[253]

A variety of scholarly approaches and methods of investigation have been used by social scientists and historians to discover and analyze long-range patterns of international conflict. By 'long-range' I mean both long-term (diachronic) and worldwide (synchronic) comparative analysis. Origins, cycles, transformations, fusions, collapses, and fluctuations are some of the social patterns that are analyzed on a large array of long-range analysis, covering thousands of years and all regions of world politics, from local to global. In spite of the acknowledged complexity of these processes and systems, however, little attention has focused on critical conceptual, empirical, and theoretical distinctions that are crucial to establish, as between different scales of conflict in long-range analysis. Conflicts that occur on different scales (e.g. the French and Indian War versus the French-British War; the Six Day War versus the Arab-Israeli War; or die Korean War versus the Soviet-American Cold War) are generally assumed to obey different causal mechanisms, unless proven otherwise (scale invariance). From an applied policy perspective, the scale of conflict also matters, often critically so, for purposes of early warning, intervention, or resolution. The Vietnam War affected an entire generation; but the Soviet-American Cold War affected several generations. Scale matters in international conflict, just as it does in all complex phenomena.

In this chapter I discuss and propose some solutions to the puzzle of scales in the long-range analysis of war. The thrust of my approach centers on several key distinctions regarding two critical scales of conflict: the behavioral scale of violent belligerence and the chronological scale of historical processes. These distinctions and some of the concepts that accompany them are new and they may help clarify not only the nature of the long-range approach to conflict, but also - more importantly - shed some light on the character of war. The main implications of these ideas are (i) new methods for systematically recording the empirical long-range incidence of war, essential for constructing more insightful theoretical explanations, and (ii) a more precise framework for developing causal explanations and a better understanding of long-range patterns. Preliminary application of these ideas already suggests that the approach is feasible and insightful, both cross-sectionally (synchronically) and cross-temporally (diachronically).

[254]

Premises The occurrence of war in human history is an ancient, complex, and on-going phenomenon, particularly when the topic is approached from a long-range comparative perspective, meaning both cross-culturally and cross-temporally. By way of measurement and conceptual background, the operational definition of war used in the Long-Range Analysis of War (LORANOW) Project is as follows:

This definition satisfies several important criteria for conducting comparative analysis � e.g. cross-temporal and cross-cultural applicability (Bartolini 1993) � as I have discussed elsewhere in detail (Cioffi-Revilla 1991; 1996). Such a definition and others with similar empirical resolution (Levy 1983; Richardson 1960;

Small and Singer 1982; Wright 1942) can record conflict events such as the First Punic War, World War I, and the Korean War. However, it fails to record similarly significant conflict events such as the Arab-Israeli War, or the Soviet-American Cold War. Why? Why cannot the latter class be measured in the same way as the former? How does the measurement instrument fail to record the latter events? More precisely, what are some of the behavioral characteristics of the latter set of conflicts that make them difficult or impossible to detect by the standard definitions of war? How can long-range analysis contribute toward the solution of this problem? Answers to these questions can provide insights for designing better measurement instruments capable of detecting and measuring the class of larger-scale conflicts exemplified by the Arab-Israeli War and the Soviet-American Cold War. For example, improved detection and measurement could have anticipated the surprising end of the Cold War.

A long-range theory of war should be based on a distinction between interrelated but different dimensions or scales of conflict phenomenon:

1 the scale of belligerence with respect to the behavior under investigation (the subject matter), and

2 the scale of process chosen as a temporal framework for the investigation (the analytical focus).

From a theoretical perspective, these scales are used to distinguish 'the trees' from 'the forest.' The first dimension of war phenomenon, what I call 'the scale of belligerence,' refers to the nature or type of conflict that is manifested by the actors within a broader system of social interactions. To define the scale of belliger-

[255]

ence one must first answer a set of substantive questions, such as: What constitutes belligerence? Which kinds of events are to be considered and what are to be excluded? Is it possible to avoid the traps of nominalism or common language conventions used in historiography and move beyond uninformative designations such as 'the War of X against Y,' 'the X-Y Crisis,' 'X's Conquest of Y,' or 'the War of Y's Disintegration'? As I explain in greater detail below, defining the appropriate scale of belligerence in a systematic fashion is essential for a rigorous treatment and theoretical explanations of the long-range dynamics of war.

The second dimension of the war phenomenon, what I call 'the scale of process,' refers to the chronological focus or degree of temporal resolution that one wishes to have on the sequential detail of history. This is important to define because historically some events of belligerence take place over the relatively short time span of months or even (e.g. the Franco-German War, World War I, the Korean War), even with high magnitude, whereas others can last decades (e.g. the Great Peloponnesian War, the Punic Wars, the Thirty Years War, the Soviet-American Cold War). Does the endurance characteristic of some wars make any difference for understanding their causes and properties? Theoretically, we know that duration alone affects the set of opportunities available to belligerents in carrying out repeated engagements (Cioffi-Revilla and Starr 1995). Defining the appropriate scale of process is therefore just as important for conducting a valid and systematic long-range analysis of war.

Given these premises, I propose to differentiate between two distinct phenomena on the scale of belligerence - 'war' and 'warfare' - and two distinct phenomena on the scale of process - 'macroprocesses' and 'microprocesses.' The differences between these phenomena in each scale, as well as their corresponding theoretical implications, are both qualitative (structural) and quantitative (measurable). Accordingly, as detailed in the next sections, macroprocesses cover the universal history of organized lethal belligerence, from origins to the present, whereas microprocesses cover particular or detailed histories of warfare, from one war to the next.' These relative differences in analytical scales are analogous to the differences that exist in the following disciplines and phenomena:

[256]

The hierarchical nature of macro and microprocesses is obvious and familiar. What may not be so obvious is that - unfortunately - the social science of international relations (or social science in general for that matter) lacks what cosmology and climatology provide for astronomy and meteorology: long-range context to understand basic principles and system change. This is the central scientific objective of the long-range analysis of war: to discover and understand dia-chronic and synchronic patterns of war through comparative analysis. As with all real world complex phenomena that evolve in historical time, the long-range analysis of war must account for phenomena in both macroprocesses and micro-processes, the latter embedded within the former.

In common language there is no well-established distinction between the terms 'war' and 'warfare'; they are synonyms. In long-range analysis, however, the scale of belligerence matters and refers to theoretical and empirical aspects of violent social conflict having to do with a set of key dimensions, such as: (1) behavioral units of analysis, (2) type or degree of organization of the conflict, (3) complexity or size, (4) chronological duration, and (5) multidimensional consequences of belligerence (political, economic, cultural). Table 13.1 illustrates the defining features of these two types of belligerence along the common set of characteristic dimensions just described, distinguishing between 'war' and 'warfare.' As a taxonomy, the distinction is admittedly idealized and the comparative dimensions (i.e. units of behavior, organization, complexity, and so on) are complementary, not mutually exclusive. This theoretical and empirical classification of violent conflicts by scale of belligerence (i.e. war versus warfare) is intended to be taxonomic; the various characteristic dimensions are not.

In terms of scale of belligerence, and consistent with the definition given earlier, a war (as opposed to 'warfare') is therefore conceptualized as a relatively disaggregated and simple event that generally lasts no longer than some years (10² - 10³ days). Wars are commonly classified as either civil (internal), interstate, or internationalized conflict (e.g. Small and Singer 1982; Geller and Singer 1998), as well as other types discussed later. Most social science and historical research on violent social conflict has focused on this type of belligerence - wars.²

Warfare, on the other hand, is an aggregate, relatively more complex process that generally lasts at least decades and is far less systematized than wars (10³ � 10⁴ days). Social scientists and historians have focused relatively less attention on this larger scale belligerence - 'warfare' - although arguably this is the scale that has greater impact on the long-range development of societies and civilizations. If war is what �forges' a state, warfare is what forges an entire civilization.

War and warfare differ along the following dimensions of scale of belligerence:

Unit of behavior War and warfare differ by the referent analytical unit of behavior being investigated. As suggested by the historical examples in Table 13.1, a war is

[257]

Table 13.1 Scale of belligerence � war and warfare

Source: Prepared by the author.

more like an event that occurs over a relatively short term (e.g. the French and Indian War or the Korean War). By contrast, warfare is a process that tends to involve a long-term interaction, as an enduring conflict or rivalry (e.g. the French-British war or the Soviet-American Cold War). Since several events constitute a process, it can therefore be said that several wars can constitute warfare. Thus, in the examples given in Table 13.1, the Spanish-Aztec War was part of a larger process of warfare called the Spanish Conquest of the New World. Similarly, the Korean War was part of the Cold War, the latter viewed as a case of warfare that also involved several other wars (e.g. the Angolan War, the Vietnam War), and even sub-war events (e.g. the 1948 Berlin Crisis, the 1962 Cuban Missile Crisis, the 1973 Alert Crisis, etc.) and civil wars (Cuban Revolution, Hungarian Revolt, Czech Uprising, Angolan Civil War). Wars are to warfare as events are to a process; the former composes the latter. Warfare can assume the form of a rivalry (Diehl 1998), when wars and disputes, occur over time among the same belligerents. However, as I discuss later, not all forms of warfare are rivalries (e.g. a succession of wars of expansion or conquests that produce an empire).

Organization Wars and warfare also differ by degree of organization. Compared to warfare, a war is a more disaggregated conflict, consisting of battles or engagements and usually involving fewer belligerents. By contrast, warfare has aggregate organization, sometimes involving numerous belligerents. Thus, the Maya-Spanish War, the Aztec-Spanish War, the Inca-Spanish War, and other separate wars fought between Spain and the various Indian belligerents aggregate to form the warfare known as the Spanish Conquest of the New World. Each of the New World wars had separate termination. The aggregate warfare, however,

[258]

ended when Spain finally conquered the last sovereign Indian enclave - at the Battle of Tayasal, Guatemala, in AD 1697, marking the end of the Maya-Spanish War. Similarly, England, France, Portugal and - later - the United States also carried out warfare for conquest or expansion in the New World. The recent Cold War was also a case of warfare involving the aggregate belligerence between Western allies, the Soviet allies, and sometimes Third World 'neutrals.' A war constitutes a relatively disaggregated form of belligerence, while warfare is relatively more aggregated, with a conceptual relation similar to that between trees and forests, such that

Wars : warfare :: trees : forest.

Complexity War and warfare also differ by degree of complexity. A war is relatively simple, compared to warfare, because a war is most frequently organized as a two-sided conflict (Richardson 1960; Wilkinson 1980), whereas in warfare the conflict commonly involves many sides, not just two, and sometimes involves alliances (Neilson and Prete 1983; Starr 1972). Spain's New World belligerence between the late fifteenth century and the late seventeenth century was a far more complex, many-sided conflict than any of the individual wars between Spain and the Indians. Similarly, Franco-German warfare (and rivalry) during the past two centuries has involved many complex interactions, including shifting systems of alliance, territorial changes, repeated crises, and repeated European and colonial wars. Empirically, the greater complexity of warfare can be recorded by appropriately designed graph-theoretic parameters capable of measuring adjacency, centrality, thickness, connectivity, vertex-degree and other significant dimensions.

Duration The preceding differences imply that wars and warfare must necessarily also differ in terms of their duration. The duration of a war is most frequently measured in months or years. In antiquity many wars consisted of single-battle engagements (Ferrill 1985/1997; Gabriel and Metz 1991; Hackett 1989; Hassig 1992; Humble 1980; Keeley 1996; Marcus 1995; Montgomery 1968). In more recent modern times few wars have lasted longer than a few years (Beer 1983; Cioffi-Revilla 1995; Levy 1983; Small and Singer 1982).

By contrast, warfare generally lasts decades and sometimes can endure for centuries. The following are some well-established cases of multi-century warfare in antiquity: Sumerian-Elamite, Egyptian-Hittite, Egyptian-Israelite, Assyrian-Elamite, Chinese-Xiong Nu, Japanese-Korean, Greek city-states, Roman-Carthagenian, Maya city-states, Maya-Toltec, Aztec and Peruvian Coastal and Highland warfare. For modern times, cases of long-duration warfare include the following: Franco-German, Anglo-German, Russo-German, Sino-Russian, Russo-Japanese, and Iranian-Iraqi warfare. Were it not for several fundamental changes in the national identity of the belligerents, the following are perhaps some of the closest parallels that can be traced between antiquity and modernity in terms of protracted or long-range warfare:

[259]

The purpose of the above comparison is not to draw any premature inferences, but simply to suggest that warfare - not just isolated wars - has a long-range record in human history, even if the nature or official institutional name of some belligerents has undergone considerable change (e.g. from archaic states with primitive economies to industrialized states with modernized economies). Warfare is also quite often a multigenerational phenomenon, thereby engaging a larger population of belligerents for a much longer period of time and social mechanisms for reproducing belligerence. Wars often have only a transitional impact on a society; whereas warfare usually shapes a society; affecting its basic political culture, defining friends and enemies through cognitive schema, and not merely affecting the polity and economy.

Typology Finally, given the preceding theoretical and empirical distinctions, wars and warfare also differ by their typology. Wars are most often classified into civil wars and interstate wars. In turn, civil wars include purely domestic wars and internationalized wars, whereas interstate wars include major-major power wars, minor-minor wars, and major-minor wars. Other categories of wars are also used, such as great power wars, hegemonic wars, world wars, colonial wars, and others (Cioffi-Revilla 1995; Levy 1990; Midlarsky 1988; Thompson 1988). Warfare, on the other hand, is not as systematized and - regrettably - typologies of warfare have not yet attracted the same amount of attention as typologies of war. A typology of warfare is long overdue, given the emerging long-range record of warfare in history (Cioffi-Revilla and Lai 1995). A typology of warfare is also necessary for measurement and modeling purposes.

I propose the following typology of warfare based on the measurement and analysis experience thus far developed at the Long-Range Analysis of War (LORANOW) Project: (1) protracted warfare, (2) integrative warfare, (3) disintegrative warfare, and (4) sporadic warfare. Each pattern of warfare is illustrated by the graphic models in Figure 13.1.

Protracted warfare is defined as a sequence of recurring wars between the same belligerents fighting for similar objectives (same ostensible causes) (e.g. Mesopotamian�Elamite warfare, Greek-Persian warfare, Punic Wars, Chinese-Xiong Nu warfare, Franco-German warfare, Arab-Israeli warfare, Iran-Iraq warfare). Interstate rivalries (Diehl 1998; Thompson, 1999) often consist of this

[260]

Figure 13.1 Typology of warfare

[261]

type of warfare � when they manifest not just disputes but actual combat as well.

Integrative warfare - a violent conflict process that begins with several belligerents and eventually ends with one victor - is a sequence of wars in which the same belligerent wins in repeated wars with different belligerents, annexing or conquering the territory of the vanquished side (e.g. Alexander's campaigns, wars of Roman expansion, Aztec warfare, Spanish Conquest of the New World, American�Indian wars, War of Italian Unification). Many states and perhaps all empires (Taagepera 1968) have been forged by this type of warfare.

Disintegrative warfare - the reverse process of integrative warfare, beginning with one belligerent actor and ending with many - is a sequence of wars in which the original unitary belligerent becomes increasingly fragmented by secessions (e.g. warfare of Roman imperial disintegration, wars of Spanish, British, and French imperial disintegration, warfare of Yugoslavian fragmentation). Integrative and disintegrative types of warfare involve opposite processes of political consolidation (fusion) and dispersion (fission), respectively - a complex dual process that remains poorly understood but can be readily identified through long-range analysis.

Sporadic warfare is a series of unrelated wars fought among different belligerents for a period of time, typically over different objectives. The 1991 Gulf War may represent a case of sporadic warfare (assuming it ended with the 1991 cease-fire agreements), although as long as relations between Iraq and the UN allies remain hostile it may be considered part of an ongoing protracted conflict with fluctuating hostility and occasional outbreaks of violence (cruise missile attacks, punitive, bombings, economic warfare, etc.). Until extensive long-range measurement is undertaken we will not know the composition of total warfare in terms of sporadic or related types (protracted, integrative, disintegrative).

Note that the graphic structure of each pattern of warfare is distinct. This property - the characteristic graphic signature of warfare - is important because it means that the graphic structure itself can be used to identify and classify the appropriate pattern regardless of conventional or nominal historical designations (nominalism), as shown below.

Figures 13.2 and 13.3 illustrate the use of this typology for systematically describing warfare in ancient China and Mesoamerica using long-range analysis. Figure 13.2 describes the observed patterns of Chinese warfare in the East Asian system, from the Legendary period to the end of the Western Zhou dynasty, a period of two millennia, from 2700 BCE to 722 BCE (Cioffi-Revilla and Lai 1995). Figure 13.3 describes the observed patterns of Maya warfare in the Mesoamerican system, based on a preliminary data set used here only for illustrative purposes (Cioffi-Revilla, Chupik and Resnick 1998). This type of graphic representation is called a chronographic model and it can be used for representing a variety of long-range conflict patterns extending over millennia and including many (or all) belligerents in a given system. A chronograph can also be derived from a geographic information system (GIS) containing time-dependent war data (Jones 1997; Starr 1998).

[262-263]

Figure 13.2 Chronograph of Chinese wars and emergent long-range patterns of warfare in the ancient East Asian system, 2700-722 BC.

[264�265]

Figure 13.3 Chronograph of Maya wars and emergent long-range patterns of warfare in the ancient Mesoamerican system ca. 800 BC-AD 700.

[266]

Although vastly different in their cultural context and civilizations characteristics, some long-range patterns of warfare in these ancient systems are strikingly comparable in several significant respects - an empirical property that cannot be easily detected without appropriate conceptual categories and measurement instruments that help distinguish 'the forest' from 'the trees.' As shown in Figure 13.2, the Chinese patterns for the eastern Asian system exhibits several cases of protracted warfare (the Chinese fighting against the barbarian Xiong Nu), integrative warfare (conflicts that unified China under the Xia, the Shang, and the Zhou), and cases of sporadic warfare (Yin-Pi Shi War of 1945 BCE, Xu-Jing War of 966 BCE). In Figure 13.3, the Maya pattern for the Mesoamerican system also shows cases of protracted warfare (several Maya states engaging in repeated wars), little if any integrative warfare (unlike the Chinese, the Maya never attained a unified political system), and several cases of sporadic warfare.

Similar charts describing the long-range structure of war and warfare - parallel regional chronographs - are being developed for other space-time regions of the world historical system. Additional recurring types of warfare beyond the above four elementary types may be discovered in the future, and such a set of patterns may eventually constitute a true taxonomy (exhaustive and mutually exclusive), as opposed to just a typology (as is now for interstate wars). A scale of such types (minimally nominal in level of precision) can then be used in long-range analysis to measure and model warfare in new ways that would advance our understanding.

Beyond its intrinsic importance for developing a long-range theory of war, the preceding discussion of the scale of belligerence - viz., the distinction between war and warfare - also provides an essential premise for discussing the other conceptual dimension of conflict, the scale of process. From a theoretical perspective, here again I distinguish between two different phenomena that are distinct in terms of several key dimensions, such as (1) analytical focus, (2) referent systems of actors, (3) historical process, (4) time scale, and (5) coevolutionary mechanisms. Table 13.2 uses this set of dimensions to define the two types of processes according to their scale, distinguishing between 'microprocesses' and 'macroprocesses.'

Macroprocesses have a universal or global focus on the world system, or largest regional system, examining the history of polities from their earliest known origins to the present. The time scale of a macroprocess is marked by centuries and millennia. On such a scale, theoretical attention must focus on the origins and subsequent development of warfare, with less emphasis on individual wars. The original areas of warfare correspond to the Fertile Crescent (Palestine, Mesopotamia, and Egypt), East Asia (Yellow River and Blue River basins), Mesoamerica (Gulf of Mexico Coast and Oaxaca Valley), and South America (Peruvian Coast and Highlands).³ In the scale of macroprocesses it is large-scale social entities such as civilizations (Quigley 1979; Wilkinson 1991), international or regional systems (Scott 1967), peer�polity interactions systems (Renfrew 1986), macro social

[267]

Table 13.2 Scale of process - macroprocesses and microprocesses

[268]

systems (Cioffi-Revma 1998a), world-systems (Chase-Dunn and Hall 1997), and supraregional groupings of cultural horizons (Willey 1991), or archaeological traditions (Peregrine 1998) that undergo the developmental processes of 'rise and fall,' accompanied by warfare and other defining processes. The pleogenic model - human history experienced multiple pristine origins in terms of the rise of polities and warfare � provides a paradigm for understanding macroprocesses of conflict and political development (Cioffi-Revilla 1991, 1996, 1998b). The principal puzzles in long-range macroprocesses concern the antiquity of the processes themselves (when did they begin?), multiple origins (pleogenesis: where did they begin?), and the understanding of parallel evolution, coevolution, fusion points, post-fusion turbulence, dynamics of political dominance, and similar topics. To date, relatively few social scientists and historians have examined this macro scale, given the challenging empirical and theoretical difficulties.

By contrast, microprocesses are more detailed or localized, focusing on systems with fewer belligerents, examining history from one war to the next, and using a shorter time scale. Here the social entities involved are mostly tribes, chiefdoms, states, alliances, empires, and other similar ('sovereign') belligerents that rise, fall, and are affected by wars. The detailed chronographic models used in Figure 13.2 and 13.3 earlier (Chinese and Maya wars in East Asian and Mesoamerican systems, respectively) provide graphic representations of microprocesses. The main puzzles concerning microprocesses are patterns of war escalation, onset, magnitude, duration, contagion, diffusion, and other similar phenomena. A relatively large community of social scientists has focused many investigations on these microprocesses, producing a growing corpus of new scientific knowledge (Bremmer and Cusack 1994; Cioffi-Revilla 1995; Diehl 1998; Geller and Singer 1998; Midlarsky 1989, 1998; and Vasquez 1993).

Macroprocesses and microprocesses of conflict differ along the following dimensions:

Focus Macroprocesses and microprocesses of conflict are phenomena that differ primarily by their theoretical and empirical focus. The focus of macroprocesses is on the universal or global history of social systems on the largest possible scale, including their belligerence. The macro-level focus is sometimes - not always -identified with a primarily diachronic perspective. By contrast, the focus of microprocesses is far more detailed and is sometimes identified with a primarily synchronic perspective. Numerous investigations exemplify the micro-level focus, such as individual works on the Persian wars, the Punic wars, the Napoleonic wars, or the wars of the French Revolution. Note that both levels - macro and micro � can take a long-range approach, but the details of the referent process will vary with the focus. In terms of relative historical granularity, the macro-level is coarse-grained whereas the micro-level is fine-grained.

System Macroprocesses and microprocesses also differ by the kind of referent system that is the object of long-range investigation. The type of system examined in macroprocesses refers to the entire world or at least its major regions or macro

[269]

social systems (e.g. Fertile Crescent, East Asia, Europe, America, Polynesia). The 'world historical system' approach is also in this macro-level tradition (chapters in this volume; Chase-Dunn and Hall 1997; Dark 1998; Frank and Gills 1993; Thompson 1983), following earlier landmark work by historians and social scientists that provided foundations for long-range analysis (Braudel 1987; Dawson 1958; McNeill 1993; Sorokin 1937; Toynbee 1934-61; and Wright 1942). By contrast, microprocesses examine smaller systems of belligerents that experience wars and other events during their history (e.g. Chinese warring states, Rome and Carthage, Classic Maya Lowlands, Soviet-American Cold War system). The identification of the referent system is essential for defining the appropriate scale of process.

History Macroprocesses and microprocesses differ by the relative length of the historical process being examined. In the case of a macroprocess, the time domain covers from the earliest origins of warfare to its present state. In terms of individually identifiable historical events, this is a process with a history of at least 6,000 years, beginning with the earliest Sumerian wars (Cioffi-Revilla and Sommer l993). Considering the first evidence of warfare Jericho pre-pottery B level) the process would extend back at least another 5,000 years (Kenyon 1979; Roper 1975). By contrast, in a microprocess the time scale is such that the process ranges from one war to the next and is therefore much more fine-grained (Cioffi-Revilla 1997; Vasquez 1987).

Time scale Because of the preceding differences in scale, macroprocesses and microprocesses are also measured according to different clocks or time scales. In general, macroprocesses are marked by events that are spaced apart by centuries or millennia (e.g. the 'ages' of human history associated with stone, copper, iron, industry, and information). By contrast, microprocesses use a time scale that is marked by more discrete events (wars, political transitions, etc.) that are usually spaced apart by years or decades (e.g. the 'epochs' of human history marked by various dynasties and regimes). These different time scales make a great difference in terms of the conflict process being considered.

Rise and fall Keeping in mind the relative nature of macroprocesses and micro-processes, the nature of the social units that evolve through the process of rise, development, and fall represents another major difference. In macroprocesses what rise and fall are the largest identifiable social units that have ever existed in human history: civilizations, cultures, international systems and other macro social systems identified earlier. By contrast, what rise and fall in a microprocess are much smaller social units, although sometimes they can define a larger area as a dominant center: tribes, chiefdoms, states, alliances, federations, empires, and other components of the larger macro social systems within which these entities interact. Clearly, the types of social units that are involved in conflict � rising, enduring, and eventually declining - make a significant difference in terms of the scale of process being examined.

[270]

Belligerence Significantly, macroprocesses and microprocesses also differ in terms of the scale of belligerence examined in each process. Macroprocesses provide a more appropriate framework for the investigation of warfare, given the coarser focus, the broader system of belligerents, and the longer chronological timeline. Most individual wars can look like minute events in the context of a macroprocess, except for those that have large system-transforming effect (e.g. some world wars). By contrast, microprocesses focus more effectively on wars, given the more detailed focus, the relatively smaller sets of belligerents, and the shorter chronological timeline. Considering the two scales combined, macroprocesses are to warfare as microprocesses are to war.

Graphics Given these differences in scales, different descriptive approaches are required for graphically modeling the two types of processes (Tufte 1983, 1990, 1997). Macroprocesses may be best represented by the pleogenic model containing multiple warfare origins (pleogenesis), parallel lines of development, several points of fusion, fissions and other developmental characteristics of warfare, as discussed earlier (see Table 13.2; also Cioffi-Revilla 1998b: Figure 1). Micro-processes are best represented by a chronographic model containing the set of belligerents, alliances, different types of wars, and other defining characteristics of microevolution (see Figures 13.2 and 13.3 and Table 13.2). Unfortunately, effective graphic methods for scientific investigation are still notoriously lacking in social science in general, and in fields like international relations in particular. Long-range analysis can contribute significantly toward developing such new tools.

Puzzles Logically, macroprocesses and microprocesses must differ by the kind of research puzzles being addressed. Macro-level puzzles include the following: how ancient is warfare? When, how, and why did warfare first occur in human history? Did warfare first occur in one location and subsequently diffuse to others, or did warfare initiate spontaneously (ex novo) in several pristine areas (pleogenesis)? Why did political complexity arise sooner but slower in the Old World, as opposed to later but faster in the New World? When and how does fusion between two or more systems occur? What determines one belligerent dominating others at a fusion point? How do various segments or subsystems (Mesopotamia, China, Mesoamerica) compare in terms of their patterns of war and political change? From an interdisciplinary perspective, these puzzles are similar to those raised in fields like cosmology or climatology, as opposed to the puzzles raised in astronomy or meteorology, respectively. Puzzles are scale-specific. By contrast, micro-level puzzles are more synchronically focused and include the following: How can we account for patterns of war onset, magnitude, or duration? What are the correlates of war? What makes some disputes escalate to war and others not? How can we account for the war-to-war process? Conversely, from an interdisciplinary perspective, these puzzles are like those raised in astronomy and meteorology, as opposed to cosmology and climatology. These and other puzzles depend on the scale being considered.

[271]

Coevolution Finally, macroprocesses and microprocesses differ by the type of coevolutionary mechanisms that may operate within each process. At the macro-level the main coevolutionary process of warfare consists of (a) parallel systems undergoing their own history elsewhere in the world (literally, 'parallel worlds') and (b) other macro-level processes within each evolving area (e.g. changes and transformation in the economic system or in political culture). Some parallel worlds of the ancient past (e.g. West Asia and East Asia at 3000 BCE) may have had weak and remote ideational contact by means of a diffused long-distance 'information net' (McNeill, this volume). However, such systems were otherwise disconnected in terms of any significant political, economic, or military interactions. Other parallel worlds in the Old World and the New World did not even share such weak interactions (e.g. Mesoamerica and Mesopotamia). Parallel worlds merged occasionally - thereby ending their separation - at a small number effusion points.

By contrast, in the scale of microprocesses that which coevolves with warfare are the associated political, economic, cultural, and demographic processes that are defined within a referent system of belligerents. For instance, in a process of integrative warfare (Figure 13.1, column 2), the political system grows, as does the economy and - usually but not always - the territorial reach of the expanding polity Warfare and politics also coevolve at this level, producing the first complex forms of political development and organized warfare (Brumfiel and Earle 1987; Brumfiel and Fox 1994; Cioffi-Revilla 1998b; Marcus and Flannery 1996).

In this chapter I identified the puzzle of scales in the long-range analysis of conflict in world historical systems and proposed some solutions. To address the puzzle, I focused on the critical distinction between different levels of belligerence (scale of belligerence) and the different levels of historical evolution (scale of process). Different scales of belligerence - the 'war' versus 'warfare' distinction -not only highlight the significance of these different phenomena but also offer a systematic typology for empirical classification, measurement, and modeling of conflict. The value of this approach is enhanced in the area of comparative analysis, both cross-culturally and cross-temporally; an area that is essential in any truly long-range investigation.

Similarly, different scales of process - the 'macro' versus 'micro' distinction -focus attention on significantly different historical mechanisms that must also obey different principles because their scale is so different. Moreover, the two scales of process - different clocks of human history - also present vastly different sets of research puzzles. For instance, whereas traditionally political scientists have been more interested in the microprocesses of wars, archaeologists, anthropologists, and world historians have dedicated more attention to the macroprocesses of warfare. The approach I propose in this chapter may also contribute to me type of interdisciplinary collaboration mat is becoming increasingly necessary to achieve real scientific advances in our understanding of the long-range evolution of war, its causes and effects on human history.

[272]

I am grateful to Jacek Kugler and William R. Thompson for their comments on an earlier version of this chapter.

1 Military historians and military operations research analysts often examine the next, lower level of detail in terms of scales of analysis, given by campaigns or field opera-dons (Battilega and Granger 1984; Dockery and Woodcock 1993; Fen-ill 1985/1997; Hughes 1984; Jones 1987; Margiotta 1984). Thus, consistent with the scales of long-range analysis: (i) several batdes or engagements compose a campaign; (ii) several campaigns compose a war; and (iii) several wars compose warfare.

2 See, e.g. Blainey (1975), Bremer and Cusack (1994), Cioffi-Revilla (1995), Dupuy and Dupuy (1993), Eckhardt (1992), Geller and Singer (1998), Keegan (1993), Levy (1983), Margiotta (1984), Midlarsky (1988), Richardson (1960), Small and Singer (1982), Vasquez (1993), and Wright (1942).

3 For recent authoritative works on these pleogenic regions see Burger (1995), Chang (1980), Frangipane (1996), Haas el al. (198 7), Keightley (1983), Liverani (1988), Marcos and Fl'annery (1996), Rothman (1999), Sharer (1994), and Willey (1991).