|
|
|
|
Change and Process In a Complex World: Using Complexity Theory to Understand World Politics
*
Matt Hoffmann and David Johnson **
International Studies Association
March 18-21, 1998
[Scholar's at the 1997 meeting of the International Studies Association] seem almost as unclear as the rest of us about the implications of phenomena such as globalization and the retreat of the nation state.
Paul Knox 1
The long and the short of it is, then, that we are not very good as a discipline at studying the possibility of fundamental discontinuity in the international system. . . . We lack even an adequate vocabulary; and what we cannot describe, we cannot explain.
John Gerard Ruggie 2
It is something of an understatement to say that the discipline of international relations was surprised by the end of the Cold War and the study of world politics appears to be at a crossroads in its aftermath. It seems to us, as the journalist and the scholar quoted above suggest, that scholars of world politics continue to be confused by this extraordinary event. Indeed, it's not even clear what we mean when we say "the end of the Cold War." 3 We believe the passing of the Cold War was marked by three important occurrences (or "non-occurrences"). First, the dramatic shift in the perception of the Soviet Union of its security needs which led to a cascade of changes in its strategic position (from its withdrawal from Afghanistan to the fall of the Berlin Wall and the stand-down of its strategic missile forces). Second, the failure of the NATO alliance to press the military and strategic advantage arising from the Soviet retrenchment. Finally, the general absence of a major war as the catalyst for these changes. These three aspects are at the heart of the surprise and confusion that continues to exist in world politics about the end of the Cold War.
In the aftermath of the Cold War, scholars are perceiving an increasingly fast-paced, dynamic and complex world. Every generation sees their situation this way 4 , but today there are possibly qualitative differences and, in any case, it is the perception of the "pace" of events which partially drives the evaluation of understanding. Furthermore, traditional tools -- perspectives and methods -- are not well suited for investigating this dynamic change nor for dealing with the uncertainty inherent in it. Again, this dynamism may not be new but its increased "pace" (and/or the heightened recognition of it) is making the conventional analysis of world politics more and more difficult.
We believe the confusion in world politics about the end of the Cold War has something fundamental in common with puzzles which have been baffling other disciplines in the natural and social sciences. How did complex inorganic molecules manage to transform themselves into the complex organic molecules that represent the emergence of life on Earth? Why is it that drawing energy and material resources from the environment to be used for human development threatens human survival through degradation of the ecosystem? Why do stock market traders use today's information to guide their trading decisions for tomorrow despite the contention of neoclassical economists that it is useless to do so? At first glance, these puzzles would seem to be unrelated; however, the common thread in each case is that the puzzle concerns a complex, dynamic system. The emerging science of complexity has been successful in addressing each of these puzzles in the natural and social sciences and we believe that complexity science can also be used to explore the end of the Cold War and other emerging puzzles in world politics.
The new science of complexity is a promising and potentially useful tool that incorporates an epistemology and methodology which may enable scholars of world politics to embrace the newly-apparent dynamism and uncertainty. Complexity is a "cutting-edge" science being explored across a wide range of disciplines which offers a new set of conceptual and theoretical tools to students of world politics. We believe the use of complexity science in pursuit of understanding in world politics ought to be explored because we believe world politics is what is known in complexity science as a "complex adaptive system." The tools of complexity science should help us to better understand world politics particularly by providing advantages over conventional approaches which are ill-equipped to address dynamic change.
Let us return to the three puzzles in the natural and social sciences we mentioned earlier to illustrate briefly how complexity theory works. The emergence of complex organic molecules was the result of the processes of self-organization and aggregation occurring among complex inorganic molecules. Ecological degradation results from a development model that ignores the complex interactions between human systems and other ecological systems. Stock market traders are inductive, boundedly-rational agents which adapt their preferences and strategies on the basis of experience with the dynamic behavior of the market. Each of these situations are examples of complex, adaptive behavior.
This introduction to the application of complexity theory to world politics is designed to accomplish three tasks. First, it serves as a general introduction to the concepts that form the basis of complexity science and lays the groundwork for the application of complexity theory to the study of world politics. Second, it presents a discussion of the two major advantages that complexity theory offers to the study of world politics: rejection of the conventional, equilibrium focus in favor of an emphasis on process and the capability to account for dynamic change. Finally, the paper seeks to highlight the potential pitfalls we anticipate in adopting complexity theory in order to preempt misunderstanding of what the perspective offers for the study of world politics.
A general discussion of complexity science is presented first. This brief introduction is followed by an examination of the reasons for inadequacies which have become apparent in the aftermath of the Cold War in traditional and recent alternative perspectives on world politics. The discussion then turns to the application of complexity science to the study of world politics, followed by a brief warning about the ways in which complexity theory might be misapplied. Our purpose here is not to present the definitive case for the study of world politics using complexity science but rather to make the case for adoption of a complexity approach.
The Complexity Process
Why did the stock market crash more than 500 points on a single Monday in 1987? Why do ancient species often remain stable in the fossil record for millions of years and then suddenly disappear? In a world where nice guys often finish last, why do humans value trust and cooperation? At first glance these questions don't appear to have anything in common, but in fact every one of these statements refers to a complex system.
M. Mitchell Waldrop 5
Complexity theory is a "cutting edge" scientific endeavor. It is not the Newtonian science currently emulated in our discipline. However, because of its focus on process, it can and does make generalizations across varied and historically contingent phenomena--a hallmark of scientific enterprise. The key difference, in the language of international relations, is that it makes no assumptions about the structure or the agents in world politics (and therefore does not generalize about either), but rather makes general assumptions about the processes by which both are generated. It is a wedding of the inductive and the deductive--the process is deductively assumed, and the nature of agents and structures is inductively examined. (Traditional approaches rely upon substantive assumptions about the inherent nature of agents and/or structure under equilibrium conditions for explanation, giving rise to analysis which reduces to these essential agents or structure.) A justification for the superiority of a complexity theory approach is found in the following sections. First, we articulate the premises and concepts that comprise the general pattern.
Complexity theory is about process. It is about the process that complex systems, whether they are immune systems, economies, ecosystems, or societies, use to adapt and to persist. As John Holland notes, "the question of coherence under change is the central enigma for each [complex system]." 6 In a political science context, complexity theory can help us to understand, among other things, why, in the words of James Rosenau, "states make it from Friday to Monday." 7
Complexity is inherently focused on understanding this change under continuity, and as a way to capture this phenomenon, the theory puts forth the exploration of the dynamic process of the complex adaptive system (CAS) and posits that a complex adaptive system is an apt description for a wide-range of systems not usually thought of as similar in any way (e.g. the immune system and the economy share some fundamental procedural characteristics). As Holland notes:
each of the systems involves a similar evolving structure. That is, these systems change and reorganize their component parts to adapt themselves to the problems posed by their surroundings. This is the main reason the systems are difficult to understand and control--they constitute a moving target. We are learning, however, that the mechanisms that mediate these systems are much more alike than surface observations would suggest. 8
The complexity process is what remains similar across issues and systems. According to this logic, political science has been looking for generalizations in the wrong areas--in structures and agents. But are human systems complex adaptive systems? This must be explored.
Holland has articulated characteristics that make systems complex and adaptive and has also examined in detail the basic structures and components of complex adaptive systems. (See Table 1 for a comparison of the characteristics of complex adaptive systems with those of political systems.) The deductive assumption we make is that human systems do indeed possess these characteristics and behave as complex adaptive systems. 9 This assumption leads us to a further examination of complex adaptive system features which are explained in greater detail in the Appendix.
Table 1: Characteristics of Complex Adaptive Systems and Human Political/Social Systems 10
| Complex Adaptive System | Human Systems |
| Network of many agents | There are many agents and meta agents in political systems, including people, committees, states, organizations, regional organizations, the UN, etc. |
| Agents in an environment produced by interactions | Institutions and norms (the environment) are constantly produced by interactions. |
| Control is dispersed/many levels of organization | Family, religion, race, governmental levels, international organizations, etc. (different power and authority relations) |
| Systems constantly revise and rearrange organization, generating new structures and agents | Authority and power relations shift, norms transform and importance of different organizations change over time. New organizations, authority and power relations are constantly emerging. |
| Systems anticipate the future | Goal oriented behavior and action prevalent. Systems as a whole adapt. States create and sign treaties--these actions are anticipations of the future and they constrain or shape future behavior. |
| Equilibrium does not exist | Constant dynamism is prevalent even in day to day continuity of society--systems "make it" from Friday to Monday. |
Complex adaptive systems are comprised of networks of many agents which interact and adapt in ways that cause the complex adaptive system to evolve and its macroscopic properties to emerge. The behavior of agents is governed by internal rule models which are learned from experience (experience being past results of interaction with other agents and the environment) and are used to anticipate the results of behavior. The rules change over time (the agent adapts) in response to interactions with other agents and the environment. The changes take place as the results of behavior are evaluated in light of the interactions with other agents and the environment. The processes of adaptation, self-organization and co-evolution are the mechanisms through which agent-level micro interactions lead to system-level macro outcomes.
Self-organization is the result of interactions. Agents aggregate information from stimuli into categories by way of their rule models. This allows them to describe a situation, predict the consequences of behavior, and prescribe a choice of action. The actions that agents take interact and the system as a whole aggregates the behaviors of the agents. This second aggregation leads to the emergent properties of the system (often its structure). This is self-organization--no outside force determines the nature of the structures generated. Co-evolution is the aspect of the system's aggregation which constrains the interactions of the agents. Agents react and tailor their behavior to the conditions created by the existence and behavior of other actors (not to some objective optimum). With the concepts of adaptation, self-organization, and co-evolution a complex adaptive system can be modeled as agents constantly adapting and interacting with other agents while that system is similarly adapting and interacting with other complex adaptive systems.
These concepts make up the complexity theory process posited to be similar across a diverse array of systems, including political ones. Agent (actor) interactions, based on rule models built through experience and evaluated through feedback, are aggregated into macro behavior and outcomes. This leads to the self-organization of structures (institutions, organizations, etc.) which form the external "environment" for agent interaction. Agents and meta-agents, and systems themselves, co-evolve as the agent interaction and aggregation proceed. This is the process deemed to be similar across systems, and is the one that is assumed in the emergent world view.
Complexity theory does not merely facilitate case study analysis with new vocabulary. Complexity theory also offers tools to analyze the very messy relationships and concepts posited above. Non-linear mathematics forms the core of these tools and enables scholars to posit and analyze non-equilibrium systems. Techniques such as neural network analysis offer a way to rigorously examine relationships between variables without positing a functional form a priori -- thus allowing the flexibility necessary to observe and understand pattern change over time.
Finally great strides are being made in agent-based modeling. Such modeling offers social scientists a "lab" to simulate social systems. Essentially these models start with the assumption of agents possessing simplified and stylized rule models (importantly the rule models are malleable) in an "environment." Social scientists can introduce stimuli into the environment and observe as the processes of self-organization and co-evolution unfold. Such modeling techniques have been able to simulate such social phenomena as cooperation, war, colonization, pollution (and reactions to it) and even culture and cultural change. Granted, these simulations are simplified and stylized. However, they do capture the underlying dynamics of behaviors observed in the real world. In conjunction with well researched case studies, they can potentially provide compelling evidence of the processes driving world politics.
Further empirical work makes up the research agenda for the emergent world view--to determine if this assumption matches observed phenomena. We envision using complexity theory to analyze a diverse set of phenomena in world politics. For instance complexity theory holds the potential to enlighten the debate between the neorealists and the neoliberals over the type of gains (absolute or relative) that are determinant for behavior. Thinking from the standpoint of the complexity theory perspective and process, we would posit that strategies to secure different types of gains are emergent phenomena to be explained in and across many, varied systems. We would then explore how the complexity process proceeded in various systems. We would look for initial conditions, trace the co-evolution of rule models for the agents in the system, observe how the adapting agents changed the structure of the system itself, and study how feedback and interaction of agents drove the adaptation process. Thus we would be able to explain how the strategies for gains arose across systems, understand how strategies to secure gains change over time and across systems, and (perhaps more importantly) explain discrepancies across systems.
Similarly, complexity theory has potential to answer questions generated in the constructivist research agenda. One of the main concerns within that agenda is to discern the important factors that facilitate norm development and acceptance. A complexity theory analysis would again look at norm development and acceptance as an emergent phenomenon of the system, and trace the development/acceptance through the adaptive process. As will be discussed below, the constructivists are having difficulty finding general results. It is apparent that different factors are important for different norms -- complexity accepts this, and posits a way to explain discrepancies and change across systems.
How might complexity theory help us to understand the end of the Cold War? There have been post hoc explanations for the end of the Cold War but we are interested in why our discipline was unable to anticipate it or to explain it while it was happening. We believe we missed the end of the Cold War because our approaches prevented us from seeing it--they led us to believe it wouldn't happen the way it did! 11 This is because our theories were (and continued to be) based on rigid and static--and therefore fragile--explanatory assumptions about the structure of the system (i.e. "anarchy") and about agents (i.e. having "static, exogenous preferences"). A complexity approach to the Cold War--based on assumptions about process--could explain (and may have anticipated): changes in the Soviet Union's perception of its security needs; changes in NATO's perception of what constituted a strategic advantage; and alternative mechanisms to major war as the catalyst for change in the international system. These aspects of the end of the Cold War can be explained by focusing on the process of complex adaptation that is the core of complexity theory.
These hypothetical examples of research to be done with complexity analysis and a brief discussion of the tools it uses, show the potential of the theory to increase understanding of puzzles in international relations. Empirical research is beyond the scope of this project however with these examples in mind we now turn to theoretical justification for the emergent world view--a first step toward acceptance of complexity theory in world politics.
An Epistemological Shift: How Can We Understand a Messy World?
[A]t heart, science is about the telling of stories--stories that explain what the world is like, and how the world came to be as it is. And like older explanations, such as creation myths, epic legends, and fairy tales, the stories that science tells help us understand something about who we are as human beings.
Doyne Farmer 12
In the introduction we asserted that traditional theories and methods in international relations are not up to the task of understanding a dynamic world--a world made evident by the end of the Cold War. Why do we assert this? We are not putting complexity theory forward because it is a new methodological toy with which we can answer the same questions that have traditionally been asked. Rather, we are advocating the use of complexity theory because it offers a fundamental epistemological and methodological challenge to traditional theories of world politics. We believe that the end of the Cold War and many other recent and/or long overlooked trends in world politics are making such a challenge imperative.
We believe that the end of the Cold War, among other trends, highlights the fact that there is an underlying dynamism in world politics that traditional approaches do not incorporate or understand. We argue that this is due to assumptions about how humanity can best understand the world. As such in this section we will trace the epistemological premises driving traditional theorizing and explore the methodological and empirical ramifications of such premises. We will contrast these with an examination of complexity theory to highlight what we consider to be the advantages of a complexity theory approach.
At the outset, it must be remembered that the theories and approaches that now dominate our discipline were developed and refined during the Cold War--a period of exceptional stability, 13 and the approaches understandably reflect an emphasis on continuity. However, the end of the Cold War and the subsequent collapse of stability highlight the failings and suggest the need for a new epistemology that considers dynamic change 14 --something that the traditional approaches are prohibited from doing because of their underlying epistemological premises.
So what are such premises? They are perhaps best summed up by a statement from the methods book by Gary King, Robert Keohane and Sydney Verba, who argue that, "political scientists would learn a lot if they could rerun history with everything constant save one investigator-controlled explanatory variable." 15 This quote is symptomatic of a philosophy that posits that the best way to understand the world is to create models that are linear in nature, have equilibrium predominate, and have variables with discrete, constant effects. 16 This world view was reinforced not only by the stability of the Cold War, but also by the tools and methodologies available to international relations scholars during the Cold War (such as linear mathematics and serial computers).
Because of this world view, not only have these analyses been linear in nature but the theories on which they were based and the explanations they offered have focused on separating out independent variables and measuring discreet effects. Assumptions are made (and conclusions are drawn) about which factor has the most explanatory salience (ceteris paribus). The interaction of variables, the change of patterns over time and the cycling of dependent and independent variables are not considered. Similarly, rational choice and game theoretic approaches assume equilibrium, optimality and rationality in the absence of more rigorous methods equipped to deal with non-linear interactions and change over time. (The stability of the Cold War contributed to the accuracy of static analyses which used these methods.) These assumptions are not inherently wrong but the static analysis which they engender leads to a problem: they can not meet the challenges of understanding the dynamic, complex patterns we find in world politics in the aftermath of the Cold War. 17
We argue that traditional epistemological premises are problematic because they lead to the construction of misleading assumptions and the inability to recognize and understand change. Utilizing equilibrium-based, discrete cause and effect, linear models is a very convenient abstraction. It allows for the easy deduction/construction of what we call substantive assumptions. These are assumptions, derived from epistemological premises that articulate general behavioral attributes.
For example, starting with an equilibrium-based world view, scholars derive basic assumptions--generally about the nature of the international system or actors within the system--and assumptions deduced from these basic assumptions drive explanations of behavior. For example, equilibrium assumptions allow neorealists to assume a constant anarchical international system. From this basic assumption--a "fact" or "definition"--assumptions about effects on states are deduced (i.e. a "security dilemma," problems of collective action, the priority of the "national interest," concerns about "relative gains," etc.). Explanations are then built on these assumptions and research is conducted to ascertain whether or not observations of world events match expected outcomes. This approach has inherent limitations.
First, discrepancies in the observed and expected outcomes lead to debates conducted around disagreements with assumptions rather than over explanatory power. For instance, in the debate between the neorealists and neoliberals about gains--absolute or relative--the protagonists are arguing more about the nature of the system and the implications that arise from their assumptions about the system than actually explaining the situation or outcome at hand. They are arguing about which set of assumptions more closely match "reality" rather than which perspective or theory can better explain the outcome.
The second problem is that explanation in international relations scholarship is hindered by the fact that assumptions are often incommensurate across theoretical perspectives. This is most evident in the split between the rationalist perspectives (neorealism and neoliberal institutionalism) and the non-rationalist perspectives (constructivism and cognitivist and post-structuralist theories). Neither accepts the explanations of the others, not because of explanatory differences per se but rather because they do not accept the bases from which the explanations are made. Nevertheless, at a core level both traditional and recent alternative approaches share a commitment to equilibrium-based methods that emphasize the importance of assumptions about structural factors and agent composition. 18
Finally, the assumptions adopted in traditional perspectives tend to be static while those adopted in alternative approaches are often not applicable in different contexts. This leads to difficulties in generalization because it limits the explanatory power to only the cases that match the assumptions. The assumptions of the traditionalists are rigid across time--if the patterns of interaction change from one period to the next then the explanation based on the initial assumptions will no longer hold. Similarly, the assumptions of the alternative perspectives can not be applied from one situation to the next. These assumptions are also limited in scope: behavior outside that considered by the assumptions can not be examined. 19 The constructivists have a similar problem. They trace developments such as norms' rise to prominence, yet have trouble generalizing their findings beyond the historically contingent details of the issue examined. 20
The empirical ramifications of the epistemological premises and methodological constructions are important as well, and very clear when viewed from a substantive assumption standpoint. During the Cold War making assumptions about the nature of the international system and about the preferences, interests and behaviors of two states (the U.S. and the Soviet Union) was often adequate for the task of understanding most of the important events--this importance based on the subjective determination of international relations scholars--that took place in world affairs. However, now not only is there confusion about what is important (other issues that existed but were not considered important suddenly command a lot of attention), but there is uncertainty in the patterns of interaction we observe in world politics. (Partially this can be explained by the focus of academia--security concerns were deemed to be the important concerns and the patterns of interaction that surrounded the security realm were fairly stable.) In this new dynamic and complex world, theories designed to explain (and indeed reify) the stable world of the Cold War will not suffice. 21 When the conditions that allowed rigid, substantive assumptions to flourish disappear, the empirical weaknesses of the entire equilibrium-based theoretical enterprise become apparent. Indeed, we argue that their focus of generalization--structure and agents--is misplaced, and that even traditional attempts at dynamic analysis such as Robert Gilpin's hegemonic stability theory or George Modelski's long-cycle theory fall short because their "dynamic" patterns are unchanging. 22 Because of their equilibrium underpinnings, these approaches provide no means of understanding--or even considering--systemic change.
Other evidence of struggle with dynamism is found in the way that traditional approaches to world politics ignore changes and new situations or reapply old methods to analyze them. For example, events relating to environmental politics are examined using traditional means (with varying degrees of success) without seeking to explain why or how the environment has become a relevant new issue. More importantly, the impact of the ascension of environmental issues on the patterns of interaction in world politics is also ignored. This willingness to approach new issues with traditional methods and perspectives while ignoring their emergence in the first place is apparent in a variety of issues from human rights to ethnic conflict and regionalism. It arises from a world view that posits that one mode or pattern of interaction determines outcomes.
We are not the first to question the premises on which traditional theorizing rests, however, we find that recent, alternative approaches suffer related difficulties as well. Alternative theories to the traditional approaches to world politics such as constructivism and cognitive studies and alternative methods such as historical and norms analysis stress description and an historicist approach. 23 These analyses are usually case-specific and often find making general conclusions to be problematic. Like the traditional approaches their focus reduces to agents or structure as the essential sources of behavior in world politics. Nevertheless, they have on occasion been criticized by traditionalists for being non-positivist, labeled post-modern (or even unscientific) and dismissed out of hand. While these more recent approaches have often provided an effective and compelling critique of traditional approaches to international relations, the traditionalists correctly point out (and many of the alternative theorists admit -- and some celebrate) that the alternative approaches have largely failed to develop comprehensive or generalized explanations for behavior in world politics.
While some alternative approaches such as constructivism do not reject systematic understanding they have largely been unsuccessful at developing general explanations for behavior. 24 The theory-building ability of these interpretive approaches has been characterized by a necessary emphasis upon historical description and "storytelling." 25 Unfortunately, this emphasis has largely perpetuated the approach of more traditional perspectives of reducing to explanations of behavior based in agents or system structure (even though these agents or structures may be social constructions). 26 They are still attempting to discern generalizations. 27 However, they are doing so inductively--a difficult task as the details of cases from the agent or structure viewpoint often appear idiosyncratic and historically contingent.
Most of the more recent, alternative approaches to understanding world politics are eager to engage the uncertainty implied by context-specific and historically contingent behavior. This acceptance of uncertainty changes the focus of the intellectual enterprise from efforts of description and explanation aimed at predicting behavior to efforts of description and explanation that hold the possibility of prescribing action--in the absence of prediction--that is contextually contingent but aimed at achieving a particular end (such as responsiveness in the face of uncertainty). This is an important shift in the study of world politics but it remains hampered by the continued reductionist approach representative of traditional perspectives that ignores the effects of process on behavior and instead is content to focus upon the characteristics of agents and structures.
So what are we left with? The inadequacy of traditional approaches to understanding world politics rests in their inability to account for the dynamic change that has become particularly apparent in the aftermath of the Cold War. More recent, alternative approaches to world politics have made some progress in accounting for change but suffer from an inability to develop generalizable theories largely due to their similar reduction to agents and structure and lack of attention to process. It is in response to these trends that we advocate the use of complexity theory.
If traditional and recent alternative approaches to understanding world politics are ill-equipped to understanding the dynamic events that have become apparent in the aftermath of the Cold War where should we turn? We suggest the emerging new science of complexity is better suited to understanding the newly-apparent dynamic world for two main reasons. First, complexity theory has a fundamentally different set of assumptions that do not reduce to agents or structure and instead are based in process which result in different questions about world politics. Second, complexity theory offers the discipline the opportunity to understand change through its focus on process which should lead to more comprehensive explanations of behavior (i.e. generalizable across time and/or across context). Complexity theory holds the promise of bridging the gap between the traditionalist and alternative approaches. It provides a rigorous approach to answer the critiques of traditional perspectives raised by constructivists and others while retaining the ability to provide the systematic explanations sought by rationalists and other traditionalists. 28 Complexity theory can improve our understanding of world politics in a variety of ways.
The source of our optimism rests in the epistemological premises which ground complexity theory. Like traditional approaches, complexity theory begins with the belief that the world is a complex, messy place. However, the approaches diverge from there. Complexity theory does not believe that the best way to understand such a world is to use an equilibrium-based, linear epistemology. Instead complexity theory starts with the belief that equilibrium is not the status quo, and is rare and fleeting at best. It further assumes that the social world and social characteristics are irreducible--more than the sum of their parts.
One result of this difference is that complexity theory replaces the explanatory assumptions that are prevalent in international relations theories with procedural assumptions. We acknowledge that assumptions are a necessary part of the scientific process. They are explicit statements about how we think that the world works and how we can understand that world. Uncertainty is inherent in all assumptions. (If we were certain about how the world works no assumptions would be necessary). Complexity theory is no different in this respect. The theory makes assumptions in order to simplify the world and as a means to organize information just as other theories do. The type of assumptions made in complexity theory is what sets it apart from traditional and recent alternative perspectives on world politics.
In contrast to the substantive assumptions of the traditional and alternative perspectives that define what is important (i.e. "relative gains" or "humanitarian norms") precluding alternative explanations, 29 the assumptions of complexity theory -- which are made about processes--do not identify the important substantive characteristics before the research is conducted. They do however posit the common thread missing in alternative approaches--the process is similar across issues and outcomes. The behaviors and outcomes to be explained, as well as the relevant factors to the explanation, are treated as emergent phenomena which will be recognized through observation--the deductive insight of similar process is wedded to the inductive research on historically contingent details.
Complexity theory then begins with the assumption of a complex adaptive system of world politics. As discussed above in the section on complexity science, this assumption rests on descriptions of the processes within the system. Substantive information about what the system or agents within the system will do in specific situations are conspicuously (and advantageously) absent. Prejudgements about important explanatory factors (again such as "relative gains" or "humanitarian norms") are not made. Indeed, these factors are conceived of as emerging from the interactions within the system. 30 These emergent properties lend flexibility and accuracy to explanations by not excluding alternatives which were precluded by initial assumptions. Thus, important attributes and behaviors can be discovered rather than assumed.
Instead of providing explanatory information, the assumptions used in complexity theory give an idea of how to frame different and more important questions, as well as providing a general framework within which to put details and data. This distinction suggests that complexity theory offers a fundamentally deeper understanding of the system in question. Complexity theory posits that the answers to the explanatory "why" questions are intricately linked to the answers to the "how" questions. It provides a framework for understanding how the system behaves so that explanations for why outcomes happen can be discovered. Because complexity theory focuses on processes it looks for answers in a fundamentally different--and more accurate--way. We argue that this "why-how" link produces more comprehensive explanation.
Besides its emphasis upon process which avoids the pitfalls of precluding unanticipated explanations, complexity theory also offers a means of understanding change in a systematic manner--a capability that is sorely lacking among both traditional and alternative approaches to world politics. Complexity science is inherently about the study of dynamism--the study of systems which persist through change. By taking advantage of advanced simulation techniques complexity theory offers a framework within which the dynamics of change may be examined and explained through the modeling of the processes of the complex adaptive system. This capability to address change offers a major advantage over traditional approaches to world politics. The systematic approach of complexity theory offers the promise of explaining behaviors in a variety of contexts thereby overcoming the limitations of many of the recent alternative approaches.
Earlier we noted some hypothetical examples of puzzles that complexity theory has the potential to address. The theoretical advantages noted above have not been ignored by scholars interested in world politics. Robert Axelrod has begun to explore the use of agent based modeling -- the experimental lab for complexity theory -- for explaining cooperation. He has applied a complexity theory perspective to various games in order to capture cooperative dynamics and has been able to simulate notions of norms and culture. 31 Joshua Epstein and Robert Axtell have used similar techniques to explore such social phenomena as war, migration, and environmental degradation. 32 Lars-Erik Cederman has used complexity theory to analyze the development and dissolution of states and nations with what he noted to be emergent models -- these are again agent-based models. With these he was able to analyze polarity and various aspects of nationalism through conceptions of path dependence and complex adaptation. 33
Complexity theory can increase understanding of world politics in two important ways. It avoids the agent and/or structure reductionism of traditional and recent alternative approaches through a focus on process which precludes the exclusion of unanticipated explanations and it allows us to describe and explain dynamism and change through advanced modeling techniques that simulate underlying processes of world politics. Nevertheless, complexity theory is not a panacea for understanding world politics. 34 We now turn our attention to the pitfalls of adapting a new perspective like complexity science to the study of world politics.
A Warning About Adopting Complexity Theory
In short, all the circumstances are in place for an eventual disillusionment with complexity theory. For despite the strides, there are severe limits to the extent to which such theory can generate concrete policies that lessen the uncertainties of a fragmented world. And as these limits become increasingly evident subsequent to the present period of euphoria over the theory's potential utility, a reaction against it may well set in and encourage a reversion back to simplistic, either/or modes of thought. Such a development would be regrettable.
James N. Rosenau 35
Even as we call for the adoption of the new, emergent world view, perspective and methodology of complexity theory, we would be remiss if we did not also offer a warning about such adoption. The discipline of international relations has a propensity to "discover" and co-opt new ideas and methodologies from other disciplines in the continuing attempt to further knowledge of world politics. Unfortunately this adoption is often undertaken without a thorough examination of the potential applicability of the new concepts to the study of world politics, and/or the new ideas are translated into analytic frameworks with too little understanding of their basic concepts and premises. Often, the new ideas become wildly popular or faddish for a short time and then disappear. Complexity theory's immense potential for increased understanding, and the striking differences from traditional approaches makes us cautious about the potential for repeating these mistakes. The above sections detail the premises of complexity theory and highlight the potential utility of the approach. Now it is incumbent upon us to turn to the pitfalls and potential misunderstandings to be avoided in the application of complexity theory to the study of world politics. The two potential problem areas discussed below examine the meaning of certain complexity terms, and address the prediction/explanation quandary.
Certainty, Order, Stability, Dynamism, Adaptation
It seems only natural that people who find themselves confronted with uncertainty, as scholars of world politics surely do, will look for ways in which to reduce, limit or even eliminate that uncertainty. This is often the perspective with which many people initially approach complexity theory--particularly in the policy world. Part of this no doubt comes from the prior appearance of chaos theory. 36 Complexity theory, following in the wake of chaos theory (indeed people often mistakenly equate the two), seems to hold the promise of dealing with the uncertainty of chaos. Problems arise from the meaning of "dealing with uncertainty." Complexity theory rests on the tenet that uncertainty is an inherent trait of complex systems--the uncertainty is not something to be minimized or eliminated, but rather to be embraced and understood. This is indeed why trends, events, and outcomes that are deemed historically contingent or idiosyncratic can be deemed similar in the complexity theory rubric. They are historically contingent, but they may follow a similar process that leads to evolving, uncertain, yet recognizable patterns. This leads to confusion because complexity theory also has the ability to explain the behavior of a diverse array of complex systems. How is this possible if uncertainty is the "order" of the day?
The answer is in the contrast between the meaning of order in traditional approaches (in both the natural and social sciences) and in complexity theory. Order has traditionally been associated with stability, a lack of uncertainty, and predictability. 37 Therefore when scholars are confronted with an increasingly uncertain world, they naturally gravitate toward a theory that offers to make sense of it. However, complexity theory is not about certainty, stability, or predictability--in fact certain (stable, static) systems are not complex systems and indeed not the human systems relevant to the study of world politics. This potential misinterpretation finds its roots in the confusion of definitions. Complex systems are stable in the sense that the patterns that underlie the systems remain recognizable, while still being malleable. Stable or predictable systems are characterized by patterns that do not change--human systems are rarely, if ever, stable in this manner.
The power of complexity theory is inherent in its explanation of the process that generates patterns, not in its ability to discern discrete future outcomes--indeed complexity theory disavows the possibility of predicting discrete future outcomes because it rests on the uncertainty of the future. As one commentator has noted, "complexity has demolished notions of control and certainty in science." 38 The process is held to be similar across complex systems, and therein lies the utility of the theory for the pursuit of world politics: many, different system outcomes or behaviors, which are historically contingent at some level can be explained at a general level--the level of the process that creates the dynamic, malleable, evolving patterns of interaction. Complexity theory holds promise for understanding and explaining that process and the trends and patterns that it creates, enabling agents to influence their direction or prescribe certain actions, rather than predicting what the direction of the trends will be.
Complexity theory holds that evolving patterns of interaction are not predictable but are adaptive. However, does this does not lead to the conclusion that by determining optimal utility, one can predict the path that systems will follow--adapting ever towards the final optimized state? 39 These two questions highlight a second potential misunderstanding that could lead to disillusionment. For complexity theory, adaptation is a process. As Holland notes:
the aggregate behavior of the system is usually far from optimal, if indeed optimality can even be defined for the system as a whole. For this reason, standard theories in physics, economics, and elsewhere, are of little help because they concentrate on optimal end-points, whereas CAS 'never get there.' They continue to evolve, and they steadily exhibit new forms of emergent behavior. History and context play a critical role, further complicating the task for theory and experiment. Though some parts of the system may settle down temporarily at a local optimum, they are usually 'dead' or uninteresting if they remain at that equilibrium for an extended period. It is the process of becoming, rather than the never-reached end points, that we must study, if we are to gain insight. 40
Complexity theory holds that the political systems do adapt rather than calling for them to adapt at a certain, specified pace. The meaning of "adapt" is different in the two formulations. Adaptation from a complexity viewpoint is the general process of receiving information, acting on information, and updating the rules that determine action based on feedback from evaluation of the system's action (behavior). 41 The claim is made that all complex systems (a category that includes political systems) follow this general process--it is indeed what makes the systems "complex," dynamic, and often unpredictable. Adaptation in this sense is not a judgment about how well the system is able to keep pace with changes in other systems.
These are no mean pitfalls, and they represent large obstacles to the acceptance of an emergent theory of world politics based on complexity theory. However, the potential of complexity theory for clarifying puzzles in world politics behooves us to present the theory with both its advantages and disadvantages in the hopes that this will lead to fruitful discussion and avoid the fad/famine pattern of past incorporation of ideas.
Conclusion: A New Perspective for World Politics
So what's the connection with economic and political policy [and complexity theory]? Well in a policy context, it means that you observe, and observe, and observe, and occasionally stick your oar in and improve something for the better. It means that you try to see reality for what it is, and realize that the game you are in keeps changing. . . .
Brian Arthur 42
Problematic traditional theories developed in the unusually static context of the Cold War and recent, alternative theories which have yet to construct their own frameworks for understanding world politics are indications of an increasingly complex world. 43 Complexity science, which is at the forefront of theoretical activity in the physical and biological sciences and in economics, has the potential to overcome the failure of these conventional approaches to address world politics as they have unfolded in the aftermath of the Cold War. Complexity science is uniquely suited to dealing with the newly-apparent complexities of world politics. It recognizes, embraces and utilizes change, uncertainty and dynamism while retaining the hope of achieving systematic understanding through its emphasis on process and rejection of irreducible agents and/or structure.
Biologists, physicists and economists now work with the knowledge that the world is dynamic and changing and that the interesting -- and important -- questions must be answered in a way that accepts and embraces such "messiness." Scholars of world politics must make the same leap, accepting uncertainty and dynamism without abandoning the hope of understanding. Complexity science holds the potential of doing just that.
We have attempted to present theoretical justification for a shift away from static perspectives of world politics that make agents or structures primitive. We are instead arguing for a new approach which replaces the focus on agents or structures with an emphasis on the complex adaptation apparent in our inherently dynamic world. The task ahead is to apply this theoretical approach to the study of world politics. The methods, models and tools of complexity science need to be used in the pursuit of answers to the puzzles of world politics. Two lines of research are suggested by such a shift in perspective. One examines world politics as a CAS itself, seeking to explain events such as the post-war transition to a "bi-polar world," the end of the Cold War and global system evolution. The other approach is "bottom-up," viewing world politics as the interaction of many different complex adaptive systems such as international environmental regimes, transnational commerce and NATO in the aftermath of the Cold War. The research agenda that a complexity theory world view calls for is indeed ambitious. It calls for a fundamental restructuring of perspectives toward a process based understanding of human behavior and interaction. However, we feel that the ambition is justified by the increased understanding of world politics that complexity theory promises.
Appendix A: From Whence Complexity Theory
Complexity theory is. . . . This sentence is not easy to complete, though the completion is crucial. It is difficult because complexity science does not have a set of natural laws, nor is it suited to one-sentence descriptions. The completion is crucial because an understanding of complexity theory is important if it is ever to be fruitfully applied to new fields by practitioners who did not take part in the formation of the "theory" itself. Perhaps then the science of complexity should be looked upon first as a perspective or world view. In this sense, the science of complexity is concerned with looking at and for connections, interactions, and dynamism. Complexity science expressly examines problems that exhibit "non-reducibility"--phenomena that are more than the sum of the parts. 44 A complexity world view or perspective leads to the asking of very different questions than traditional perspectives, and it fundamentally transforms the manner in which answers are developed.
While the previous comments did not reveal the completion of the "Complexity is . . ." statement, it has perhaps led to a better sentence to finish: "Complexity theory is about. . . ." This sentence can be completed in many ways, and all seem to be important. Complexity theory is about patterns that change (or evolve) while remaining coherent. Complexity theory is about the process of adaptation and learning. Complexity theory is about how simple agents interacting can cause the emergence of complex macroscopic behaviors/structures. 45 Complexity theory is about the idea of multiple causes, the importance of initial conditions, and the dynamics of path-dependency. Complexity theory is about evolution to "nowhere." 46 The list can go on for pages. The very nature of complexity theory is holistic, and it would indeed be antithetical to the idea of complexity to narrow the theory down to a one-sentence definition. The messiness of the definition of complexity science matches the messiness of the phenomena that it attempts to study. 47 Indeed, it matches the nature of the interesting questions in all of the sciences--natural and social.
Complexity theory can be considered a response to two trends in the natural sciences. 48 First there has been a marked increase in uncertainty and decrease in acceptance of traditional models of scientific analysis. 49 In the social sciences this is suggested by the proposition that social systems are no longer "controllable" (as if they ever truly were) or that they are more unpredictable. As economist and complexity theorist, Brian Arthur notes life, language, chemistry, physics, and economics are inherently messy. 50 This is in great contrast to the Newtonian clockwork metaphor that has held in most fields of science (social science as well) for the last few centuries. 51 Complexity is a reaction to the growing dissatisfaction with the inaccurate "orderly" assumptions in Newtonian science.
In addition, complexity can be seen as a response to the development of chaos theory in math and physics. Though complexity accepts and indeed embraces the messiness that is inherent in chaos, it also offers hope of wading through (some of) it. The key contrast is that in chaos, the patterns are random (and therefore not really patterns at all) and though chaos is deterministic, it is inherently unpredictable and unknowable. 52 In complexity, the messiness remains, but patterns emerge. The dynamics appear to be knowable, though likely not predictable. Complexity is the "science at the edge of order and chaos." 53
Beyond noting what complexity is a response to, it is useful to examine the motivation that led to the formation/evolution of complexity theory. 54 Interestingly, from the beginning, complexity theory has been an interdisciplinary endeavor. Scientists and thinkers from physics, chemistry, biology, economics, and computer science have found common strains running through their most challenging puzzles. From economics, the puzzles of increasing returns, path dependency, and technological emergence were all defying the "laws" of neoclassical economics. In biology and chemistry, natural selection is being found to be inadequate in explaining the rise of life, and the dynamics of ontogeny have remained a puzzle. 55 In computer science, those working toward artificial life, are realizing that "expert systems" will never be adaptive enough to be considered life-like. Throughout the sciences, uncertainty has become pervasive, control illusive, and the need for change in thought more and more obvious.
The Complex Adpative System (CAS): Aggregation, Self-organization and Co-evolution
Complex adaptive systems are comprised of networks of many agents. It is the interactions of the agents within complex adaptive systems as well as their adaptive natures that cause the complex adaptive system to evolve and cause the macroscopic properties of the complex adaptive system to emerge. As Holland notes, aggregate behavior is the behavior of interest, and to understand it, "we must understand how the aggregate behavior emerges from the interaction of the parts." 56 Who or what are these agents and what do they do, where in fact do they come from? Obviously this depends upon the system in question. They can be T-cells in the immune system, people and firms in the economy, or a multitude of other manifestations. Complexity posits that at some level they do similar things, no matter what they are, and the theory provides a method for understanding how agents behave and what behavior agents exhibit. Murray Gell-Mann describes the concept that facilitates this a schema, while Holland describes it as an internal rule model. 57 He argues that "it is useful to think of an agent's behavior as determined by a collection of rules." 58
This sounds suspiciously close to a rational choice perspective, but a closer reading reveals that the rules are not deductively arrived at in the presence of known utility functions and perfect information. Instead the rules are learned from experience and are used to anticipate the future. Most importantly, the rules change over time as situations change and the environment is altered--there is no set utility and hence the agents are adapting or evolving to "nowhere." 59 As Holland observes:
Overall, then, we will view CAS as systems composed of interacting agents described in terms of rules. These agents adapt by changing their rules as experience accumulates. In CAS, a major part of the environment of any given adaptive agent consists of other adaptive agents, so that a portion of any agent's efforts at adaptation is spent adapting to other adaptive agents. 60
Does an agent focus help us to understand world politics? The answer will be yes if there is a mechanism to understand how micro interactions lead to macro outcomes. Complexity theory does this through the mechanisms of aggregation, self-organization, and co-evolution.
Aggregation is something that both agents within complex adaptive systems, and complex adaptive systems, as a whole, do. Agents, through their internal rule models, aggregate information or stimuli from the environment into categories. Holland notes that this is done through mechanisms of tagging and building blocks. 61 The system itself displays aggregation as complex large-scale behavior emerges from the aggregate interactions of the agents. Understanding this relationship between agent interaction and emergent behavior is the defining goal of complexity. As Holland argues, "the study of CAS turns on our ability to discern the mechanisms that enable simple agents to form highly adaptive aggregates." 62 An understanding of aggregation is the first step in understanding complexity theory. 63
Aggregation on the agent level contributes to the ability of the agent to adapt and alter its internal rule models. Agents aggregate information into categories. By combining this aggregation with the use of internal rules, agents anticipate the future and act/interact. This action/interaction has effects on the agent's environment which in turn affects the agent itself. 64 The effects on the agent help the agent to evaluate how well the internal rule model performed and what changes need to be made. In this way individual agents adapt and learn and new agents are generated. 65
Aggregation at the system level has two effects. First, the aggregation of agent interactions causes the emergence of structure and the system-level behavior that is most often of interest. Second, the aggregation of agents themselves into meta-agents causes the diversity and many levels of organization characteristic of complex adaptive systems. 66 How agent interactions cause the emergence of behavior is the subject of intense scrutiny in complexity circles and is being addressed mostly through the use of computer modeling, and non-linear mathematics. 67
A result of aggregation, self organization is the process of systems' "incessant attempts to organize [themselves] into ever more complex structures." 68 Self-organization is both a product of aggregation, and a path to adaptation. Self-organization is a result of agents adapting to conditions--it emerges from the interactions and adaptations of agents--and, importantly, this happens in the absence of central control. It is also a path to macro-adaptation, because new forms of organization and patterns of interaction arise and emerge, altering the form of the complex adaptive system itself. Agents within a complex adaptive system self-organize into new and varied forms of organization, and this self-organization at the agent level then aids the evolution and adaptation of the complex adaptive system as a whole as building blocks of organization are created and revised.
Finally, complexity theory includes the notion of co-evolution. Stuart Kauffman describes co-evolution as the process whereby, "as we evolve, so do our competitors; to remain fit, we must adapt to their adaptations." 69 Co-evolution is the notion that organisms, or indeed systems of any kind, do not evolve in a vacuum. Organisms and systems occupy "niches" that are created by the existence and action of other organisms or systems. In essence, organisms and systems are constrained by their environment, while also adding to the environment. This phenomena has been studied in detail in the context of technological development. New technologies create niches for further new technologies. As one complexity theorist notes, "Entry of the car drove the smithy to extinction and created the market for motels." 70 Systems or agents are constantly reacting and adapting to other systems or agents and to the environment, this helps to create the environment, which then feeds back to the adaptive process, creating and recreating the system and agents. 71 This path dependence has often been noted in international relations, however, complexity theory offers a way to generalize the process by which path dependence is important and determines outcomes across a variety of cases. 72
Notes:
*: The authors would like to acknowledge the insightful comments on previous drafts from Martha Finnemore, James Goldgeier, David Kinsella, James Rosenau, and Susan Sell. In addition we would like to acknowledge the material support provided by the National Science Foundation and the George Washington University. Back.
**:
Department of Political Science
The George Washington University
Office of the University Professor of International Affairs
2130 H. St., NW, Room 613
Washington, DC 20052
E-mail: david@gwu.edu or mjhoff@gwu.edu
Paper available on the World Wide Web at
http://gwis2.circ.gwu.edu/~david/Complex. Back.
Note 1: Paul Knox, Globe and Mail (Toronto), quoted in the International Studies Assocation newsletter, (v24/n3), May/June 1997. (http://csf.colorado.edu/isa/newsletter/883209/may97.html). Back.
Note 2: Ruggie, John Gerard, "Territoriality and beyond: problematizing modernity in international relations," in International Organization, (v47/n1), Winter 1993, p. 143-144. Back.
Note 3: Soviet leader Mikhail Gorbachev began his glasnost and perestroika reforms in 1987, signing the Intermediate-range Nuclear Forces (INF) Treaty with the United States in December. Soviet forces began their withdrawal from Afghanistan in April 1988. Free elections in Poland in June 1989 brought Solidarity Prime Minister Tadeusz Mazowiecki to office in August. The Berlin Wall was opened in November 1989 and Germany was reunited in October 1990. The Soviet Union supported the U.S.-led military offensive against its former client, Iraq, in January 1991 after Iraq's occupation of Kuwait. Georgia declared independence from the Soviet Union in April 1991. After an abortive coup against Gorbachev, Russia, Ukraine and eight other republics (including the Baltics which were recognized by several Western states) declared their independence in August 1991; the Soviet Union was dissolved in December. The "beginning of the end" of the Cold War might be dated as early as the election of Pope John Paul II (who, as a Polish cardinal, was critical of the Soviet-backed regime) in October 1978 and the formation of Solidarity in August 1980 more than a decade before the dissolution of the Soviet Union. Back.
Note 4: Here, for example, is Robert Gilpin writing in 1981: "During the 1970s and early 1980s a series of dramatic events signaled that international relations were undergoing a significant upheaval." War and Change in World Politics, Cambridge University Press, 1981, p.1. Back.
Note 5: Waldrop, Mitchell, Complexity, New York: Simon and Schuster, 1992, back cover. Back.
Note 6: Holland, John, Hidden Order, New York: Addison Wesley Publishing Company, 1995, p. 4. Back.
Note 7: Rosenau is fond of this question and uses it often in lectures. It gets at the nature of complexity because states don't usually fall apart over the weekend (The USSR being a glaring exception to prove the rule), and political science doesn't have much understanding of why that is. Back.
Note 8: Holland, John "Complex Adaptive Systems," Daedalus, Winter 1992, (v121/n1), p. 17. Back.
Note 9: One agenda for further research is to examine this assumption. Agent-based modeling is a powerful technique being developed to do just that. See the discussion of Sugarscape in Epstein, Joshua, and Axtell, Robert, Growing Artificial Societies, Washington, DC: Brookings Institution Press, 1996. Back.
Note 10: This is not necessarily a comprehensive list of CAS characteristics. Waldrop points out that the lack of equilibrium does not mean that CAS's are unstable. See Waldrop, Mitchell, Complexity, New York: Simon and Schuster, 1992, p. 145. Back.
Note 11: Again, for example, here is Gilpin writing in 1981: "On all sides, centrifugal forces could pull at this last of the great multiethnic empires as neighbors make demands for revision of the territorial status quo and as subordinate non-Russian peoples seek greater equality and autonomy. Such external and internal challenges could give rise to powerful defensive reactions on the part of the Soviet governing elite." War and Change, p. 243 (emphasis added). Back.
Note 12: As reported in Waldrop, p. 318. Back.
Note 13: The continuity observed by scholars was as much (if not more) a product of the perceptions in the discipline of international relations of what was considered important (i.e. security, 'high' politics) than it was an "objective" view of "reality." In many issue areas, the world was far from stable. The geometric growth in the number of states and international organizations -- governmental and non-governmental, as well as the rise of diverse new issues such as the environment and human rights during the Cold War period belie the continuity or stability that scholars "observed." Back.
Note 14: See John Lewis Gaddis, "International Relations Theory and the End of the Cold War," in International Security, (v17/n3), Winter 1992/93, pp. 5-58, Rosenau, James N., Turbulence in World Politics, Princeton, NJ: Princeton University Press, 1990. Back.
Note 15: King, Gary, Keohane, Robert, and Verba, Sydney, Designing Social Inquiry: Scientific Inference in Qualitative Research, Princeton, NJ: Princeton University Press, 1994, p. 82. Back.
Note 16: We do not claim that scholars of the rationalist tradition naively assume the world to be a simple or linear place, but rather as we argue below, that they believe equilibrium-based, linear analysis is the best way to understand a complex, messy world. Back.
Note 17: In this context, "dynamic" refers to patterns of interactions which have changed and appear to be more uncertain over time than previously thought. "Complex" refers to a state where many actors, relationships and issues are new or are interacting in new ways. Back.
Note 18: This critique may seem somewhat unfair when directed at constructivists but Wendt's attempt to explain a socially-constructed system of international anarchy relies upon assumptions about agents ("predator states") which "may become predisposed toward aggression" for "whatever reasons--biology, domestic politics, or system victimization. . . ." Wendt, Alexander, "Anarchy is what states make of it: the social construction of power politics," International Organization, Spring 1992 (v46/n2), pp. 407-408. Back.
Note 19: For instance Gilpin cannot explain events that do not fit the assumptions of hegemonic stability. Keohane cannot explain events when organizations do not matter, or matter more than he thinks they should. Grieco cannot explain events where relative gains are not a concern. See Grieco, Joseph M., Cooperation among Nations: Europe, America, and Non-tariff Barriers to Trade, Ithaca, NY: Cornell University Press, 1990. Back.
Note 20: The "global norms of modernity" which Eyre and Suchman suggest explain the purchase of similar types of weapons by developing nations which may face very difference security challenges have no apparent relationship to the "civilizational norms" which Price and Tannenwald suggest explain international prohibitions against the use and proliferation of nuclear and chemical weapons. See Eyre, Dana P., and Suchman, Mark C., "Status, Norms, and the Proliferation of Conventional Weapons: An Institutional Theory Approach," and Price, Richard, and Tannenwald, Nina, "Norms and Deterrence: The Nuclear and Chemical Weapons Taboos," in Katzenstien, Peter J., ed., The Culture of National Security: Norms and Identity in World Politics, New York: Columbia University Press, 1996. Back.
Note 21: Cederman, Lars-Erik, "Emergent Actors in World Politics: How States and Nations Develop and Dissolve," doctoral dissertation, University of Michigan, Ann Arbor, MI, 1994. Back.
Note 22: See Gilpin, War and Change; and Gilpin, Robert, The Political Economy of International Relations, Princeton: Princeton University Press, 1987.; and Modelski, George, "Is world politics evolutionary learning?," in International Organization, (v44/n1) Winter 1990. Back.
Note 23: Post-modern or post-structuralist approaches to world politics reject the possibility of comprehensive understanding and generalizable conclusions about behavior. These approaches rest on philosophical perspectives that are incommensurate with those that seek to develop systematic understanding of world politics. While the post-modern perspective holds some appeal for the authors we are not willing to abandon the capabilities (or hope) for abstraction and generalization that remain fundamental foundations of complexity theory. Thus, the incommensurability of our approach with that of the post-modernists can merely be recognized and accepted. For some examples of post-modern approaches to world politics see Der Derian, James, and Shapiro, Michael, International/Intertextual Relations, Lexington, MA: Lexington Books, 1989. Back.
Note 24: Kowert Paul, and Legro, Jeffrey, "Norms, Identity, and Their Limits: A Theoretical Reprise," in Katzenstien, Peter J., ed., The Culture of National Security: Norms and Identity in World Politics, New York: Columbia University Press, 1996, p. 469. Back.
Note 25: We certainly do not mean to downplay the importance of such works. Some impressive theoretical strides have been made in this area. See for example, Finnemore, Martha, National Interests in International Society, Ithaca, NY: Cornell University Press, 1996, and Sikkink, Kathryn, Ideas and Institutions, Ithaca, NY: Cornell University Press, 1991. Back.
Note 26: A good proponent of this view, and indeed the theorist that began the constructivist agenda, is Alexander Wendt. See Wendt. Back.
Note 27: Cognitive approaches attempt to do this as well, as evidenced by Robert Jervis's insightful work with (mis)perception in world politics. See, Jervis, Robert, Perception and Misperception in International Politics, Princeton, NJ: Princeton University Press, 1976. Back.
Note 29: Again, Kowert and Legro say of the constructivists that they "tend to treat their own core concepts as exogenously given" and that "about the process of identity [or norm] construction [they] have relatively little to say." Kowert and Legro, op. cit. Back.
Note 30: Again, while neorealists take the importance of relative gains to be a given, and then use that fact to explain outcomes, complexity theorists would take concerns about relative gains to be an emergent property of the system, and one that impacts the patterns of interaction within the system. Similarly, an orientation toward humanitarian codes of conduct would also be an emergent property that influenced patterns of interaction. Back.
Note 31: Axelrod. See especially chapters 1, 3, 7. Back.
Note 32: Epstein, and Axtell, 1996. Back.
Note 33: Cederman. See especially, chapters 2, 3, 4, 5. Back.
Note 34: Rosenau, James N., "Many Damned Things Simultaneously: Complexity Theory and World Affairs," in Alberts, David S., and Czerwinski, Thomas J., eds., Complexity, Global Politics, and National Security, National Defense University, Washington, 1997. Back.
Note 35: Rosenau, "Many Damned Things Simultaneously," p. 74. Back.
Note 36: Chaos theory enjoyed a short, faddish period in the pursuit of understanding in world politics following the uproar initiated by the publication of Gleick's Chaos. See Gleick, James, Chaos: Making a New Science, New York: Viking, 1987. Back.
Note 37: That social scientists, and especially international relations scholars, desire this kind of order is evidenced by the influence that Realism has had on the discipline for so long--a paradigm that holds that explanatory patterns have changed little (if at all) since the time of Thucydides. Back.
Note 38: Sardar, Ziauddin and Ravetz, Jerome, "Complexity: Fad or Future?," Futures, July/August 1994, (v26/n6), p. 564. Back.
Note 39: Brian Arthur contemplates this notion of optimality. He argues that, "If we, as humans, try to take action in our favor without knowing how the overall system will adapt -- like chopping down the rain forest -- we set in motion a train of events that will likely come back and form a different pattern for us to adjust to, like global climate change. . . . You can't talk about optimization, because it becomes meaningless. . . . You have to talk about accommodation and coadaptation." (quoted in Waldrop, p333). Back.
Note 40: Holland, "Complex Adaptive Systems," p. 20. Back.
Note 41: Hoffmann, Matt, "Complexity and International Relations," unpublished manuscript, George Washington University, Washington, DC, 1997. Back.
Note 42: As quoted in Waldrop, p. 331. Back.
Note 43: It is our contention that the current state of international relations scholarship reflects what Kuhn refers to as a "crisis" or "pre-paradigmatic" state in the discipline. Kuhn points out that "when confronted by even severe and prolonged anomalies . . . [scientists] do not renounce the paradigm that has led them into crisis. They do not, that is, treat anomalies as counter-instances, though in the vocabulary of philosophy of science that is what they are." Kuhn, Thomas S., The Structure of Scientific Revolutions, 2nd. ed., Chicago: University of Chicago Press, 1970, p. 77. Back.
Note 44: Stein, Daniel, "Preface," Lectures in the Sciences of Complexity, New York: Addison Wesley Publishing Company, 1989, pp. xiii-xv. Back.
Note 45: Simple here is used in comparison to the more complex aggregations. These "simple" agents are often themselves complex adaptive systems. Back.
Note 46: Holland discusses this in various places. See, for example, Hidden Order, and "Complex Adaptive Systems," p 17-30. Back.
Note 47: There is also research going on to define complexity itself, though this does not seem relevant to the pursuit of international relations. Knowing or believing that political systems are complex is enough, and measuring the degree of complexity does not seem to be fruitful in adding information. Back.
Note 48: Herbert Simon argues that studies of complex phenomena have a long and distinguished lineage. See Sciences of the Artificial, Cambridge: The MIT Press, 1996. However, for the purposes of this paper, we will concentrate on the most recent manifestation of complexity studies. Back.
Note 49: Einstein's theory of relativity, the Heisenberg principle, and quantum mechanics have fostered this feeling. See Gribbin, John, In Search of Schrödinger's Cat: Quantum Physics and Reality, New York: Bantam Books, 1984. Back.
Note 50: Waldrop, p. 329. Back.
Note 51: As noted in Waldrop, p. 329. Popper also mentions the "clock" metaphor although in a context that admits the alternative viewpoint of messier "clouds" as well. See Popper, Karl, "Indeterminism and Human Freedom," in Miller, David, ed., Popper Selections, Princeton, NJ: Princeton University Press, 1985, pp. 247-249. Back.
Note 52: Simple equations govern the dynamics of deterministic chaotic systems. However, they are so sensitive to initial conditions that the slightest change causes the system to change completely. Back.
Note 53: Waldrop, to our knowledge, first described complexity in this way, although he attributes the term "edge of chaos" to Christopher Langton. See Waldrop, p. 230. Back.
Note 54: For a comprehensive and very readable account of the history of complexity theory, especially as manifested at the Santa Fe institute, see Waldrop. Back.
Note 55: Ontogeny refers to the life cycle of a single organism from conception to maturity. Back.
Note 56: Holland, "Complex Adaptive Systems," p. 20. Back.
Note 57: Gell-Mann, Murray, The Quark and the Jaguar, New York: W.H. Freeman and Co., 1994. Back.
Note 58: Holland, Hidden Order, p. 7. Back.
Note 59: Holland discusses this notion of evolving not to set optimal points, but rather evolving in a very context and path dependent manner. See, for example, Hidden Order, and "Complex Adaptive Systems," p 17-30. Back.
Note 60: Holland, Hidden Order, p. 10. Back.
Note 61: Tags facilitate selective interaction and provide a basis for filtering, specialization, and cooperation. Examples of tags are chemical signatures on molecules, activation sites on enzymes, or flags for human groups. Tags, "enable us to ignore certain details, while directing our attention to others." Building blocks are used to parse scenes and information into known elements. Rather than having a list of all possible situations, agents understand information by way of pieces of things already known. For instance when we see a red Volvo broken down by the side of the road, we may have never seen that particular scene before, but we understand it because we have building blocks from past experience (i.e. red, Volvo, flat tire, etc.) that help us to piece together the scene as a whole. Holland, Hidden Order,, p. 34-37. Back.
Note 62: Holland, Hidden Order, p. 12. Simple in this case relates to the comparison between agent and systems. A T-cell's behavior is simpler than the behavior of the immune system as a whole. A single human being's behavior is simpler than the behavior of society. In addition "simple" points to the fact that agent behavior cannot be simply translated into aggregate behavior--the whole is more than the sum of the parts. Back.
Note 63: Aggregation in the complexity theory sense is contrasted with aggregation in the neoclassical economics sense. In the former, interactions and behaviors are viewed as aggregated in a non-linear fashion, while in the latter, behaviors are summed up. It is the focus on interaction and non-linear relationships that distinguishes complexity theory aggregation from that found in neoclassical economics. Back.
Note 64: This is the feedback effect characteristic of complex systems. In addition, the effect of the environment on the agents can also be considered external rules. Back.
Note 65: The process of complex adaptation generates new agents as well as structures. Agents can group together to form meta-agents which are agents that are comprised of other agents (i.e. humans aggregating to form an organization), or agent rule models can shift -- changing the way in which an agent is identified. Back.
Note 66: As noted, meta agents are agents comprised of other agents. For instance in the economy, an agent could be a human being. Humans aggregate to form meta agents of firms and businesses, firms and businesses might aggregate to form meta agents of corporations and conglomerates. In essence agents at one level are systems at another (i.e. a human is a CAS, but he/she is also an agent in other CAS's such as states, firms, families, etc. For a graphic that illustrates this point see, Hoffmann, Matt and Johnson, David "Action in a Complex World: Observations on Complexity, Global Politics, and National Security, unpublished manuscript submitted as commentary on the Complexity, Global Politics, and National Security Conference Washington, DC, November 13-14, 1996. Sponsored by the National Defense University and RAND Corporation. Back.
Note 67: This subject will not be briefly addressed below. See Epstein and Axtell and Weisbuch, Gerard, Complex Systems Dynamics, New York: Addison Wesley, 1991, and Axelrod, Robert, The Complexity of Cooperation: Agent Based Models of Competition and Collaboration, Princeton, NJ: Princeton University Press, 1997, for detailed discussions of computer modeling techniques and the application of non-linear mathematics to the study of CAS's. Back.
Note 68: Waldrop, p. 102. Back.
Note 69: Kauffman, Stuart, At Home in the Universe, New York: Oxford University Press, 1995, p. 27. Back.
Note 70: Kauffman, p. 280. Back.
Note 71: These examples of co-evolution also highlight the relationship between co-evolution and path dependency. Co-evolution and the importance of initial conditions lead to path dependency. Brian Arthur has discussed this in terms of the "QWERTY" keyboard. The QWERTY keyboard was developed initially to slow down early typists who typed too fast, thereby entangling the early typewriter mechanisms. Typists--and typewriter manufacturers--co-evolved with this new technology and technique until all other types of keyboards became "extinct." Now the QWERTY keyboard is the only path left--even though it is not necessarily the most efficient keyboard design. There is no way by just looking at the QWERTY keyboard today to determine why it is configured the way it is. The initial conditions, co-evolution, and path-dependency are too important. However, in other situations (social and political as well) analysts attempt to assess why things are the way they are by making substantive or "deductive" assumptions--this may prove to be inaccurate. Back.
Note 72: The difficulty of generalization has plagued constructivists in particular. In a recent collection of constructivist work, two sympathetic critics point out that "some of the authors in this volume . . . often seem to view the generative processes of norm building and evolution as unique to the cases they examine. And yet there is considerable material in these essays from which to construct plausible hypotheses about norm creation that might be applied to other cases." Kowert and Legro, op. cit. Back.
Email this citation