Sie sind auf Seite 1von 222
Complexity Complexity in in World World Politics Politics Concepts and Methods of a New Paradigm
Complexity Complexity
in in World World
Politics Politics
Concepts and
Methods of a
New Paradigm
Edited by
Neil E.


SUNY series in Global Politics

James N. Rosenau, editor


Concepts and Methods of a New Paradigm

Edited by Neil E. Harrison

State University of New York Press

Published by State University of New York Press, Albany

© 2006 State University of New York

All rights reserved

Printed in the United States of America

No part of this book may be used or reproduced in any manner whatsoever without written permission. No part of this book may be stored in a retrieval system or transmitted in any form or by any means including electronic, electrostatic, magnetic tape, mechanical, photocopying, recording, or otherwise without the prior permission in writing of the publisher.

For information, address State University of New York Press, 194 Washington Avenue, Suite 305, Albany, NY 12210-2384

Production by Judith Block Marketing by Michael Campochiaro

Library of Congress Cataloging-in-Publication Data

Complexity in world politics : concepts and methods of a new paradigm / edited by Neil E. Harrison. p. cm. — (SUNY series in global politics) Includes bibliographical references and index. ISBN 0-7914-6807-0 (hardcover : alk. paper)

1. International relations—Philosophy.

2. International relations—Methodology.

3. Complexity (Philosophy)

I. Harrison, Neil E., 1949–

II. Series.

JZ1305.C657 2006



ISBN-13: 978-0-7914-6807-4 (hardcover : alk. paper)

10 9 8 7 6 5 4 3 2 1


1 Thinking About the World We Make Neil E. Harrison


2 Complexity Is More Than Systems Theory Neil E. Harrison with J. David Singer


3 Complexity and Conflict Resolution Dennis J. D. Sandole


4 Understanding and Coping with Ethnic Conflict and Development Issues in Post-Soviet Eurasia Walter C. Clemens, Jr.


5 Beyond Regime Theory: Complex Adaptation and the Ozone Depletion Regime Matthew J. Hoffmann


6 Agent-Based Models in the Study of Ethnic Norms and Violence Ravi Bhavnani


7 Alternative Uses of Simulation Robert Axelrod


8 Signifying Nothing? What Complex Systems Theory Can and Cannot Tell Us about Global Politics David C. Earnest and James N. Rosenau






When Worlds Collide: Reflections on the Credible Uses of Agent-Based Models in International and Global Studies Desmond Saunders-Newton



Complex Systems and the Practice of World Politics Neil E. Harrison




List of Titles, SUNY series in Global Politics






Thinking About the World We Make

Neil E. Harrison

Despite nearly a hundred years of theorizing, scholars and practitioners alike are constantly surprised by international and global political events. The abrupt end of the much-studied Cold War was widely unanticipated, as were the conse- quences of the collapse of communism in Europe. The defining characteristics of four decades of international politics were erased in a few short years, but the globalization of economic and social life has continued. The 1997 Asian finance crisis rattled the US and European stock markets, civic strife in Venezuela influ- ences the price of oil, and the needs of AIDS patients in South Africa challenge international agreements on intellectual property. Out of the blue, terrorists at- tacked within the United States one sunny September morning. A year earlier, in the space of a few months the global economy lurched from rapid expansion to recession and flirted with deflation. After so much ink has been spilt, we still know so little about international relations and world politics that events continue to surprise us. There is no agree- ment on the cause of this failure. Some believe that international theorists think too small and fail to synthesize relevant insights from a range of disciplines (Buzan and Little 2001); others criticize the emphasis on positivist methods (Smith, Booth, and Zalewski 1996); and postmodern scholars reject the ahistorical, ratio- nalist foundations of most international theory (Der Derian and Shapiro 1989; George and Campbell 1990). This book takes a different tack. It argues that the reality of world politics is more complex than dreamt of in current theories. Current theories of world politics assume that the social world is appro- priately modeled as a simple system; this book proposes that it should instead be viewed as a complex system. In this book my colleagues and I describe, and demonstrate the benefits of, a paradigm of system emergence from complex




agent interactions that we call “complexity”. The study of complexity in systems

is “complexity science” and descriptive, explanatory, or predictive theories—

formal statements that generate empirically testable hypotheses—based in com- plexity ideas and concepts are “complex systems theories.” 1 Like realism, complexity is a thought pattern, set of beliefs, or ideological

orientation about the essence of political reality that organizes theorizing about and empirically investigating events in world politics. Realism assumes that es- sential human characteristics drive political behavior within fixed structures; complexity views politics as emerging from interactions among interdependent but individual agents within evolving institutional formations. So world politics

is a more or less self-organizing complex system in which macroproperties

emerge from microinteractions. This and the next chapter outline a taxonomy of the central ideas and concepts of a complexity paradigm of world politics from which useful theories or models of complex world politics may be constructed. This ontological shift from simple to complex systems opens new paths to knowledge and understanding yet incorporates much current knowledge; it val- idates novel research methods; and theories founded in this approach will gen- erate radically different solutions to policy problems. In the next section, I compare basic concepts of simple and complex systems and thereby frame a com- plexity paradigm. Following that, I show how complexity concepts can be used in theories of world politics. In the final section of this chapter I outline the rest of the book.


A system is a portion of the universe within a defined boundary, outside of

which lies an environment. An atom is a system, as is an animal or a country. Usually, the definition of the boundary is a convenience used to assist human analysis, as when scientists define for study an individual ecosystem. A pond is only arbitrarily separated from its shoreline, the air, and the Sun. Similarly, a de- finition of “country” may be in terms of its recognized sovereign territory, its ter- rain and ecosystems, its economy (where the distinction between gross national and gross domestic product is important), or its state or government. A system is simple if the units and their relations are relatively fixed, per- mitting reasonable prediction of future system states. An automobile may be complicated, but it is a simple system. Each of the parts has a specific role in the system, and the actions of all the parts are centrally coordinated toward a collec- tive outcome. The existence of workshop manuals further illustrates the sim- plicity of the system: they identify all potential problems and explain how to



remedy them. They also illustrate problems and solutions, define the character- istics of each part, and the range of relations between them in exhaustive detail. As table 1.1 shows, a living system is complex in many ways that an auto- mobile is not. The two primary differences between complex and simple systems are diversity and decentralization. In an automobile there are many diverse parts constructed for very specialized roles, but there is centralized coordination of their operations through mechanical or electronic management systems. In liv- ing systems, not only are the units diverse but each has a range of freedom of choice denied to parts in a mechanical system. Because units in a complex system have discretion in their choice of behavior, they are commonly called “agents.” Decentralized decision-making increases complexity. One measure of com- plexity is the length of the shortest possible message that fully describes the sys- tem (Gell-Mann 1994, 30–38). Description of a jaguar in the jungle is longer than of a quark (a unit within an atom). If all the units of a system are identical, system description is shorter; only one unit need be described in detail. Thus, heterogeneity among the units increases description length. 2 But if the units also have behavioral discretion, system description requires description of the units (perhaps by class), of the range of their available choices, and of the rules of be- havior that each will follow in making their individual choices. Centralization of decision-making simplifies complicated systems. Mod- ern automobiles have sophisticated management systems that use miniature


Simple Systems

Complex Systems

Few agents Few interactions Centralized decision-making Decomposable Closed system Static Tend to equilibrium Few feedback loops Predictable outcomes Examples:

Many agents Many interactions Decentralized decision-making Irreducible Open system Dynamic Dissipative Many feedback loops Surprising outcomes Examples:


Immune systems




Molecules in air

Boyle’s law


Gravitational system




computers to govern feedback cycles and responses to environmental changes. Although these systems respond almost instantaneously to multiple indicators, there is only a single programmed response to any change in system condition. These centralized management systems prohibit freedom of choice in the units. In living systems, decision-making is decentralized, and units can choose their actions. Bacteria have fewer choices of behavior than ants, which are, in turn, more regimented and less “free” than animals. Mammal societies are more complex than anthills or bacterial infections. As the “degrees of freedom” of choice for individual members in a system increase, the range of individual behaviors increases, making the system more complex. The common assumption, usually implicit, that a system is simple rather than complex simplifies analysis. If the system is simple, it can be decomposed into its parts. It is nothing more than its parts and their defined relationships. The auto- mobile can be disassembled and reconstructed and work just as well as it did before. Disassembling a living system, or removing any part of it, can destroy the system or, at least, make it much less than it was previously. Only Victor Frankenstein has yet been able to deconstruct and reconstruct a human and breathe life into it. The desire to simplify analysis also leads to the common assumption that the system under study is closed to other systems, does not exchange energy with them, and is not affected by them. The desire among social scientists for closed systems reflects their common admiration for the analytical control of the labo- ratory sciences. The laboratory is designed to close the system under study. Boyle’s law that pressure and temperature are inversely related can be demon- strated to be true only within a closed cylinder within a controlled environment. Unfortunately, social systems are always open, and wishing them closed often makes assumption of closure unreasonable. Simple systems usually are static and tend to equilibrium; complex systems are always dynamic and they are dissipative. This is most clearly illustrated by the “arrow of time” (Prigogine 1997). Without an input of energy, a simple system can remain largely unchanged for long periods. It declines only marginally by interac- tion with its environment (to that extent, it is an open system). The automobile is a static system that remains in equilibrium if no energy (for example, gasoline and human control) is added to the system. In contrast, a living system perpetually changes. Humans age and die, a dynamic process of constant change in the cells within our bodies and the relationships between them. And we are “dissipative structures” because we have to draw energy from our environment in the form of oxygen, food and water merely to stay alive (Prigogine and Stengers 1984). Even in simple systems, effects can feedback on their causes. Negative feed- back slows down processes, and positive feedback speeds them up. The thermo- stat is the classic example of a simple system with a negative feedback loop. As



the air cools below the set-point temperature, an electrical circuit closes to turn on the furnace and blow hot air into the room. When the air is returned to its set point, the circuit opens and the furnace shuts down. The homeostatic be- havior of animals reflects feedback from activity (hunger, hunt, satiation, sleep). Environmental selection operates on the individual agent as a form of feedback; behavior can change from punishment/reward contact with the environment. Complex systems usually have multiple feedback loops. Positive feedback loops strengthen the cause and the subsequent effect in an ever increasing cycle that can lead to nonlinear transitions and system collapse. For example, atmos- pheric scientists hypothesize that positive feedback loops caused Venus’s swirling toxic mists and 900-degree surface temperatures (Schneider 1989). Some scien- tists fear that climate change on Earth could also progress with a nonlinear shift in the system (Ocean Studies Board et al. 2001). Complex systems are unpredictable. By its nature, nonlinearity is unpre- dictable and difficult to represent mathematically, and most complex systems are potentially nonlinear. In complex systems, prediction as a path-dependent ex- trapolation of historical processes runs the risk of nonlinear change. Beyond the very short term, the range of possible system paths for a complex system widens dramatically. Decentralized decision-making and diversity among agents permits a wide range of agent actions and openness to changes in environmental condi- tions (the state of another complex system), and the prevalence of positive feed- back loops inject further uncertainty into the system under study. Complex systems may not be predictable, but they may be simulated with interacting rules for agent behavior. These rules may be few and simple, yet the outcome of their interaction can simulate complex systems in which agent be- havior appears random and system order seems accidental. For example, the flocking behavior of birds looks random and disorganized but can be modeled with three rules (Waldrop 1992, 241–43). The location of water temples in Bali can be simulated with a few rules of kinship and farming practice (Lansing, Kre- mer, and Smuts 1998). The collapse of the Anasazi civilization in the American Southwest has been explained by the interaction of social rules and environ- mental changes (Axtell et al. 2002). In comparison to an automobile, the game of checkers seems uncomplicated. Yet it “provides an almost inexhaustible vari- ety of settings (board configurations)” (Holland 1998, 76). Because complexity emerges from the simple rules of checkers, we should expect that “complexity will be pervasive in the world around us” (76) both natural and social. But it also “gives hope that we can find simple rule-governed models of that complexity.” That hope is partially fulfilled by simulations of social systems with agent-based models in which systems are modeled from the interactive behavior of essential agents, as described throughout this book.



The characteristics summarized in table 1.1 and described in this section are most commonly associated with each genus of system, simple or complex. No single descriptor defines either simple or complex systems. For example, simple systems may have many and diverse parts and complex systems (e.g., of bacteria) may have homogeneous units; and complex systems can be, at least temporarily, in equilibrium, while some simple systems appear dynamic. However, the more descriptors of one system genus that can be attributed to a specific system, the greater is the probability that that system is of that genus. Thus, complexity is an accumulation of the characteristics of complex systems. The next section shows how complexity concepts can be used to construct a complex systems taxonomy of world politics (that is further elaborated in chapter 2).


Intuitively, the social world seems complex in the sense described here, but cur- rent theories of world politics model it as a simple system. As Ruggie (1993) com- ments, world politics theories are “reposed in deep Newtonian slumber.” Newton described a universe formed out of particles that were all made from the same material and whose movements in absolute space and time were governed by forces that followed unchanging and universal laws. These laws could be expressed exactly through mathematics (Capra 1982, 65–67; Ruggie 1993). For example, the properties of gases can be reduced to the mathematically describable motion of their atoms or molecules. Thus, the image is of a universe constructed like a perfect mechanical watch. Science, aided by mathematics, was the method for prizing open the watch case to see the workings inside (Hollis and Smith 1990, 47). Locke and other early political and social theorists enthusiastically emu- lated Newton and attempted “to reduce the patterns observed in society to the behavior of its individuals” (Capra 1982, 69). A fixed human nature was pre- sumed to determine human behavior, and “natural laws” governed spontaneous human society: “As the atoms in a gas would establish a balanced state, so human individuals would settle down in a society in a ‘state of nature.’” Natural laws included freedom, equality, and property rights (Locke 1980, 123–27; Kym- licka 1990, 95–159). The shadow of Newton’s universe continues to obfuscate knowledge in the social sciences. For example, while neoclassical economics remains the dominant explanation of economic phenomena, it is “an economic science after the model



of mechanics—in the words of W. Stanley Jevons—as ‘the mechanics of utility and self-

interest’” (Georgescu-Roegen 1975, emphasis in the original). Economic actors are

assumed to be rational in their pursuit of undefined, subjective self-interest. Their behavior is assumed to be an objectively rational response to external forces such

as the level of supply and demand of goods and services. If supply exceeds de-

mand and prices fall, economic actors will increase their purchases. In such a model, agency is limited to only economic interests and programmed responses

to external stimuli. Recent debates about agency and structure do not hide the similarly mech-

anistic paradigm that still drives orthodox theories of world politics. Essentially identical units—interests and identities are assumed to be exogenously formed— are driven by “natural laws” to behave predictably in response to exogenously de- termined conditions. A rational-choice approach, borrowed from neoclassical economics, is used in an attempt to generate ahistorical, universal explanations

of relations between states. The result is several significant simplifications of re-

ality. For example, concentrating on the state as the unit of analysis causes an an- alytically convenient but arbitrary separation of international and domestic politics, and the theoretical focus on “explaining constancies, not change” privi- leges structure over agency (Smith 2004). Constructivist theories—the most recent incarnation of liberalism—posit that state interests and identities are intersubjectively malleable at the margin through interaction with other states. While it is now historically located within international society, as in rational-choice theories, the state remains the unit of analysis. Thus, I start with the state to better illustrate the primary concepts of a complexity taxonomy of world politics.


A complex system is commonly described as more than the sum of its parts.

That is, properties of the system are emergent, created by the interaction of the

units. The basic unit of any social group is the individual. In biological terms, the human body is a system; socially, each human is an essential unit within sev- eral systems, and any social group, including the state, is an emergent system. Social and political institutions emerge from the interaction of individual hu- mans and human groups. Groups may be local or national; they may be loose- knit coalitions or adhesive groups of fervent followers, and may be more or less centrally organized. Out of the interactions among this mélange of groups and individuals emerges the set of institutions, people, and practices that scholars call the “state.”



Open Systems

The state is not a closed system: it is open to other natural and social systems. For example, defined as a political system, it is open to technological, cultural, and economic systems that influence political choices and processes (Skolnikoff 1993; on political economy, Gilpin 1996 and Strange 1994, 1996 among many others). 3 The state also is open to other states and, as constructivism argues, is in- fluenced by interactions with them. Some social systems are both within and outside the state. For example, unions, major corporations, and nongovernmental organizations (NGOs) cross boundaries and operate in several jurisdictions simultaneously (Goddard, Passé- Smith, and Conklin 1996; Korten 1995; Keohane and Nye 1971). Although the state is evidently an open system, theories of world politics conventionally assume that all systems are closed to their environment much as optimal natural science experiments are controlled and isolated from unwanted external influences. Despite occasional attempts to bring in domestic politics (Evans, Jacobson, and Putnam 1993; Putnam 1988), the state is usually modeled as a unit with exogenous identity and objective interests. This greatly reduces the range of possible causal explanations for any perceived social event, simplifying causal analysis and hypothesis generation and testing. The assumption of closure thereby permits historical theorizing and sup- ports the widespread belief among scholars that general laws can be found. This would be impossible if social systems were modeled as open, because “constant conjunctions (empirical regularities) in general only obtain under experimentally controlled conditions”—that is, under closure (Patomäki and Wight 2000). Open systems are “susceptible to external influences and internal, qualitative change and emergence” (232) and “outcomes might be the result of many different causes and the same cause might lead to different outcomes” (229). Small changes that can initiate a radical system shift may come from a change in environmental con- ditions, or from inside, from interactions among its constituent agents. The non- linearity of open systems prevents the theorist from mapping specific causes to observed effects. Thus, open dynamic systems are inherently unpredictable (Doran 1999). But that is no reason to model them as closed systems.


The state is both an emergent system and a unit within the international system of states. In Holland’s (1995) terminology, they are “meta-agents” whose “inter- nal models” (discussed below) emerge from the interaction of domestic agents. State behavior then results from the interaction of internal model and external



reality, and feedback is available on whether internal system processes and state behavior “fit” within the environment, not unlike the concepts elaborated in Putnam’s (1988) two-level game, though in a more fluid and dynamic relation- ship. The concept of meta-agents can be used in any issue area in which agents and actions at more than one level of aggregation are involved. In contrast, orthodox international relations (IR) theory usually takes the state as the primary unit of interest, while recognizing in passing the potential in- fluence of substate and nonstate actors. Constructivism and other cognitive the- ories treat states as subjects, but the state still is assumed to be a unitary actor whose identity and interests change primarily as a result of interaction with other states (Wendt 1994). The extent to which states also may self-consciously change their interests and identity is debated, but the potential for change as a result of domestic political discourse is usually disregarded (Hasenclever, Mayer, and Ritt- berger 1997, 186–92).

Internal Models

Each human agent, the essential unit of any social system, has an internal model of his or her desires and beliefs about how to achieve those desires in the world. 4 If their beliefs are out of synch with reality, they will act inappropriately, fail to achieve their goals, and may be punished. Agents who learn from such an expe- rience, change their internal models and, thus, their behavior. Individual agents’ behaviors are responsive and purposeful but not objec- tively rational. According to Elster (1986, 16), an action is rational if it is the best way for an actor to satisfy his or her desire based on beliefs that are optimal given the available evidence and as much information as possible, given the desire. Be- liefs and desires must be free of internal contradictions. Finally, actions must be the intended result of beliefs and desires. This is substantially the same description of rationality used by Green and Shapiro (1994, 6) to explain the foundations of rational-choice theory. However, by assuming diversity among agents, complexity does not make the simplifying jump to an assumption of objective rationality. Each agent can have unique desires and unique beliefs about how to achieve them. The alignment of behavior with desires and beliefs indicates agent rationality, but there is no assumption that the outcomes of an agent’s choices will be individually or col- lectively rational or will match agent intent. This is not Simon’s “substantive con- ception of rationality” quoted and approved by Keohane (1988): “‘behavior that can be adjudged objectively to be optimally adapted to the situation.” Because agents cannot predict the effects of their actions in complex systems, behaviors of individual agents are “optimally adapted” to their situation only accidentally. Rationality is subjective—within the agent—rather than objective.



Constructivism broadens “the array of ideational factors that affect inter- national outcomes” and introduces “logically prior constitutive rules alongside regulative rules” (Ruggie 1998). The concept of internal models potentially ex- tends the ideational content of world politics theories, while at the same time making analysis of agent motives more difficult. However, simulation of agent behavior now is possible. Internal models drive agent behavior, but those models may change when

tested in a selective environment. Agents that consistently act in ways that are se- lected by their social environment as suboptimal face eradication. Because states are themselves systems, the process of matching internal model to external reality

is one of trial and error. As Putnam (1988) has suggested, the state may be not be

able to move its internal model—particularly in terms of its (causal) beliefs about what is possible—to accord with the reality of the international system. If all states are adaptive complex systems, then the international system emerges from coevo- lution. International norms influence behavior through the internal process of in- ternal model formation, one component of which is the desire to participate in a society of nation-states. In chapter 5, Hoffmann investigates how states changed their beliefs during negotiations over regulation of ozone depleting substances and how the internal model of the United States adapted to these changes.

Dynamic Systems

Superficially, complexity appears to have some affinity with other world politics systems theories, like neorealism and world systems. As Waltz (1979, 91) de- scribes the international system, it is “formed by the coaction of self-regarding

units,” and its structure is “formed by the coaction of their units” and “emerge[s]

International-political systems, like economic

markets, are individualist in origin, spontaneously generated, and unintended.” However, the similarity is more perceived that real, as shown in more detail in chapter 2. The distinction is in the details: in conventional systems theories, struc- ture is a fixed or only slowly changing determinant of agent behavior. In complex systems, structure is dynamic but “organization” is fixed. The “organization of a

from the coexistence of

living system is the set of relations among its components that characterize the system as belonging to a particular class (such as bacterium, a sunflower, a cat, or


human brain)” (Maturana and Varela 1980, 18). To describe the organization,


is only necessary to describe the relationships and not the components. For ex-

ample, self-organization is “a general pattern of organization, common to all liv- ing systems, whichever the nature of their components.” The structure of a complex system is the actual relations among actual physical components: “[I]n



other words, the system’s structure is the physical embodiments of its organiza- tion.” While organization is static—a cat cannot become a dog—structure is dy- namic. Thus, structure is not fixed but a fleeting embodiment—in social systems manifested by institutions—of the deep organization within apparent chaos. As the momentary embodiment of prior agent interactions, complex system struc- ture changes dynamically. In complex systems, structure has a social role but no purpose. In func- tional social theories like constructivism and neoliberalism, “history is path- dependent in the sense that the character of current institutions depends not

only on current conditions but also on the historical path of institutional devel- opment” (March and Olsen 1998, 959). Because “rules, norms, identities, orga- nizational forms, and institutions that exist are the inexorable products of an

surviving institutions are seen as uniquely fit to the environ-

ment, thus, predictable from that environment” (958). Complexity science makes no such assertions: it does not assume or judge the fitness or efficiency of emergent institutional arrangements. Institutions and rules are the consequence of history but may not fit agents’ purposes. The common (usually implicit) assumption that the international system is homeostatic is a stronger version of the orthodox presumption that events in dif- ferent spatiotemporal locations may be compared. It is equally untenable. Simple dynamic systems find a point of equilibrium that is “sustained by micro-mecha- nisms operating in finely attuned and compensating ways” (Elster 1983, 31–32). Despite its “balance of power” bromide, classical realism is really about the processes of systemic change from dynamic forces. Realism presumes that just as the neoclassical market continually returns to an equilibrium between demand and supply, the international system returns to a balance between many forces. Complex social systems are never homeostatic: in both markets and world politics the frequent and temporary equilibrium points are always distinct phe- nomena. Each state of balance, like a human standing still through tensions be- tween opposing muscles, is a fleeting event within a specific set of conditions, a point on a path of change. The dynamic European system has found several mo- mentary points of balance between myriad forces. Tudor England understood the need to change alliances to continually balance power in Europe. Though power was balanced in Europe before World War I and in the Cold War, the conditions were unique to each period.

efficient history


The uncertainty of complex social systems calls into question conventional world politics assumptions about causation. Conventional world politics



theories presume that causation is proximate and proportionate. Like most of social science, they have adopted Hume’s rules for causal explanations (Hume 1975). 5 These rules require that the cause can be shown to precede the effect, that cause and effect are contiguous (there was no intermediate event), and that there is a “necessary connection” between events such that this cause can be shown to always precede this effect under consistent conditions. For at least four reasons, these rules are not appropriate causal explanations in complex so- cial systems. First, they only apply in closed systems in which conditions can be controlled. But if social systems are open, it is unlikely that conditions will re- main constant or be comparable between different states of affairs. In an open system, a cause may have different effects at different times due to changed con- ditions. Therefore, it is not surprising that no general laws of world politics have been found. Second, social systems are so complex that parsimonious the- ories that attempt to isolate single (or few) causes for observed effects may dan- gerously oversimplify models. In complex social systems, the events noted at the start of this chapter (among others) are surprising only when we expect to find a singular cause. Understood as the emergent consequence of multiple inter- acting prior events, such events are less astounding. The events of September 11, 2001, may be the result of all of the explanations commonly offered: failures of collection, coordination, and distribution of intelligence; a clash of cultures; hatred by fanatics; and so on. But each of these “causes” were themselves caused by multiple prior events. Osama Bin Laden is the product of his family, Islam, the Saudi culture, and personal experience defending Afghanistan against the Soviets. The clash of cultures (or civilizations: Huntington 1993) is as much a consequence of U.S. actions as of Muslim choices. Intelligence failures resulted, in part, from decisions that restricted human intelligence gathering, decisions made by successive US governments after several high-profile misadventures in the 1970s. Thus, September 11 could have emerged from a plethora of choices and events across the globe over decades, not as an inevitable consequence of any of them but as a path-dependent phenomenon. And if it was not path de- pendent, it was a symptom of a nonlinear system shift that cannot be predicted or explained. In neither case is conventional thinking about causation useful. Third, the immediate cause of an effect may, as part of a higher-order Markov chain, itself be the effect of an earlier, and possibly more important, cause. If an earlier cause is more important than later ones in the chain, this implies action at a distance in space/time that both Newton and Hume reject (Elster 1983, 26–30). Fourth, causation may be simultaneous as in open, emergent systems where the interaction of parts of the system constitute the system. In addition to these limits to the normal rules of causal explanation, as- sumptions of the proportionality of cause and effect are often erroneous. As dis-



cussed above, in open, emergent systems, small perturbations in the system may have very large effects, making identification of the connection between cause and effect nearly impossible and explanation problematic. Was the fact of Kaiser Wilhelm’s withered arm or his relationship with his English nanny a sufficient cause of World War I (Röhl 1998)? In a complex system, many factors symbiotically cause an effect. Theorists should look to the evolution of the system, not to individual events, for causes of observed effects. Patomäki and Wight (2000, 230) argue that “ontologically, the social world can only be understood as a processual flow that is intrinsically open and subject to multiple and at times contradictory causal processes.” Uninten- tionally, this is a fair exposition of complex systems. Social phenomena only occur because agents act within an existing and real context that is “not re- ducible to the discourses and/or experiences of the agents,” as constructivists argue. As Maturana and Varela (1980, 98) wrote: “[O]ur problem is the living or- ganism and therefore our interest will not be in properties of components, but in processes and relations between processes realized through components.” In social systems, processes are not as automatic as they are in insects and bacteria. Humans and social groups are conscious and self-aware entities (that is, their in- ternal models are more elaborate and complex) who, therefore, may act strategi- cally toward some goal within their perception of their environment.


Most social sciences have begun to embrace complex systems concepts. Ideas from thermodynamics coupled with a concern for economic systems’ environ- mental effects (Georgescu-Roegen 1975, 1971) led to the development of eco- logical economics that specifically models the economy as an open system (Barbier, Burgess, and Folke, 1994; Krishnan, Harris, and Goodwin 1995; Costanza 1991; Daly 1991). Brian Arthur and others have identified the pres- ence and effect of feedback loops in economic systems (van Staveren, 1999; Arthur, Durlauf, and Lane 1997; Arthur 1990; Arthur 1989; Anderson, Arrow, and Pines 1988; Romer 1986). Complex systems approaches have attracted soci- ological interest (Luhmann 1998, 1990; Eve, Horsfall, and Lee 1997; Knapp 1999; Hanneman 1988; Collins, Hanneman and Mordt 1995) and touched pub- lic administration and organization studies (Griffin 2002; Stacey 2001; Marion 1999; Elliott and Kiel 1999). Even political science is not immune (Richards 2000; Axelrod 1997; Jervis 1997; Cilliers 1998; Cederman 1997; Cederman and Gleditsch 2004), though efforts are disparate and inchoate. This book is de- signed to drive forward the complexity research agenda as a viable alternative to orthodox theories of world politics by establishing the central concepts and ideas



needed for the development and empirical assessment of complex systems theo- ries of issue-areas in world politics. The next nine chapters further develop the concepts outlined in this chap- ter and illustrate their application to several world politics issue areas. Chapter 2 begins to sketch out a taxonomy of complexity by comparing complex systems concepts to those developed more than three decades ago for a general systems taxonomy. Systems theories that were relatively popular in the early days of the Cold War have, in recent years, have fallen into disrepute as overly “grand” in purpose. Harrison, with Singer’s aid, compares and contrasts conceptual de- scriptions between general systems and complex systems taxonomies. Several concepts are common to the two approaches, but this chapter also identifies the important differences between the two taxonomies. Complexity is not a warmed- over version of general systems theory but builds on its ideas to generate theories that better explain issue-areas in world politics. As this is a new approach to understanding world politics/IR, this book does not attempt to illustrate its application to the whole range of possible is- sues. The next four chapters show how complexity can generate new insights and hypotheses when applied to selected issue areas. They are arranged from the least to most technical in their use and application of complexity con- cepts. Because this book is an introduction to complexity in IR that is in- tended to initiate research rather than to develop applications adapted to all issue-areas of international relations, these chapters are only exemplars of the application of the complexity paradigm. None formally models their case but they all describe how their hypotheses might be further elaborated or empirically tested. In chapter 3, Dennis Sandole argues that complexity creates opportunities to integrate and synthesize apparently opposing worldviews. He reconsiders the- ories of identity-based conflict in the post-9/11 world and proposes a theoreti- cal framework to demonstrate that traditionally competing Realpolitik and Idealpolitik (conflict resolution) approaches can coexist. Not only can they co exist, but more robust guides to identify conflict and formulate policy responses can be constructed by integrating both approaches into a single framework. In chapter 4, Walt Clemens attacks a knotty puzzle that has emerged from the collapse of the Soviet empire: why have some ex-Soviet states fared far better than others? Natural resources, education, and ethnic homogeneity do not ex- plain why the Baltic states and Slovenia are joining the European Union, while oil-rich and more-homogeneous states are embroiled in factional fighting or war, or have stagnated in neo-Stalinism. Using complexity concepts, Clemens pro- poses an innovative explanation of why some newly freed states appear to have failed while others are joining the EU.



Drawing on complex adaptive systems theories (a version of complexity that uses more life science concepts), Clemens notes that some states were “fit- ter” than others and so better able to exploit opportunities that opened for them after the collapse of the Soviet empire. Seeking the sources of that fitness, he finds that long-standing, religiously inspired institutions in the Protestant coun- tries developed internal models in the population that reduced ethnic tensions and increased acceptance of democratic virtues. He also shows that his marker for fitness correlates with measures of development and describes how to empir- ically test his hypothesis. Matt Hoffmann looks at the coevolution of states’ internal models in chapter 5. He considers two puzzles in the formation of the interna- tional regime designed to protect the stratospheric ozone layer: why did the norm governing participation change and why did the United States accept this new norm? Hoffmann shows that rational explanations are deficient and that complex systems concepts can help us to unravel both puzzles. From a complexity perspec- tive, evolution of the universality norm is a simple story of complex adaptation. As some Southern Hemisphere states’ internal models changed to universal par- ticipation, the flux in the system eventually led other states to adapt to the new in- ternational norm. Hoffmann shows that when the United States reconsidered its internal model (with some pressure from domestic groups), it recognized it would have to accept the universality norm and negotiate in good faith with the South to achieve its goal of an effective treaty. He concludes with suggestions for theo- retical, empirical, and methodological development of these ideas. In recent years, genocide within a country has become an international issue. The stimulus to this international interest in domestic interracial relations was the terrifying genocidal violence in Rwanda in 1994 that killed possibly as many as eight hundred thousand people. In addition to the moral implications, since Rwanda it is now clear that genocide in one country has serious conse- quences for its neighbors, making it a legitimate concern for the international community (“The Road Out of Hell,” Economist, March 27, 2004, 25–27). In chapter 6, Ravi Bhavnani shows how complexity concepts can help us to under- stand why the speed and magnitude of the killing was so much greater than in all previous ethnic attacks in that country. Conventional explanations of the scale of the Rwandan violence are inad- equate. Bhavnani shows how bottom-up simulations can generate new hypothe- ses about the spread of ethnic violence. Building on evidence from the field and reasonable assumptions about relationships between extreme and moderate Hutus, he describes a simulation of how the killing rampage took hold so quickly and led to murder even within families.



The next three chapters explain the empirical validity of simulations, discuss potential problems with constructing complex systems theory, and show how multiple ABMs may be used to improve forecasting and decision-making. Chapter 7 is a reprint—used with permission—of part of Robert Axelrod’s chapter entitled “Advancing the Art of Simulation in the Social Sciences,” in Simulating Social Phe- nomena, edited by Rosario Conte, Rainer Hegselmann, and Pietro Terna (Berlin:

Springer-Verlag, 1997), 21–40. Axelrod argues that simulation is best thought of as a new way of doing social science. Inductive methods are needed to find patterns in, for example, opinion surveys and macroeconomic data, and sometimes in in- ternational interactions. If social agents are assumed, as in conventional theories, to be objectively rational actors, deductive methodology suffices. Simulation is the third way—the only way, if agents are assumed to be adaptive. In the social sciences, the most common form of simulation is agent-based modeling (ABM), which builds systems from the bottom up rather than, as with deductive methods, from the top down. Like deduction, simulation starts with explicit assumptions, but it cannot prove theorems. Like induction, it looks for patterns, but it uses data gen- erated from defined rules rather than the real world. Axelrod argues that, in social science simulations, simple is better: like thought experiments simulations can deepen understanding of fundamental processes. David Earnest and Jim Rosenau in chapter 8 question whether political systems are complex systems, as commonly understood, and argue that simula- tion of political systems begs the questions it attempts to answer: who are the ac- tors and who has authority? They reject complexity as a theory, because it fails the standard of theory in positivist epistemology and offers no alternate episte- mology; and implicitly they reject more limited applications of complex systems theory. While Axelrod describes simulation as a third way, Earnest and Rosenau argue that it is no way: it lacks both the empirical appeal of induction and the disconfirmative value of deduction. For them, thought experiments are “much ado about nothing.” They acknowledge that complexity is an attractive paradigm but argue that more development is required before it may generate viable theo- ries of world politics. Chapter 9 is an indirect response to Earnest and Rosenau’s critiques. In Desmond Saunders-Newton’s opinion, while there are epistemological problems with ABMs, these are neither insurmountable nor critical problems. As scholars debate the fine points of ontology and epistemology, complex systems thinking and ABMs already are being put to use in the service of policymakers to generate and assess multiple policy options. Saunders-Newton argues that complex systems thinking and computa- tional methods that emphasize agent-level phenomena are part of a new trans- discipline that allows analysts to rigorously consider increasingly complex



phenomena in an interdisciplinary way. In addressing the epistemological issues surrounding computational methodologies, he argues that efficacy or usefulness is more important than the quality of model isomorphism and method. He then describes how several computational social science models (including ABMs), in- tegrated with computer-assisted reasoning methods, are being used in the Pre- Conflict Management Tools Program being tested at the National Defense University for its ability to assess social vulnerability. The concluding chapter draws some general lessons from the four cases and shows how they illustrate important complexity concepts. It then assesses the va- lidity of simulations and computational models and the epistemological ques- tions they raise. It also shows that, from complexity concepts and ideas, complex systems theories for issue-areas can be specified and models for specific problems generated. Yet, because political systems are complex—and becoming more com- plex—a new epistemology and new methods are needed to understand them. Fi- nally, it shows that recognition of complexity in politics suggests radically new policies for addressing international problems and pursuing national interests.


1. The terms “world politics” and “international relations” are used in-

terchangeably throughout this book. I use “world politics” to better reflect the multilevel structure of the political world to which complex systems thinking is so well adapted. Patomäki and Wight (2000, 232–33) opine that the “key error” of much international theorizing is “to treat levels of the state and the interna- tional system as related as agents to structures” instead of as “layers” within world politics. The terms “complex system” and “complex adaptive system” are often used interchangeably; the concepts described here principally derive from the latter. I use the term “paradigm” in the sense of a set of assumptions, con- cepts, values, and practices that comprise a view of reality, and in that sense it is quite comparable to “worldview” (Hughes 2000).

2. Common knowledge also shortens description. “Bicycle” conveys to

most people a clear image of the system. Imagine how much more complex would be a description to a Martian who is completely unfamiliar with a bicycle or any of the common parts used in its assembly.

3. Smith (2004) comments that world politics/IR theorists err in thinking

of the state as solely political. Whether the state is modeled as political interact- ing with other subsystems of society or as a political unit of a social system

among economic and cultural agents depends on the question being addressed.

4. In this book, the terms “internal model,” “mental model,” and “schema”

are used interchangeably.



5. As a skeptic, Hume also argued that causation is a human construct. All we ever see is the conjunction of events, and we impute a causal relationship. But without an explanatory force linking cause and effect, causation cannot be “real” (Patomäki and Wight 2000).


Anderson, Philip W., Kenneth J. Arrow, and David Pines, eds. 1988. The Econ- omy as an Evolving Complex System. Santa Fe Institute Studies in the Sci- ences of Complexity, 5. Redwood City, CA: Addison-Wesley.

Arthur, W. Brian. 1989. “Competing Technologies, Increasing Returns, and Lock-in by Historical Events.” Economic Journal 99 (March 1): 116–31.

———. 1990. “Positive Feedbacks in the Economy.” Scientific American 262, no. 2 (February): 92–99.

Arthur, W. Brian, Steven N. Durlauf, and David A. Lane, eds. 1997. The Econ- omy as an Evolving Complex System II. Reading, MA: Addison-Wesley.

Axelrod, Robert M. 1997. The Complexity of Cooperation: Agent-Based Models of Competition and Cooperation. Princeton, NJ: Princeton University Press.

Axtell, Robert L., Joshua M. Epstein, Jeffrey S. Dean, George J. Gumerman, Alan C. Swedlund, Jason Harburger, Shubha Chakravarty, Ross Ham- mond, Jon Parker, and Miles Parker. 2002. “Population Growth and Collapse in a Multiagent Model of the Kayenta Anasazi in Long House Valley.” Proceedings of the National Academy of Science of the United States of America, supp. 3, vol. 99, no. 10:7275–79.

Barbier, Edward B., Joanne C. Burgess, and Carl Folke. 1994. Paradise Lost: The Eco- logical Economics of Biodiversity. London: Earthscan for the Beijer Institute.

Buzan, Barry, and Richard Little. 2001. “Why International Relations Has Failed.” Millennium: Journal of International Studies 30, no. 1:19–39.

Capra, Fritjof. 1982. The Turning Point. New York: Simon & Schuster.

Cederman, Lars-Erik. 1997. Emergent Actors in World Politics: How States and Na- tions Develop and Dissolve. Princeton, NJ: Princeton University Press.

Cederman, Lars-Erik, and Kristian Skrede Gleditsch. 2004. “Conquest and Regime Change: An Evolutionary Model of the Spread of Democracy and Peace.” International Studies Quarterly 48, no. 3 (September): 603–29.

Cilliers, Paul. 1998. Complexity and Postmodernism: Understanding Complex Systems. New York: Routledge.



Collins, R., R. Hanneman, and G. Mordt. 1995. “Discovering Theory Dynamics by Simulation.” Sociological Methodology 25: 1–46.

Costanza, Robert, ed. 1991. Ecological Economics: The Science and Management of Sustainability. New York: Columbia University Press.

Daly, Herman E. 1991. Steady-State Economics. 2nd ed. Washington, DC: Island Press.

Der Derian, James, and Michael J. Shapiro, eds. 1989. International/Intertextual Relations: Postmodern Readings of World Politics. Issues in World Politics, ed. James N. Rosenau and William C. Potter. Lexington, MA: Lexing- ton Books.

Doran, Charles F. 1999. “Why Forecasts Fail: The Limits and Potential of Fore- casting in International Relations and Economics.” International Studies Review 1, no. 2 (Summer): 11–41.

Elliott, Euel, and L. Douglas Kiel, eds. 1999. Nonlinear Dynamics, Complexity and Public Policy. Commack, NY: Nova Science Publishers.

Elster, Jon. 1983. Explaining Technical Change. Cambridge: Cambridge University Press.

———. 1986. Introduction to Rational Choice, ed. Jon Elster, 1–33. New York:

New York University Press.

Evans, Peter B., Harold K. Jacobson, and Robert D. Putnam, eds. 1993. Double- Edged Diplomacy: Institutional Bargaining and Domestic Politics. Studies in International Political Economy, ed. Stephen D. Krasner and Miles Kahler, 25. Berkeley: University of California Press.

Eve, Raymond A., Sara Horsfall, and Mary E. Lee, eds. 1997. Chaos, Complexity and Sociology: Myths, Models, and Theories. Thousand Oaks, CA: Sage.

Gell-Mann, Murray. 1994. The Quark and the Jaguar: Adventures in the Simple and the Complex. New York: W. H. Freeman and Company.

George, Jim, and David Campbell. 1990. “Patterns of Dissent and the Celebra- tion of Difference: Critical Social Theory and International Relations.” International Studies Quarterly 34, no. 3 (September): 269–93.

Georgescu-Roegen, N. 1971. The Entropy Law and the Economic Process. Cam- bridge, MA: Harvard University Press.

———. 1975. “Energy and Economic Myths.” Southern Economic Journal 41, no. 3 (January): 347–81.

Gilpin, Robert G. 1996. “The Nature of Political Economy.” In International Political Economy: State-Market Relations in the Changing Global Order, ed.



C. Roe Goddard, John T. Passé-Smith, and John G. Conklin, 9–24. Boulder, CO: Lynne Rienner.

Goddard, C. Roe, John T. Passé-Smith, and John G. Conklin, eds. 1996. Inter- national Political Economy: State-Market Relations in the Changing Global Order. Boulder, CO: Lynne Rienner.

Green, Donald P., and Ian Shapiro. 1994. Pathologies of Rational Choice Theory:

A Critique of Applications in Political Science. New Haven, CT: Yale Uni- versity Press.

Griffin, Douglas. 2002. The Emergence of Leadership: Linking Self-Organization and Ethics. New York: Routledge.

Hanneman, R., 1988. Computer-Assisted Theory Building. Newbury Park, CA: Sage.

Hasenclever, Andreas, Peter Mayer, and Volker Rittberger. 1997. Theories of International Regimes. Cambridge: Cambridge University Press.

Holland, John H. 1995. Hidden Order: How Adaptation Builds Complexity. Read- ing, MA: Perseus Books.

———. 1998. Emergence from Chaos to Order. Cambridge, MA: Perseus Books.

Hollis, Martin, and Steve Smith. 1990. Explaining and Understanding International Relations. Oxford: Clarendon Press.

Hughes, Barry B. 2000. Continuity and Change in World Politics: The Clash of Per- spectives. 3rd ed. Englewood Cliffs, NJ: Prentice-Hall.

Hume, David. 1975. Enquiries Concerning Human Understanding and Concerning the Principles of Morals. 3rd. ed., with revisions and notes by P. H. Nidditch. Oxford: Clarendon Press.

Huntington, Samuel P. 1993. “The Clash of Civilizations?” Foreign Affairs 72, no.


Jervis, Robert. 1976. Perception and Misperception in International Politics. Prince- ton, NJ: Princeton University Press.

———. 1997. System Effects: Complexity in Political and Social Life. Princeton, NJ:

Princeton University Press.

Kaplan, Morton A. 1969. “Variants on Six Models of the International System.” In International Politics and Foreign Policy, ed. James N. Rosenau. Toronto: Free Press.

Keohane, Robert O. 1988. “International Institutions: Two Approaches.” Inter- national Studies Quarterly 32, no. 4 (December): 379–96.



Keohane, Robert O., and Joseph S. Nye, eds. 1971. Transnational Relations and World Politics. Cambridge, MA: Harvard University Press.

Kiel, L. Douglas. 1994. Managing Chaos and Complexity in Government. San Fran- cisco: Jossey-Bass.

Knapp, Peter. 1999. “Evolution, Complex Systems and the Dialectic.” Journal of World-Systems Research 5:74–103.

Korten, David C. 1995. When Corporations Rule the World. West Hartford, CT:

Kumarian Press and Berrett Koehler.

Krishnan, Rajaram, Jonathan M. Harris, and Neva R. Goodwin, eds. 1995. A Survey of Ecological Economics. Washington, DC: Island Press.

Kymlicka, Will. 1990. Contemporary Political Philosophy: An Introduction. Oxford:

Clarendon Press.

Lansing, J. Stephen, James N. Kremer, and Barbara B. Smuts. 1998. “System- Dependent Selection, Ecological Feedback and the Emergence of Func- tional Structure in Ecosystems.” Journal of Theoretical Biology 192:377–91.

Locke, John. 1980. Second Treatise of Government. Ed. C. B. Macpherson. Indi- anapolis: Hackett.

Luhmann, Niklas. 1990. Essays on Self-Reference. New York: Columbia University Press.

———. 1998. Observations on Modernity. Stanford, CA: Stanford University Press.

March, J. G., and J. P. Olsen. 1998. “The Institutional Dynamics of International Political Orders.” International Organization 52, no. 4 (Autumn): 943–69.

Marion, Russ. 1999. The Edge of Organization: Chaos and Complexity Theories of For- mal Social Systems. Thousand Oaks, CA: Sage.

Maturana, Humberto, and Francisco Varela. 1980. Autopoiesis and Cognition. Dordrecht, Holland: D. Reidel.

Ocean Studies Board, Polar Research Board, and Board on Atmospheric Sci- ences and Climate. 2001. Abrupt Climate Change: Inevitable Surprise. Washington, DC: National Academy of Sciences.

Patomäki, Heikki, and Colin Wight. 2000. “After Postpositivism? The Promises of Critical Realism.” International Studies Quarterly 44, no. 2 (June):


Prigogine, Ilya. 1997. The End of Certainty: Time, Chaos, and the New Laws of Na- ture. In collaboration with Isabelle Stengers. New York: Free Press.



Prigogine, Ilya, and Isabelle Stengers. 1984. Order Out of Chaos: Man’s New Dia- logue with Nature. New York: New American Library.

Putnam, Robert D. 1988. “Diplomacy and Domestic Politics: The Logic of Two- Level Games.” International Organization 42, no. 3 (Summer): 427–60.

Richards, Diana. 2000. Political Complexity: Nonlinear Models of Complexity. Ann Arbor: University of Michigan Press.

Röhl, John C. G. 1998. Young Wilhelm: The Kaiser’s Early Life, 1859–1888. Cam- bridge: Cambridge University Press.

Romer, Paul. 1986. “Increasing Returns and Long Run Growth.” Journal of Political Economy 94, no. 5: 1002–35.

Ruggie, John Gerard. 1993. “Territoriality and Beyond: Problematizing Moder- nity in International Relations.” International Organization 47, no. 1 (Winter): 139–74.

———. 1998. Constructing World Polity: Essays on International Institutionalization. London: Routledge.

Schneider, Stephen H. 1989. Global Warming: Are We Entering the Greenhouse Cen- tury? New York: Vintage Books.

Skolnikoff, Eugene B. 1993. The Elusive Transformation: Science, Technology, and the Evolution of International Politics. Princeton, NJ: Princeton University Press for the Council on Foreign Relations.

Smith, Steve. 2004. “Singing Our World into Existence: International Relations Theory and September 11.” International Studies Quarterly 48, no. 3 (September): 499–515.

Smith, Steve, Ken Booth, and Marysia Zalewski, eds. 1996. International Theory:

Positivism & Beyond. Cambridge: Cambridge University Press.

Sprout, Harold, and Margaret Sprout. 1971. Toward a Politics of the Planet Earth. New York: Van Nostrand Reinhold.

———. 1978. The Context of Environmental Politics: Unfinished Business for America’s Third Century. Lexington: University Press of Kentucky.

Stacey, Ralph D. 2001. Complex Responsive Processes in Organizations: Learning and Knowledge Creation. London: Routledge.

Strange, Susan. 1994. States and Markets. 2nd ed. London: Frances Pinter.

———. 1996. The Retreat of the State: The Diffusion of Power in the World Economy. Cambridge: Cambridge University Press.



van Staveren, Irene. 1999. “Chaos Theory and Institutional Economics: Meta- phor or Model?” Journal of Economic Issues 33, no. 1 (March): 141–67.

Waldrop, M. Mitchell. 1992. Complexity: The Emerging Science at the Edge of Order and Chaos. New York: Simon & Schuster.

Waltz, Kenneth. 1979. Theory of International Politics. Reading, MA: Addison- Wesley.

Wendt, Alexander. 1994. “Collective Identity Formation and the International State.” American Political Science Review 88:384–96.



Complexity Is More Than Systems Theory

Neil E. Harrison with J. David Singer

The previous chapter emphasized the differences between conventional theories

of world politics and international relations and complexity. In 1971 Singer pro-

posed a general systems taxonomy that potentially supported both systemic ex- planation and more limited theorizing. This chapter shows how the important concepts of general systems taxonomy compare with concepts in a complex sys- tems taxonomy. In addition, we argue that complexity’s modifications to general systems taxonomy create a more flexible taxonomy that incorporates current knowledge and integrates theories that hitherto have been considered incom- mensurate views of world politics. We anticipate that the complex systems tax- onomy can generate radically new hypotheses about world politics and develop innovative ways of testing them and thereby increase knowledge and improve



A taxonomy “has no descriptive, predictive or explanatory power, since it contains

only definition propositions. But to serve as the basis for building models and the- ories, it must specify two things quite clearly: the basic constructs by which the rel- evant domain is to be described and the definitional relationships between and among these constructs” (Singer 1970, 3). As a framework for theorizing about world politics and international relations, complexity modifies and expands on Singer’s general systems taxonomy, which itself built upon current knowledge from orthodox theories (such as it was) and corrected the epistemological and methodological defects found in most general systems theories. Despite correcting




faults of general systems theories and a well-argued explanation of the benefits of

a general systems approach, in the last two decades Singer’s taxonomy has attracted

little theoretical interest. This section shows that some of the concepts from the general systems ap- proach appear only slightly modified in the complex systems taxonomy. In both taxonomies the international system is modeled as comprising multiple, hierar- chical systems of interacting agents, and the systems are open to their environ- ments but can be distinguished for theoretical purposes. Each system can contain multiple feedback loops and may be susceptible to path dependencies, and none is likely to display rational behavior. Yet, as described in the next sec- tion, complexity also changes, develops, and adds to general systems concepts in ways that increase its potential utility for building effective theories of world pol- itics and international relations.

Nested Systems

Textbooks and scholars agree that “there are different ontological layers in the world” (Patomäki and Wight 2000, 232). Most textbooks recognize several levels of analysis from individuals to the international system in which to seek expla- nations of global or international events. Commonly, at least five are identified:

system, state, society, government, and individual. Yet, in pursuit of a false sim- plicity, international theory has been largely confined to competing, singular lev- els of analysis. Singer (1970) comments, for example, that the “perhaps fatal flaw lies in the general tendency to focus on only one level of analysis, rather than treat the interactions that occur across the several relevant levels. The com-

mon focus on a single level of analysis blinds theorists to influential processes op- erating at other levels of analysis.” Both the general systems and complex systems taxonomies explicitly model the ontological layers in world politics as interrelated systems. In the gen- eral systems taxonomy the world political system is modeled as “a hierarchy of nested sets of subsystems, each embraced by those at the next higher level of analysis and embracing those at all lower levels. It follows from this that any sys- tem or set of systems at one level of analysis constitutes the environment of all the entities existing at any lower level” (Singer 1971, 12). For each state, the in- ternational system is only “real” as an environment within which it operates. Singer argued (1971, 17) that the nation-state remains a “useful object of analy- sis, but that at the same time the many entities comprising those social coalitions known as nations may also serve as useful objects of analysis.” But the state is not

a solid body; it is a “coalition of all social entities at the individual, primary, and secondary levels” (1971, 17), and government agencies are only components in



the aggregation known as the state. Similarly, complexity views the state as a “meta-agent” (Holland 1995) that forms its variable internal model out of the on- going interactions of social aggregations within its domestic political processes. However, as described below, complexity and general systems approaches diverge in their conception of cause-and-effect relations across levels of analysis.

Agents, Not Actions

Singer (1971, 7–11) identifies two schools of general systems theories: system-of- action and system-of-entities. In the former school, systems are identified around “actions, behavior, interaction, relationship, or role,” largely ignoring the entities that “participate and experience them.” The latter school models systems “around individuals, groups, associations, or aggregations of people”—that is, so- cial entities. In contrast to the “actions” school, scholars in this tradition usually explicitly posit that systems “will show rather similar patterns and processes as well as a fair degree of structural isomorphism.” He then argues persuasively that systems must be conceived of in terms of the characteristics of their constituent entities rather than in terms of agent actions. He shows that this approach is methodologically more tractable; it permits more effective separation of subsys- tems (e.g., political from economic) where this would be theoretically more use- ful; it more clearly distinguishes the system from its environment (is the social system part of the political or vice versa?); and it clarifies levels of analysis (at what level are individual actions of decision-makers?). Yet, in international rela- tions and world politics theories, scholars often focus on the actions and abstract away from agents. For example, Kaplan’s (1957) model describes a system of ac- tions and interactions between states, and Waltz (1979) elaborates a model of the international system that he compares to a market and in which, beyond crude power measures, the characteristics of the actors are of no interest. His- torical-materialist theories focus on the structural forces that dictate state behav- ior, regardless of the characteristics of the states. As in Singer’s general systems taxonomy, complexity assumes that the characteristics of social entities generate agent actions and participate in con- structing system structure. Complexity posits that internal models cause agent ac- tions and the pattern of agent behavior reflects the interaction of the agent’s internal model with environmental constraints. Thus, agent-based models as- sume that agents choose actions that are consistent with their individual desires and their beliefs about how to satisfy those desires. In social systems, institutions are the environmental selection rules that govern punishment and reward for agent actions. Institutions are the dynamic, path-dependent consequences of prior agent interactions through earlier patterns of institutions. 1



Open Systems

Political systems are conventionally assumed to be closed and homeostatic: dis- turbances are temporary and the system tends to return to equilibrium. Yet, in his 1971 monograph Singer simply states that all social systems are open—that is, their boundaries are “permeable to information and energy from [their] envi- ronment”—and that no social system can realistically be treated as closed. There is, he writes, “clearly no such thing as a completely closed social system” (13). 2 If so, assuming closure to simplify analysis and reduce empirical effort is not a rea- sonable parsimony but a gross distortion of reality. While every model is neces- sarily an “ersatz” reality, it must retain a recognizable link to the portion of reality it purports to represent or it will generate inaccurate “knowledge.” Complexity similarly treats all social systems as open, but, as Singer has commented, it is an empirical question as to how open each is (1971, 13). Some scholars approvingly describe rural communities’ openness to their “natural” en- vironment, especially in poorer countries. However, such open communities often are both somewhat closed to their social environment and much simpler than more complex, modern societies that are more open to other social systems but less open to ecological systems (Harrison n.d.). Openness is not a measure of complexity, but complex systems usually are open.


Conflict recurs in the international system because the conflicting incentives and temptations within nations and the lack of effective constraints between nations support positive feedbacks to conflict: “[I]ntra-national and inter-national events all impinge on one another in a cyclical and ongoing process within which the self-aggravating propensities frequently exceed the self-correcting ones by an un- acceptably large amount” (Singer 1970, 165). National elites use rhetoric for do- mestic political consumption that can incite potential enemies, the public and military desire the psychological comfort of discernible superiority, media am- plify internation conflicts, and the benefits of participation in the ideological mainstream preserve the distribution of power and inhibit changes in the his- torical patterns that transform inevitable conflicts into costly rivalries. Self- restraint within political elites and the media has diminished with the increase in the number of competing media companies, their geographical coverage, and the diminishing time lag between event and report, and corrective mechanisms within the international system have atrophied. Technologically induced imme- diacy reduces opportunities for editorial restraint. Similarly, technology has



reduced the time available for consideration of alternatives by decision-makers, and raw data crowd information and reduce the quality of its assessment. But positive feedback also can be beneficial. Constructivists would point to positive feedbacks operating in the formation of norms that underpin inter- national environmental treaties (as Hoffmann describes in chapter 5). The Mon- treal Protocol is often cited as a precedent for the 1992 signing of a global treaty to mitigate emission of gases believed to fuel climate change. Earlier examples come from the literature on the formation of the European Community, in which functional links or communications links were thought to increase trust and lead to eventual integration between countries (Mitrany 1966; Haas 1964; Deutsch 1953). Drawing on models of ecological systems, complexity posits that the in- teraction of multiple independent, volitional agents allows positive feedback loops to develop that can drive the system to “flip” to a new state (Levin 1999). Revolution might be an example of a flip in a social system. Like ecosystems, so- cieties have often collapsed from runaway internal feedback loops (Mäler 2000; Tainter 1988).

Path Dependence and Randomness

Path dependence is the idea that system development from time t to t 1 is not wholly random and can only fall within limits created by the prior state of the system. Living systems are considered path dependent: the current system state is related to and is, in part, determined by the prior state of the system, and that to its prior state all the way back to its nascence. Similarly, social systems evolve continuously, and the international system may change its structure without be- coming another system. The Cold War was a period in the evolution of the in- ternational system that was in part caused by all of history that preceded it. It was not a discrete system and cannot be separated from its history. But to state that the Cold War evolved out of prior history is not to claim that it was an inevitable effect of historical causes. The choices of multiple discretionary agents (from in- dividuals to states) inject randomness into outcomes. Thus, in complex systems the arrow of time is not reversible (Prigogine 1997). Studies show that in conflict prior experience matters in at least two ways. First, conflict stimulates innovation in search of increased military capa- bility. Armed conflict is an incentive to modernize both equipment and tactics (Smith 1985). If generals are always fighting the last war, it is better for them if the last war was more recent and a success. A major defeat may eliminate the state or so emasculate it that future aggression is militarily impractical (e.g., the



Austro-Hungarian and Ottoman Empires after World War I). Second, there is institutional memory: how decision-makers and the public perceive the ben- efits and costs of conflict. Success in conflict tends to bolster militarily adven- turous groups, and recent failure may cause caution. Failure, however, may encourage a desire to regain position and respect. For example, defeat in World War I supported German aggression in the 1930s. The Correlates of War (COW) project is a major effort to overcome a common cause of the evident failure to understand and better control interna- tion conflict. By collecting data for more than a century of interstate and civil wars, the project seeks patterns and commonalities among conflicts and avoids the historical fallacy that defines each conflict as a discrete and separable event. If the experience of conflict influences later conflict choices, feedback mecha- nisms within the state are the likely link. States have memories that influence fu- ture perceptions and choices. It is conventional wisdom, for example, that a “Vietnam syndrome” influenced decades of decision-makers, causing an appar- ent US aversion to armed conflict. Research has failed to find statistically significant linkages between war ex- perience and later conflict choices (Singer and Cusack 1981). Analysis of the COW data showed that none of the usual hypotheses about learning from war experience is supported: “[T]he probability of the major powers getting into war is statistically independent of when and with what effects they experienced their prior wars.” This does not mean that each conflict must be treated as a discrete event, but it does show that the feedback mechanisms within states are signifi- cantly more complex than is commonly believed. There is a randomness to the influences of memory and history that is not captured by simple theories. Through the concept of emergence (discussed further below), a complex systems theory of national security potentially allows for both path dependency through experience (state memory and capability) and randomness. Because state behavior emerges from domestic interactions, current conditions and in- stitutions, and the variable distribution of power between politically influential groups, influence state internal models. But current social conditions and power relations are themselves historical artifacts. Thus, historical experience is per- ceived through the ever-changing lens of the present, which itself emerges from the past.


Rationality assumptions are used as a convenient simplification in both orthodox theories and general systems approaches. But in the latter rationality is assumed only at the lowest level of aggregation: the individual human. At this level, rational-



ity is consonance of behavior with desires and beliefs about how to achieve those de-

sires within the perceived environment (primarily the social world). In orthodox the- ories, rationality is more often assigned to higher levels of aggregation. Realism, liberalism, and constructivism theorize at the level of the international system and assume that states and other important agents in that system are rational actors. 3 This greatly simplifies analysis but is inherently misleading. For social aggregations rationality has several meanings. It can mean that the outcome “was (or might turn out to be) desirable, successful, or functional for the perpetrators” (Singer 1990, 6). But positive feedbacks can magnify the preferences, and states rarely learn from past wars: “[K]nowledge may be neces- sary for rational human intervention, but the bloody pages of international his- tory remind us that it is hardly sufficient” (Singer and Cusack 1981, 417). While it is claimed to be rational for both parties to defect in the prisoner’s dilemma, both suffer individually suboptimal outcomes. Rationality also can mean that the decision process was rational. But successful or functional outcomes may em- anate from thoroughly irrational processes. Conversely, “the most careful, thor- ough, and rational process can, with some frequency, culminate in disaster, even

to be a positive relationship between high rationality in

the process and the desirability of the outcome” (6–7). Singer draws three conclusions. First, rationality in social aggregations can only describe the processes they follow, not the outcomes of those processes. Second, reality is too complex to call behavior rational if agents pur- sue outcomes that coincide with their individual preferences. President Bush may have wanted to install democracy in the Middle East through Iraq, but out- comes are always unpredictable and small causes, like pictures of prisoners tor- tured by US military police, may derail the most laudable policies. Third, the rationality of processes must be judged in relation to specific social aggrega- tions. Even if individuals and groups only minimally respond to their private in- terests, which may be rational for them, the effect is “extrarational” for the aggregation. Establishment and celebration of military organization, positive feedback in elites and media, social rewards to conformity, sunk costs (“they shall not have died in vain”), an inability to consider all options, and protecting the individual credibility of leaders are among the factors that make social de- cision processes extrarational. In the language of complexity, only the individ- ual can be rational (again, in process only) and all behaviors of social aggregations emerge from social interactions: “the ‘invisible hand’ of parochial sub-system interests is ubiquitous, virtually assuring that deviation from rational choice and the implied prudent pursuit of collective interests will remain the norm” (Singer 1990, 18). Social system behaviors are neither rational nor irra- tional. Singer’s terminology is exact: they are extrarational.

though there tends



Related to the orthodox presumption that social aggregations are rational

is the belief, sometimes implicit, that systems have a purpose or function in a

teleological sense. If systems are emergent, they have no purpose beyond the in- tentions and preferences of the subsystems. As the subsystems (for example, do- mestic interest groups) interact in pursuit of their preferences, the system

emerges (legislation is crafted and enacted). 4 As Singer comments (1971, 13), so- cial systems are not “inherently supposed to perform and survive, or seek to do so.” Social systems collapse when they no longer serve the needs of their con- stituent agents and the costs of belonging exceed the benefits though authority permits formal institutions (such as government agencies) to preserve themselves beyond the limits of social acceptance (Tainter 1988). While aggregations may not be rational, complexity diverges from gen- eral systems theory by accepting that a social aggregation like a nation-state may have preferences and interests. Singer’s general systems taxonomy specifically “denies that any social entity other than a human being can think, hope, pre-

also insists that any social entity can behave

(Singer 1971, 19, emphasis in original). But if the nation-state is conceived as

a meta-agent, it may, like an individual human agent, construct an internal

fer, expect, perceive, and

model of its environment that guides its behavior. Unlike realism, the internal model is not assumed to be objectively rational and, therefore, relatively un- changing, but the process by which it is constructed may be assessed for its ra- tionality. In chapter 5 below, Hoffmann explains how the US internal model of the ozone issue changed over time, responding to changes in international negotiations and domestic political interactions. 5


While the previous section discussed the similarities between several concepts used in the general systems taxonomy and concepts in complexity, the latter is more than a warmed-over version of the general systems approach. Several com- plexity concepts are additional to general-systems concepts and others are modi- fications of concepts used in the earlier approach. This section highlights the important differences between general systems theory and complexity.


Although the “interaction of individual properties (both within and among sin- gle humans) may produce emergent effects” (Singer 1968), emergence was re- jected by most general systems scholars as “unnecessary and scientifically misleading” (Singer 1971, 18). The effects are “neither structural or behavioral,”



and if they are cultural they can be observed as individual psychological proper- ties. He rejected the “bromide” that “a social system is more than the sum of its parts” because the cultural properties of large social systems can be better described in a “strictly aggregative fashion, by observing the distribution and con- figuration of individual psychological properties” (Singer 1968, 144). Complex- ity also relates system culture to individual psychological properties but models culture as emerging from the interplay of diverse agent internal models within institutional strictures. Singer also argued that “a system is nothing more than the sum of its parts and the relationships and interactions among them” (Singer 1971, 19) and that a system is “not composed of [external] systems, or of any other phenomena beyond its own component units and the relationships and interac- tions among them.” In complexity a social system is more than merely the aggre- gation of its parts: the system is modeled as emerging from the relationships and interactions between member agents. In contrast to Singer (1968), the emergent effects of agent interactions are both dynamic and important and cannot be cap- tured by observation of structure, behavior, or individual psychology. Although the emergent properties of a system cannot be captured by study- ing the system’s parts, emergence is real. It would be unscientific to reject a the- oretically useful concept merely because accepted scientific methodologies cannot record the phenomenon. Fortunately, as elsewhere in the science, devel- opments of new methods permit new thinking and empirical testing of novel hy- potheses. With George Zweig, Gell-Mann had developed a theory of quarks. He was always skeptical that such partially charged subatomic particles would ever be found, and by the end of the 1960s it seemed that he was right; no evidence had been found for their existence (Riordan 2004). But by 1973, with new research techniques, “everything seemed to be coming up quarks.” It took most of the decade for the theory of quarks to become generally accepted, and it was later recognized with a Noble Prize. Similarly, one of the predictions of Einstein’s gen- eral theory of relativity was tested by astronomical observation for the first time in 2004, nearly ninety years after the theory was first proposed. Developments in computer modeling now permit simulation of emergent systems and avoid “metaphysical pursuits” that attempt to isolate and measure emergence as a definable property of a complex system. When Singer wrote his critique of emergence, it was not practical to model complex systems. The mod- eling for the Club of Rome project Limits to Growth was a massive effort on mainframe computers by specialist programmers at the Massachusetts Institute of Technology (Meadows et al. 1972). Not only are the predictions of that project now largely discredited, but also the design of the model, limited as it was by in- complete data sets and computing power, is considered crude. 6 Social scientists who are not computer specialists can now program ABMs that model “mystical”



emergent properties as an integral part of the whole system. Indeed, the concept of emergence is central to ABM construction, and social scientists can now experiment as never before.


The “natural” sciences—most notably physics, mechanics, and chemistry—are still the measure of “scientific” for most social scientists. Their laboratory experimen-

tation allows repeatable, controlled manipulation of isolated potential causal fac- tors, a technique rarely available to the social scientist. Field experiments sometimes are possible, but they permit less control than in the laboratory. Singer (1977, 3) proposed that “the historical experiment is a perfectly legitimate mode of research” that may offer advantages over laboratory or field experiments. Increased control over principal factors in laboratory experiments allows more accurate observation and measurement, and an increased ability to ascer- tain covariation, permitting causal inference. In the social sciences, control over factors may be increased with comparative case studies, statistical manipulation

(as in COW), and simulations. In the “all-machine simulation

and variation of every input is fully controlled by the researcher” (Singer 1977). Inputs may range from the “purely speculative to the thoroughly grounded.” Computer simulations can test out myriad ideas against history until a good fit is found. In this way social research replicates the level of control of the classic laboratory experiment. Atmospheric scientists use computer simulations of cli- matic history to explain historical and to project future climate change. The de- velopment of desktop computers as powerful as supercomputers of twenty years ago, object programming, and concepts such as neural networks now permits (at least in principle) such historical experiments, as well as more speculative simu- lations designed to generate as much as to test hypotheses, as discussed further in chapters 7 through 9.

the magnitude

Cause and Effects

Complex-systems concepts encourage theory that covers multiple levels of analy- sis, but can it support construction of theories of cause-effect interactions across levels of analysis? We believe that complexity provides a conceptual framework for theories that can accommodate both causes working from below and from above the system under study. Within a system, emergence connects causes at lower level of analysis with effects at higher levels of aggregation. As discussed earlier, a state as meta-agent may form its internal model from the interaction of domestic constituencies and



the interplay among participants in decision-making groups within the executive. For example, the ability of an individual to shape state policy depends on the in- stitutional arrangements that regulate the individual’s influence on state behav- ior and the acceptance of his or her internal model by others in the decision processes. The chief executive is usually accorded more influence on state policy than other participants, and his or her internal model will tend to dominate. But in the process of negotiating policy, internal models may change, especially in terms of causal beliefs of what is possible. The “butterfly effect” of less institu- tionally gifted individuals also may emerge up through levels of aggregation to influence the state internal model. Beyond the emergence of behavior from internal interactions is the greater problem of theorizing links between causes at higher levels with effects at lower levels. The question is: how does environment affect system behaviors? For ex- ample, how does the state of the international system influence state policies, or how do national policies determine individual behavior? Singer (1961) argued that explanation of behavior at each level of analysis was problematic. For example, explanation from the state level requires several, often implicit assumptions, the most important of which is that state decisions are not influenced, even in part, by the perceptions held by individual decision-makers of the conditions in the system. Yet Jervis (1976), among others, has detailed the many ways in which individual decision-makers’ perceptions of the state’s environment are formed and how they influence state interests and behaviors. In complexity, environment affects system behavior in two ways. First, it con- strains what is possible and “selects” behaviors that are most appropriate within current institutional arrangements. Second, perceptions of environment influence agents’ internal models. And there may be interaction among both processes. Institutional arrangements in the environment create selection processes that act on system behavior. If misperception leads to maladapted behavior, the agent will be “punished,” at a minimum by being prevented from moving toward satisfaction of its desires. Selection means that agents adapt or are eliminated; coadaptation implies dynamic recursive adaptive responses between multiple agents. Agent learning is the cognitive adjustment that increases behavioral sur- vivability in a selective environment. Agents “learn” when they change their desires or their beliefs about how to achieve their desires. The post-9/11 War on Terror is a realignment of pref- erences in decision-makers’ internal models. Its prosecution through armed at- tacks on Afghanistan and Iraq and formation of the Department of Homeland Security was the result of negotiation among decision-makers’ causal beliefs. Small groups within the state (especially those closely associated to decision- makers, perhaps energy companies or conservative Christian groups) or outside



(al Qaida or Abu Ghraib prison guards) also may influence the behavior of states through their impact on the internal models of decision-makers. International organizations also may influence state choices in several ways, as through infor- mation dissemination, rules of due process (in the UN Security Council, for ex- ample), peacekeeping missions, state-building activities, or as the locus of norm construction and treaty negotiations.


In his A General Systems Taxonomy for Political Science (1971), Singer sets out six criteria for a good taxonomy that are equally applicable here. First, it should have theoretical relevance to the phenomena we hope to account for. And if it can be relevant to theory for a broad range of phenomena, so much the better. In com- plexity the dependent variable or outcome is the behavior of social systems; in world politics it is the emergence of political events (for example, policies and agent actions) and institutions. But it may be hoped that regularities will be found among social systems at all levels of aggregation and in all issue-areas of world politics. The anticipated predictor variables are rules, both internal and institutional, respectively within and between agents. Second, knowledge should be transferable between empirical domains. To achieve this, constructs must be sufficiently abstract to embrace “concepts that are substantially identical” (1971, 5), allowing for idiosyncrasies of differ- ent fields. But, where possible, they also should include current knowledge within the field. By linking concepts from conventional theories to complexity, this and the previous chapter show that this paradigm can include much that is known about world politics and, as discussed below, it may integrate views that are usually considered incompatible. Third, a good taxonomy indicates what is not known and needs to be learned. Recognizing the many weaknesses and gaps in our knowledge, a taxon- omy of world politics should be liberal and allow a range of eclectic approaches to many different phenomena. Historical experimentation with agent-based models is a flexible method adaptable across issue areas and levels of analysis. Fourth, a taxonomy need not be parsimonious, especially if parsimony would prevent many plausible models: “Parsimony is a virtue only when well advanced to- ward the verification stage of the discipline and may often be a liability when we are still in the discovery stage” (6). Fishing expeditions are permissible. One way to achieve sufficient permissiveness of testable hypotheses “is to develop a minimum number of classes of variables so that, while many options remain open, the taxon- omy also remains conceptually clean and manageable” (6). Complexity entails only



a few broad concepts that may be adapted to theory goals. So, complex systems the- ories may be more parsimonious than competing orthodox theories. Fifth, coverage of all levels of aggregation is needed: the taxonomy must

“be able to deal with several levels of analysis and

levels [must] not be a source of slippage and confusion” (6). We have argued that, with concepts such as emergence, environment, and internal model, com- plexity is eminently and uniquely able to satisfy this demand. Sixth, constructs must be operational. Chapters 3 through 6 of this vol- ume demonstrate how complex systems concepts can be operationalized in vari- ous issues and at different levels of analysis. Finally, semantic clarity is essential:

the interface between these

it is “preferable to select words that do convey generally accepted meaning and then, if necessary[,] specify the restricted or expanded definition intended” (6). This volume is intended to begin this work.

Encompassing and Improving Orthodoxy

Certainly the first, and perhaps most important, test of a new taxonomy is its ability to open new research agendas by better integrating existing theories and knowledge and thereby explaining some of what was previously inexplicable. Or- thodox theories may be classified in several ways; one useful approach is to dis- tinguish them by their “view” of reality, which may be external or internal. The complexity paradigm should support theories that fully or partially integrate these two apparently incommensurate views. The external or “scientific” approach assumes that the social world, and the natural world in which it exists, is an environment, independent of human agents and potentially predictable (S. Smith 1994; Hollis and Smith 1990). Be- havior is then assumed to be explicable using methods borrowed from the nat- ural sciences. The expectation is that there are regularities in behavior that may be explained by universal causal “laws.” “Behavior is generated by a system of forces or a structure” (Hollis and Smith 1990, 3) and decision-makers are re- placeable and only represent their position in the system with little personal vo- lition. In Singer’s general-systems taxonomy, whole systems are only the sum of their parts and could, therefore, be disaggregated and comprehended by analyz- ing the parts and their relations. Orthodox theories like constructivism pursue an alternative, “inside” approach that views the social world as constructed of rules and meaning through human interaction. Each agent tries to pick an intelligent course through multiple social engagements in which other agents bring their individ- ual characteristics to their social roles (Hollis and Smith 1990, 6). Here the goal



is to understand behavior, and the means are often hermeneutical, examining “human action from within, seeing it as intentional and meaningful behaviour” (S. Smith 1994, 400). The two approaches usually are assumed to be incompatible. Explanation

from the outside, however scientific, is incomplete without consideration of the

units: “The anarchical character of the international system

that the units affect the shape of relations, however firm the shove [from struc- tures]” (Hollis and Smith 1990, 198). It also is difficult to see “how the system changes its structure in a closed system without a change in the units and purely functional explanations are bound to be suspect, unless they include a causal contribution from the units.” The inside approach is rejected by some as inter- pretive and, therefore, inherently unscientific. And the two levels cannot not be combined to achieve an “overarching theory which explained how system-level and unit-level factors interacted to produce state behaviour” (100). Complex systems theories potentially integrate outside and inside ortho- dox views. The central concept of emergence marks complexity as favoring the inside view, yet its experimental methods are potentially as scientific as those of the revered “hard” sciences. The agent/structure problem is a manifestation of levels of analysis that turns on “the ‘reality’ of systems or on the need to feature them in explanations” (Hollis and Smith 1990, 197–98). If the system is real and must be analyzed as a whole, it must be shown that “wholes are more than their parts and that science is capable of establishing such a proposition” (198). ABMs can simulate behaviors of whole systems from the inside without consciously in- terpreting behavior and with no presumption of motives or meaning at any level by using randomized internal models and rules of interaction. As discussed, ABMs also may be constructed to simulate historical reality. Only theoretical de- velopment and empirical and experimental application will demonstrate if world politics theories based in complexity can overcome the incompatibility of the two views of orthodoxy and open useful new research agendas in issue-areas. The following chapters begin that task.

strongly suggests


This chapter was written by Neil Harrison based on David Singer’s selection from among his published works of those that anticipate complexity, and his comments and advice on earlier drafts.

1. This formulation echoes the recursive interaction of biological entity with its environment that is well accepted in biology and discussed in detail in Levins and Lewontin 1985.



2. Except for “lost” tribes in places like Amazonia or New Guinea, this

would seem a fair assessment. Autarky, once valued by “developmentistas” (see, for example, Palma 1978 and Gunder Frank 1969), is probably impossible in the modern world.

3. In critical theories derived from Marxian analysis, social aggregations of

classes and states are assumed to rationally pursue their interests.

4. The equating of policy-making with sausage production reflects the

inherently extrarational outcome of the policy process.

5. Also see Harrison 2000, which shows how domestic politics influenced

U.S. policies on climate change.

6. Even programming of linear-world models is much more sophisticated,

and the models are more accessible to social scientists not trained in the arcane tricks of effective computer programming. For example, see Hughes 1996, which comes with a computer model and relevant data on a CD.


Deutsch, Karl W. 1953. Nationalism and Social Community. Cambridge, MA:


Gunder Frank, André. 1969. Latin America: Underdevelopment or Revolution. New York: Modern Reader.

Haas, Ernst B. 1964. Beyond the Nation-State. Stanford, CA: Stanford University Press.

Holland, John H. 1995. Hidden Order: How Adaptation Builds Complexity. Read- ing, MA: Perseus Books.

Harrison, Neil E. 2000. “From the Inside Out: Domestic Influences on U.S. Global Environment Policy.” In U.S. Climate Change Policy, ed. Paul G. Harris, 89–109. New York: St. Martin’s.

———. N.d. “Complex Systems, Social Complexity, and the Environment.” Draft manuscript.

Hollis, Martin, and Steve Smith. 1990. Explaining and Understanding International Relations. Oxford: Clarendon Press.

Hughes, Barry B. 1996. International Futures: Choices in the Creation of a New World Order. 2nd ed. Boulder: Westview.

Jervis, Robert. 1976. Perception and Misperception in International Politics. Prince- ton, NJ: Princeton University Press.



Kaplan, Morton A., 1957. System and Process in International Politics. New York:


Levin, Simon A. 1999. Fragile Dominion: Complexity and the Commons. Reading, MA: Perseus Books.

Levins, Richard, and Richard Lewontin. 1985. The Dialectical Biologist. Cam- bridge, MA: Harvard University Press.

Mäler, Karl-Göran. 2000. “Development, Ecological Resources and Their Man- agement: A Study of Complex Dynamic Systems.” European Economic Review 44 (2000): 645–65.

Mitrany, David. 1966. A Working Peace System. Chicago: Quadrangle.

Palma, Gabriel. 1978. “Dependency: A Formal Theory of Underdevelopment for the Analysis of Concrete Situations of Underdevelopment.” World Development 6:881–924.

Patomäki, Heikki, and Colin Wight. 2000. “After Postpositivism? The Promises of Critical Realism.” International Studies Quarterly 44, no. 2 (June):


Prigogine, Ilya. 1997. The End of Certainty: Time, Chaos, and the New Laws of Na- ture. In collaboration with Isabelle Stengers. New York: Free Press.

Riordan, Michael. 2004. “Science Fashions and Science Facts.” Viewed at on March 22.

Singer, J. David. 1961. “The Level of Analysis Problem in International Rela- tions.” In The International System: Theoretical Essays, ed. K. Knorr and S. Verba, 77–92. Princeton, NJ: Princeton University Press.

———. 1968. “Man and World Politics: The Psycho-Cultural Interface.” Journal of Social Issues 24, no. 3:127–56.

———. 1970. “Escalation and Control in International Conflict: A Simple Feed- back Model.” General Systems 15:163–73.

———. 1971. A General Systems Taxonomy for Political Science. New York: General Learning.

———. 1977. “The Historical Experiment as a Research Strategy in the Study of Politics.” Social Science History 2, no. 1 (Fall): 1–22.

———. 1990. “The Invisible Hand, Extra-Rational Considerations and Decisional Failure.” Paper presented at the Annual Convention of the American Political Science Association, San Francisco, September 2.

Singer, J. David, and Thomas Cusack. 1981. “Periodicity, Inexorability, and Steermanship in International War.” In From National Development



to Global Community: Essays in Honor of Karl W. Deutsch, ed. Richard L. Merritt and Bruce M. Russett, 404–22. Boston: George Allen & Unwin.

Smith, Merritt Roe. 1985. Military Enterprises and Technological Change. Cam- bridge, MA: MIT.

Smith, Steve. 1994. “Rearranging the Deckchairs on the Ship Called Modernity:

Rosenberg, Epistemology and Emancipation.” Millennium: Journal of International Studies 23, no. 2:395–405.

Sterling-Folker, Jennifer. 2002. “Realism and the Constructivist Challenge:

Rejecting, Reconstructing, or Rereading.” International Studies Review 4, no. 1 (Spring): 73–97.

Tainter, Joseph A. 1988. The Collapse of Complex Societies. Cambridge: Cambridge University Press.

Wallerstein, Immanuel. The Capitalist World Economy. Cambridge: Cambridge University Press, 1979.

Waltz, Kenneth. 1979. Theory of International Politics. Reading, MA: Addison- Wesley.



Complexity and Conflict Resolution

Dennis J. D. Sandole

The events of September 11, 2001, undermined much conventional analysis in world politics and international relations (IR). Much as the fall of the Berlin Wall was not anticipated by IR scholars, the terrorist attacks in New York and Washington did not neatly fit within conventional explanations of international conflict. In this paper, therefore, I attempt to (1) respond theoretically and prag- matically to the events and aftermath of September 11, 2001; (2) deal with Realpolitik (and one of its concomitants, ethnocentrism) and conflict resolution as tra- ditionally contending, but potentially complementary, approaches to dealing with threats to order and security at the domestic and international levels; and (3) pro- vide a theoretical and pragmatic basis for further research, theory building, and practice in domestic and world affairs, with a view to dealing effectively with both the deep-rooted causes and the very clear symptoms of the “new” terrorism and re- lated identity-based conflicts fueled by the ending of the Cold War. To make sense of the attacks and to illustrate the potential complementarity of Realpolitik and conflict resolution approaches, I draw on several complexity concepts, including emergence; nonlinear, “catastrophic” responses to initial conditions; and syner- gistic coexistence of traditionally competing frameworks and ideas.


Realpolitik is the traditional power paradigm governing efforts to manage the uncertainty and disorder inherent in “Hobbesian space.” At its most virulent extreme, it is expressed as dictatorship domestically and as imperialism inter- nationally, with all the attendant manifestations of structural, cultural, and




physical violence—including genocide—implied by the defense and perpetuation of a preferred status quo at the expense of those who do not benefit from it (see Galtung 1969, 1996). Realpolitik has a long lineage, going back in recorded history to at least 416 BC, the midpoint of the Peloponnesian War, when Athens attempted to nego- tiate control over the neutral island state of Melos, a situation chronicled elo- quently by Thucydides:

[S]ince you know as well as we do that right, as the world goes, is only in question between equals in power, while the strong do what they can and

the weak suffer what they must

but a question of self-preservation and of not resisting those who are far

of men we know, that by a necessary law of their

nature they rule wherever they can. And it is not as if we were the first to make this law, or to act upon it when made: we found it existing before us

all we do is make use of it, knowing that you and everybody else having

the same power as we have, would do the same as we do

that those who do not yield to their equals, who keep terms with their su- periors, and are moderate towards their inferiors, on the whole succeed best. (1951, 331–36)

it is certain

stronger than you are

the contest not being an equal one

This is clearly an old story, which has been repeated thousands of times up to the present day, with Hans Morgenthau (1973, 4) being one of the more “re- cent” successors to Thucydides and reminding us all about the “laws” that gov- ern human behavior to Realpolitik effect. He says,

• Political realism believes that politics, like society in general, is governed by objective laws that have their roots in human nature.

• Human nature, in which the laws of politics have their roots, has not changed since the classical philosophies of China, India, and Greece endeavored to discover these laws.

In other words, for Morgenthau and other realists, human nature—which makes “statesmen think and act in terms of interest defined as power” (1973, 5)— has not changed since Thucydides made his observations in 416 BC. Hence, the “key concept of interest defined as power is a objective category which is univer- sally valid” (8). In the modern Westphalian world, power as interest is usually re- served for the protection of the nation-state, but it has also been used in defense of the tribe and the ethnic group.




Ethnocentrism is a natural corollary of Realpolitik: power is used by the privileged to maintain themselves and their groups at the expense of others. According to William Graham Sumner (1906), who coined the term, ethnocentrism is

the technical name for this view of things in which one’s own group is the center of everything, and all others are scaled and rated with reference to Each group nourishes its own pride and vanity, boasts itself supe- rior, exalts its own divinities, and looks with contempt on the most important fact is that ethnocentrism leads a people to exaggerate and intensify everything in their own folkways which is peculiar and which differentiates them from others. It therefore strengthens the folkways. (quoted in LeVine and Campbell 1972, 8).

Sumner also generalized “that all groups show this syndrome.” In other words, according to him and extensive research carried out by Henri Tajfel (1978, 1981) and others, ethnocentrism—following Thucydides’ and Morgen- thau’s characterizations of Realpolitik—is the universal tendency for humans to divide humankind into two groups: “them” and “us.” The criteria for doing so are not fixed and can be based on, among other things, nationality, eth- nicity, religion, race, class, region, or gender—criteria for which Realpolitik can mobilize defenses. Accordingly, ethnocentrism enhances intragroup community, especially under threat from out-groups (see Simmel 1955; Coser 1956), and in-group eth- nocentrism works against intergroup community. Indeed, it is safe to say that, es- pecially within a Realpolitik frame, ethnocentrism makes for a zero-sum relationship between peace at the intragroup level and war at the intergroup level. Again, according to Sumner (1906):

The insiders in a we-group are in a relation of peace, order, law, govern-

ment, and industry, to each other. Their relation to all outsiders, or others- groups, is one of war and plunder, except so far as agreements have modified The relation of comradeship and peace in the we-group and that of hos- tility and war towards others-groups are correlative to each other. The exi- gencies of war with outsiders are what make peace inside, lest internal

discord should weaken the we-group for

reacted on each other and developed each other, one within the group,

.Thus war and peace have

the other in the group relation. (quoted in LeVine and Campbell 1972, 7–8, emphasis added)



So pervasive is the tendency to subdivide humanity in this way, even under minimal intergroup differences, that fascinating experiments have been conducted with children, with the resulting “them-us” hostility between the contrived groups bordering on the remarkable. In their famous “Robbers Cave” experiments, for example, Sherif and Sherif (1953) were able to stimulate the development of hos- tile relationships between two groups of boys who had originally been friendly members of one and the same group. And in the famous (or to some, “infamous”) Blue Eyes/Brown Eyes exercises conducted by Jane Elliott, originally with her fourth-grade pupils in Riceville, Iowa, shortly after Rev. Dr. Martin Luther King had been assassinated:

Elliott told her children that brown-eyed people were superior to blue- eyed, due to the amount of the color-causing chemical, melanin, in their blood. She said that blue-eyed people were stupid and lazy and not to be trusted. To ensure that the eye color differentiation could be made quickly, Elliott passed out strips of cloth that fastened at the neck as col- lars. The brown eyes gleefully affixed the cloth-made shackles on their blue-eyed counterparts. Elliott withdrew her blue-eyed students’ basic classroom rights, such as drinking directly from the water fountain or taking a second helping at lunch. Brown-eyed kids, on the other hand, received preferential treat- ment. In addition to being permitted to boss around the blues, the browns were given an extended recess. Elliott recalls, “It was just horrifying how quickly they became what I told them they were.” Within 30 minutes, a blue-eyed girl named Carol had re- gressed from a “brilliant, self-confident carefree, excited little girl to a frightened, timid, uncertain little almost-person.” On the flip side, the brown-eyed children excelled under their new- found superiority. Elliott had seven students with dyslexia in her class that year [1968] and four of them had brown eyes. On the day that the browns were “on top,” those four brown-eyed boys with dyslexia read words that Elliott “knew they couldn’t read” and spelled words that she “knew they couldn’t spell.” (Kral 2000, 2, emphasis added) 1

Elliott conducted the original exercise “to demonstrate to her fourth-grade students how harmful the myth of White superiority is and what, as a result of this myth, it meant to be Black in America” (1). Since then, she has appeared on television, e.g., The Tonight Show, Oprah Winfrey Show, and PBS’s documentary Frontline (see PBS 1985) in the United States. She has also gone on to conduct these exercises for adults, including police officers, around the world, letting them “experience” for themselves that



We learn to be racist, therefore, we can learn not to be racist. Racism is not genetical. It has everything to do with power. (quoted in Coronel 1996, 2, emphasis added)

In the language of complexity, Elliott has demonstrated with a simple experi- ment how mental models are socially constructed and can adapt to interactions with others, especially those with authority. Related to the Realpolitik-ethnocentrism nexus is the seductive totality and simplicity of the clash of civilizations idea of Samuel Huntington (1993, 1996), and earlier of Benjamin Barber’s (1992) jihad. Since September 11, 2001, some commentators have been using the “clash” or “jihad” to characterize the polariz- ing global relationship between Judaic-Christian and Islamic “civilizations”:

probably the ultimate global expression of “us-them” hostility in the history of humankind. This idea may become, self-fulfillingly, more fact than fiction: a de- velopment that, especially if accompanied by nuclear weapons, would certainly not be in the interests of the United States or anyone else.

Sources of Ethnocentrism

Members of the conflict/conflict resolution community tend to be humanists, liberals associated with flexible, optimistic views of human nature. They tend to agree with Albert Bandura (1973) and others that whatever humans do in conflict situations is a function of learning: change what they learn and change their behavior! This view, which is associated with the Idealpolitik paradigm (see Sandole 1999a, 110–13), tends to ignore—on ideological, political, emo- tional and practical grounds—the role of biology in conflict, especially violent conflict behavior. Elsewhere (1990), I have argued that biological factors play a role in human behavior—as part of a complex constellation of social, political, economic, and other factors—and adherents to that view can now “come out of the closet” without fear of being ostracized as purveyors of Nazi eugenic philosophies and programs. I have also argued (1999a, 180–85) that it is not simply a question of “nature” or “nurture,” or indeed, in some simple additive sense, of “na- ture” and “nurture,” but of both interacting in complex ways, such that each may be affected by the other. 2 Given the observations of, among others, Ed- ward O. Wilson (1979), John Pfeiffer (1984), and Joseph Montville (1988) that our brains seem to be preprogrammed to bifurcate everything, including fellow human beings, into membership in in-groups and out-groups, it seems reasonable to conclude that “nature” has invested Homo sapiens with this



particular kind of “hard wiring” to protect us from one another in Hobbes’s infamous “state of nature,” where

in that

condition which is called Warre; and such a warre as is of every man

men live without a common Power to keep them all in awe,

against every man

brutish, and short. (Hobbes 1950, 103, 104)

where the life of man [is] solitary, poore, nasty,

Clearly, learning, culture, and other aspects of “nurture” can impact sig- nificantly whom an agent defines as “threatening,” and how he or she responds to them; but the biological predisposition to bifurcate fellow members of the species into “them” and “us” nevertheless seems to be there, ready to interact with cul- ture to create certain “histories,” certain “facts on the ground,” that then become the bases of violent conflict spirals, including the genocidal ethnic cleansing that has returned to Europe in the wake of the ending of the Cold War. In this regard, R. Paul Shaw and Yuwa Wong (1989) argue that the moti- vations that predispose human beings toward defense of their in-groups are part of “human nature”; that is, the “seeds of warfare” lie in ultimate (in contrast to proximate) causes—inclusive fitness and kin selection. Inclusive fitness has two parts: (1) “increased personal survival and in- creased personal reproduction (classical Darwinian fitness)”; and (2) “the en- hanced reproduction and survival of close relatives who share the same genes by common descent (a kinship component)” (Shaw and Wong 1989, 26). 3 Kin selection “implies that assistance, favors or altruism would be directed at individuals who were genetically related enough to give the common gene pool greater survival advantages. Genetic relatedness would be greatest with members of one’s lineage and one’s own kin or nucleus ethnic group” (Shaw and Wong 1989, 27). Here we have the crux of the matter concerning ethnocentrism for evolu- tionary psychologists: “[P]roviding an ultimate, evolutionary rationale for coopera- tion and civility among genetically related individuals also provides an ultimate rationale for anticipating origins of reduced cooperation among less related indi- viduals” (41, emphasis added). This amounts to a “sociobiology of ethnocentrism” (44–45, emphasis added) underpinning “we-them” distinctions, including those as framed in the “clash of civilizations.” Perhaps the ultimate example of complexity in human affairs is that

Humans have outfoxed themselves. They have learned to maximize inclu-

sive fitness—through ethnocentrism, out-group enmity, nationalism and patriotism—to the extent that they have created the means to destroy the

very inclusive fitness they seek to foster and

unless some kind



of action is forthcoming

will escape nuclear devastation, if not extinction. (197)

there is no reason to believe that Homo sapiens


Concerned members of the international community could join with Jane El- liott and start to teach children in the schools, not that racism, anti-Semitism, and other isms are “normal,” but that they are learned, oftentimes dysfunctional ex- pressions of our biological predisposition to bifurcate people into friend and foe. Given that the predisposition is part of our “wiring,” that is, originally meant to have survival value, we are sort of stuck with it. We are not, however, stuck with the culturally/experientially determined referents of that predisposition. Indeed, as implied, some of those definitions may be counter to our survival, either as members of in-groups or as an entire species. Hence, it would be in our best in- terests to work on changing those definitions, and on changing the mental mod- els through which individuals comprehend the world around them and in terms of which they choose their behaviors. Imagine classrooms at all levels, up to university level, where pupils and students are actively encouraged, by conflict resolution–trained facilitators, to brainstorm the kinds of emotions they experience when they think about, dis- cuss, or interact with members of certain groups. (They would thereby make it exceedingly difficult to do what Roger Fisher and William Ury (1983, chap. 2) counsel: to “separate the people from the problem.”) They would brainstorm where those feelings come from, the consequences of those feelings, examples throughout the country and the world where those kinds of feelings have trans- lated into violent conflict situations, how to work on changing those feelings, dif- ficulties in doing so, and so on. This is a complex tall order: the feelings that we experience have a “natural” base; they are, therefore, part of our “human nature.” However, the culturally de- fined targets of those feelings are not part of our nature: they may be wrong, un- fair, self-fulfillingly counterproductive and dangerous and, therefore, should be—and can be—changed! This would be quite a challenge to bring into any level of classroom, but it is a necessary one if we are to make a dent on the levels of violence that have, for example, brought genocide back to Europe, motivated nineteen young Arab Muslim males with box cutters to turn passenger-filled aircraft into cruise mis- siles against the World Trade Center and the Pentagon, or turned the United States into the most violent country in the industrialized world (see Sandole 1999a, 4). Recent examples of its violence include the Washington, DC–area



sniper incident (October 2–24, 2002) and the murder of three professors at the University of Arizona by a failing student (October 28, 2002); and as Stepp

(2002, A3) points out, “The homicide rate for U.S. infants

equal to the murder rate for teenagers, according to a new analysis of govern- ment data that revealed a surprising demographic milestone.” But what about the adults, some of whom may be killing their kids (A17)? To what extent can nurture close the nature-nurture gap of ethnocentrism for them, especially in the post-9/11 world? The War on Terror is currently being waged within a Realpolitik frame- work, again, elevating the level of analysis to a more global version of us-them hostilities. President George W. Bush’s strident declaration that “you are either with us or the terrorists” has radicalized Muslims all over the world. It has also made many Americans feel a closer sense of community, but at the expense of the security of many American Arabs and Muslims, who feel threatened and vic- timized by governmental security services as well as by purveyors of hate crimes (see Pierre 2002a, 2002b). In other words, we are returning to the dangerous simplicity of a bipolar world, where, given the Bush administration’s continu- ing war rhetoric to keep the patriotic fervor flowing beyond the ebbing impact of military successes in Afghanistan and Iraq, and the worsening insurgency in Iraq, the Realpolitik-ethnocentrism nexus and the intragroup peace versus intergroup war dynamic are taking on a more global, civilizational, “jihadic” character. Hence, although Realpolitik has been conceived as a rational approach to the defense of individual and national interests, it has, in practice, tended to be- come more a part of the problem than of the solution: more and more it has been revealed to be a significant source of self-stimulating and self-perpetuating con- flict systems (see Vasquez 1993; Sandole 1999a).

[is] now virtually

Complex Systems

Complexity offers insights in this regard. One of its major assumptions is that, among other things, everything is connected to everything else (see Waldrop 1992). Accordingly, any attempt at problem solving must be at least multi- if not interdisciplinary. But few of us have been educated that way: we receive our de- grees usually in only one discipline, and therefore we as analysts may also be more a part of the problem than we are of the solution.

Conflict Resolution

As an interdisciplinary field, conflict resolution is intuitively similar to complexity and provides a conceptual basis for capturing the complexity of complex con-



flicts. Part of its appeal is that it does not replace Realpolitik as such, because on occasion we need the military to prevent or stop atrocious acts of violence such as the genocidal conflicts of recent times in Rwanda and Bosnia. Also, we could have used our police or military as armed marshals on board the four hijacked aircraft on Tuesday, September 11, 2001, to prevent the planes crashing into their intended targets. However, to be effective in the long run, Realpolitik must always be included in a larger frame, a metaparadigm, where it coexists and co- evolves with, for example, Idealpolitik, Marxism, and something I call “non- Marxist radical thought,” which focuses on basic human needs (see Sandole 1993; Sadole 1999a, 110–13, 117–20, 137–40).


Given small differences in the start-up conditions of biological, economic, physi- cal, and other systems, the consequences may be catastrophically or otherwise rad- ically different. 4 To paraphrase Heisenberg’s uncertainty principle from quantum mechanics (see Nagel 1961, 293–305), in such cases neither analysts nor policy- makers would be able to predict a system’s behavior with unlimited precision. Nevertheless, there would also be discernible patterns underlying chaos, thereby keeping alive the possibility of prediction. The distributed decision-making in complex systems and their consequent dynamism and tendency toward nonlin- earity make them unpredictable: though patterns may hold for a period of time, their sensitivity makes them liable to change out of all proportion to any stimulus. As with many, if not all, innovations in thought, complexity had been around awhile before it was conceptualized as such (see Saperstein 1995). For instance, Kenneth Boulding remarked that in conflict analysis and resolution,

Human beings are moved not only by immediate pressures but by distant goals that are contemplated in the imagination. These goals are susceptible of change, often of dramatic change, as a result of apparently slight changes in current information. On the other hand, they also have a good deal of sta- bility, and this gives a stability to the system in the large that it may not have in the small. (1962, 24, emphasis added)

Lewis F. Richardson’s (1939, 1960a) work on the dynamics of an arms race is another source of ideas in orthodox conflict studies that show similari- ties with those of complexity: in a dyadic relationship, depending upon each actor’s sensitivities to the other’s arms levels (mutual fears), plus the constraints of each on further arms spending (limiting factors), and underlying grievances, there could be a stable balance of power with regard to “rate[s] of rearmament or



disarmament”; or there could be an unstable equilibrium in which either complete

disarmament or a runaway arms race is possible. There could also be radical shifts between stable and unstable systems (in either direction), “for relatively modest

assumptions” regarding mutual fears, limiting factors, and griev-

ances (Nicholson 1989, 152). Or, within the unstable condition, there could be radical shifts from complete disarmament to a runaway arms race (or vice versa), resulting from “small shift[s] in the position of the initial point” of armament ex- penditures at “time zero” (152; also see Boulding 1962, chap. 2; Rapoport 1960, chap. 1; and Saperstein 1995). 5

variations in



It is clear that conflict researchers and policy-makers cannot predict with cer- tainty what kinds of conflicts-as-process will emerge from various kinds of conflicts- as-startup conditions (see Sandole 1999a, 129–31), or predict the course of any particular conflict-as-process. The danger in this, of course, is that conflict re- searchers may be paralyzed into recommending nothing and policy-makers para- lyzed into doing nothing, or at least nothing of major significance: witness Bosnia and Herzegovina, at least up to the Dayton Peace Accords of October– December 1995. But we should be fair: the danger of paralysis derives from the possibil- ity that conflicts-as-process could, unpredictably, and because of very small shifts in existing conditions, escalate out of control (a continuing risk in Iraq). In other words, beyond some threshold, conflicts-as-process could escalate into self-stimulating/self-perpetuating spirals, where attempts to deal with them could backfire, leading to destruction of the conflict systems themselves. In such cases, we can talk of entropic conflicts: conflicts that approach entropy, or progressive disorder. The danger that, unpredictably, conflicts can assume an entropic charac- ter (as Iraq may already have)—what Gregory Bateson (1973, 98) refers to as a schismogenic “regenerative causal circuit or vicious circle”—is implicit in Realpoli- tik: the use of a “measured” amount of force, even as part of an Idealpolitik strat- egy to achieve negative peace as a necessary (but not sufficient) condition of positive peace, could backfire, making matters worse. This may explain why, with the exception of the NATO bombing campaign that, in part, led to the Dayton Peace Agreement, Robert Axelrod’s (1984) fascinating theory of cooperation has not been applied to Bosnia.



Axelrod has argued that in all situations involving the prisoner’s dilemma—a classic game of Realpolitik analysis and prescription that applies to the interpersonal as well as international levels—the best way to act is in terms of the TIT FOR TAT strategy:

TIT FOR TAT’s robust success [in prisoner’s-dilemma situations] is due to being nice, provocable, forgiving, and clear. Its niceness means that it is never the first to defect, and this property prevents it from getting into un- necessary trouble. Its retaliation discourages the other side from persisting whenever defection is tried. Its forgiveness helps restore mutual coopera- tion. And its clarity makes its behavioral pattern easy to recognize; and once recognized, it is easy to perceive that the best way of dealing with TIT FOR TAT is to cooperate with it. (176, emphasis added, also see 54)

For TIT FOR TAT to work, however, “the future must have a sufficiently large shadow”; that is, it “requires that the players have a large enough chance of meet- ing again and that they do not discount the significance of their next meeting too greatly” (174). Extending Axelrod’s theory to the wars in former Yugoslavia during 1991–95 leads to the following scenario:

1. Slovenian and especially Croatian declarations of indepen- dence from the Yugoslav Federation in June 1991 resurrected Serbian fears (especially among Serbs living in Croatia) of Croa- tian defection from the stable TIT FOR TAT equilibrium that had existed up to that point.

2. Serbian military successes, plus the “nonprovocability” of the international community, stimulated the development and ex- acerbation of a violent, asymmetrical conflict-as-process, that is, “ethnic cleansing,” which was prosecuted by the Serbs against the major victims of the wars in former Yugoslavia, Bosnian Slavic Muslims.

3. In the absence of the “provocability” of the Bosnian Slavic Mus- lims, the international community was effectively shamed into becoming “provocable” and retaliating against the Serbian “defection” from the previously stable TIT FOR TAT equilibrium, although in a very restrained way (as in the live-and-let-live sys- tem of trench warfare during World War I; see Axelrod 1984, chap. 4). Subsequently, the international community was “for- giving” toward the Serbs to avoid stimulating new or exacerbat- ing ongoing violent conflict spirals.



4. The international community embarked on a “train-and-equip” program for a joint Bosnian Muslim–Croat army (see Pomfret 1996a) so that Bosnian Muslims in particular could, in the fu- ture, become appropriately “provocable.”

In terms of this analysis, the “provocability” of the international commu- nity (more so than of the Bosnian Muslims) was the issue. Until Dayton, the in- ternational community had not been sufficiently “provocable” in Bosnia and

Herzegovina, perhaps because of paralysis associated with the unpredictability of the consequences of even minor adjustments in complex systems capable of gen- erating entropic conflict processes. As Michael Lund (1996, 111) puts it: “From

1990 into 1992, it may be remembered, a major obstacle to European and U.S.

involvement in the Yugoslavian imbroglio was considerable uncertainty as to the wider ramifications of the gathering storm” (emphasis added). But even with Dayton, the “provocability” of the international community remained an issue; for instance, during the summer of 1996 demands by the in- ternational community and threats of sanctions were followed by vague promises by the Bosnian Serbs to comply, and their failure to do so was followed by a breakdown on sanctions. This only emboldened the Serbs. “It was, according to many western diplomats, a humiliating retreat and one that was greeted with ju- bilation in the self-styled Republic of Srpska” (Hedges 1996b). While Serbs celebrated the first “anniversary” of the fall of the UN “pro-

tected safe area” of Srebrenica, war crimes investigators were sorting through the remains of men and boys captured and shot after the Muslim enclave fell. Serbs

and reiterated their goal of keeping the territory “ethni-

marked their victory

cally pure.” 6 Taken together with the observation that the conditions specified by Day-

ton for “free and fair elections” in Bosnia—freedom of movement, freedom of ex- pression, freedom of press, and freedom of association—had not been met, even though the Organization for Security and Cooperation in Europe (OSCE) de- clared that national elections could nevertheless take place on September 14,

1996 (see Hedges 1996a), and municipal elections a year later, then it is clear

that, nearly one year following the cessation of hostilities in Bosnia, with the ex- ception of the NATO bombing campaign leading up to Dayton Axelrod’s theory remained basically untried and untested in former Yugoslavia. Paul Stern and Daniel Druckman (1994/1995, 114) view Axelrod’s theory

as an example of the strongest evidence of the “hegemonic position of realism” in US international relations thinking and practice, because it effectively legiti- mates cooperation within the Realpolitik paradigm. Axelrod’s theory is certainly appropriate for Realpolitik-defined realities, such as the wars in former



Yugoslavia. In Bosnia and Herzegovina, for instance, Anthony Lewis (1996, 11) had concluded that “The only thing that ever moved the Bosnian Serbs to more than empty promises during the war [there] was force.” Also, a Bosnian Serb of- ficial in the city of Brcko characterized the Dayton Peace Agreement’s call for the return of refugees as “a clear attempt to change the biological structure of the city.” He went on to assert, with Muslim refugees in mind, that “We will defend our frontiers biologically” (Dobbs 1996a, emphasis added), thereby implying a continuation of the doctrine and practice of “ethnic cleansing.” But of the four elements of TIT FOR TAT, only two—provocability and clar- ity—reflect Realpolitik as such. TIT FOR TAT’s other two elements—niceness and for- giveness—locate it in a more “complex” constellation of options, very much like that suggested by Stern and Druckman’s own “contours of a new paradigm” (1994/1995, 115–17) and by my own “4 2 framework,” which combines Realpolitik, Idealpolitik, Marxist, and non-Marxist radical definitions of reality, plus cooperative and competitive means for dealing with conflict (see Sandole 1999a, 110–13). This “complex” orientation shares with Fisher and Keashly’s (1991) “contingency model” the prescription of using what is necessary under one set of conditions, but of using other tools as well when those conditions have changed (also see Fisher 1993; Fisher 1997, chap. 8). Indeed, TIT FOR TAT is a re- sponse to “complexity”: it can encourage, through learning, the development of cooperation out of the “coevolutionary dance of competition and cooperation” (see Waldrop 1992, 259–60, 262–65, 292–94). TIT FOR TAT is an example of a process of agent interactions in which each agent learns to coevolve with others. But for TIT FOR TAT to be ultimately successful, there must be, in addition to a “sufficiently large shadow” of the future (which, admittedly, ethnic cleansing had eroded), stability in the sense of Richardson’s (1939, 1960a) “balance of power”—another Realpolitik aspect!—between the “coevolving” parties in their re- spective capabilities to inflict pain on each other. Unless a stable balance exists, the parties may engage in what Lewis Coser (1956, 136) refers to as a “trial by or- deal,” in which “conflict may be an important balancing mechanism” designed to achieve the very equilibrium that may be absent to begin with:

Conflict consists in a test of power between antagonistic parties. Accom- modation between them is possible only if each is aware of the relative strength of both parties. However, paradoxical as it may seem, such knowl- edge can most frequently be attained only through conflict, since other mechanisms for testing the respective strengths of antagonists seem to be unavailable. Consequently, struggle may be an important way to avoid conditions of disequilibrium by modifying the basis for power relations. (137, emphasis added)



Apropos less lethal forms of conflict handling (e.g., mediation or arbitra-

encounter the difficulty that the assess-

ment of the actual power relations between the contenders can hardly be made before their relative power has been established through struggle” (135–36). The US-led effort to arm the Bosnian Muslims was designed to “make pos- sible a reassessment of relative power and thus serve as a balancing mechanism which helps to maintain and consolidate societies” (137) and to provide a mate- rial basis for increased Muslim “provocability,” especially in the relative absence of such on the part of the international community, thereby establishing a stable balance of power and ensuring that TIT FOR TAT succeeds in Bosnia without fur- ther international intervention. Indeed, according to James Pardew, the official originally in charge of the US program, the weapons “would be used for Bosnia’s defense and would contribute to stability in the region. The purpose of the train- and-equip program [therefore] is to prevent war by creating a military balance in Bosnia” (Pomfret 1996b, emphasis added; also see USIP 1997). There is, however, a problem with “balance of power,” as there is with Realpolitik in general. As Shaw and Wong (1989, 47) imply, “trials by ordeal” to determine “relative strength,” as manifested in former Yugoslavia, are associated with “groups as forces of selection [that] represent an emergent, proximate, environmental cause [of war]”:

tion), Coser tells us that such “Efforts

Since failure to maintain a balance of power could have resulted in ex-

tinction, groups and their expansion figure as forces of selection in our the- ory. Motivated by resource competition, conflict, and warfare, struggles to maintain balances of power [have given] rise to more complex societal

units which [have] continued the legacy of intergroup

It is by

this process that out-group enmity and ethnocentrism have been rein- forced and carried over from nucleus ethnic group to band, to tribe, to chiefdom, to nation-state. (45)

Applying complexity concepts, therefore, involves more than stable bal- ances associated with negative peace of an enforced temporary respite from violent conflict; it also involves building upon and transcending these and, in positive peace fashion, exploring the dynamic of deep-rooted processes and conditions that make, in the shorter run, the balances a “natural” consequence of social processes (see Galtung 1969, 1996). Seven years after the U.S. intervention, Bosnia was unable to function as a sovereign state. It depended on a North At- lantic Treaty Organization (NATO) force of twelve thousand foreign troops, and the functions of government relied on Western representatives with sweeping powers. And the same nationalist parties that incited the conflict had been



reelected (WP 2002). In December 2004, the NATO force was replaced by a British-led EU peace keeping force of some 7,000 personnel. The problem with the Dayton Peace Agreement for Bosnia is not only that the physical and emotional reconstruction of the country (positive peace) has lagged behind the enforced prevention of violence (negative peace), but also— with provocability still an issue—that the negative peace is not a stable one. TIT FOR TAT, therefore—and with it, complexity in general—still remain to be fully ap- plied to Bosnia.


As implied thus far, Realpolitik philosophers, theorists, and practitioners tend to re- spond to the disorder, unpredictability, and insecurity inherent in “Hobbesian space” by advocating and/or pursuing the enhancement of predictability, regularity, and stability (the “PRS needs,” see Sandole 1984)—and, therefore, of order and secu- rity—in their domestic and international environments through the simplistic bifur- cation of the species into “them” and “us” and by the threatened or actual use of force against “them” whenever circumstances within the Realpolitik/ethnocentric frame call for such. Hence, Kenneth Waltz’s (1964) earlier defense of a “bipolar” in- ternational system as inherently more conducive to stability than a multipolar sys- tem; John Mearsheimer’s (1990a, 1990b) lamenting of the end of the Cold War and its simplicity; and following the terrorist attacks of September 11, 2001, Presi- dent George W. Bush’s strident declaration to the entire world that “[y]ou are either with us or the terrorists!” This desire for simplicity in the face of real complexity is the foundation of most conventional perspectives on conflict and its resolution. At the extreme “right-wing” end of the Idealpolitik-Realpolitik continuum, we find authoritarians who have a low threshold for uncertainty and insecurity and who, therefore, tend to find democracy too chaotic. But the irony here is that, as “extreme” Realpolitik practitioners implement more and more antidemo- cratic and threat- or force-based measures in pursuit of order and security, their ef- forts tend to become more and more counterproductive and self-defeating (see Burton 1972, chap. 6), generating “security dilemmas” (Herz 1950) and the even- tual collapse of their own systems. If such policymakers are alive and well at the end of the day and respond to the “cognitive dissonance” (Festinger 1962) generated by their failed Realpoli- tik-based policies and expectations with a “paradigm shift” (Kuhn 1970) to Ide- alpolitik-based norms and polices, then we might have a situation as we did following the termination of World War II in Europe, when the erstwhile mor- tal enemies Germany and France established the basis for what has become the



European Union: one of the most illustrative, successful experiments in peace- building (still ongoing!) in pursuit of “positive peace.” But there is irony here as well: even Idealpolitik may contain the seeds of its own destruction, because “too much stability” may lead to boredom and atrophy. As Paul Sites (1973) and Kenneth Boulding (1962), among others, have argued, whatever else we as humans may “need” to develop effectively—physically, emotion- ally, psychologically, and socially—we also have a need for stimulation, for “drama.” This is the crux of complexity: the “need” to nudge systems at the “edge of chaos” so that neither chaos nor order prevails at the zero-sum expense of the other. Quite a challenge, especially when people are stressed by threats at home and abroad. One of the major lessons of complexity, therefore, is to never take anything for granted for too long on either end of the Idealpolitik/voluntary order– Realpolitik/force-based order continuum. Hence, Viktor Frankel (1985) was able to survive the horrors and brutalities of a Nazi concentration camp (an incredi- bly negative setting) by discovering “meaning” in his adversity (a remarkably posi- tive occurrence). Similarly, former Yugoslavia was “able” to implode genocidally in the 1990s (an incredibly negative event) after years of intergroup stability and relative prosperity (a positive setting). Frankel’s feat is an example of the potential for diversity in human mental models, Yugoslavia’s flip to violent conflict may re- flect complex system sensitivity to initial conditions and small events. Given that complex systems can shift “catastrophically” from one end of the continuum to the other with apparently little effort, one implication here for the architects of globalization is the need to creatively influence the balance be- tween order (“McWorld”) and disorder (“Jihad”), so that, for example, those in the developing world who have traditionally borne the brunt of colonialism and imperialism have a chance to close the gap between the “haves” and the “have- nots,” lest frustration/aggression-based cycles of violence degenerate further into the “new” terrorism (see Sandole 2002)! One of the interesting aspects of the “new” terrorism—where terrorists are quite prepared to die in the execution of their acts to inflict catastrophic damage and destruction on their symbolic and human targets—is that the terrorists, moti- vated by fundamental ideology more than by a political agenda, are not deterred by traditional Realpolitik threats or the actual use of force. Consequently, their ulti- mate intention and effect are to generate maximum unpredictability, instability, and, therefore, disorder and insecurity in the West and those supported by the West. Once the “bite-and-counterbite” dynamic of terrorism versus counterter- rorism reaches some critical threshold, another consequence is each side’s over- perception of and overreaction to the actions of “the Other” (see Zinnes, North, and Koch 1961; Holsti, North, and Brody 1968). Hence, the self-fulfilling confirma- tion of Huntington’s (1993, 1996) otherwise contentious “clash of civilizations”



or of Barber’s (1992) “jihad,” with weapons of mass destruction in the bargain, all further enhanced by the U.S. rush to war against, and apparently long-term occu- pation of, another developing country in the Arab/Muslim world, Iraq.

Three-Pillar Framework

While complexity may have generated paralysis over Bosnia, complex-systems con- cepts have given new meaning to a possible antidote: frameworks that can poten- tially integrate most if not all disciplines in an effort to explain and to facilitate dealing with the foci of any one of them. I have developed the “Three-Pillar” com- prehensive mapping of conflict and conflict resolution (see Sandole 1998a; 1999a, chap. 6; 2003) as one such framework for identifying and integrating factors associ- ated with traditionally competing frameworks. I have used this framework as a basis for developing a “new European peace and security system” (NEPSS) potentially rel- evant to preventing “future Yugoslavias” (see Sandole 1999a, chap. 7 and below). Basically, the three-pillar framework comprises pillar 1, conflict—latent con- flict (pre-MCP); manifest conflict processes (MCPs); or aggressive manifest con- flict processes (AMCPs)—while pillar 2 deals with conflict causes and conditions, and pillar 3, conflict (third-party) intervention (see table 3.1).


Pillar 2

Pillar 1

Pillar 3



Conflict Causes

(Latent [Pre-MCP])


and Conditions




Parties Issues Objectives Means Conflict-handling Orientations Conflict Environments

Third-Party Objectives Conflict Prevention Conflict Management Conflict Settlement Conflict Resolution Conflict Transformation




Third-Party Approaches Competitive and/or Cooperative Processes Negative and/or Positive Peace Orientations Track 1 and/or Multitrack Actors and Processes



Under pillar 1 (the “Middle Kingdom”), we have the parties, the issues about which they are in conflict, the long-term objectives they hope to achieve by waging conflict over certain issues, the means they are employing; their preferred conflict-handling orientations, and the conflict “spaces” within which their con- flict is occurring. Pillar 2 comprises four levels of explanation—individual, societal, interna- tional, and global/ecological—that capture potential causes and conditions of the conflict occurring in the “conflict spaces” of pillar 1. Finally, pillar 3 deals with third-party objectives such as violent conflict prevention, management, settlement, resolution, and/or transformation; plus the means for achieving any of these objectives: competitive (confrontational) and/or cooperative (collaborative) processes; “negative peace” and/or “positive peace” orientations; and “track 1” (official, governmental) and/or “multitrack” (nongovernmental, unofficial, and other) actors and processes. Reflecting the complexity perspective, the three-pillar framework maps the conflict, its causes, and potential interventions at multiple levels. Complex prob- lems are characteristics of whole systems. So, they do not have simple solutions, and it is not possible to anticipate the effects of interventions. The working hy- pothesis of the three-pillar framework is that to design and implement an effec- tive intervention into any particular conflict “space” under pillar 1, a potential third party under pillar 3 will have to “capture the complexity” of the conflict as represented by all four levels of potential “drivers” under pillar 2. Complexity is especially relevant here, as it provides the essential concep- tual basis for combining into a coherent whole traditionally competing frame- works and ideas. For example, if we were to ask an anthropologist, an economist,

a historian, an international relations specialist, a political scientist, a psycholo- gist, and a sociologist for their views on why former Yugoslavia imploded into a genocidal frenzy during the 1990s, we would likely get radically different re- sponses. Similarly, if we were to ask a businessperson, a citizen activist, a diplo- mat, a humanitarian aid worker, a journalist, a military officer, and a religious leader about how to deal with Yugoslav-type conflicts, we would also get differ- ent responses. All these, however, can be accommodated within the three-pillar framework. This is precisely what I have attempted to do with “NEPSS.”

The New European Peace and Security System (NEPSS)

I have used the three-pillar framework as a basis for designing the NEPSS: an in-

tervention into post–Cold War Europe that just might be relevant to preventing “future Yugoslavias” (see Sandole 1998b; 1999a, chap. 7; 1999b) and that, appro- priately adapted, also may be relevant to conflict interventions outside Europe.



NEPSS comprises descriptive and prescriptive elements—that is, develop- ments that are actually occurring as well as those that could or should occur. De- scriptively, NEPSS makes use of existing international organizations in Europe such as OSCE, the European Union (EU), the Council of Europe (CoE), and the North Atlantic Treaty Organization (NATO). NEPSS also employs the basic structure of the OSCE as a conceptual and operational framework for enhancing the complementarity and synergy of all mechanisms working together on common problems. Within this framework, NATO represents an example of political and military aspects of a reframed, more comprehensive sense of security, the European Union (EU) an example of economic and environmental aspects, and the Council of Europe (CoE) an example of humanitarian and human rights as- pects of comprehensive security. More importantly, each of these heretofore Cold War institutions has been reaching out to its former enemies, inviting them to become members or join together in constituting new, post–Cold War institutions. For example, at its November 2002 summit in Prague, NATO, which had already taken in the Czech Republic, Hungary, and Poland as members, issued invitations to seven other former members of the communist world—Bulgaria, Estonia, Latvia, Lithuania, Romania, Slovakia, and Slovenia—all of which be- came members by March 2004. Subsequently, at its December 2002 summit in Copenhagen, the EU issued invitations to Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia, and Slovenia, all of which became members by May 2004. Quite simply, these developments are nothing short of revolutionary, facilitating a genuine paradigm shift from Realpolitik, zero- sum national security to Idealpolitik, positive-sum common security. But revolutionary though these developments are, all these organizations are basically interstate in nature, while the problems posed by conflicts in former Yugoslavia and elsewhere are essentially intrastate in nature. Hence, there has been a need for something else to deal with the conflicts of the post–Cold War world. This is where the prescriptive element enters the picture. Prescriptively, NEPSS is characterized by integrated systems of conflict resolution networks, with vertical and horizontal components. Under the vertical, we would have a mapping of Europe in terms of the local, societal, subregional, regional, and global levels of analysis, with track 1–9 actors and processes—governmental/official; nongovernmental/professional; business; private citizen; research, training, and education; activist; religious; funding; and media—corresponding to each level (see Diamond and McDonald 1996). The idea here is that “all conflicts are local.” And, assuming an early warning system to activate the preventive diplomacy envisaged by Michael Lund (1996) and others (e.g., Peter Wallensteen [1998] and Walter Kemp



[2001]), conflicts developing at any local level could be responded to by a synergis- tic combination of track 1–9 resources at that level—plus, to the extent necessary and possible, societal, subregional, regional, and global levels as well. Should the vertical dimension fail to prevent “the house from catching on fire,” then there may be a need for the horizontal dimension to become operational. This would in- volve the judicious use of Realpolitik force, but basically within an Idealpolitik frame- work, to achieve negative peace (suppression of the fire) but only as a necessary (not sufficient) condition for achieving positive peace: the elimination of the (pillar 2) un- derlying causes and conditions. While some recent developments in Europe are suggestive of progressive re- inforcement of NEPSS’s descriptive character and the “vertical” dimension of its pre- scriptive character—such as the emergence from the November 1999 OSCE Summit in Istanbul of the Charter for European Security, inclusive of the Platform for Co-operative Security (see OSCE Istanbul 1999a, 1999b)—other developments are suggestive of the sole narrow use of Realpolitik force (e.g., the destruction of Grozny and killings of tens of thousands of Chechen civilians in the Russian Fed- eration). Even the 1999 NATO air war against Serbia over Kosovo—albeit clearly for the humanitarian purpose of preventing further genocidal ethnic cleansing of Kosovar Albanians—falls more into the category of the narrow use of Realpolitik force basically within a Realpolitik (instead of an Idealpolitik) framework. These and other developments—the relentless Israeli-Palestinian carnage; the “new” terrorism; the U.S.–Iraq war; the possibility of a nuclear war between India and Pakistan over Kashmir, with one or both sides trying to “preempt” the other; and the escalating development of a “clash of civilizations” or “jihad” be- tween the Judaic-Christian and Islamic worlds against the background of easily available weapons of mass destruction—are not only tragic but, via the law of unintended consequences, potentially very destabilizing. The primary message of a complexity approach to conflict analysis and res- olution is that there may be a need for a policeman to pull the attacker off a vic- tim—for example, for NATO to stop genocide. But this is not the same as declaring or insinuating, for example, that all Afghans, all Chechens, all Pales- tinians, all Saudis, all Wahhabis, all Arabs, or all Muslims are “terrorists,” and then proceeding to eliminate (or be perceived to be eliminating) the entire pop- ulation and its culture as a way to deal with the “terrorist” problem. Paradoxi- cally, this problem is being, in part, created self-fulfillingly by this perspective and corresponding behavior! Within the terms of the argument posed here, therefore, Realpolitik force must always take place within a bigger picture, a framework that also allows for and encourages conflict resolution (dealing with the underlying causes of the fire



at hand) and conflict transformation (dealing with the long-term relationships among the survivors of the fire), as well as (violent) conflict prevention (preventing the house from catching on fire in the first place), conflict management (if initial conflict prevention fails, preventing the spread of the fire), and conflict settlement (if management fails, forcefully putting out the fire). If peace is not positive as well as negative—if it does not ultimately deal with the underlying “conflicts-as-startup conditions”—then “conflict-as-process” will never be far from the surface, always available to come back to haunt us time and time again! This is the ultimate message and categorical imperative of a com- plexity approach to conflict analysis and resolution: not only to think and act outside the Realpolitik-only box, but to combine it synergistically with other, usu- ally competing, ways of knowing and acting.


It would take extremely enlightened leadership—in the United States, Europe, Rus- sia, China, and Japan, among others (e.g., Israel, Iraq, North Korea, Palestine, Saudi Arabia)—to pursue “positive” as well as “negative peace” in coordinated response to as- saults to the “global commons”: superordinate goals that no one state can achieve on its own, but only in collaboration with others (see Sherif 1967). But for some inex- plicable (perhaps, in part, “biological”) reasons, ecological degradation, exponen- tial population growth, and a growing gap between haves and have-nots, among other compelling elements of the global problematique (e.g., AIDS), have failed to rise to the status of William James’s (1989) “moral equivalent of war.” Perhaps, “if we have time,” we can leave it to the children: the next gener- ation of decision-makers. In the meantime, however, especially after the Bali bombings (October 12, 2002), Moscow Chechen hostage crisis (October 23–26, 2002), Madrid train bombings (March 11, 2004), and London transit bombings (July 7, 2005), global terrorism itself just might provide the motivation for the in- ternational community to come together, and not just to “root out” terrorism (Realpolitik), but to deal with its root causes as well (Idealpolitik). This would be a truly superordinate undertaking that could galvanize the international community into developing a culture of global problem-solving that transcends traditional ethnocentrism and a reliance on Realpolitik-only perspec- tives and measures, paving the way for a new definition of “the enemy” as any and all assaults to the global commons: a truly complex approach to a set of com- plex problems at the “edge of chaos.” Among the conceptual tools that could facilitate movement in this constructive (albeit ambitious) direction is the Three-Pillar Framework (3PF) or



3PF-generated new European peace and security system (NEPSS) discussed briefly in this chapter. Analysts working together with policy-makers could use either or both to capture the complexity of complex conflict situations. In this way, they could deal with relationships that have gone wrong and the underlying causes and conditions driving negative developments in those relationships as well as the symp- toms (indicators) of those negative relationships. To a very large extent, the US-led invasions and occupations of Afghanistan and Iraq (especially the latter) seem to be addressing only the symp- toms of the conflicts that have torn these Muslim countries apart; furthermore, those interventions may actually be exacerbating the causes of 9/11-type terror- ism. Such counterproductivity is the price that policy-makers—and the rest of us— might continue to pay for rejecting or otherwise avoiding conceptual tools that transcend symptoms and capture the complexity of complex conflicts. 7


An earlier version of this paper was presented at the 44th Annual Convention of the International Studies Association (ISA), Panel on “Global Complexity:

Agent-Based Models in Global and International Studies,” Portland, Oregon, February 25–March 1, 2003. The author gratefully acknowledges comments and suggestions made by Neil Harrison and Patrick James.

1. Also see Rosenthal and Jacobson’s (1968) classic study “Pygmalion in

the Classroom.”

2. For other discussions in this regard, see Cowley (2003) and Oldham


3. “Inclusive fitness thus equals an individual’s Darwinian (egoistic) fit-

ness augmented by an allowance for the effect that the individual can have on the reproductive success of those who share identical genes by common descent” (Shaw and Wong 1989, 26–27, emphasis in the original).

4. This section reflects and builds upon parts of chapter 8 (especially

pp. 193–201) of Sandole (1999a).

5. Sensitivity to initial conditions is a characteristic of complex systems

that can also be found in some simple systems. In complex systems, however, small changes in initial conditions may lead to nonlinear system changes, a flip in the system to something quite different—as when a forest becomes a desert. But it does not follow that a system in which small changes in initial conditions “cause” large system changes is necessarily complex.



6. “There is no place for Turks [the derogatory term the Serbs use for

Bosnian Muslims, whose ancestors adopted the Islamic faith of Turkish in- vaders] in Republika Srpska,” said General Milenko Zivanovic, the regional com- mander, who led the final assault on Srebrenica” (AP 1996, emphasis added). (Also see Honig and Both 1996 and Rohde 1997.)

7. Thus far, I have applied the 3PF to an analysis of the causes of 9/11-

type terrorism (see Sandole 2002) and the 3PF-generated NEPSS to a design for an EU intervention into post-NATO Bosnia (see Sandole 2004).


AP. 1996. “Serbs Mark a ‘Victory,’ Victims Are Exhumed.” Associated Press. International Herald Tribune, 12 July: 10.

Axelrod, Robert. 1984. The Evolution of Cooperation. New York: Basic Books.

———. 1997. The Complexity of Cooperation: Agent-Based Models of Competition and Collaboration. Princeton, NJ: Princeton University Press.

Bandura, Albert. 1973. Aggression: A Social Learning Analysis. Englewood Cliffs, NJ: Prentice-Hall.

Barabasi, Albert-Laszlo, and Eric Bonabeau. 2003. “Scale-free Networks.” Scien- tific American, 288, no. 5 (May): 60.

Barber, Benjamin R. 1992. “Jihad Vs. McWorld.” Atlantic Monthly 269, no. 3 (March): 53–65.

Bateson, Gregory. 1973. Steps to An Ecology of Mind: Collected Essays in Anthro- pology, Psychiatry, Evolution, and Epistemology. Frogmore, England:


Bertalanffy, Ludwig von. 1973. General System Theory. Harmondsworth, England:


Boulding, Kenneth E. 1962. Conflict and Defense: A General Theory. New York:

Harper and Row.

Burton, John W. 1972. World Society. Cambridge: Cambridge University Press.

———. 1997. Violence Explained. Manchester, England: Manchester University Press.

Coronel, Mercita. 1996. “Interview with Jane Elliott.” Contrast, 28 March, 1–2.



Coser, Lewis A. 1956. The Functions of Social Conflict. New York: Free Press; Lon- don: Collier-Macmillan.

Cowley, Geoffrey. 2003. “Our Bodies, Our Fears (Anxiety and Your Brain: How Living With Fear Affects the Mind and Body).” Newsweek, 24 February,


Diamond, Louise, and John W. McDonald, Jr. 1996. Multi-Track Diplomacy: A Systems Approach to Peace. 3rd ed. West Hartford, CT: Kumarian Press.

Dobbs, Michael. 1996a. “The Story of Brcko Is the Story of Bosnia.” International Herald Tribune, 28 June, 2.

———. 1996b. “U.S. Ready to Participate in New Force in Bosnia: Gore Signals Willingness, NATO Officials Say.” Washington Post, 14 November,


Festinger, Leon. 1962. A Theory of Cognitive Dissonance. Stanford, CA: Stanford University Press.

Fisher, Roger, and William Ury. 1983. Getting to Yes: Negotiating Agreement With- out Giving In. London: Penguin Books.

Fisher, Ronald J. 1993. “The Potential for Peacebuilding: Forging a Bridge from Peacekeeping to Peacemaking.” Peace & Change 18, no. 3 (July):


———. 1997. Interactive Conflict Resolution. Syracuse, NY: Syracuse University Press.

Fisher, Ronald J., and Loraleigh Keashly. 1991. “The Potential Complementar- ity of Mediation and Consultation within a Contingency Model of Third Party Intervention.” Journal of Peace Research 28, no. 1 (February):


Frankl, Viktor E. 1985. Man’s Search for Meaning. Revised and updated. New York: Washington Square Press.

Galtung, Johan. 1969. “Violence, Peace and Peace Research.” Journal of Peace Research 6, no. 3:167–91.

———. 1996. Peace By Peaceful Means: Peace and Conflict, Development and Civiliza- tion. London: Sage Publications, for the International Peace Research Institute, Oslo (PRIO).

Gleick, James. 1987. Chaos: Making a New Science. New York: Penguin Books.

Hedges, Chris. 1996a. “Bosnian Vote in the Fall Wins a Wary Approval.” Inter- national Herald Tribune, 26 June, 6.



———. 1996b. “West’s About-Face on Karadzic: Familiar and Sad.” International Herald Tribune, 2 July, 1, 10.

Herz, John H. 1950. “Idealist Internationalism and the Security Dilemma.” World Politics 5, no. 2 (January): 157–80.

Hobbes, Thomas. 1950. Leviathan. New York: E. P. Dutton; London: J. M. Dent and Sons. (Orig. publ. in 1651.)

Holsti, O. R., R. C. North, and R. A. Brody. 1968. “Perception and Action in the 1914 Crises.” In Quantitative International Politics: Insights and Evidence, ed. J. David Singer. New York: Free Press; London: Collier-Macmillan.

Honig, Jan Willem, and Norbert Both. 1996. Srebrenica: Record of a War Crime. New York: Penguin Books.

Huntington, Samuel P. 1993. “The Clash of Civilizations?” Foreign Affairs 72, no. 3 (Summer): 22–49.

———. 1996. The Clash of Civilizations and the Remaking of World Order. New York:

Simon & Schuster.

James, William. 1989. “The Moral Equivalent of War.” In International War: An Anthology and Study Guide, ed. Melvin Small and J. David Singer. 2nd ed. Homewood, IL: Dorsey Press. (Orig. publ. in McClure’s Magazine, August 1910, 463–68.)

Jervis, Robert. 1997. System Effects:Complexity in Political and Social Life. Princeton, NJ: Princeton University Press.

Kemp, Walter A., ed. 2001. Quiet Diplomacy in Action: The OSCE High Commis- sioner on National Minorities. The Hague: Kluwer Law International.

Kral, Brigitta. 2000. “The Eyes of Jane Elliott—Blue Eyes Brown Eyes Exercise.” Horizon, 1–5.

Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago:

University of Chicago Press.

LeVine, Robert A., and Donald T. Campbell. 1972. Ethnocentrism: Theories of Conflict, Ethnic Attitudes and Group Behavior. New York: John Wiley.

Lewis, Anthony. 1996. “Karadzic and Co. Have to Be Tried.” International Herald Tribune, 9 July, 11.

Lund, Michael S. 1996. Preventing Violent Conflicts: A Strategy for Preventive Diplo- macy. Washington, DC: U.S. Institute of Peace Press.

Mearsheimer, John J. 1990a. “Back to the Future: Instability in Europe After the Cold War.” International Security 15, no. 1 (Summer): 5–56.



———. 1990b. “Why We Will Soon Miss the Cold War.” Atlantic Monthly, August, 35–50.

Montville, Joseph V. 1988. Foreword to Vamik Volkan’s The Need to Have Ene- mies and Allies: From Clinical Practice to International Relationships, by Vamik Volkan. Northvale, NJ: Jason Aronson.

Morgenthau, Hans J. 1973. Politics Among Nations: The Struggle for Power and Peace. 5th ed. New York: Alfred Knopf.

Nagel, Ernest. 1961. The Structure of Science: Problems in the Logic of Scientific Explanation. London: Routledge and Kegan Paul.

Nicholson, Michael. 1989. Formal Theories in International Relations. Cambridge Stud- ies in International Relations, 3. Cambridge: Cambridge University Press.

Oldham, Pamela. 2003. “A Flight of Mind: The Act of Dissociation Can Protect Children Emotionally from Trauma, but Repeated Use May Cause Lasting Harm.” Washington Post, 18 February, F1, F5.

OSCE Istanbul. 1999a. OSCE Istanbul Summit Declaration. 18–19 November ’99. Istanbul: Organization for Security and Cooperation in Europe.

———. 1999b. OSCE Istanbul Summit. 18–19 November ’99: Charter for European Secu- rity. Istanbul: Organization for Security and Cooperation in Europe.

Pagels, Heinz R. 1988. The Dreams of Reason: The Computer and the Rise of the Sci- ences and Complexity. New York: Simon & Schuster.

PBS. 1985. “A Class Divided.” Frontline (PBS documentary produced by W. Peters).


Pfeiffer, John. 1984. “Human Nature: The Universe Inside Your Skull.” Science Digest 92:92.

Pierre, Robert E. 2002a. “Fear and Anxiety Permeate Arab Enclave Near Detroit:

Muslim Americans Feel They Are Targets in War on Terror.” Washing- ton Post, 4 August, A3.

———. 2002b. “Victims of Hate, Now Feeling Forgotten: Family of Man Killed After 9/11 Finds Little Charity, But Much Hardship.” Washington Post, 14 September, A1, A10.

Pomfret, John. 1996a. “Croats and Muslims Edge Toward a Joint Army.” Inter- national Herald Tribune, 8 July, 5.

———. 1996b. “U.S. Starts Delivery of Heavy Weapons to Bosnia’s Muslim-Croat Forces.” Washington Post, 22 November, A44.



Rapoport, Anatol. 1960. Fights, Games, and Debates. Ann Arbor: University of Michigan Press.

Richardson, Lewis F. 1939. “Generalized Foreign Politics: A Study in Group Psy- chology.” British Journal of Psychology, Monograph Supplements 23, no. 7 (June): 1–91.

———. 1960a. Arms and Insecurity. Chicago: Quadrangle Books; Pittsburgh: Box- wood Press.

———. 1960b. Statistics of Deadly Quarrels. Chicago: Quadrangle Books; Pitts- burgh: Boxwood Press.

Rohde, David. 1997. Endgame: The Betrayal and Fall of Srebrenica, Europe’s Worst Massacre Since World War II. New York: Farrar, Straus and Giroux.

Rosenau, Pauline Marie. 1992. Post-Modernism and the Social Sciences: Insights, Inroads, and Intrusions. Princeton, NJ: Princeton University Press.

Rosenthal, Robert, and Lenore Jacobson. 1968. Pygmalion in the Classroom. New York: Holt, Rinehart and Winston.

Sanders, T. Irene. 2002. “To Fight Terror, We Can’t Think Straight.” Washing- ton Post, 5 May, B2.

Sandole, Dennis J. D. 1984. “The Subjectivity of Theories and Actions in World Society.” In Conflict in World Society: A New Perspective on International Relations, ed. Michael Banks. New York: St. Martin’s; Brighton, England: Wheatsheaf.

———. 1987. “Conflict Management: Elements of Generic Theory and Practice.” In Conflict Management and Problem Solving: Interpersonal to International Applications, ed. Dennis J. D. Sandole and Ingrid Sandole-Staroste. Lon- don: Frances Pinter; New York: New York University Press.

———. 1990. “The Biological Basis of Needs in World Society: The Ultimate Micro-Macro Nexus.” In Conflict: Human Needs Theory, ed. John W. Burton. New York: St. Martin’s; London: Macmillan.

———. 1993. “Paradigms, Theories, and Metaphors in Conflict and Conflict Res- olution: Coherence or Confusion?” In Conflict Resolution Theory and Practice: Integration and Application, ed. Dennis J. D. Sandole and Hugo van der Merwe. Manchester, England: Manchester University Press; New York: St. Martin’s Press.

———. 1998a. “A Comprehensive Mapping of Conflict and Conflict Resolution:

A Three Pillar Approach.” Peace and Conflict Studies 5, no. 2 (Decem- ber): 1–30.



———. 1998b. “A Peace and Security System for Post-Cold War Europe: Prevent- ing Future Yugoslavias.” In Encyclopedia of the European Union, ed. Desmond Dinan. Boulder, CO: Lynne Rienner.

———. 1999a. Capturing the Complexity of Conflict: Dealing with Violent Ethnic Con- flicts of the Post-Cold War Era. London: Pinter/Cassell (Continuum International).

———. 1999b. “A Design for Peace and Security in Post-Cold War Europe.” Eth- nic Conflict Research Digest (University of Ulster) 2, no. 1 (February):


———. 2002. “The Causes of Terrorism.” In Terrorism: Concepts, Causes, and Con- flict Resolution, ed. R. Scott Moore. Fort Belvoir, VA: U.S. Defense Threat Reduction Agency.

———. 2003. “A Typology.” In Conflict: From Analysis to Action, ed. Sandra Cheldelin, Daniel Druckman, and Larissa Fast. London: Continuum International.

———. 2004. “Building Peace in Post-NATO Bosnia: A Recommended Action Plan.” In From Peace Making to Self Sustaining Peace—International Presence in South East Europe, ed. Predrag Jurekovic, Frederic Labarre, and Ernst M. Felber- bauer. Vienna: National Defence Academy.


Saperstein, Alvin M. 1995. “War and Chaos.” American Scientist 83, no. 6 (November–December): 548–57.

Shaw, R. Paul, and Yuwa Wong. 1989. Genetic Seeds of Warfare: Evolution, Nationalism, and Patriotism. Boston: Unwin Hyman.

Sherif, Muzafer. 1967. Group Conflict and Cooperation: Their Social Psychology. Lon- don: Routledge and Kegan Paul.

Sherif, Muzafer, and Carolyn W. Sherif. 1953. Groups in Harmony and Tension. New York: Harper and Row.

Simmel, Georg. 1955. Conflict and the Web of Group-Affiliations. Trans. Kurt H. Wolff and Reinhard Bendix. New York: Free Press; London: Collier- Macmillan.

Sites, Paul. 1973. Control: The Basis of Social Order. New York: Dunellen.

Stepp, Laura Sessions. 2002. “Infants Now Murdered As Often As Teens: Actual Rate May Be Higher, Experts Say.” Washington Post, 10 December, A3,


Stern, Paul C., and Daniel Druckman. 1994/1995. “Has the ‘Earthquake’ of 1989 Toppled International Relations Theory?” Peace Psychology Review 1, no. 2 (Autumn/Winter): 109–22.



Sumner, William Graham. 1906. Folkways. New York: Ginn.

Suplee, Curt. 1992. “The Order of Things.” Review of Complexity: Life at the Edge of Chaos, by Roger Lewin, and Complexity: The Emerging Science at the Edge of Order and Chaos, by M. Mitchell Waldrop. Washington Post Book World 22, no. 51 (20 December): 1, 10.

Tajfel, Henri, ed. 1978. Differentiation between Social Groups: Studies in the Social Psychology of Intergroup Relations. New York: Academic Press.

———. 1981. Human Groups and Social Categories: Studies in Social Psychology. Lon- don: Cambridge University Press.

Thucydides. 1951. The Peloponnesian War. The unabridged Crawley translation with an introduction by John H. Finley, Jr. New York: Modern Library.

USIP. 1997. “Dayton Implementation: The Train and Equip Program.” Special Report of the Dayton Implementation Working Group, September. Wash- ington, DC: U.S Institute of Peace.

Vasquez, John A. 1993. The War Puzzle. Cambridge Studies in International Relations, 27. Cambridge: Cambridge University Press.

Waldrop, M. Mitchell. 1992. Complexity: The Emerging Science at the Edge of Order and Chaos. New York: Simon & Schuster.

Wallensteen, Peter, ed. 1998. Preventing Violent Conflicts: Past Record and Future Challenges. Report no. 48. Uppsala, Sweden: Department of Peace and Conflict Research, Uppsala University.

Waltz, Kenneth N. 1964. “The Stability of a Bipolar World.” Daedalus 93 (Sum- mer): 881–909.

Wilson, Edward O. 1979. On Human Nature. New York: Bantam Books.

Woodcock, Alexander, and Monte Davis. 1978. Catastrophe Theory. New York:

E. P. Dutton.

WP. 2002. “After the War.” Editorial. Washington Post, 12 November, A24.

Zeeman, E. C. 1977. Catastrophe Theory: Selected Papers, 1972–1977. Reading, MA: Addison-Wesley.

Zinnes, D. A., R. C. North and H. E. Koch, Jr. 1961. “Capability, Threat, and the Outbreak of War.” In International Politics and Foreign Policy: A Reader in Research and Theory, ed. James N. Rosenau. New York: Free Press; London: Collier-Macmillan.



Understanding and Coping with Ethnic Conflict and Development Issues in Post-Soviet Eurasia

Walter C. Clemens, Jr.

Generated by scholars from various disciplines, complexity science integrates concepts from many fields to produce a new slant on evolution. 1 Its exponents seek a general theory able to explain many different types of phenomena—social as well as biological and physical. If complexity fulfills this goal, it should also help us to understand ethnic and other problems in post-Soviet Eurasia and other troubled regions. The contributions of complexity to this understanding are evaluated in this paper. This chapter contends that basic complexity concepts do much to explain the movement toward or away from resolution of ethnic problems in newly in- dependent states. These concepts do not contradict explanations centered on the success or failure of movement toward democratization (Snyder 2000), but rather enrich them and offer linkages to other fields of knowledge. Complexity starts with a wider lens than democratization but includes it. The concept of so- cietal fitness, a major concern of complex systems theories, subsumes political, economic, and cultural strengths. The precise role played by each strength in shaping societal fitness becomes an important but secondary question. The analysis here suggests that ideas and concepts from complexity can en- hance our ability to describe and explain the past and present. But for several reasons discussed in chapter 10 and elsewhere in this volume, complex systems theories have much less utility for projecting alternative long-term futures or pre- scribing international strategy. Still, ideas and concepts from complexity can enlarge our vision and complement other approaches to social science.




All the hypotheses discussed here are pitched at the macrolevel: they focus on emergent properties of state and society, or on the international system as an emergent phenomenon of the interactions of states. As discussed in the intro- duction to this volume, complexity produces bottom-up theories and models. However, this chapter does not specifically address the ultimate actor—individu- als, often the decisive factors in tipping the balance of forces one way or the other. A full assessment of the past, present, and future of any social system would have to analyze the key individuals and groups who shape it. Having registered these caveats, let me summarize the essence of complex- ity and then apply it to explain divergent policy outcomes in the former Com- munist states of Eastern Europe and the USSR.


Nonlinearity and complexity are hallmarks of human social networks. Com- plexity theorists endeavor to explain the process of complex adaptation within complex systems—whether they be ecosystems, the Internet, or political systems. 2 The version of complexity used here—derived from the interpretation of complex adaptive systems (or CAS) developed by Stuart Kauffman and others at the Santa Fe Institute—is anchored in eight basic concepts. Three of these—emergence, agent-based systems, and self-organization—were described in chapter 1. This sec- tion considers in more depth the ideas related to coevolution, fitness, criticality, and punctuated equilibrium that are particularly relevant to the discussion in this chapter. Coevolution. 3 No organism evolves alone. Every individual, species, and so- ciety coevolves with others and with their shared environment. A change in any one actor or environment can alter the environment of multiple actors and chal- lenge their fitness. The more variables shape a system, the harder it is to antici- pate how change in one element will affect others (the “butterfly effect”). Fitness. CAS defines fitness as the ability to cope with complexity. To sur- vive challenges and make the most of opportunity, a fit organism can process in- formation about and deal with many variables. The theory posits that all life forms exist on a spectrum ranging from instability (chaos) to ultrastability (or- dered hierarchy). Fitness is found in the middle ranges of this spectrum between rigid order and chaos—not in a crystal, where every atom resides in an ordered hi- erarchy, nor in gases whose molecules move at random. Move too far toward ei- ther pole, and you lose fitness. Creative and constructive responses to complex challenges, however, are more likely to be found close to the edge of chaos than toward the other end of the spectrum.



The behavior of social systems emerges from the interactions of their members. This means that their fitness is a function of the interaction of indi- viduals within the social and political parameters of the society. Because the fit- ness of countries is an emergent property, it is not possible to predict with precision how countries will react to changes in the environment. Up to a point, countries that are more decentralized are expected to be more adaptable and, therefore, fitter. The fitness of the United States hovers close to the edge of chaos, while that of Singapore teeters on the brink of rigidity. Fitness Landscapes. Coevolution of units within a complex system can be mapped as a rugged landscape in which the relative fitness of each organism is shown as a peak rising or falling as a consequence of coevolution. As in an arms race, the peaks of a predator and its prey may gain or decline according to changes in their offensive and defensive capabilities. If attackers acquire more lethal weapons, the fitness peak of the prey will drop. If individuals among the prey population acquire characteristics that reduce their vulnerability, their peaks will rise. Self-organized Criticality. Balanced between order and chaos, a fit being is like a sandpile that, if one more grain of sand is added, may collapse in an avalanche. This fragile equilibrium is called self-organized criticality. The sand- pile metaphor, however, is not universally accepted and is not essential to com- plexity theory. Punctuated Equilibrium. The concept of punctuated equilibrium under- scores that evolution is often marked by surges of speciation and avalanches of extinction (Gould 2002). 4 Species often develop quickly, endure with little change for a long time, and then die out suddenly—not gradually. Thanks to mu- tation and self-organization, members of the species find their niche and hang on to it. When their environment changes, they must adapt or disappear. How long the system is stable and endures is difficult to predict—especially in politics. Scientists in many fields noticed in the 1990s that critical events occur more often—both earlier and later than forecast by the model of a bell- shaped curve.



The huge area to which we shall try to apply CAS is Eastern Europe and the for- mer USSR. Adapting Snyder’s (2000) analysis, we identify four large domains that took shape in Eurasia after the breakup of the USSR in 1991—each distin- guished by the way it dealt with ethnic and development issues. In zone A was a



set of countries that benefited from ethnic calm and enjoyed gradual economic and political development; in zone B, a shatterbelt of ethnic conflict and mater- ial regress; in zone C, a region virtually frozen in time—with little ethnic conflict and stagnant economic life (except in countries where the promise of carbon fuels brought injections of outside capital). Finally, we may distinguish a hybrid zone D where major countries—Russia and Ukraine—shared some but not all characteristics of the other regions. Zone A consists of societies and states that have experienced almost no ethnic violence and have made strong progress toward democratic institutions and economic development through market economics. From the former Yu- goslavia, the exemplar is Slovenia. From erstwhile Soviet allies in Eastern Eu- rope, the leaders are the Czech Republic, Poland, Hungary, and a late rising star, Slovakia. 5 Of former Soviet republics, only Estonia, Latvia, and Lithuania belong in zone A (Clemens 2001). Zone B comprises societies that became embroiled in severe ethnic fight- ing in the 1990s—Chechnya, most of former Yugoslavia, and the erstwhile Soviet republics of Armenia, Azerbaijan, Georgia, and Moldova. Each showed a very low capacity for coping with ethnic differences and the problems of establishing a viable economy and a stable democracy. In each case, as Snyder says, partial de- mocratization probably aggravated ethnic tensions. Thus, “democracy” made it harder for Armenia’s leaders to negotiate any kind of compromise with Azerbai- jan over Nagorno-Karabakh, because nationalist firebrands could mobilize votes against them. 6 Zone C refers to Central Asia and Belarus, where dictators suppressed eth- nic or other challenges to their rule. In the 1990s Tajikistan experienced much fighting between political rivals, but ethnic differences were not at issue. In the former Soviet republics of Central Asia, erstwhile Communist leaders became dic- tators claiming to be both nationalist and democratic. Kyrgystan had a free press for a time, but this ingredient of a true democracy disappeared in the mid-1990s. 7 President Aleksandr Lukashenko tried to russify Belarus and negotiate its union with the Russian Federation. His opponents sought to establish and maintain a clear Belarusian identity, but Lukashenko repressed them with lit- tle overt violence. Where to place the other states not clearly in one of these three zones? By the early twenty-first century Slovakia had clearly moved into zone A. There were signs that Bulgaria, Romania, Croatia, Montenegro, and perhaps even Serbia might follow suit. But the scales teetered. Each of these countries could readily drop into zone B or C. Thus, Serbia made major strides toward real democracy and peace with Montenegro in 2001–2, but could still become embroiled in



more ethnic warfare with Kosovars or the Hungarian-speakers of Vojvodina. By 2005 Bulgaria and Romania were becoming difficult to classify. Neither had suf- fered much ethnic violence in the previous fifteen years, but each had a low HDI ranking compared to, say, Slovenia or Slovakia. Each had been admitted to NATO, on the hope that they would contribute to George W. Bush’s War on Terrorism, but neither came close to qualifying for membership in the EU. As this chapter is designed to illustrate the uses of CAS for understanding ethnic conflict and development in post-Soviet Eurasia, the precise allocation of these countries to group B or group C is not crucial. Their location on the A-to-D spec- trum may well change, influenced, for example, by accession to or distance from the European Union. The two largest Slavic states emerging from the USSR comprised the hy- brid zone D. By the early twenty-first century neither Russia nor Ukraine had achieved a real democracy or a strong market economy. But neither suffered from outright ethnic violence, except for Russia’s wars against Chechnya (1994–96 and again after 1999). Russia’s ethnic nationalism was qualified by civic nationalism. Thus, Moscow recognized Tatarstan’s “sovereignty” within the Russian Federation (Rossiskaia Federatsiia, where rossiskaia is more inclusive than the term russkaia, as “British” takes in more diversity than “English”). There were signs early in the century that the Russian Federation’s Duma and President Putin might require that any would-be Russian citizen be fluent in Russian. Ukraine achieved a kind of civic nationalism incorporating native Russian and Ukrainian speakers. Kyiv avoided war with Russian irredentists in the Crimea and with Moscow over its claims to ships and naval facilities in Sevastopol. Like Russia, however, Ukraine failed to use effectively its vast natural resources and highly educated work force (D’Anieri 1999). Transparency International placed Russia and Ukraine among the world’s most corrupt countries in the late twenti- eth and early twenty-first centuries. Ukraine’s “orange revolution” in 2005 promised fundamental changes. President Viktor A. Yushchenko’s administration followed a Western orientation even as it labored to overcome the misgivings of diffident Russian-speakers in Ukraine. By year’s end, however, many of the coun- try’s old problems had reemerged, albeit with new faces.


Adopting the language of CAS, this chapter argues that countries such as Slove- nia, the Czech Republic, and Estonia in the 1990s demonstrated a high level of























n 133






















155 (RE)

151 (RE)


72 (MU)

19 (MF)

28 (MF)

18 (MF)

33 (MF)

35 (MF)

56 (MF)

62 (MF)

52 (MF)

68 (MF)

n 153


6 (F)

6 (F)




Partially Free

Index 2003


Partly free

Partly free

Not free

Not free

Not free












n 175





















n 175





















Norway Belgium United States Canada Japan Israel Greece Cyprus Singapore Slovenia Korea, Republic of Brunei Darussalam Argentina Estonia Cuba Belarus Malaysia


dom Index” is from Freedom House at “Economic Freedom” is from Heritage Foundation at “Honesty Rank” is from Trans- parency International at

Sources: “HDI” and “GDI” are from U.N. Development Programme, Human Development Report 2003 (New York: Oxford University Press, 2003), tables 1 and 22. “Free-











n 133









127 (MU)

119 (MU)



89 (MU)

99 (MU)

72 (MU)

40 (MF)

n 153


Code: For economic freedom, F free; MF mostly free; MU mostly unfree; RE repressed.

Index 2003


Partly free

Not free

Not free




n 175










n 175










TABLE 4.1. (continued)

Cape Verde









fitness. As we see in table 4.1, they scored much higher on the UN Human De- velopment Index and in Freedom House ratings for political and civil liberty than did comparable peers such as Serbia, Romania, and Belarus. Each country in zone A joined both NATO and the European Union. In zones C and D, by contrast, few countries showed much interest in or had much prospect of join- ing NATO or the EU in the foreseeable future. Societies in zone A achieved high levels of fitness on many fronts after the demise of the Soviet empire. Success in one domain helped them cope with problems in others. Ethnic peace made it easier to raise living standards, consol- idate democracy, and nourish creativity. Economic advances in Estonia, for ex- ample, make it easier for Tallinn to provide welfare benefits for Russian-speakers residing in Estonia but who were not citizens. Countries in zones B, C, and D displayed low levels of overall fitness even though many possessed assets lacking in zone A. Thus, Azerbaijan, Kazakstan, and Russia possess energy resources far superior to those in any lands in zone A. Parts of Ukraine and Russia have better soil as well as much richer mineral deposits than any country in zone A. Georgia, Armenia, Ukraine, and Russia have evolved from states and cul- tures dating back more than a thousand years. Slovenia, by contrast, was never an independent state before 1992. Estonia and Latvia had only two decades of independence between the two world wars. Most countries in zones B, C, and D faced simpler ethnic challenges than in many zone-A countries, because they were more homogeneous. Ethnic mi- norities were very small in Belarus, Moldova, the South Caucasus, and in most of Central Asia (except for Kazakstan). About four-fifths of the Russian Federa- tion’s population was Russian but most other groups in the federation spoke Russian. A million or so Chechens occupied only a dot on the federation’s pe- riphery. Still, the governments in zones B, C, and D experienced great difficulty in dealing with ethnic minorities. By contrast, Estonia and Latvia in the 1990s faced minorities of Slavic speakers that made up more than one-third of the res- ident population. Estonian and Latvian leaders espoused a kind of ethnic na- tionalism tempered by civic moderation. They instituted a naturalization process that required aspiring citizens to pass residency, language, and civic tests. By the early twenty-first century—more than a decade since indepen- dence—few of either country’s Slavic speakers had acquired a working knowl- edge of the official state language. Children and young adults, of course, learned Estonian or Latvian more readily than most of their elders. Still, eth- nic tensions produced no deaths in the Baltic. Estonia even permitted nonciti- zens to vote in local elections. Indeed, the city councils in Riga as well as Tallinn were sometimes dominated by coalitions of old leftists and “unity” parties devoted to the interests of Russian-speakers.




Self-organization takes in more than democratic politics. It entails also a market economy and a social system that, from the bottom up, produces innovation and ways to meet needs and exploit opportunities. The centralized regimes in zones B, C, and D attempted to direct economic and cultural life as well as politics from the top down. As in Soviet times, they repressed newspapers and news media that contradicted the official line. President V. V. Putin was designated acting president by his predecessor before a snap election that confirmed the ap- pointment—bolstered by a then popular war against ethnic aliens. Privatization in Russia and most other countries in zones B, C, and D permitted privileged in- siders to seize public resources at low cost. So great was the plunder that by 2004 there were more billionaires in Moscow than in New York.


This concept explains several features of post-Soviet Eurasia. Most countries close to Western Europe have coevolved with the West more quickly and thoroughly than those that are more distant. Thus, the Czech Republic is more “First World” than is Kyrgyzstan. But if a country shuts itself off or is otherwise isolated from global trends, its overall fitness will suffer. Thus, Albania abuts Greece, but its Communist rulers sought autarky. Belarus abuts Poland and Lithuania, but the government’s orientation toward Moscow serves to minimize productive exchanges with the West. Kazakstan “coevolves” with foreign oil drillers, but this is a very limited facet of coevolution. In many respects Kazakstan and other Central Asia states resemble Communist Albania—cut off from the West by government fiat.


Nowhere in the formerly Communist lands did there emerge strong patterns of cooperation. Instead, it was more like “every state for itself”—indeed, “every na- tional and subnational group for itself.” Even in zone A, each state focused on joining Western Europe and NATO—not on cooperating for shared ends with its immediate neighbors. Rivalries persisted in the Caucasus even though both Georgia and Armenia needed the energy that Azerbaijan could provide and for which it needed buyers. Central Asian states proved unable, after as well as before independence, even to find ways to stop the shrinkage of the Aral Sea—an environmental disas- ter that affects the whole region. The Commonwealth of Independent States had many accords registered on paper but never executed. Subgroups meant to either to resist or to strengthen the commonwealth also achieved little.



Whatever the shortfalls of the European Union, it is a triumph of coop- eration compared to the beggar-thy-neighbor behaviors of ex-Communist soci- eties. Indeed, it was EU and NATO demands for settled borders and ethnic peace that persuaded Hungary and Romania to patch up their differences and convinced Estonia and Latvia to renounce some border regions seized by Moscow in the 1940s.

Agent-Based Systems

In zone A individual agents are free to innovate and carry on their business with a minimum of government control. The system is shaped by its members rather than by a central command. This is not quite “order for free,” which Kauffman’s version of CAS attributes to established ecosystems (such as coral reefs). Still, it resembles the positive results that Adam Smith expected if individuals were allowed to do what they do best, as if guided by an “invisible hand.”

Self-Organized Criticality

CAS warns that societies may be less fit than they appear. Fitness depends on the harmony of many factors. Just as an extra grain of sand may cause a sandpile to collapse, a new or heavier burden could seriously weaken an apparently fit so- ciety. How would Lithuanians respond if a faulty nuclear reactor shut down their energy supply or spread poison to the air and soil? Or if Russians for a pro- longed time simply turned off the oil and gas flows on which Lithuania (and many post-Soviet societies) depends? Each Baltic country endured severe stresses in the 1990s, but one cannot be sure what grain of sand—what policy in- novation or social change—may start an avalanche that radically changes a soci- ety and its fitness.

Punctuated Equilibrium

The concept of punctuated equilibrium warns us not to expect steady progress in fitness. West European unification did not emerge gradually but in sharp jumps and with some steps backward. Meaningful social change often requires a period of preparation. New generations can be educated. In Estonia and Latvia many young persons who speak Russian at home are learning the official state lan- guage. Accumulating experiences may tip even middle-aged Russian-speakers to- ward integration with native Balts. Long plateaus without improvement may drive some people to depart or to take drastic steps to effect change. But regress



is also possible. How long will displaced persons in Bosnia and other parts of the former Yugoslavia wait until they return to their homes?

Fitness Landscapes

The relative fitness of a fruit fly and a frog population may be portrayed as “peaks” that rise and fall with coevolution. Can we graph changing patterns of fitness among the societies of post-Soviet Eurasia? This is not a simple task, if only because fitness among humans is multidimensional. The UN Human De- velopment Index provides a solid starting point to measure public health, edu- cation, and material living standards. 8 If we focus on ethnic problems, we would also study measures of ethnic harmony and its opposite—injury, dislocations, and deaths caused by ethnic un- rest. We expect that low fitness in this domain will tend to correlate with low scores in overall human development, lack of political and civil liberties, low technological achievement, and corruption. Though it is difficult to show all these variables in a single peak, a cobweb graphic could illustrate the correlations suggested here. 9


Theories of complex adaptive systems provide useful concepts for analyzing ethnic issues and other ingredients of societal fitness. But they offer only general princi- ples for anticipating future outcomes or prescribing constructive policies. In this re- gard, however, it does no worse than most competing theories—few of which provide useful handles for predicting or shaping the future. Indeed, if CAS is cor- rect about the role of self-organization in fitness, social Darwinists and ultrareal- ists are wrong: success in politics does not derive from raw power plus cunning. The fundamental insight of CAS is its prediction that fitness will be found along the middle of the bell curve ranging from rigid order to random instabil- ity, though high creativity is most frequently found close to the edge of chaos. This insight helps explain why Central Asia is frozen in time, why the Caucasus explodes, and why Russia resorts to an iron fist to overcome chaos, and why Slovenia and Estonia adapt well to their new freedoms. This insight has clear policy implications: avoid the extremes of dictator- ship and anarchy. To generate a healthy and innovative society, cultivate self- organization—not a system steered from on high. Western policymakers and in- vestors should not count on authoritarian regimes in Kazakstan or Azerbaijan to



maintain order forever. They should not prop up local dynasties in the hope of securing privileged access to oil and gas. Outsiders cannot compel internal re- forms but should do what they can to nudge these societies toward greater self- organization. Countries such as Azerbaijan suffer not only from top-down controls but also from a rent-seeking mentality among many well-educated per- sons who will eventually play major roles in business and politics. Their attitudes as well as formal structures will determine whether Azerbaijan and Kazakstan use their petrodollars to create values for the entire community (as in Norway) or fol- low more closely the Saudi Arabian or Nigerian models. CAS attention to independent actors agrees with the growing conviction among political scientists that formal and informal institutions of civil society help to buffer the ravages of free markets and curb the excesses of willful govern- ments. The stronger and more diverse the independent agents shaping the for- merly Communist societies, the healthier and fitter they will be. Constructive policies will cultivate creative individuals, businesses, and NGOs that enlarge pub- lic goods and are not dominated by government. These independent agents face a difficult struggle against the moral legacies of Communism—corruption, group- think, and a welfare mentality that discourage initiatives from the bottom up. Even if the goal of self-organization seems clear, questions arise about the road to this goal. What if democracy terminates democracy—as happened in Ger- many in the 1930s? Is self-organization desirable if the majority votes against the minority, as happened in Sri Lanka and as Serbs feared would happen in a ma- joritarian Bosnia? And what if the majority brings in a government that imposes the laws and mores of one religion, as in parts of Nigeria?


Culture matters. 10 All the societies in zone A became oriented toward universal literacy, free thought, and open debate (relative to most other societies) long ago. The societies in zones B, C, and D moved toward universal literacy only in the past 100 years. 11 Many regimes in these zones still discourage or try to prevent open debate on policy and other important issues. Following the leads of John Wycliffe, Jan Hus, Martin Luther, and other re- formers, each society in zone A acquired its sacred religious texts in the vernacular between the fifteenth and seventeenth centuries. For the first time in history, some princes and religious leaders also urged individuals—female as well as male—to read and interpret sacred texts on their own. This twin revolution helped to liberate all who experienced it (Clemens 2005). After the Peasants’ Revolt, however, Luther feared that he was provoking chaos. He then wrote his Short Catechism instructing people what to believe. But Luther could not stop the transformation he had



ETHNIC CONFLICT AND DEVELOPMENT 85 Figure 4.1. Date the Bible Published in Vernacular Correlated with HDI

Figure 4.1. Date the Bible Published in Vernacular Correlated with HDI Rank

unleashed. The synergies of literacy and individualist thinking were empowered by the printing press, the Renaissance, the discovery of the New World, and growing refinement of scientific methods. Catholic France and Italy had Bibles in the ver- nacular even before Luther’s challenge to Rome. In the seventeenth century Swe- den’s monarchy and state church wanted their subjects—even servant girls—to read and discuss the Bible. Bibles in the vernacular also helped cultivate a sense of na- tional identity (Hastings 1997; Lepore 2002). Certainly many factors shape human development, but figure 4.1 shows a strong correlation between high HDI scores and early publication of Bibles in the vernacular. Where Orthodox Christianity prevailed, Bibles in the vernacu- lar were not widely published until the late nineteenth century or the twentieth century. (The sole exception was Romania, which published both the New and Old Testaments in the seventeenth century.) 12 Wide-scale literacy came to the Orthodox countries much later than in Protestant and Catholic countries or in Jewish communities.



Unlike the Christian Bible, the language in which the Qur’an was first written is regarded by Muslims as sacred—the only truly accurate way to express God’s message. Islamic societies did not encourage mass literacy or, as a rule, individual interpretation of sacred texts. For Arabs as well as non-Arab Muslims, memorization and recitation of the Qur’an have been far more important than discussion. Few Bosnians, Azeris, or Central Asians have been able to read clas- sical Arabic. 13 Translations of the Qur’an into Persian, Turkish, and Chinese were for many years largely in the form of paraphrase and commentary. 14 By the 1950s Communism had brought near universal literacy to the USSR and Eastern Europe—even to Albania. But Communist regimes and schools discouraged freethinking. Centralized controls channeled thought and discouraged debate. Even when Communist regimes sought to foster technolog- ical innovation, this proved difficult, because of state secrecy and communica- tions networks that ran vertically but not horizontally. The Soviet dissident Andrei Sakharov lost his security clearance and was sent into internal exile; many other dissidents suffered worse fates.


In the early twenty-first century most governments in zones B, C, and D still do not encourage free thought and debate. Until they do, they will not possess a nec- essary ingredient of social fitness. Comparatively unfit, they will lag their more westernized neighbors in many ways. In the language of CAS, these countries— even erstwhile superpower Russia—will wander in the valleys of a fitness land- scape, looking for ways to propel their peak(s) upward. Lacking self-organized economies and polities, they will have great difficulty dealing with ethnic issues within and across borders. Democratic in form but authoritarian in substance, they will tend to repress dissent rather than create solutions for mutual gain. Elections held in 2000 and 2004 suggested that most Russians still hoped that a vigorous leader, Vladimir Putin, like a legendary vozhd, would unite and mobilize the people for a better life. Having won many votes by intensifying the war against Chechens, Putin proceeded to silence independent media, jail the country’s richest man when he sought to shape political life, and pushed through reforms permitting the Kremlin to appoint regional governors instead of having them directly elected by their subjects. In the mid-1990s Georgians welcomed Eduard Shevardnadze back from Moscow to Tbilisi, counting on him to end a reign of chaos. But reliance on top- down leadership did not end turmoil in Georgia. Rather, it added to the already heavy burdens of corruption at the center. Shevardnadze was ousted in 2003 in a popular revolt led by an American-trained lawyer, Mikheil Saakashvili, who



promised to replace corruption and chaos with a rule of law. By 2004, however, he was promoting his own cult of personality. Saakashvili had learned the rhetoric of democracy, but—in a society that wants a strong, charismatic leader— he gravitated toward the national norm. Would closer ties with America improve fitness in former Soviet republics? For countries such as Georgia and Uzbekistan, closer ties with the U.S. hyper- power might bring material gains but could also weaken self-organized fitness. Lacking internal strength, each people’s capacity to cope with ethnic diversity might then decline—especially if exploited by political entrepreneurs hoping to gain power and wealth from others’ differences (Singer 1999, 57).

Where Can We Go from Here?

Complexity cannot generate precise algorithms for analyzing ethnic conflict. But it does provide valuable conceptual tools for this task—principles, metaphors, models. Thus, a major insight of CAS is the concept of societal fitness. Unlike neorealists who believe that relative material power—missiles and GDP—is the best guide to world politics, CAS suggests each actor’s most basic need is a ca- pacity to cope with challenges at home and abroad, including ethnic diversity. How could we operationalize these concepts? Let us assume that HDI rank is an approximate indication of societal fitness, and that societal fitness in a large, modern society depends on universal literacy and free expression. Let us assume also that the onset of universal literacy and free thought can be traced to the dates when the most sacred books of that society were published and when conditions were established in which they could be subjected to individual interpretation. The graphics in this essay suggest a correlation between HDI rank and the date when the Bible was published in the vernacular of the countries that later became units of the Soviet bloc and Yugoslavia. To really understand these rela- tionships, however, each variable would have to be studied in depth. Here are a few of the tasks:

1. Identify the conditions in which the Bible was rendered in the vernacular and published in each country. When? By whom? Why? How many copies? Were they repressed (as in Russia in the 1820s)? Did they sell? When did subsequent printings take place?

2. Trace the evolution of literacy in these countries. Develop a common standard for measuring literacy. How much literacy was there before the printing press and Martin Luther? How did it evolve in the decades and centuries after Luther? What forces and institutions resisted or facilitated the growth of literacy?




Trace the growth of independent thinking. What indicators— in science, the arts, politics, and economics—show indepen- dent thinking? Such indicators are more evident in a relatively open metropole such as England than in a repressed depen- dency such as the places we now call Estonia or Slovakia.


Adapt the approach used to study traditionally Christian coun- tries to those in the Muslim and other religious traditions.


Determine a way to prove causation rather than mere correla- tion. How should we weigh the contribution of one factor, such as literacy, against others, such as growing wealth?


Analyze the chicken-and-egg. Ask which came first: individual freedom or the twin revolution? Long before Luther, condi- tions favoring individual freedom were stronger in some re- gions (such as Bohemia) than in Byzantium or Russia.


Distinguish the kinds of ethnic/national consciousness that existed in previous centuries (for example, among Bohemia’s Hussites) and that of the last century or two.


Distinguish technology from the culture where it is applied. Why did most Europeans respond with alacrity to the print- ing press while Islamic cultures did not?


Learn from outliers: Romania (Orthodox but closer to Rome than to Russia) had the Bible relatively early but nonetheless has a low HDI ranking. Slovaks got the Bible relatively late but achieved a fairly high HDI score in the late 1990s. Rich data on individual countries is available in the human devel- opment reports produced by local social scientists in many East European and former Soviet states; they are accessible online from the U.N. Development Programme.


Consider the shortcomings of the twin revolution. If literacy and independent thinking conduced to human development, why has the West shown so much intolerance, violent nation- alism, and war? Perhaps the twin revolutions were necessary but not sufficient for overall fitness.


Consider the policy implications: If high levels of human devel- opment may be traced to a twin revolution that began five hun- dred years ago, how can they be fostered in societies that have experienced one or both revolutions only in recent decades?



The argument here is that to understand present trends and develop construc- tive ways to deal with ethnic diversity, we must review not just decades but cen- turies of history—cultural, political, economic. The tasks in such work are complex and vast, but can be made more manageable by using the conceptual tools developed by CAS for studying societal fitness.


1. The following interpretation of complexity theory is based largely on

the work of Stuart A. Kauffman (1993, 1995, 2000) and other scholars—from the Nobel physics laureate Murray Gell-Mann to the Nobel economics laureate Ken- neth Arrow—who have interacted at the Santa Fe Institute. For early work at the Santa Fe Institute, see Roger Lewin (1992). The Santa Fe Institute publishes the journal Complexity and working papers such as Martin Shubik, “Game Theory, Complexity, and Simplicity Part I: A Tutorial” (98-04-027); and Melisa Savage and Manor Askenazi, “Arborscapes: A Swarm-based Multi-agent Ecological Dis- turbance Model” (98-06-056). Robert M. Axelrod (1997a, 1997b) and Axelrod and Michael D. Cohen (1999) have used a variety of methods to resolve complex problems. For an application of complexity theory by a former student of Axel- rod, see Lars-Erik Cederman (1997). For a book that blends historical analysis, international relations theory, and systems analysis, see Hendrik Spruyt (1994). In the same vein, Robert Jervis (1997) examines the complex interactions of so- cial units, but says little about self-organization. Compare with James N. Rose- nau (1990). For related work by IR specialists, see papers given by Michael Lipson (1996) and Matthew J. Hoffmann (1999). For applications to manage- ment, see Roger Lewin and Birute Regine (2000). The utility of complexity the- ory is assessed by Hayward R. Alker and Simon Fraser (1996) and continued in Alker (1996). For a skeptical view of complexity theory, see John Horgan (1996). For a more balanced appraisal, see “Edge of Chaos” and many relevant entries in Ian Marshall and Danah Zohar (1997).

2. Many aspects of nonlinearity are examined in Diana Richards (2000),

where nonlinear models are applied to federalism, alliance formation, epochs in political economy, environmental regimes, and outbreaks of war. Richards says that nonlinear modeling can build directly on existing economic theory. In political science, however, nonlinear modeling must invent a method unique to

the problem at hand—from dynamical systems to spatial voting models to time- series analysis.

3. On coevolution, see also Charles J. Lumsden and Edward O. Wilson

(1981); Martin A. Nowak et al. (1995); and Edward O. Wilson (1998).

4. But “punctuation” may result from an incomplete fossil record;

also, Gould may have confused “individual,” “class,” and “species.” See Mark



5. See the studies done for Freedom House by Adrian Karatnycky,

Alexander Motyl, and Aili Piano (2001) and the country reports published regu- larly by Human Rights Watch, Amnesty International, Transparency Interna- tional, the United Nations Development Programme, and the U.S. Department of State—all accessible on the Internet.

6. Citing several studies, Snyder concludes that the ethnic content of the

Moldovan conflict was ambiguous. The Moldovan government in the early 1990s was nationalistic, but Russian-speakers in the breakaway “Transdniestr Re- public” were driven more by nostalgia for the Soviet empire than by nationalism (Snyder 2000, 250–51).

7. Abutting the former USSR, Moscow’s one-time client state Mongolia

is a special case. In the 1990s Mongolia moved quickly toward democracy, even though it was poorer than most parts of the USSR and had a weak infrastructure for education and communication. The country had few internal ethnic prob- lems (90 percent of the population is Mongolian; 4 percent Kazak; 2 percent Russian; 2 percent Chinese; 2 percent other) and did not clash with China

despite the potential for expansionist claims by each side.

8. For discrepancies between the UNDP Human Development Report

published annually in New York and country reports published by UNDP offices in Baltic and East European capitals, see Clemens 2001, 110–11.


For a model, see Maruca 2000, 24.


For a range of viewpoints, see the essays in Lawrence E. Harrison and

Samuel P. Huntington (2000); Mariano Grondona (1996); and Dominique Jacquin-Berdal et al. (1998).

11. Literacy rates are difficult to track and measure, but estimates for

many formerly Communist countries are at Snyder 2000, 200–202.

12. Dates of Bible publication in many languages, including those

native peoples of Siberia and North America, are given in, which also reproduces the opening lines of St. John’s Gospel in each language. This survey is so detailed that it notes the very different years for Bible publications in Tartu Estonian (no longer spoken) and Tallinn Estonian and for Eastern and Western Livonian, each spoken now by only a hundred or so persons.

13. An Azeri in Moscow showed me his family Qur’an written in Arabic,

for him a completely unknown tongue.

14. Some translations into Malay have been so literal that they were not

intelligible without prior knowledge of Arabic. But Bosnians could have read a translation into Serbo-Croat in 1875 (Swartz 2004).




Alker, Hayward R. 1996. Rediscoveries and Reformulations: Humanistic Methodolo- gies for International Studies. Cambridge: Cambridge University Press.

Alker, Hayward R., and Simon Fraser. 1996. “On Historical Complexity: ‘Natu- ralistic’ Modeling Approaches from the Santa Fe Institute.” Paper de- livered at the American Political Science Association Annual Meeting, San Francisco.

Axelrod, Robert M. 1997a. “Advancing the Art of Simulation in the Social Sci- ences.” Paper delivered at the International Conference on Computer Simulation and the Social Sciences, Cortona, Italy.

———. 1997b. The Complexity of Cooperation: Agent-Based Models of Competition and Cooperation. Princeton, NJ: Princeton University Press.

Axelrod, Robert M., and Michael D. Cohen. 1999. Harnessing Complexity: Orga- nizational Implications of a Scientific Frontier. New York: Free Press.

Brown, Michael E., ed. 1996. Debating the Democratic Peace: An International Security Reader. Cambridge, MA: MIT Press.

Bunce, Valerie. 2004. “Status Quo, Reformist, or Secessionist Politics: Explain- ing Minority Behavior in Multinational States.” Paper prepared for the Workshop on Nationalism, Secession and Inter-Ethnic Cooperation and Conflict, Cornell University, April 23–24.

Cederman, Lars-Erik. 1997. Emergent Actors in World Politics: How States and Nations Develop and Dissolve. Princeton, NJ: Princeton University Press.

Clemens, Walter C., Jr. 2001. The Baltic Transformed: Complexity Theory and Euro- pean Security. Lanham, MD: Rowman & Littlefield.

———. 2004. Dynamics of International Relations: Conflict and Mutual Gain in an Era of Global Interdependence. 2nd ed. Lanham, MD: Rowman & Littlefield.

———. 2005 “The Culture of Democracy: Literacy, Individual Freedom, and Human Development.” Work-in-progress at Boston University and Harvard University.

D’Anieri, Paul J. 1999. Economic Interdependence in Ukrainian-Russian Relations. Albany: State University of New York Press.

Doyle, Michael W. 1997. Ways of War and Peace: Realism, Liberalism, and Social- ism. New York: W. W. Norton.

Elman, Miriam Fendius, ed. 1997. Paths to Peace: Is Democracy the Answer? Cambridge, MA: MIT Press.

Epstein, Joshua M., and Robert Axtell. 1996. Growing Artificial Societies: Social Sci- ence from the Bottom Up. Cambridge, MA: MIT Press.



Friedman, Edward, and Barrett L. McCormick, eds. 2000. What If China Doesn’t Democratize? Implications for War and Peace. Armonk, NY: M. E. Sharpe.

Gould, Stephen Jay. 2002. The Structure of Evolutionary Theory. Cambridge, MA:

Harvard University Press.

Grodona, Mariano. 1996. El mundo en clave. Buenos Aires: Planeta.

Harrison, Lawrence E., and Samuel P. Huntington, eds. 2000. Culture Matters:

How Values Shape Human Progress.