gong, a ruler of a state during the Spring and Autumn period in the seventh century b.c., was defeated after he refused to attack a vulnerable enemy force as it crossed a river. He argued that it was immoral to attack before the enemy had fully formed up. For his chivalry he suffered a defeat and the opprobrium of later generations of "realist" strategists. See also Lan, "Mao Zedong junshi sixiang," p. 68. Back.

Note 37:  Mao, "On Contradiction," p. 322. Back.

Note 38:  Mao's essay "On Policy" is a good example of how to deal with an actor in the second cell. In it he argues for a united front with the gmd even in the face of increased anti-Communist actions by the ccp's erstwhile anti-Japanese allies. The essay was re-released in the early 1970s to justify the necessity of rapprochement with the U.S. in the face of a growing Soviet threat. See Mao Zedong, "On Policy" [1940], in Selected Works of Mao Zedong, 2:441-49. Back.

Note 39:  This did not mean that Mao advocated the wholesale slaughter of enemy forces or political enemies. Annihilation required the elimination of the enemy's military power. This could entail the application of such overwhelming military force that enemy forces capitulated en masse. What Mao opposed was attrition, whereby enemy forces would be routed or bloodied such that they could still fight another day. As he remarked, "Injuring all of a man's ten fingers is not as effective as chopping off one, and routing ten enemy divisions is not as effective as annihilating one of them" (Mao Zedong, "Problems of Strategy in China's Revolutionary War" [1936], Selected Works of Mao Zedong, 1:248). Back.

Note 40:  Lan, "Mao Zedong junshi sixiang," p. 69, and He and Zhang, Mao Zedong Zhanfa, p. 11. It is often suggested that Mao came to Clausewitz through Lenin. This may indeed have been Mao's first exposure to the notion of war as a continuation of politics, but recent new materials on Mao suggest that Mao also studied Clausewitz directly, beginning in the spring of 1938. He even organized a Clausewitz study group joined by other Party military leaders. This sparked a small "Clausewitz fever" in the Communist base area at Yenan over the next couple of years as translations and commentaries on "On War" appeared in the Communist press. See Sun Baoyi, ed., Mao Zedong de du shu shengya [The book reading life of Mao Zedong] (Beijing: Knowledge Press, 1993), pp. 79-80. Back.

Note 41:  This is consistent with the ancient Chinese notion of "using the military instrument only under unavoidable circumstances" (bu de yi er yong bing). Whether or not these circumstances were avoidable depended wholly on the enemy. Since the environment was, in the main, conflictual, and since the conflict with the enemy was, in the main, zero-sum, the likelihood that force would have to be used frequently was quite high. For a good contemporary example of this type of argument, see Mao's discussion of the use of coercion against Taiwan under unavoidable circumstances. See Mao Zedong, "Dui Zhonggong ba da zhengzhi baogao gao de pi yu he xiugai" [Comments on and revisions of the draft political report to the 8th Congress of the ccp] [1956], in Jianguo yilai Mao Zedong wen gao, 1956-1957 [Mao Zedong's manuscripts since the founding of the nation] (Beijing: Central Documents Publishing House, 1992), 6:142-43. The pla marshal Nie Rongzhen argued in a speech to leading cadres in military industry in 1963, "Whether we fight or not is definitely not up to us. We are not the general staff of imperialism" (Nie Rongzhen, "Zai jun gong lingdao ganbu huiyishang de jianghua" [1963], in Nie Rongzhen, Nie Rongzhen junshi wenxuan [Selected military works of Nie Rongzhen] [Beijing: Liberation Army Press, 1992], p. 497). Back.

Note 42:  Mao, "Jiang tang lu," p. 585. Back.

Note 43:  Mao Zedong, "Min zhong de da lian he--yi" [The great alliance of the masses--1] [1919], in Mao Zedong zao qi wen gao, pp. 293, 341. Back.

Note 44:  Song Shilun, Mao Zedong junshi sixiang de xingcheng ji qi fazhan [The formation and development of Mao Zedong's military thought] (Beijing: Military Sciences Press, 1984), pp. 9-16; see also Mao Zedong, "Zhengquan shi you qiang ganzizhong qu de de" [Political power is obtained from the barrel of a gun] [1927], in Mao Zedong, Mao Zedong junshi wen xuan (neibu ben) [Selected military essays of Mao Zedong (internal edition)] (Beijing: Liberation Army Press, 1981), p. 4. Back.

Note 45:  Mao Zedong, "Muqian xingshi he dang zai yi jiu si jiu nian de ren wu" [The present situation and the Party's tasks for 1949] [1949], in Mao, Mao Zedong junshi wen xuan, p. 328. Back.

Note 46:  A word about notation. The first numeral refers to the page number of the text, the next refers to the paragraph number, and the last to the assigned number of the cause-effect relationship found in that paragraph. The cause-effect statements were entered into a simple database program that allowed me to search for particular effects or particular causes using key words. In each text the cause and effect concepts were collapsed and clustered into self-evident categories, and I tried to keep these categories as consistent as possible across texts. For each causal argument, I have listed only a couple of representative cause-effect statements here, rather than providing notation for each of the 810 cause-effect statements relating to security, statecraft, strategy, and tactics that I coded. On coding, see Margaret Tucker Wrightson, "The Documentary Coding Method," in Axelrod, Structure of Decision. My own coding adapted and simplified some of her procedures.

Back.

Note 47:  As Mao wrote, "With the slogan of defending the revolutionary base areas and defending China, we can rally the overwhelming majority of the people to fight with one heart and one mind, because we are the oppressed and the victims of aggression. In every just war the defensive not only has a lulling effect on politically alien elements, it also makes possible the rallying of the backward sections of the masses" (Mao Zedong "Problems of Strategy in China's Revolutionary War" [1936], Selected Works of Mao Zedong, 1:207). Back.

Note 48:  Ibid., p. 187. There are a slew of cause-effect relationships in the text in which the cause concept is some formulation of the notion of absolute flexibility or the exploitation of changing circum-stances and the effect is military utility or military victory. See, for example, 184.2.6; 187.4.1; 190.2.5; 191.1.6; 205.3.4; 240.7.1. Back.

Note 49:  Liu, Si Ma Fa zhi jie, p. 33. Back.

Note 50:  Academy of Military Sciences, eds., Tou Bi Fu Tan  (Beijing: Academy of Military Sciences Press, 1984), p. 87. Back.

Note 51:  This is a common characterization found in both Chinese and Western analyses of Mao's thought. See, for instance, Ralph Powell, "Maoist Military Doctrines," Asian Survey  (1964): 239; Zhang Shuguang, Deterrence and Strategic Culture: Chinese-American Confrontations, 1949-1958  (Ithaca: Cornell University Press, 1992), p. 278; Lin Chongpin, China's Nuclear Weapons Strategy  (Lexington, Mass.: Lexington Books, 1988), p. 24; Ellis Joffe, The Chinese Army After Mao  (Cambridge: Harvard University Press, 1987), p. 4. Unfortunately the discussion of the role of the "human" element in Mao's thought has missed some of the subtlety in his writings. To be sure, for Mao a highly motivated population or military was preferable to a poorly motivated one. The question is why he argued this. I would suggest that his de-emphasis on material capabilities--technology, weaponry, and so on--was a part of a strategy for mobilizing morale under conditions of relative military weakness. Mao believed that in the face of strength it made sense to appear cavalier and unafraid so as to combat fatalism on one's own side and to make the enemy think twice about the wisdom of unleashing a conflict over which it might have little control. It is also clear that one reason that Mao stressed the application of massively overwhelming numerical superiority was that in the face of a technologically superior enemy, only a quantitatively superior force could concentrate quantitatively superior technological capabilities. Mao was no strategic Luddite; he did not believe that a military should deliberately eschew technological modernization or the development of superior capabilities. Rather, the human element and technology had synergistic multiplier effects on each other. Christensen makes the important argument that one of the key goals of the Great Leap Forward in the late 1950s was to mobilize the Chinese economy to produce more advanced military technology to match (eventually) the capabilities of other major powers; see Christensen, Useful Adversaries; see also Song, Mao Zedong junshi sixiang, pp. 214-15. Back.

Note 52:  Mao did concede, consistent with his notion of absolute flexibility, that when a state lacked sufficient power it should avoid direct battle with larger states and instead adopt "ingenious methods" and stratagems (192.1.5). Back.

Note 53:  Here, as elsewhere in his works, Mao was not constrained by a narrow consistency. He also argued that the basic factor leading to socialism's victory was the desire or will of the people, the human factor (119.7.1). At the same time, however, the speech stressed the causal importance of steel production, the stuff of industrial and military power, in the superiority of socialism. Mao was at this time apparently quite influenced by Stalin's views on economics, which took steel as the "key link" in the development of socialism. I thank Thomas Christensen for this last point. Back.

Note 54:  Mao uses a couple of historical analogies to make this point. In "Problems of Strategy in China's Revolutionary War," he cites the example of the states of Lu and Qi in the Spring and Autumn period of the Zhou dynasty (twelfth century b.c. to third century b.c.) to illustrate the axiom "When the enemy tires, we attack" (p. 211). In an essay on strategy against Japan, he uses the analogy of "Relieving the state of Chao by attacking the state of Wei" to argue that if the enemy's forces were diverted elsewhere, then an offensive deep inside enemy territory where it was weakest was a legitimate strategy (Mao, "Problems of Strategy in Guerrilla War Against Japan" [1938], in Selected Works of Mao Zedong, 2:104). Back.

Note 55:  Open Chinese sources do not mention this early exposure. Sun suggests only that during the revolutionary wars against the gmd Mao read Sun Zi over several times, along with another Warring States text, Guan Zi. From Guan Zi, Mao is said to have absorbed the principle of attacking the enemy's weak points first and avoiding his strong points. See Sun, Mao Zedong de du shu shengya, pp. 78-79; and Lan, "Mao Zedong junshi sixiang," p. 66. Back.

Note 56:  Mao, "Jiang tang lu," pp. 595-96. Back.

Note 57:  Mao Zedong, "Zhanzheng he zhanlue wenti" [Problems of war and strategy] [1938], in Mao Zedong xuanji, p. 512 (Beijing: People's Press, 1967). Back.

Note 58:  Interview with researchers at the Academy of Military Sciences, Beijing, March 1991; Guo Huaruo, " 'Sun Zi yi zhu' qian yan" [Preface to "Translation and annotation of Sun Zi"] [1983], in Gua Huaruo, Guo Huaruo junshi lunwen xuanji [Guo Huaruo's selected essays on military affairs] (Beijing: Liberation Army Press, 1989), p. 427. Back.

Note 59:  I have discussed the congruence in preference rankings across the Seven Military Classics elsewhere (Johnston, Cultural Realism, ch. 4). Suffice it to say that there is not perfect statistical congruence across all seven texts. Depending on how the preferences of a couple of the more ambiguous texts are inferred, the coefficient of congruence (Kendal's W) ranges from .39 to .43. The former is significant at the 0.09 level, the latter at the 0.05 level. This moderate consistency is in large measure attributable to the effects of one text's preference rankings--the San Lue. This text, in contrast to the other six, places more emphasis on the Confucian-Mencian central paradigm. Consequently, it tends to prefer accommodationist and defensive strategies over clearly offensive ones. Back.

Note 60:  Liao Guoliang, Li Shishun, and Xu Yan, Mao Zedong junshi sixiang fazhan shi [The history of the development of Mao Zedong's military thought] (Beijing: Liberation Army Press, 1991), p. 450. Back.

Note 61:  Mao Zedong, "Gei di yi canmie yu gei di yi canmie xing da ji bixu tongshi zhuzhong" [We must pay simultaneous attention to annihilating the enemy and attacking him with annihilating effects] [1947], in Mao Zedong junshi wen xuan, p. 314. Back.

Note 62:  Interestingly, Chinese sources indicate that Mao's decision in 1948 to shift to the strategic offensive in the war with the gmd came as material conditions were shifting clearly in the ccp's favor. In 1946 the ratio of regular gmd to ccp troops stood at 3.5:1. By 1948 this figure had dropped to 1.3:1, but of these the ccp's first-line forces outnumbered the gmd's first-line forces by a ratio of 1:0.62. See Song, Mao Zedong junshi sixiang, pp. 190-91. Back.

Note 63:  For a good explication of Mao's doctrine of active defense and the political uses of "defense," see Peng Dehuai's report to the National Defense Commission in July 1957; Peng Dehuai, "Junshi jianshe gaikuang" [The general situation in military construction] [1957], in Peng Dehuai junshi wenxuan [Selected military works of Peng Dehuai] (Beijing: Central Documents Publishing House, 1988), pp. 588-91. See also Song, Mao Zedong junshi sixiang, pp. 166-208. The deliberate use of "defense" to describe a doctrine that allowed, indeed required, offensive operations points to another similarity between Maoist and traditional Chinese strategic cultures, namely, the presence of a distinct symbolic discourse disconnected from operational decision rules. While Maoist Chinese strategic culture and behavior exhibited parabellum tendencies, it was clothed in the rhetoric of people's war. In principle the doctrine of people's war is quintessentially defensive--an attacking enemy should be "lured in deep" into Chinese territory and enveloped by an armed population and a mobile professional military. I would argue that people's war per se did not reflect Mao's most preferred strategy, in that he believed defensive strategies were only one stage in a broader process of shifting to the offensive. Indeed, in practice, at no time in the post-1949 period did the prc fight a people's war. The pla's use of force was invariably at or beyond China's borders. For a more detailed discussion of the hypothesized roles of the symbolic discourse in strategic decision making, see Johnston, Cultural Realism, ch. 5. Back.

Note 64:  Wang Sanxin, Mao Zedong junshi sixiang yanjiu [Studies in Mao Zedong's military thought] (Beijing: National Defense University Press, 1988), p. 24. Back.

Note 65:  This did not exclude the initiation of violence when it appeared that conflict was imminent. In the parlance, Mao eschewed preventive war but not preemptive war. In his view, China's initial offensive in the Korean War was a second strike because the initial basis of conflict had been established by the U.S. presence on Taiwan and its threatening operations on the Korean peninsula; see Song, Mao Zedong junshi sixiang, p. 222. Similarly, according to a recent analysis, China's operations against Vietnam in 1979 were militarily offensive but politically defensive. "Concerning the counter-aggression nature of the national revolutionary warfare, a strategic counterattack carries the implications of a strategic offensive. From a political perspective, it makes more sense and is more advantageous not to call it 'attack' but to call it a 'counterattack.' For example, the February 1979 self-defense counterattack against Vietnam, from the perspective of military operations, we adopted offensive actions, but the essence of this type of offense was a self-defense counterattack" (Zhang Jing and Yao Yanjin, Jiji fangyu zhanlue qianshuo [An introduction to the active defense strategy] [Beijing: Liberation Army Press, 1985], p. 137). Back.

Note 66:  See Peng, "Junshi jianshe gaikuang," pp. 588-91. Back.

Note 67:  Ibid. Back.

Note 68:  Jonathan Wilkenfeld, Michael Brecher, and Sheila Rosen, Crises in the Twentieth Century: Handbook on Foreign Policy Crises  (New York: Pergamon, 1988), 2:161. Back.

Note 69:  On the crucial role of territorial disputes in the outbreak of interstate war, see Vasquez, The War Puzzle. Back.

Note 70:  The finding that the Chinese have been more crisis-prone than many other states would suggest, logically, that we should not expect the realpolitik strategic cultures of these other major powers to be as "hard" as that of the Chinese. Since I have not done a cross-national study of strategic cultures, I can't show this empirically. That China, compared to other states, appears to be more willing to consider political/diplomatic crises to be high-threat situations indirectly suggests, however, that many other states would not rank as high as it does on estimates of the inherent zero-sum nature of the strategic environment. This is a testable hypothesis. Back.

Note 71:  Indeed, in almost every variable relating to the type and scope of states' responses to a foreign policy crisis in the Brecher, Wilkenfeld, and Rosen data set, the Chinese cases indicate a higher level of militarized behavior than one finds in the population of states as a whole. (1) In 56.5% of foreign policy crises triggered by nonmilitary events, states as a whole responded with nonmilitary responses (N = 216). In 9.7% of the cases of a nonmilitary trigger, states responded with a violent response. In 0% of China's foreign policy crises triggered by nonmilitary events, China responded with nonviolent responses (N = 2). In 50% of the cases of a nonmilitary trigger, China responded with a violent response. (2) In 42.3% of foreign policy crises categorized as military-security-territory related, states as a whole responded with a nonmilitary response (N = 281). In 20% of China's cases categorized as military-security-territory related, China responded with a nonmilitary response (N = 5). (3) In 10% of foreign policy crises categorized as military-security-territory related, states used violence as the preeminent response (N = 281). In 40% of China's crises categorized as military-security-territory related, China used violence as the preeminent response. (4) In 11% of foreign policy crises categorized as political-diplomatic related, states used violence as the preeminent response (N = 127). In 33.3% of China's crises categorized as political-diplomatic, China used violence as the preeminent response. Back.

Note 72:  Thomas J. Christensen, "Threats, Assurances, and the Last Chance for Peace: The Lessons of Mao's Korean War Telegrams," International Security 17, no. 1 (1992): 122-57; Zhou Enlai, "Kang Mei huan Chao, baowei heping" [Resist America, Support Korean, protect peace] [950], in Zhou Enlai, Zhou Enlai Xuanji [Selected writings of Zhou Enlai] (Beijing: People's Publishing House, 1984), pp. 50-54. Back.

Note 73:  Thomas Stolper, China, Taiwan, and the Offshore Islands  (Armonk: M. E. Sharpe, 1985); Zhang, Deterrence and Strategic Culture, pp. 193-99. Back.

Note 74:  The secondary literature about China's military disputes also suggests that the Chinese began large-scale military operations against the target in most of these cases. See Christensen, "Threats, Assurances, and the Last Chance for Peace; Harlan Jencks, "China's Punitive War on Vietnam: A Military Assessment," Asian Survey 19, no. 8 (August 1979): 568-84; Gerald Segal, Defending China  (London: Oxford University Press, 1985). Back.

Note 75:  Mao Zedong, "Speech of 18 November 1957" [1957], in Michael Schoenhals, "Mao Zedong: Speeches at the 1957 'Moscow Conference,' " Journal of Communist Studies 2, no. 2 (June 1986): 115-24; Zhang, Deterrence and Strategic Culture, p. 229. Christensen argues that at that point Mao did not really believe socialism was prevailing over imperialism in material capabilities terms. Rather, the "East Wind" statements were projections of a future state of affairs if current technological and industrial trends continued. Moreover, Mao's upbeat pronouncements were part of a pep talk to the socialist camp in order to deflate the threatening nature of U.S. imperialism. See Christensen, Useful Adversaries. I think the evidence on this point is somewhat murky. To be sure, Mao believed that a certain bravado in the face of the adversary was necessary to rally popular morale and prevent fatalism and pacifism. This is in part what he meant by "despising the enemy strategically." On the other hand, in his November 1957 speeches he clearly puts socialism's material superiority in the present tense (Mao, "Mao Zedong: Speeches," p. 118). Later, in November 1958 after the Quemoy-Matsu crisis, Mao again disparaged imperialism's strength, remarking, "All evidence proves that imperialism adopts a defensive stance and has not undertaken the slightest offensive" (cited in Allen S. Whiting, "Mao, China, and the Cold War," in Yonosuke Nagai and Akira Iriye, eds., The Origins of the Cold War in Asia, p. 260 [New York: Columbia University Press, 1977]). Back.

Note 76:  In the U.S.-prc cases the U.S.-prc power ratio shifted from 2.58:1 to 2.33:1 in 1949-1950, from 2.11:1 to 1.91:1 in 1953-1954, and from 1.68:1 to 1.4:1 in 1957-1958. In the ussr-prc cases, the Soviet-Chinese power ratio shifted from 1.21:1 to 1.13:1 in 1968-1969 and from 1.2:1 to 1.16:1 in 1978-1979. The capabilities data come from the Correlates of War major powers capabilities data set. The percent shares are on a world base. These data are suggestive only; the shifts in ratios are quite small in some instances (though these can translate into substantial raw power resources). In some cases, however, these data do reflect Mao's unambiguous subjective calculation of whether the superpower involved was on the defensive or on the offensive. Back.

Note 77:  The U.S.-prc power ratio increased from 1.74:1 in 1960 to 1.84.1 in 1961. Back.

Note 78:  See Martin Wight, International Theory: The Three Traditions, ed. Gabriele Wight and Brian Porter (Leicester: Leicester University Press, 1991), for a discussion of realpolitik as one of the three Western traditions in international relations. Back.

Note 79:  This was a point made by Stephen Walt in his comments on this book project at the American Political Science Association meeting in New York, September 1994. Back.

Note 80:  In this respect, China's realpolitik strategic culture is not merely permissive, since it specifies exactly how one should react under different material conditions. Absent this realpolitik and we should expect very different reactions to similar measures of the material variables. Back.

Note 81:  Bruce Russett, Grasping the Democratic Peace: Principles for a Post-Cold War World  (Princeton: Princeton University Press, 1993); Bruce Bueno de Mesquita and David Lalman, War and Reason: Domestic and International Imperatives  (New Haven: Yale University Press, 1992); Randall L. Schweller, "Domestic Structure and Preventive War: Are Democracies More Pacific?" World Politics 44 (1992): 250-51. Back.

Note 82:  On the Spring and Autumn interstate system, see Richard L. Walker, The Multi-State System of Ancient China  (Hamden, Conn.: Shoe String Press, 1953). On the Han, see Yu Ying-shih, "Han Foreign Relations," in Denis Twitchett and Michael Loewe, eds., The Cambridge History of China, vol. 1, The Ch'in and Han Empires, 221 B.C.-A.D. 220  (New York: Cambridge University Press, 1986); and Thomas J. Barfield, The Perilous Frontier: Nomadic Empires and China  (Cambridge: Blackwell, 1989). On the Tang, see ibid. On the Five Dynasties, see Edmund Worthy, "Diplomacy for Survival: Domestic and Foreign Relations of the Wu Yueh, 907-978," in Morris Rossabi, ed., China Among Equals: The Middle Kingdom and Its Neighbors, 10th-14th Centuries  (Berkeley: University of California Press, 1983). On the Song, see Forage, "The Struggle for the Northwestern Frontier"; and Wang Gungwu, "The Rhetoric of a Lesser Empire: Early Sung Relations with Its Neighbors," in Rossabi, China Among Equals. And on the Ming, see Arthur Waldron, The Great Wall of China: From History to Myth  (Cambridge: Cambridge University Press, 1990); and Johnston, Cultural Realism. Back.

Note 83:  This is where Fischer's critique of Kratochwil's analysis of the feudal system runs into problems. See Markus Fischer, "Feudal Europe: Discourse and Practice," International Organization 46, no. 2 (Spring 1992). The feudal and Westphalian systems, like the unipolar imperial and multipolar state systems in China, cannot both be usefully classified as similarly anarchical. This stripped-down definition of anarchy implies that all historical "state" systems have been anarchical, from imperial, unipolar systems to sovereignist, multipolar systems, because even in the former there is no overarching authority governing relations between the empire and smaller states and tribes. As Mueller implies, this kind of conceptual stretching reduces the utility of the concept; see John Mueller, "The Impact of Ideas on Grand Strategy," in Richard N. Rosecrance and Arthur A. Stein, eds., The Domestic Bases of Grand Strategy, pp. 48-62 (Ithaca: Cornell University Press, 1993). Back.

Note 84:  Some realists will argue that the key link between anarchy and realpolitik behavior is uncertainty about the intentions and capabilities of others. This leads states to fear that others will exploit them and thereby threaten their survival, hence their reliance on self-help measures. So regardless whether the system is uni-, bi-, or multipolar, or some form of mixed-actor system, these are all technically anarchical, hence they all produce realpolitik behavior. Yet presumably fear is a variable--that is, it varies as uncertainty varies. Not all anarchical systems ought to produce the same degree of uncertainty and fear. Indeed, the more asymmetrical the distribution of power--e.g., the more the system resembles a unipolar imperial system--the less the imperial state needs to worry about the capabilities other states have (even if their intentions remain constant), and thus it can be more certain about its ability to survive in the face of external threats. One therefore ought to see less "self-help" realpolitik behavior on the part of the empire the more powerful and system-dominant it becomes. If one does not, this suggests that neither anarchy nor the (asymmetrical) distribution of power variables explains the persistence of realpolitik impulses. Back.

Note 85:  Indeed, the construction of an "old-fashioned" realist model is not as unproblematic as structural realists might assume, as a whole generation of excellent literature by scholars trained in neorealist analysis has shown. See Stephen M. Walt, The Origins of Alliances  (Ithaca: Cornell University Press, 1987); and Jack Snyder, Myths of Empire: Domestic Politics and International Ambition  (Ithaca: Cornell University Press, 1991). On the problems that realism has conceptualizing the relationship between structure and behavior, see Alexander Wendt, "Anarchy Is What States Make of It: The Social Construction of Power Politics," International Organization 46, no. 2 (Spring 1992): 391-425; Fareed Zakaria, "Realism and Domestic Politics: A Review Essay," International Security 17, no. 1 (Summer 1992): 177-99; Randall L. Schweller, "Neorealism's Status Quo Bias: Bringing the Revisionist State Back In" (paper presented at the National Security Seminar, Olin Institute for Strategic Studies, Harvard University, November 1993); and Robert Powell, "Anarchy in International Relations Theory: The Neorealist-Neoliberal Debate," International Organization 48, no. 2 (Spring 1994): 313-44. Back.

Note 86:  In this sense, anarchy for structural realists is analogous to norms, identities, ideas, and cultures for some constructivists. To the extent that these are constant they cannot in and of themselves explain variation in behavior. But both require the intervention of changing material variables to provide variation in the composite independent variable. For the former, fear, bred from anarchy, provides a particular interpretation of the meaning of these changes. For the latter, the interpretation depends on the content of the ideational constant. Realpolitik strategic cultures will give these changes a realpolitik hue; non-realpolitik strategic cultures, a non-realpolitik hue. Structural realists, however, will only predict realpolitik interpretations because anarchy can only breed uncertainty and fear. Thus they mispredict a great deal of non-realpolitik behavior in the world. This suggests that anarchy, then, does not have the determining effect on interpretations of relative capabilities, and thus, by definition, cannot account for the realpolitik behavior either. Back.

Note 87:  Mearsheimer, "The False Promise of International Institutions," pp. 9-10. Back.

Note 88:  Robert H. Jackson, "The Weight of Ideas in Decolonization: Normative Change in International Relations," in Goldstein and Keohane, Ideas and Foreign Policy, pp. 112-13. Back.

Note 89:  Goldstein and Keohane implicitly accept this conceptualization when they suggest that the way to test for the influence of ideas on behavior is to test the null hypothesis that behavior conforms to "egoistic interests in the context of power realities" (Goldstein and Keohane, "Ideas and Foreign Policy," p. 26). Their approach seems to me to be contradictory. On the one hand, they recognize that most material-interest arguments are indeterminate because of the problem of multiple equilibria and uncertainty. This is precisely why, they argue, ideational factors are important (e.g., in establishing focal points, for instance). But on the other hand, they insist that one has to construct an interest maximization null hypothesis to show that ideas are not epiphenomenal. Back.

Note 90:  G. John Ikenberry, "Creating Yesterday's New World Order: Keynsian 'New Thinking' and the Anglo-American Postwar Settlement," in Goldstein and Keohane, Ideas and Foreign Policy. p. 59. Back.

Note 91:  Wendt, "Anarchy Is What States Makes of It." Back.

Note 92:  A fruitful direction may be to examine the relationship between the formation of in-groups and out-groups, the requirements for legimating power structures within the in-group, and the devaluation of the out-group. If the out-group is viewed as dispositionally likely to act in devalued ways, then the shift from viewing an out-group as being somewhat suspect to seeing it as a hardcore enemy is a shift in degree, not kind. In-groups are predisposed to view out-groups as potential adversaries, and this provides the ideational basis for realpolitik behavior. Variation in the strength of in-group/out-group differences and changes in the composition of the in-group will explain variation in strength of realpolitik ideas and axioms and hence in realpolitik behavior directed at other actors. See Donald T. Campbell and Robert A. Levine, "Ethno-centrism and Intergroup Relations," in Robert Abelson et al., eds., Theories of Cognitive Consistency: A Sourcebook, pp. 551-64 (Chicago: Rand McNally, 1968); Murray Edelman, Political Language: Words That Succeed and Policies That Fail  (New York: Academic Press, 1977); Ernest G. Bormann, "Symbolic Convergence: Organizational Communication and Culture," in Linda L. Putnam and Michael E. Pacanowsky, eds., Communication and Organizations: An Interpretative Approach, pp. 99-122 (Beverly Hills: Sage, 1983); Jonathan Mercer, "Anarchy and Identity," International Organization 49, no. 2 (1995): 229-52; Daniel Druckman, "Nationalism, Patriotism, and Group Loyalty: A Social Psychological Perspective," Mershon International Studies Review 38, supp. 1 (1994): 47-48; and Marc Ross, The Culture of Conflict: Interpretations and Interests in Comparative Perpsective  (New Haven: Yale University Press, 1993), pp. 11-12, 40. Back.

Note 93:  Vasquez, The War Puzzle, p. 113. Back.

Note 94:  John A. Vasquez, The Power of Power Politics: A Critique  (New Brunswick, N.J.: Rutgers University Press, 1983), p. 216; Vasquez, The War Puzzle, p. 89. Back.

Note 95:  Evidence for the presence of non-parabellum strategic cultures and behavior suggests that the null hypothesis for a cultural realpolitik model is not structural realism but cultural liberalism, or cultural institutionalism, or some formal model that aggregates domestic political preferences, themselves ideationally rooted. If structural realism is incomplete without ideational assumptions about interests, then it is not really an alternative to a cultural realpolitik model. In other words, any competitive hypothesis testing of a parabellum strategic culture model should be done with other ideational models, not against structural realism. This conclusion suggests that many of those involved in the "norms, ideas, and culture" project are right to reject the ideas-interest dichotomy in favor of an ideas-ideas dichotomy. Back.

---

The Culture of National Security