Human Studies 21: 157–186, 1998. c 1998 Kluwer Academic Publishers. Printed in the Netherlands.

157

The Theory of International Politics? An Analysis of Neorealist Theory KEITH TOPPER Department of Government and International Studies, University of South Carolina, Columbia, SC 29208, USA

Abstract. In recent years a number of writers have defended and attacked various features of structural, or neo-realist theories of international politics. Few, however, have quarrelled with one of the most foundational features of neorealist theory: its assumptions about the nature of science and scientific theories. In this essay I assess the views of science underlying much neorealist theory, especially as they are articulated in the work of Kenneth Waltz. I argue not only that neorealist theories rest on assumptions about science and theory that have been questioned by postpositivist philosophers and historians of science, but also that the failure to consider the work of these writers yields theories of international politics that are deficient in several respects: they are “weak” theories in the sense that they cannot illuminate crucial features of international politics, they presuppose and sustain a narrow view of power and power relations, they reify practices and relations in the international arena and they offer little promise of producing the sort of “Copernican Revolution” for which Waltz called (or, more modestly, even a minimally satisfactory theory of international politics). In light of these shortcomings, I sketch an alternative approach to the study of international affairs, one that has been termed “prototype studies.” I contend that such an approach provides scholars with a rigorous way of studying international politics, without being a theoretical science.

At the beginning of Beyond Good and Evil, Friedrich Nietzsche asked a series of questions that have haunted – if not always influenced – all subsequent reflections upon the nature of knowledge in the human sciences: Supposing truth is a woman – what then? Are there not grounds for the suspicion that all philosophers, insofar as they were dogmatists, have been very inexpert about women? That the gruesome seriousness, the clumsy obtrusiveness with which they have approached the truth so far have been awkward and very improper methods for winning a woman’s heart? What is certain is that she has not allowed herself to be won – and today every kind of dogmatism is left standing dispirited and discouraged. If it is left standing at all! (Nietzsche, 1886/1966: 2).1 This claim, that the unveiling of truth and the acquisition of knowledge requires the techniques of seduction rather than the seduction of technique, is particularly relevant to contemporary theorists of international politics. For despite echoes of similar sentiments almost 500 years ago, when a founding

158

KEITH TOPPER

theorist of international politics warned that fortune, being a woman, could not be mastered “by those who proceed coldly” (Machiavelli, 1532/1950: 94), current theorists have largely ignored such injunctions. Instead they have been captivated by the remarkable successes of the natural sciences and have attempted to appropriate the “ideal model” of natural scientific investigation2 as the methodological foundation for studies of international affairs. In the most general terms, this consists of the attempt to predict behavior or outcomes in the international realm by first abstracting context-free features from their particular contexts, and then using rules or laws relating these features as the basis for prediction. Obviously, though, this repudiates Machiavelli’s admonitions, because the mastering of fortuna requires precisely that contextual understanding which is explicitly rejected by most contemporary theorists of international relations.3 Now this contravention of a self-proclaimed intellectual forefather would warrant only passing interest if the putative duplication of method also produced a duplication of results. Unfortunately, this has not been the case. The science of international relations has produced few strict laws, and fewer still that are anything more than trivial; its predictions fall far short of the standards set in the mature natural sciences; and it has thus far failed to attain anything resembling the status of normal science. This, of course, does not imply that no insights have been secured, or no advances obtained. The vital questions, however, are whether current theories of international relations hold much hope of achieving the status of a rigorous science as their proponents construe it; and, if not, whether alternative approaches offer greater promise for understanding events in the international realm. These, I would argue, are the critical issues for all contemporary theorists of international politics. If we hope therefore to rescue thinkers like Machiavelli and Nietzsche from Cassandra’s fate, we must raise these questions. However, since it is obviously impossible to survey adequately the entire field of international relations in the scope of one essay, the following discussion will be confined to one influential school of thought, that school commonly called structural, or neo-realism. This focus on neorealism, although clearly narrower than the concerns sketched above, is nevertheless useful for at least two reasons. First, as perhaps the dominant school of thought in present-day international relations theory, an analysis of neorealism is inevitably also one of a substantial body of work in the field. If, accordingly, one can show that neorealism is logically and conceptually otiose, then much of what currently passes as important work in international relations must be reexamined. Second, and perhaps more significantly, neorealism’s authority among contemporary scholars of international relations rests largely upon its claims of scientific superiority. Kenneth Waltz, for example, contends that reductionist theories – and for Waltz this is the

AN ANALYSIS OF NEOREALIST THEORY

159

only serious alternative to a structural approach – focus on lower rather than higher-order configurations, and thus produce only “endless arguments that are doomed to being inconclusive” (Waltz, 1979: 65). Assuming that there is at least some foundation to this claim, an examination of neorealism might be expected to have implications that extend beyond its own specific boundaries. It might, in short, indicate more generally the likelihood of developing a theory of international relations modelled along the lines of ideal natural science. In what follows, I will examine neorealist theory both in the general context of recent work in the history and philosophy of science, and in its application to one specific international conflict, namely, the Vietnam War. In so doing, I shall argue that neorealist theories are deficient in several respects: they are generally weak theories in the first place, they rest upon and promote a narrow view of power and power relations, they reify relations in the international arena, they cannot account for change, and they offer only the slimmest possibility of ever becoming rigorously “scientific.” Moreover, I will maintain that an alternative approach, one which has been termed “prototype studies,” provides us with a rigorous way of studying international relations, without being a theoretical science. Before launching into a discussion, however, it may first be useful to say a few words about the organization and assumptions guiding this essay. The essay itself proceeds through an examination of three closely related and sometimes overlapping issues: first, the problem of neorealism’s explanatory and scientific status; second, the question of its structuralism; and, finally, its construal of power and power relations. Furthermore, it focuses principally – but not exclusively – on the writings of Kenneth Waltz. This approach, while clearly limited in some respects, has the advantage of avoiding those dangers endemic to any ideal-typical constructions, which, whatever their virtues, have the consequence of referring simultaneously to everyone and no one. By concentrating on a particular exemplar – and especially one that has been widely praised for its avowed scientific and methodological rigor – I hope to avoid this predicament, while securing the benefits of examining a single version of neorealist theory in a more detailed and holistic manner than would otherwise be possible.4 1. The Claim of Science As Waltz explains, neorealism’s appeal derives largely from reflection on the failures and apparent endemic weaknesses of “reductionist” theories. These theories, which locate the causes of international outcomes at the individual, subnational, or national level (Waltz, 1979: 18, 60), form the cornerstone

160

KEITH TOPPER

of much scientifically-oriented work in international relations theory today. Neorealist critics argue, however, that irrespective of reductionist successes in other domains of inquiry, these approaches have failed systematically, and are likely to continue to fail, to produce similar results in the study of international politics. Offering familiar assessments of reductionism as embracing a form of behavioralism that privileges description over explanation, as failing to link systemic and unit level forces (Waltz, 1979: 65), and as incapable of accounting for persistent continuities in international affairs (Waltz, 1979: 65–67), neorealists maintain that a more promising path to progress lies in a turn toward structural premises. Without a doubt, then, neorealism’s claim of scientific and theoretical superiority springs from its devotion to specifically structural commitments. Neorealists hold that a focus on structures represents an alternative to the defensive reductionist tactic of formulating ad hoc adjustments to unanticipated new evidence – a tactic that produces, in Waltz’s words, “an infinite proliferation of variables” (1979: 65). By abandoning the behavioralist faith in explaining outcomes in terms of behaving units, and granting priority to the constraining and disposing forces of structures, neorealists believe they can explain more, and with fewer variables, than has previously been possible. While the very concept of “structure” is notoriously elusive and contested (Runciman, 1978), in the case of neorealism one can locate at least a proximal meaning. Generally, it “designates a set of constraining conditions” (Waltz, 1979: 73) that indirectly produce a limited range of outcomes. Structures thus operate as constraints upon the system’s units, disposing them to act in certain ways and not in others. Units, i.e., states, that fail to heed these constraints thus become less competitive, and if they persist in doing so they are ultimately eliminated from the international system. In this way nations, like individual firms in market economies, are socialized into the international system. They either adapt to the specific requirements of structures or cease to exist (Waltz, 1979: 76–77). The structural, or systems, approach is therefore not distinctive because it a priori rules out unit level influences. It does not. Rather, it is distinctive, first, because it grants explanatory and causal priority to structures, and, second, because it demands that one carefully distinguish unit and system levels so that the effect of each on the other can be determined. Approaches which ignore these features make it impossible to answer or ask such questions as how the structure affects the characteristics of states, or how structures are affected by variations in the internal organization and behavior of states. In short, they make the establishment of causal relationships, as well as the attempt to find patterns of recurrent behavior and ranges of expected effects that may be associated with different international systems, impossible.

AN ANALYSIS OF NEOREALIST THEORY

161

Having hinged their claims of superiority on neorealism’s professed structural and scientific credentials, neorealists recommend the construction of models of international politics that reflect the elegance, completeness, predictive precision, and abstractedness of ideal natural science. Given these aspirations, it is appropriate to begin with an examination of two of the most central and vexing questions affecting all sciences, the questions of explanation and evaluation. Since it is commonly thought that explanations in the natural sciences are both powerful and intersubjectively testable, one would assume that the aspiration to model methods on those sciences would generate theories of international politics which were at least in principle both powerful and testable. Surprisingly, however, Waltz’s structural theory is neither. As we shall observe, his theory lacks explanatory scope, as well as a clear and common standard for evaluating theories. On the question of the explanatory power of structural theories, three deficiencies in particular are vitally important: (1) they can neither explain changes within a system nor predict structural change itself; (2) they cannot account for particularities; and (3) they cannot satisfy the demands of testability. Now the first deficiency is technically two separate issues, though both stem from the same source: the inability of structural theories to account for change. Structures, Waltz explains, affect outcomes by constraining the range of possible choices available to actors at the state level. By granting methodological priority to these structural limits on the range of “rational” activity,5 structural theories offer potentially powerful explanations of recurrent outcomes in the international arena, outcomes that occur independently of states’ political and economic institutions, social customs, and political ideologies. However, precisely because these structures constrain and dispose, rather than determine actions, they can – and for some reason Waltz seems rather insouciant in announcing this – be expected to explain only “recurrences and repetitions, not change” (Waltz, 1979: 69).6 Now to the extent that Waltz makes the rather modest suggestion that we should not reject a theory solely because it fails to explain all of the behavior within a given system, his point is both salient and sensible. If what a theory explains is valuable – and I assume that an explanation of recurrences within a system is – then we would be ill-advised to abandon it solely because it fails to explain other phenomena. However, it is equally clear that we would be ill-advised to exclude entirely considerations of what a theory cannot explain from our evaluation of it. Here as elsewhere any assessment must weigh the theory’s explanatory costs, as well as its virtues. In the case of Waltz’s structural theory, however, the costs are more than nominal. For a theory which in principle cannot explain the dynamics of change within a system is also one

162

KEITH TOPPER

which cannot explain structural change itself; and as Waltz himself suggests, structural change is not identical to those small incremental changes that we encounter so regularly during the course of our daily lives. Rather, structural change is “a revolution,” and like all revolutions its effects resonate throughout the international “system,” ultimately transforming the very pattern and meaning of our practices. As such, they clearly have profound consequences, and thus an ability to anticipate and respond to them is absolutely critical. To cite only two of several possible examples, the American refusal to maintain liberal economic policies during the 1920s and 1930s, as well as the unwillingness of British leaders to shift their national priorities after World War II (Blank, 1978: 89–137; Lake, 1983: 517–43), are widely viewed as catastrophic policies resulting from an inability or unwillingness to adapt to structural changes in the international political economy.7 The second deficiency – that structural theories cannot account for particularities – is a source of some acrimony among contemporary writers. At least one figure within the neorealist tradition implicitly rejects Waltz’s stance on this issue, assuming instead that such a posture places unwarranted restraints on the application of purely structural theories. By contrast, others have denounced Waltz’s position as nothing more than a clever effort to shield structural theories from potentially falsifying evidence.8 In what follows I shift the discussion in a rather different direction and suggest that although Waltz’s general methodological point is sound, it nonetheless fails to screen the structural approach from attack in the way he assumes. According to Waltz, structural theories of international politics cannot be expected to explain the particular policies of states any more than one can expect the law of gravitation to predict the “wayward path of a falling leaf” (Waltz, 1979: 121). This, he explains, is because theories must always define an appropriate level of investigation, yet the very act of definition itself implies that some phenomena fall outside of any given theory’s explanatory framework. He concludes therefore that theories “at one level of generality cannot answer questions about matters at a different level of generality” (Waltz, 1979: 121). Waltz’s basic point is certainly incontestable: just as a theory of action need not account for every detail of a person’s behavior, e.g., coughs or the precise bodily movements constituting a specific action could normally be excluded from consideration, neither must a theory of international politics account for all behavior in the international realm. Importantly, though, this alone does not vindicate Waltz’s stance. For whether a particular range of phenomena can be excluded from a theory depends upon whether those questions which a theory seeks to explain can be answered without reference to the range of phenomena in question. In less abstract language, this means that the level

AN ANALYSIS OF NEOREALIST THEORY

163

of analysis one chooses must be determined by the practical questions which one wants the theory to address. Here the practical question – or at least one of them – is whether a purely structural theory of international politics yields explanations that usefully inform practices and activities in the international realm. On this issue, however, there is considerable cause for apprehension. For although Waltz’s theory can at least in principle explain which structures are most stable, such information is rarely of any practical use to policymakers and other actors who normally lack the capacity to effect structural change. To believe that knowledge of Waltz’s theory – even if it was shown to be universally true – could play a substantive role in transforming a multipolar into a bipolar system, or could convince a rising third power to curtail its military and economic growth in the name of preserving stability, is doubtful at best. In short, a purely structural theory like Waltz’s might in a highly abstract way inform our understanding of the international system, but given the dynamics of his framework it is unclear how it could inform our practices in that system.9 The third shortcoming of structural theories – that they fail to satisfy the demands of testability – is perhaps the most disturbing, since it apparently leaves us without any clear standard of evaluation. As Waltz argues, a purely structural theory cannot be rejected on the basis of particular deviating cases. This is because, first, balance-of-power theory offers only indeterminate or “probabilistic” predictions, and, second, the internal conditions of states also shape their actions and policies. Waltz suggests that theories of this type can therefore be tested only in the most general sense – for example, if one finds that balances-of-power tend to form even in cases where these formations would otherwise be implausible, then this tends to confirm the theory. While Waltz is correct in maintaining that the tests used to evaluate a theory cannot be determined in abstraction from the theory itself, he fails to recognize that it is also the case that whatever tests are adopted must in some way facilitate our ability to intersubjectively evaluate that theory.10 Such tests may, of course, prove inconclusive. However, tests which by their very structure preclude even the possibility of doing this are simply hollow. But as Friedrich Kratochwil has incisively observed, Waltz’s probability position, when combined with the fact that tests of international relations theories usually draw upon a limited number of cases to begin with, consigns his proposed tests to just this status: Finally there is the issue of the basis on which we are justified in rejecting a law that figures prominently in an explanation sketch, especially if we deal with a very small number of “cases”(as is usual in international relations). After all, no single case can now refute the stated hypothesis. When

164

KEITH TOPPER

observed and predicted results coincide, the “hypothesis” is corroborated; when they diverge, the hypothesis can be immunized from criticism by arguing that the prediction was only probabilistic. The problem of “refutation” then becomes largely one of practice, when a given set of scientists no longer feels comfortable (subjective belief) with a standard explanation and thus rejects the “law” stated in the hypothesis (Kratochwil, 1984: 314). In sum, Waltz’s basic methodological point may be credible, but his proposed tests remain inadequate. If the preceding discussion of explanation and evaluation fails to clearly distinguish or validate the neorealist enterprise, one might hope that Waltz’s discussion regarding the nature of laws and theories would. Unfortunately, a close examination of his account indicates not only that the neorealist vision of scientific practice is in no way unique, but that in many respects it rests upon questionable and misleading assumptions about the nature of scientific practice itself. For Waltz, the debate over the standards of scientific practice is clearly a debate involving only two voices. First, there is the naive inductivist school which claims that theories are mere “collections or sets of laws pertaining to a particular behavior or phenomenon” (Waltz, 1979: 2). According to this account, theory construction is a process of “collecting carefully verified, interconnected hypotheses” (Waltz, 1979: 2). The governing assumption behind this “inductivist illusion” is the belief that through the sheer force of amassing data and correlations some sort of theory corresponding to “reality” will magically emerge. Unsurprisingly, though, this faith turns out to be nothing more than another desperate dance into a dead end. Facts, Waltz argues, never “speak for themselves.” Nor does the process of collecting endless amounts of data and establishing larger and larger sets of statistical correlations ever establish certain knowledge, since “Nothing is both empirical and absolutely true” (Waltz, 1979: 5). Such a procedure, moreover, cannot even suggest which correlations are relevant or how those which are relevant are interrelated. It cannot, in short, explain anything. Against this purely descriptive conception of theory, Waltz juxtaposes a second construal. In this view, unlike the previous one, there is a qualitative difference between theories and laws. Theories are not “mere collections of laws,” but are explanations of them. Furthermore, because the purpose of theories in this account is to explain the world rather than reproduce it, theories are never constructed through induction alone. They instead require simplification, i.e., isolation, abstraction, aggregation and idealization, and this always necessitates a movement “away from ‘reality’ ” (Waltz, 1979: 7). According to Waltz, then, theories never replicate reality, but rather picture

AN ANALYSIS OF NEOREALIST THEORY

165

the principal forces and relations in some bounded domain of reality. As Waltz writes,

the aim is to try to find some central tendency among a confusion of tendencies, to single out the propelling principle even though other principles operate, to seek the essential factors where innumerable factors are present (Waltz, 1979: 10). Now insofar as one is presented with the choice of accepting either a necessarily barren “inductivist illusion,” or Waltz’s own variation on the thought of Ernest Nagel, James Conant, and others associated with the standard “empiricist” position in the philosophy of science, Waltz certainly emerges victorious. However, given the way in which Waltz frames his discussion, it is not clear that we have a debate since it is not clear that we have any genuine choices. At a minimum, it seems doubtful that many, if any, writers self-consciously subscribe to Waltz’s caricatured account of the inductivist stance. Certainly among contemporary philosophers and historians of science this view is held in ill-repute. Moreover, in articulating his own position Waltz frequently either states the obvious, e.g., “Both induction and deduction are indispensable in the construction of a theory, but using them in combination gives rise to a theory only if a creative idea emerges” (Waltz, 1979: 11),11 or fails to offer any sustained argument in defense of highly questionable claims. One notable instance of the latter tendency is his distinction between laws, which express purely observational relations, and theories. This distinction, which is central to his argument against induction, is based on the positivist belief that observation terms and statements are different in kind and not just in degree from theoretical terms and statements. Although this view is also widely rejected by contemporary philosophers of science – principally because it has now become apparent (1) that visual experiences are not determined exclusively by either the light rays entering the observer’s eyes or the images on the retina (Gombrich, 1969; Hanson, 1965) and (2) that all observations and observation statements presuppose some theory (Feyerabend, 1975: chapter 7; Kuhn, 1970: chapter 10; Popper, 1968: chapter 5 and appendix 10)12 – Waltz does not even acknowledge these objections, much less respond to them. Perhaps even more significant is Waltz’s refusal to examine earnestly the appropriateness of the natural science model itself. For although the question of whether one can successfully appropriate this model for human affairs clearly logically precedes the question of how one appropriates it, Waltz spends considerable time explaining the latter without ever seriously consid-

166

KEITH TOPPER

ering the former. He thus categorically assumes precisely that which needs to be shown. If Waltz had not put the proverbial cart before the horse, one suspects he would have been more circumspect from the outset. For the belief that progress in the human sciences will be achieved by following the example of ideal natural science has come under sharp attack from a variety of fronts, although two in particular merit attention. On the one hand, Charles Taylor and others argue that there is an important distinction between the kind of understanding that we can expect in the human sciences and the sorts of theoretical explanations that one finds in the mature natural sciences (Taylor, 1979: 25–74).13 According to this view, the unique subject matter of the human sciences makes any effort to appropriate the model of ideal natural science misguided from the start. On the other hand, Richard Rorty (1979; 1980: 39–46; 1982: 191–210 and 1988: 49–74), Mary Hesse (1980) and many of those associated with the “strong programme” in the sociology of science hold that traditional distinctions between the natural and human sciences rest on false dichotomies.14 Far from possessing a special method which underwrites their success, the natural sciences are every bit as infected by problems of interpretation, translation, and theory-ladenness as the human sciences. Thus, after outlining five false contrasts between the natural and human sciences, Hesse claims: What is immediately striking about it to readers versed in recent literature in philosophy of science is that almost every point made about the human sciences has recently been made about the natural sciences, and that the five points made about the natural sciences presuppose a traditional empiricist view of natural science that is almost universally discredited (Hesse, 1980: 171–172).15 While this latter argument is important and interesting, for the sake of brevity I will confine discussion principally to the thesis that there is a fundamental difference between the natural and human sciences.16 In the essay, “Interpretation and the Sciences of Man,” Charles Taylor argues that there is an essential difference between the objects studied in the natural and human sciences, one that, correctly understood, explains why the aspiration to attain the explanatory and predictive power of the mature physical sciences is ultimately both misguided and futile. In Taylor’s account, the natural and human sciences are distinguished by the fact that while the former study mere objects, the latter study self-interpreting and self-defining beings, i.e., beings that are partly constituted by their individual and social self-interpretations and self-definitions.17 However, precisely because these self-interpretations and

AN ANALYSIS OF NEOREALIST THEORY

167

self-definitions can and do change over time, place, and culture, any attempt to develop general and highly predictive theories of human behavior must prove unavailing. Taylor thus concludes that it is a grave mistake to measure the human sciences “against the requirements of a science of verification: we cannot judge them by their predictive capacities” (Taylor, 1979: 71). To bring this into sharper focus it might be useful to cite Taylor’s own example, one taken directly from the discipline of political science. As Hubert Dreyfus, paraphrasing Taylor, explains:

recent political scientists assume that the political domain is essentially restricted to the adjudication of individual preferences. This definition of politics makes possible theories based upon polls to determine preferences. But this definition is just one possible interpretation of the political domain. It can change with changing social self-interpretations. The point is not that the actual techniques used to collect data can be criticized as not sufficiently controlled, etc. (such a critique could also be appropriate in physics), nor that the assumptions guiding the collection of data might be false and therefore lead to false data (this can happen in any science since all data are theory laden). Taylor’s example is meant to show, rather, that interpretation both on the part of researchers and on the part of the groups that they study, is involved in determining what are to count as political facts. Taylor claims that similar conflicts of interpretation do not occur in the natural sciences, since the objects they study are not self-interpreting (Dreyfus, 1986: 5). This stance has, I believe, important implications for both the theoretical and metatheoretical arguments which Waltz and other neorealists adduce. Setting aside the former issues for the moment, I shall begin with the latter. At the metatheoretical level, Taylor’s emphasis upon the constitutive role of selfinterpretations is critical because it forces one to recognize that Waltz’s – and neorealists’ – conception of ideal theory is only a particular interpretation of what theory is and of what a good theory should look like. The notion, for example, that there is a single paradigm of understanding which should govern not only one’s conception of the natural world, but also that of the human – and hence political – world is itself a relatively recent phenomenon. It is, moreover, a conviction based upon a specific interpretation of politics, human agency, etc., one which, for instance, Aristotle and most Greeks would have thought absurd. Yet it is an idea which most neorealists accept. Given, then, the fact that the neorealist conception of ideal theory is an interpretation and not an analytic truth, one must ask why this interpretation is to be privileged

168

KEITH TOPPER

over all others, why one standard of theory should be granted the status of arbiter over all that is called theory? Waltz offers several answers to this type of question, but clearly his principal argument is based upon his understanding of the sources of “progress” in the natural sciences. He maintains that progress in the sciences of nature is chiefly the result of the researcher’s ability to move “away from everyday reality.” By abjuring empirically-based theory construction, students of natural science have developed highly parsimonious and powerful models of the natural world. Scholars of international politics, on the other hand, have made the mistake of increasing the “empirical content” of their theories, and so have failed “to bring off the Copernican revolution that others have called for” (Waltz, 1979: 69): To construct a theory we have to abstract from reality, that is, to leave aside most of what we see and experience. Students of international politics have tried to get closer to the reality of international practice and to increase the empirical content of their studies. Natural science, in contrast, has advanced over the millennia by moving away from everyday reality and by fulfilling Conant’s previously mentioned aspiration to lower “the degree of empiricism involved in solving problems.” Natural scientists look for simplicities: elemental units and elegant theories about them. Students of international politics complicate their studies and claim to locate more and more variables. The subject matters of the social and natural sciences are profoundly different. The difference does not obliterate certain possibilities and necessities. No matter what the subject, we have to bound the domain of our concern, to organize it, to simplify the materials we deal with, to concentrate on central tendencies, and to single out the strongest propelling forces (Waltz, 1979: 68). Apart from the fact that the very notion of cumulative scientific progress is itself an issue of deep dispute within post-Kuhnian history and philosophy of science, there are several problems with this argument. One problem is that Waltz’s distinction between the empirical content of theories in the natural and human sciences largely mirrors the second of Hesse’s five false dichotomies: “In natural science theories are artificial constructions or models, yielding explanation in the sense of a logic of hypothetico-deduction. . . . In human science theories are mimetic reconstructions of the facts themselves . . .” (Hesse, 1980: 170). Against this “almost universally discredited” view, Hesse maintains that “natural science theories are not models externally compared to nature in a hypothetico-deductive schema, they are the way the facts themselves are seen” (Hesse, 1980: 172).

AN ANALYSIS OF NEOREALIST THEORY

169

In addition, Waltz’s argument suffers from an equivocation in his use of the term “simplification.” When Waltz speaks of simplification he typically uses the term in a highly restrictive sense where it designates the particular form of simplification associated with natural scientific theories. Here simplification implies the construction of abstract models which employ a minimum number of variables and offer general explanations and predictions. This form of simplification we might call theoretical simplification. Simplification, however, can be used – and Waltz at times uses the term in this way – in a much more general sense of concentration on “central tendencies” or identification of “the strongest propelling forces.” This form of simplification, however, need not be identified with theoretical simplification. Such a form of simplification, which might be called practical simplification, is consistent, for example, with the “prototype studies” employed by Clifford Geertz and Michel Foucault.18 Here their interpretations of social and cultural practices are organized around particular paradigms such as Balinese cock fights or European prisons. And while these paradigms do not pinpoint “central tendencies” or “propelling forces” in a strictly statistical sense, they contain simplifications in the sense that they bring into perspicuous focus general features and modes of organization of the cultural and background practices they explore. Indeed, prototype studies reveal propelling forces without constructing formal models; they identify central tendencies without offering conventional causal explanations; they function as analogues and benchmarks without presupposing decontextualized decision-procedures; they inform and locate what is at stake in our practices without abstracting from the shared self-interpretations which partly constitute them. As Dreyfus, drawing on Thomas Kuhn’s work, explains more fully: the function of prototypes cannot be captured in a theory. A scientific paradigm, Kuhn has pointed out, is neither rationalized nor rationalizable. An activity is recognized as good science by being seen as similar to an exemplar, without it being similar with respect to any abstractable, identical features. Thus paradigms or exemplars cannot be treated as abstract types; rules cannot replace the typical case. More generally, human beings tend to agree in their judgments of how similar an object or event is to a prototypical case, without being able to rationalize their judgments in terms of context-free features. . . . Indeed . . . what counts as a relevant feature seems to follow from judgments of similarity rather than to be presupposed by them (Dreyfus, 1986: 21). Thus, if the distinction between theoretical and practical simplification is warranted, one can agree with neorealists who argue that simplification is

170

KEITH TOPPER

necessary, yet deny that all simplification must be theoretical. Moreover, given the repeated failure of efforts to pattern human and social scientific investigations on the model of ideal natural science, one might build a prima facie case for preferring prototype studies over structural theories. In concluding this discussion I might add that although prototype studies are not held in particularly high esteem by most contemporary students of international relations, this has not always been the case. For in spite of the dubious tendency to see Machiavelli as “the first true political scientist” (Gilpin, 1984: 297),19 his method – and Thucydides’ also – is in fact much closer to the prototype studies advocated above than to neorealism or any of the other “scientific” approaches to the study of international politics. In this regard, it is worth recalling that although Machiavelli sometimes appears committed to generalization and “scientific rationalism,” these passages are almost always supplemented with endless examples. 2. Structuralism As is well-known, Waltz argues that the political structure of the international system consists of three analytic components: the principle according to which the system is organized or ordered (in the case of international, as opposed to domestic, politics Waltz claims that this principle is anarchy, i.e., the absence of hierarchical order and relations of authority [Waltz, 1979: 88– 93]); the differentiation of units and the specification of their functions; and the distribution of capabilities across units.20 While these features are widely accepted as basic components of the international system, I shall contend that both the first and third parts of this trinity are miscast in ways that deeply affect our understanding of international politics.21 Moreover, to the extent that neorealists recast their definition of structure in more adequate terms, the likelihood of generating a powerful, elegant, and parsimonious systems theory of international affairs becomes correspondingly more remote. Beginning with the first component, i.e., anarchy, and in particular the atomistic notion of interstate relations which it implies, what is clear is that the concept, taken literally, is in some sense absurd. For as the previous discussion regarding the constitutive role of self-interpretations indicated, every form of collaboration in the international realm, even if engaged in for purely instrumental purposes, presupposes shared self-interpretations on the part of the participants. The idea of negotiation, for example, is central to all sorts of activities and outcomes in the international realm: reducing tariffs, ending wars, resolving border disputes, and so forth. Yet the “process of negotiation” is literally impossible in the absence of certain shared self-interpretations on the part of those involved. Without a shared understanding of notions like

AN ANALYSIS OF NEOREALIST THEORY

171

bargaining, the distinct identity and autonomy of the parties concerned, the willed nature of their relations, etc., the vocabulary of negotiation and the distinctions it marks would be altogether inappropriate. Thus, when we speak of a “process of negotiation” we presuppose that the participants share a range of practices which are constitutive of those participants’ understanding of social reality. Now the purpose of this example is not to suggest that Waltz or anyone else who posits anarchy as a structural feature really believes that states possess nothing in common. Such a view would be preposterous, and Waltz, for one, explicitly denies it. It is, however, meant to suggest that the atomistic model of social relations implied by anarchy, when taken as anything more than a very modest heuristic, itself becomes incoherent. To seriously maintain “the theoretical primacy of individual actors” (Ashley, 1984: 243), to argue that social order and collaboration are entirely derivative relations, implies not simply the adoption of reified assumptions (after all, one might argue, assumptions qua assumptions are always to a greater or lesser extent reified), but more importantly the embrace of logical absurdities. For while such assumptions may explain certain consequences of collaboration and social order, they cannot explain how collaboration and social order emerge in the first place. In all fairness it should be noted that neorealists do not wholly ignore these implicit absurdities, even if they fail to genuinely transcend them. Waltz, for instance, admits that “all societies are mixed” (Waltz, 1979: 115) in the sense that they all contain both hierarchic and anarchic elements. Nevertheless, he defends a strict categorical separation between anarchy and hierarchy by arguing that one “who wishes to explain rather than to describe” (Waltz, 1979: 115) must resist the temptation to multiply the number of categories. The problem with such a procedure, however, is that the absence of precise categories sabotages the very effort to develop causal explanations capable of yielding precise predictions. This is because the categories themselves are likely to produce such results only to the degree that they are isomorphic with the principal forces impinging upon state behavior, but since those forces themselves change in accordance with changes in self-interpretations, the adequacy of those categories will fluctuate significantly over time. Moreover, this problem is compounded by the fact that many of the relevant forces are exceedingly difficult to measure in any but the most proximal way. The result is that theorists like Waltz are forced to moderate their claims considerably, thus immunizing themselves from direct refutation at the expense of articulating precisely the limits within which their theories can be predictive. Waltz, for example, tells us that hierarchic elements do operate “but only in ways strongly conditioned by” (Waltz, 1979: 116, emphasis added) the structure

172

KEITH TOPPER

of anarchy. Later he says that “Structures defined according to two ordering principles help to explain important aspects of social and political behavior” (Waltz, 1979: 116, emphasis added). While statements of this type manifest the virtues of prudence and caution, they are clearly incapable of operating as the basis for a stable and predictive science. This is because, unlike statements of boundary conditions in physics and chemistry, Waltz presumes that background conditions change, yet he is incapable of indicating in advance what particular conditions must hold in order for them to remain predictively reliable. The result is a theory which is intrinsically unstable in a way that theories in physics, or even in most areas of biology, are not (Dreyfus, 1986: 16). In fact, a close examination of these problems reveals that the very same tendencies cited by neorealists as reasons for jettisoning reductionist approaches themselves reappear in neorealist theories. As I have argued above, the search for parsimony precludes a rigorously causal and highly predictive analysis of international politics. Waltz claims therefore that outcomes are only “strongly conditioned” by the structure of anarchy, that reducing structures to two ordering principles “helps” explain “important aspects” of political and social behavior. However, in the absence of further specification of how, when, where, and to what degree outcomes are “strongly conditioned” by the anarchical order, the theory itself becomes lost in a fog of vague qualifiers and ceteris paribus clauses. In order to avoid this disturbing consequence neorealists such as David Lake, Stephen Krasner and others have rightly recognized the need to develop less parsimonious but more highly descriptive models of the international system. But as Waltz has argued, such attempts inevitably imply the multiplication of variables. Thus Lake, having acknowledged that “a purely structural argument” could not explain American foreign economic policy even over a period of less than fifty years, argues that “it appears that issues of high domestic political saliency may be more insulated from the effects of the international economic structure than other policies” (Lake, 1983: 539). “Ultimately,” he adds, “structure and process must be integrated in a theory of political economy” (Lake, 1983: 540). Reading this, one is reminded of nothing so much as the reductionist attempts to avoid otherwise falsifying evidence by introducing new variables ex post. 3. Power Although efforts to define the concept of power are notoriously contentious – Gilpin, Krasner, Lake, and Waltz, for example, all define power in subtly different ways – this should not deter one from recognizing that neorealists share several basic assumptions about the nature of power. Most importantly, neorealists typically construe power as relational, subjective, and tangible,

AN ANALYSIS OF NEOREALIST THEORY

173

and view force or military might as the dominant form of power. In what follows I will argue that such definitions of power misconceive the ways in which power operates in the international arena. Additionally, I shall hold that any attempt to incorporate a more complex vision of power into Waltz’s third component of structure is highly problematic, since many of these forms cannot be designated specifically in terms of the objective capabilities of states. Perhaps the most serious problem with the neorealist conception of power is its fidelity to the idea that there is a rigid hierarchy among different forms of power, one that is independent of the various local contexts in which power relations operate. According to most neorealists, political power, or force, is almost invariably the dominant form of power, with economic capability typically being secondary and what Richard Ashley labels the “competence model of power,” or what Pierre Bourdieu calls “symbolic power,” being of least import (Ashley, 1984: 259; Bourdieu, 1989: 14–25).22 This construal is thus founded on the conviction, first, that there is an unchanging hierarchy of different forms of power, and, second, that power resources are noncontextual in the sense that their capacities are not highly sensitive to the situational context in which they are employed.23 These convictions, however, are themselves problematic. At a minimum, they disregard both historical and contextual changes in the relative import of different forms of power. For instance, in this century, and especially during the last forty years, there has been a shift toward greater diversification of power resources. Unlike previous periods, power can no longer be identified with military strength alone. This means, on the one hand, that states must possess a wider range of power resources in order to act effectively across a more diverse set of foreign policy issues. It also means that power resources cannot be deployed indiscriminately in any situation where they are needed. Military power, for example, cannot always be used even if a state is militarily superior. On the other hand, it suggests that states and nonstate actors with highly specific power resources can play an important role in affecting outcomes in specific issue areas. The role of the OPEC nations during the 1970s is only one conspicuous instance of how actors without any substantial military or political power can significantly shape outcomes in the international arena.24 More pointedly, recent events in the ex-USSR, Eastern Europe, and the United States have revealed the considerable risks of assuming that military power is the principal tool for shaping outcomes in international politics. In the former Soviet Union, massive military might was such a costly affair that it eventually proved incapable of preserving – in fact, most argue that it increasingly crippled – the state itself. Similarly, the recent elevation of the US

174

KEITH TOPPER

to a position of undisputed military hegemony has not noticeably transformed its ability to shape events in the international arena. Indeed, in the several years before and the short time since the fall of the USSR, American ability to translate its enormous military power into effective control or influence of international events and issues has been manifestly limited: the situation in the ex-USSR and much of Eastern Europe is at the very least disquieting, with widespread crime, poverty, ecological despoliation, and political instability being perhaps the most likely scenarios in some states during the immediate future;25 various forms of nationalism and religious fundamentalism continue to spread across the North African littoral, Eastern and Central Europe, and other parts of the world as well; and American ability to sustain a leadership position in the international economy remains a highly precarious affair. Even in the Persian Gulf War, where superior military power unquestionably influenced events and outcomes, the use of that power was both predicated upon and constrained by an international consensus regarding its legitimate ends and scope. A second, although related, shortcoming of the neorealist conception of power is its failure to acknowledge that the “competence model of power” or “symbolic power” plays a significant causal role in shaping international outcomes. According to Ashley, the competence model of power refers to the power that an actor has by virtue of his or her recognition by a community. In order to achieve this recognition actors must first demonstrate their “competence” to act in ways consistent with the communally defined meanings, norms, and practices. The difference between this conception of power and the neorealist one corresponds, Ashley suggests, to the difference between power “as a capacity to influence (to ‘have power’) and as a designation of competent, recognized, participants (to ‘be a power’)” (Ashley, 1984: 272). Symbolic power, although similar to the competence model in that it cannot exist independently of a community that confers its status, is “the power to make groups,” or “to make things with words” (Bourdieu, 1989: 23). This power is based upon the accumulation of what Pierre Bourdieu calls “symbolic capital,” i.e., prestige or social honor deriving from “a reputation for competence or an image of respectability” (Bourdieu, 1984: 291). Bourdieu contends that symbolic capital and symbolic power are “inextricably linked” to other forms of capital and power (in the sense that one form can frequently be “converted” into another), but remain conceptually distinct from and irreducible to them. For example, parents seek to improve their children’s position on the social trajectory by sending them to prestigious colleges or universities, hoping that the symbolic capital associated with a degree from such institutions can be converted into “profits” in the form of increased power or symbolic capital, e.g., professional titles or economic capital (in

AN ANALYSIS OF NEOREALIST THEORY

175

the form of remuneration that is at times, but not universally, connected with the possession of professional titles). Likewise, public figures who have been decorated or at least have served in the military – Charles De Gaulle, John F. Kennedy, and George Bush, to name just a few – accrue a form of symbolic capital that can, through the judicious use of various “strategies,” be converted into other forms of capital or power, e.g., political positions or the political power to act as a spokesperson – in Bourdieu’s words, one who is a “producer of objectified representations of the social world” (Bourdieu, 1985: 730) – on issues of war and national security. Bourdieu maintains that in certain circumstances symbolic capital is not only a causal factor in social and political life, but can be “perhaps the most valuable form of accumulation” (Bourdieu, 1977: 179). The reason why neorealists reject these construals of power should be apparent. Neorealist theory conceives of the international system as anarchic and void of any legitimate authority, and this conceptualization eliminates all theoretical space for notions of power resting upon a “consensual understanding” on the part of a community. It is therefore precisely the condition of anarchy which renders irrelevant symbolic or competence models of power. As I have suggested previously, however, this conception of anarchy can at best be seen as a pedagogical heuristic, not an ontological principle. For if indeed relations in the international arena presuppose shared selfinterpretations, then it is inexplicable why neorealists would exclude in principle the relevance of these forms of power. Moreover, reflection upon the nature of contemporary politics lends at least some credence to these claims. If we consider breakdowns in negotiations on some important issue, e.g., arms control, territorial disputes, or trade issues, and the volley of accusations that inevitably follow, we see not only the unleashing of frustrations, but also subtle attempts by various parties to accrue symbolic capital by appealing to an international audience. Likewise, if we think of how a spokesperson becomes endowed with the power to speak as the representative of a group, we see a process which is not unlike that outlined by Bourdieu and Ashley. Of course, disagreements may exist over whether a spokesperson is in fact a legitimate representative of some group, but this alone fails to vitiate the possibility of having a consensual understanding. Indeed, genuine disagreement presupposes it: apart from the question of how much people’s beliefs converge is the question of how much they have a common language of social and political reality in which these beliefs are expressed. This second question cannot be reduced to the first; intersubjective meaning is not a matter of converging beliefs or values (Taylor, 1979: 49).

176

KEITH TOPPER

Finally, there is the neorealist emphasis upon the subjective nature of power. For neorealists, power is an objective capability which is always possessed – whether or not it is ever employed directly – by a unitary subject. Thus, power implies not only an ability to affect others, but a specific locus from which that ability derives. However, this view, though commonplace, masks the dynamics of power embodied in what Michel Foucault calls the “microphysics of power.” According to Foucault, it is a mistake to conceive of power “as a group of institutions and mechanisms that ensure the subservience of the citizens of a given state,” or as a “mode of subjection,” or even as a “general system of domination exerted by one group over another” (Foucault, 1980: 92). Most significantly, The analysis, made in terms of power, must not assume that the sovereignty of the state, the form of the law, or the overall unity of domination are given at the outset; rather, these are only the terminal forms power takes . . . power must be understood in the first instance as the multiplicity of force relations immanent in the sphere in which they operate and which constitute their own organization . . . and lastly, as the strategies in which they take effect, whose general design or institutional crystallization is embodied in the state apparatus, in the formulation of the law, in the various social hegemonies (Foucault, 1980: 92–93). For Foucault, power relations are simultaneously “intentional and nonsubjective” (Foucault, 1980: 94, emphasis added). That is, at the local level power relations are indeed constituted by intentional, volitional activity. Here individual and collective subjects make decisions, shape agendas, plot strategies, and pursue goals, oftentimes in a palpably conscious and transparent manner. At the same time, however, these disparate localized “strategies” produce broader patterns and consequences which are not, and often cannot be anticipated, willed, or otherwise coordinated in a conscious, systematic fashion. In these “strategies without a strategist” Foucault holds that it is a mistake to think that power emanates from some central point – be it the state, the ruling class, the legal code, or elsewhere. Rather, the “logic” of a situation may be “perfectly clear, the aims decipherable, and yet it is often the case that no one is there to have invented them, and few who can be said to have formulated them” (Foucault, 1980: 95). In cases like these, Foucault deems it deeply misguided to “look for the headquarters that preside over its rationality” (Foucault, 1980: 95). This account pinpoints several features of power relations that are necessarily ignored or denied by neorealists. First, the idea that power is “intentional and nonsubjective” makes sense of actions which neorealists must otherwise

AN ANALYSIS OF NEOREALIST THEORY

177

label irrational. Stephen Krasner, for instance, states that all of the traditional accounts of American participation in Vietnam “just do not make any sense” (Krasner, 1978: 321). For Krasner, what is remarkable about American participation in this debacle is precisely the absence of any clear means-ends assessments, the detachment of policy from any “prudent calculation” of costs and benefits. Instead, policy appears to have been driven by the force of deep ideological commitments, commitments which were themselves deaf to the tremendous costs – both domestically and internationally – of American involvement. Such “nonlogical” pursuit of ideological ends is, Krasner contends, intelligible only from “a statist view that sees the state as capable of defining its own autonomous goals” (Krasner, 1978: 333). Suspending for a moment the very real issue of whether cost-benefit calculations were as absent from the decision-making process as Krasner assumes, what is striking about this argument is the way in which it transforms mystery into an explanatory virtue. For if we take Krasner seriously, then what requires explanation is not only the distinctive goal of American policymakers during the post-war period, but also the disturbingly nonlogical manner in which that goal was apparently “pursued.” While Krasner offers at least plausible Hartzian arguments (Hartz, 1955) in reply to the former demand, his response to the latter remains entirely occult. We are given no genuine reasons for this peculiar behavior, and can only surmise that the spell of anticommunism ran so deep that it occluded policymakers’ very ability and/or willingness to calculate costs and benefits, to recognize that their behavior was shredding the social and ideological fabric which they intended to export around the globe. But surely, we are inclined to think, there must be more effective ways of creating the world in one’s own image. Why, then, the picture of mass autism among policymakers, why the utter inability of otherwise learned and intelligent individuals to recognize the counterproductive consequences of their policies? Here I contend that Foucault’s vision of power offers a more compelling account of the Vietnam experience, one that makes sense of particular decisions and policies, while avoiding the idea that the broader emergent patterns of behavior are necessarily the product of any systematically coordinated efforts. On the one hand, it facilitates a richer understanding of the sundry activities, aims, decisions, conflicts, and strategies taking place within those institutions constituting “the state.” Rather than assuming a symphonic unity of purpose orchestrated throughout by the ideology of Lockean liberalism, Foucault suggests that we see the state as a complex field of forces which are sometimes relatively autonomous, which at other times converge and overlap, and at still other moments diverge and collide. Within this field ideological motives would almost certainly have a place, but these motives would them-

178

KEITH TOPPER

selves be situated within the context of numerous other local considerations, e.g., Democratic leaders’ fear that losing Vietnam would cost them dearly in future elections; various bureaucrats’ desire to aggrandize their own, or their bureaucracy’s position; policymakers’ faith in the infallibility of Western technology; foreign policy managers’ inability to understand the values and practices of a foreign culture; politicians’ concern that ending the war might plunge the nation into a recession. On the other hand, this view departs dramatically from reductionist theories of international politics. For these theories – whether they be Marxist or neo-Marxist theories, e.g., the Hobson-Lenin theory of imperialism or some version of interest-group theory – share with structural realism the urge for general explanations of international political behavior. In one way or another, they assume that behavior in the international realm is a direct outgrowth of some stipulated general elements or laws operating at national or sub-national levels. However, just as Foucault rejects the statist’s propensity to detach state behavior from the field of local political struggles, so too he repudiates the reductionist’s assertion that activity in the international realm can be unproblematically “read off” of some theory operating at the national or subnational level. Such theories, he implicitly suggests, conceive of fluid, contingent, tension-ridden, heterologous, and multivocal relations as essentially formal, mechanical, homologous, and univocal ones (see for example Foucault’s discussion of the “micro-physics of power” in Foucault, 1979: 26–28). Related to this is a further point which is critical to Foucault, but likewise cannot be captured on the neorealist grid of power. This is the idea that power is exercised on both the dominated and the dominant, that “there is a process of self-formation or autocolonization involved” (Dreyfus and Rabinow, 1983: 186). For Foucault, this means that although classes or other social groups may act to establish themselves in positions of domination, they often – and this may at times even be a precondition of domination – create in the process a network of power relations which operate upon and “dominate” even their own members. In this regard, Foucault insists that it is a mistake to think of power as a property, privilege, or possession; rather, it is better understood as a “strategy,” as something that is exercised without being owned, that is relational rather than locational (see especially Foucault, 1979: 26–27).26 Now one can discern faint echoes of such an idea in Waltz’s claim that outcomes cannot be inferred from intentions and in Gilpin’s suggestion that “a capitalist international economy plants the seeds of its own destruction in that it diffuses economic growth, industry and technology, and thereby undermines the distribution of power upon which that liberal, interdependent economy has rested” (Gilpin, 1975: 260). However, even here the gener-

AN ANALYSIS OF NEOREALIST THEORY

179

al arguments move along distinctly different planes. First, Gilpin neglects entirely the fact that the very creation of a hegemonic or imperialist state presupposes the imposition of a sweeping range of disciplinary technologies upon its own citizens. Moreover, he indicates that the self-destruction itself can be predicted and the consequences minimized if only the hegemon forsakes short-term for long-term rationality. For Foucault, however, the consequences of power cannot be predicted and controlled in this way. This is partly because networks of power condition, but do not determine actions, and partly because one can never stand “in a position of exteriority in relation to power” (Foucault, 1980: 95). Finally, Foucault’s contention that power is dispersed, mobile and in some sense autonomous impugns the neorealist faith that power is a unidirectional passive force that can be measured, calculated, divided and controlled. In Waltz’s words: “I offer the old and simple notion that an agent is powerful to the extent that he affects others more than they effect him” (Waltz, 1979: 192). For Foucault, the belief that power is an instrument that is located primarily or exclusively in, and used by, the state is not just a reified assumption but a dangerous delusion. In this respect, his equivocal use of the term “conduct” to describe power relations serves as a warning to Waltz and all others who maintain the sanguine belief that power is a pure instrument reducible to scientific and technical control: Perhaps the term conduct is one of the best aids for coming to terms with the specificity of power relations. For to conduct is at the same time to “lead” others (according to mechanisms of coercion which are, to varying degrees, strict) and a way of behaving within a more or less open field of possibilities. The exercise of power consists in guiding the possibility of conduct and putting in order the possible outcome. Basically, power is less a confrontation between two adversaries or the linking of one to the other than a question of government . . . “Government” . . . designated the way in which the conduct of individuals or groups might be directed. . . . To govern, in this sense, is to structure the possible field of action of others (Foucault, 1983: 221–222). If one reflects on the neorealist conception of power in light of the preceding remarks, it should be clear that there are several aspects of power that are not factored into the neorealist equation. While some of these might be assimilated fairly easily, others, such as Foucault’s emphasis on the dispersed, nonsubjective nature of power, would seem to require a fundamental rethinking of the enterprise itself. At a minimum, however, the idea that power can be designated in terms of objective capability – an idea that is foundational

180

KEITH TOPPER

to the neorealist conception of power – must be substantially reformulated if it is to reflect accurately the multiplicitous ways in which power shapes outcomes in international affairs. 4. Conclusion My analysis has sought to demonstrate the deficiencies of neorealist theories of international relations, and to sketch, however briefly, an alternative way of studying international affairs, one that is rigorous and edifying without being based upon ideal theory. While the issues raised can ultimately only be answered empirically, the above arguments are meant to indicate obstacles which a neorealist theory must overcome, if it is to be a rigorous structural theory of international relations. Of course, in most cases these arguments did not demonstrate in principle that the neorealist enterprise must fail; however, they have identified likely conundrums and shown how these conundrums are now appearing – often in precisely the way indicated by my analysis – in the current work of neorealists. Given these considerations, perhaps it is time to begin thinking about a new analytic order in the field of international politics. Notes

1.

2.

3. 4. 5.

I would like to thank Benjamin Barber, Jeffrey Frieden, Lyle Massey, Christilla Roederer, Ronald Rogowski, Michael Wintroub, and an anonymous reviewer for insightful comments on earlier versions of this essay. Whatever errors remain should be attributed to my own stubbornness. This quote, as well as the following one by Machiavelli, is intended to highlight and query some of the assumptions characteristic of influential theories of international politics. While the passages themselves contain sexualized and gendered connotations about knowledge and action, I will not discuss them here. For some explorations of these issues, see Jean Greybeal (1990); Luce Irigaray (1991); and Paul Patton (1993). On Machiavelli, see Hanna Fenichel Pitkin (1984). In using this phrase, I refer to a model of what ideally constitutes a natural scientific theory. For most neorealists there are at least four essential features of such an ideal theory: completeness, elegance, predictability, and abstractedness. For a more detailed, but not entirely parallel discussion of ideal theory, see Hubert L. Dreyfus (1986: 11–12). When I speak of “contexts” I refer to a situational background or set of background practices which, unlike a “system,” cannot be formalized or otherwise treated thematically. However, I will, when appropriate, identify and discuss those points where Waltz’s position departs significantly from the broad stream of neorealist thought. As employed here, the term “rationality” should not be understood as either a variant of Kantian “substantive rationality” or what is sometimes called “practical reason.” Both of these terms imply a form of reason which elicits normative judgments and obligations that are independent of subjective needs and desires. These forms of reason are universally rejected by neorealists as unscientific, metaphysical conceptions. Rather, rationality is construed as instrumental, or means-ends, rationality. This definition connotes the pursuit of the most efficient course of action available for the attainment of some predesignated

AN ANALYSIS OF NEOREALIST THEORY

6.

7.

8.

9.

10.

181

end. It produces, to use Kant’s terms, not categorical but hypothetical imperatives: “Here there is absolutely no question about the rationality or goodness of the end, but only about what must be done to attain it” (Kant, 1797/1964: 82–83). Additionally, one should note that for neorealists rationality implies not only that states have, and efficiently pursue, preferences or ends, but “that they have consistent, ordered preferences, and that they calculate costs and benefits of alternative courses of action in order to maximize their utility in view of these preferences” (Keohane, 1984: 27; emphasis added). Although the assumption of rationality is characteristic of the neorealist approach (see, for example, Lake, 1983: 517–543; and 1984: 141–168; Gilpin, 1981; and Ashley, 1984: 243), not all neorealists claim full allegiance to this premise. Stephen Krasner (1978: 15–17) argues that one of the fundamental reasons for preferring a statist approach over a structural Marxist one is that statism does not presuppose logical behavior or means-ends calculations. However, as John Ruggie has pointed out, this claim effectively transforms the realist tradition itself by shifting “onto Marxist shoulders precisely that burden of proof which has always been demanded of realists, while in his version of realism meansends calculations apparently are unnecessary and the state can destroy its own society in the name of the national interest” (Ruggie, 1980: 298). Also, Waltz argues in Theory of International Politics that his theory “requires no assumption of rationality or of constancy of the will on the part of all the actors. The theory says simply that if some do relatively well others will seek to emulate them or fall by the wayside” (Waltz, 1979: 118). But presuming that Waltz does indeed dispense with the assumption of rationality, one must wonder how his claim that “States . . . try in more or less sensible ways to use the means available in order to achieve the ends in view” differs in substance from the assumption of instrumental rationality. It cannot mean simply that some states employ more perfectly instrumental rationality than others, for this is a statement of fact which would not be denied by anyone adopting the rationality assumption. Nor could it mean that not all states are power maximizers, since this is an assumption about interests, not rationality. Thus, if this is not an assumption of instrumental rationality, it is unclear just what kind of an assumption it is. For additional criticisms of Waltz on this issue, see Morton Kaplan (1979: 29–33). Waltz’s position on this issue is by no means a universally held one. Recently, there have been several interesting efforts to integrate temporal and transitive dimensions into traditional structural analysis. See, for example, Marshall Sahlins (1981 and 1982: 73– 102); Pierre Bourdieu (1977); Paul Ricoeur (1974: 79–96); and Roy Bhaskar (1979 and 1986). The foregoing point, while significant, is not entirely novel. For other discussions regarding the problem of change in systems theoretic approaches, see Ole Holsti et al. (1980); and Barry Buzan and R.J. Barry Jones (1981). David Lake (1983), for example, attempts to predict American foreign economic policy using a “purely structural argument.” For a more direct critique of Waltz’s stance on this issue, see Richard Rosecrance (1981: 705). One might, of course, argue that a theory’s value need not be judged exclusively in terms of its ability to transform our practices, that merely clarifying or explaining them is both a worthy and honorable aspiration. While I agree with this stance, it is in this case misplaced. First, though legitimate, it misses the point of the above discussion, which is only to identify the limitations of purely structural theories, not to claim that they are entirely without merit. Secondly, it is at odds with Waltz’s own stance on this question, for he himself indicates that “questions of usefulness and uselessness” are indeed relevant criteria in matters of theory choice (Waltz, 1979: 124). Significantly, in a response to his critics, Waltz appears to endorse just this claim. He says, for example, that “one way to avoid an infinite regress is to require not tests but testability” (Waltz, 1986: 334–335). Thus, if my remarks are correct, then Waltz’s theory fails to meet his own demands.

182

KEITH TOPPER

11. Sometimes Waltz also combines statements of the obvious with a rather disturbing obliviousness regarding the complexity of the issue he is addressing, e.g., when he maintains that “In the end, one sticks with the theory that reveals most” (Waltz, 1979: 124). Such phrases are both true and empty, since the issue – and an extremely rancorous one at that – revolves around the question of how one determines which theory, among competing or incommensurable ones, “reveals” the most. 12. In noting this, I do not wish to suggest – in the manner of Gestalt psychology – that the thesis of “theory-ladenness” implies that all observations are so dominated by theoretical assumptions that they ineluctably presuppose the correctness of the theory they ostensibly test or elaborate, and thus must be abandoned upon the supersession of one theory by another. As Hacking points out (1983: 167–185), the concepts of theory and observation have disparate senses and applications, something which is often obscured by (and often obscures debates between) both positivists and their opponents. 13. For a sampling of different versions of this neo-Diltheyian stance, see Hubert Dreyfus (1980: 3–23); J¨urgen Habermas (1971); John Searle (1984); and Donald Davidson (1982). 14. For a selection of work associated with the “strong programme,” see David Bloor (1976 and 1981: 199–213); Barry Barnes (1977); Barnes and Bloor (1982: 21–47); Michael Mulkay (1979); and Steven Shapin (1982: 157–211). 15. In another essay, however, Hesse (1980: 187–205) appears to adopt a somewhat different stance, arguing that the human sciences differ from the natural sciences in that the former but not the latter lack any single criterion of theory choice. 16. I attempt to sort out what is compelling and problematic with this line of reasoning in Topper (chapter 1). 17. While Taylor’s account is discussed because it is the most systematically developed variant of this argument, it is neither the only nor the earliest one. Both Giambattista Vico (1725/1966) and John Stuart Mill (1872/1988) developed accounts of science that contained close family resemblances to Taylor’s, and reached similar conclusions about the distinctiveness of the human sciences. 18. The distinction between theoretical and practical simplification is indebted to Hubert Dreyfus’ discussion of two distinct types of holism (Dreyfus, 1980), which he labels practical and theoretical holism. While the parallels between these two uses should not be overstretched, the terms do, I believe, locate an important distinction. 19. Hanna Pitkin points out that those “who read Machiavelli as social scientist or proto-social scientist regard the concept of fortune as a lamentable regression to mythical prescientific explanation” (Pitkin, 1984: 160). Pitkin herself clearly rejects the idea that Machiavelli’s insistence on the role of fortune in human affairs was just a lamentable lapse from his normal scientific rationalism (Pitkin, 1984: 160–165). In a similar vein, Bernard Crick emphasizes that Machiavelli “is not a modern social scientist,” nor is there any “formal methodology” to be found in his writings (Crick, 1970: 37, 53). For a detailed examination of neorealists misappropriation of Thucydides’ thought, see, Daniel Garst (1984). 20. Waltz, of course, goes beyond this and argues that since the international arena is characterized by an absence of both central authority and functional differentiation, the “generative grammar” of international authority derives solely from the distribution of capabilities among units. 21. For the purposes of brevity I will not treat the second component of structure. However, for a penetrating critique of this component see John Ruggie (1984: 261–285). 22. I should emphasize that none of this suggests that neorealists exclude in principle factors such as leadership, prestige, or competence. Rather, a number of them – and Gilpin especially has made important gestures in this direction – explicitly maintain that these elements constitute part of the category of power. Nevertheless, the structure of neorealist theory forces one to discount the import of these factors. For the hallmark of the structural approach is the belief that structure tends to select motivation, goals, policies, and even the organization of domestic politics according to the necessities imposed by the environment. This means that while there is always some allowance for indeterminacy of outcome, and

AN ANALYSIS OF NEOREALIST THEORY

23.

24. 25. 26.

183

hence for the causal relevance of process-oriented factors like leadership, diplomatic skill, etc., these processes themselves occur only within the objective boundaries imposed by the structure. Thus, while the less tangible, less measurable aspects of capability are not wholly irrelevant, the fact that they operate within a highly restricted domain tends to minimize their role in constituting the category of capability. This further implies, of course, that states lacking the more fundamental constituents of power, e.g., military strength and economic capability, are deemed incapable, except in extraordinary circumstances, of significantly affecting the system even with the most astute and dynamic leadership. This position is not held by every neorealist. Robert Keohane (1986: 184), for example, emphasizes repeatedly that military power is not as fungible as neorealists – and particularly Waltz – sometimes suggest. Rather, he argues that the fungibility of military power varies over different “issue-areas”; thus, “different sets of capabilities will qualify as ‘power resources’ under different conditions.” Moreover, he observes correctly that such a stance is at odds with the neorealist’s demand for parsimonious theories. For if fungibility varies in the way that Keohane suggests, then one would require not a single, elegant theory, but a number of different theories, with each being appropriate only to a particular type of issue-area. Accordingly, since issue areas are almost never entirely autonomous of one another, but overlap and intersect in extremely complex and dynamic ways, this would require a further, “grand” or “global” theory, one that explains the relationships between various issue-areas, as well as the underlying dynamics generating shifts in hierarchies over time and place. Such a theory, of course, would almost certainly be wildly unparsimonious. In this regard, Waltz may be mistaken on the question of fungibility, but correct in recognizing that the proposed alternatives undermine neorealism’s own methodological canons. A number of these points are discussed in greater detail in Stanley Hoffman (1978: 182–190 and chapter 3). On the less than rosy prospects for genuine democracy in the near future, see McFaul et al. (1994). On environmental conditions, see Murral Feshbach and Alfred Friendly, Jr. (1992). Though I shall not discuss the issue, Foucault’s “relational” and “dispersed” conception of power challenges a further assumption common to much neorealist thought, namely, the assumption that power is connected to territory. Indeed, a major weakness of neorealism lies in its “territorial” conception of power, one that fits well with the neorealist accent on military power, but prevents an adequate theorization of the implications and consequences of globalization.

References Ashley, Richard. (1984). The Poverty of Neorealism. International Organization. 38 (Spring): 225–286. Barnes, Barry. (1977). Interests and the Growth of Knowledge. London: Routledge and Kegan Paul. Barnes, Barry and David Bloor. (1982). Relativism, Rationalism, and the Sociology of Knowledge. In Martin Hollis and Steven Lukes (Eds.), Rationality and Relativism. Cambridge: MIT Press. Bhaskar, Roy. (1979). The Possibility of Naturalism. Brighton, Sussex: Harvestor Press. Bhaskar, Roy. (1986). Scientific Realism and Human Emancipation. London: Verso. Blank, Stephen. (1978). The Politics of Foreign Economic Policy, the Domestic Economy, and the Problem of Pluralist Stagnation. In Peter J. Katzenstein (Ed.), Between Power and Plenty: Foreign Economic Policies of Advanced Industrial States. Madison, WI.: University of Wisconsin Press.

184

KEITH TOPPER

Bloor, David. (1976). Knowledge and Social Imagery. London: Routledge and Kegan Paul. Bloor, David. (1981). The Strengths of the Strong Programme. Philosophy of the Social Sciences 11 (June): 199–213. Bourdieu, Pierre. (1977). Outline of a Theory of Practice. Cambridge: Cambridge University Press. Bourdieu, Pierre. (1984). Distinction. Cambridge, MA: Harvard University Press. Bourdieu, Pierre. (1985). The Social Space and the Genesis of Groups. Theory and Society 14 (November): 723–744. Bourdieu, Pierre. (1989). Social Space and Symbolic Power. Sociological Theory 7 (Spring): 14–25. Buzan, Barry and R.J. Barry Jones (Eds.). (1981). Change and the Study of International Relations: The Evaded Dimension. London: Frances Pinter Publishers. Crick, Bernard. (1521/1970). Introduction. In Niccol`o Machiavelli, The Discourses. New York: Viking Penguin. Davidson, Donald. (1982). Essays on Actions and Events. Oxford: Clarendon Press. Dreyfus, Hubert L. (1980). Holism and Hermeneutics. Review of Metaphysics 34 (September): 3–23. Dreyfus, Hubert L. (1986). Why Studies of Human Capacities Modeled on Ideal Natural Science Can Never Achieve Their Goal. In J. Margolis, M. Krausz, and R.M. Burian (Eds.), Rationality, Relativism and the Human Sciences. Dordrecht, Boston and Lancaster: Martinus Nijhoff Publishers. Dreyfus, Hubert L. and Paul Rabinow. (1983). Michel Foucault: Beyond Structuralism and Hermeneutics. Chicago: University of Chicago Press. Feshbach, Murral and Alfred Friendly, Jr. (1992). Ecocide in the USSR: Health and Nature Under Siege. New York: Basic Books. Feyerabend, Paul. (1975). Against Method. London: Verso. Foucault, Michel. (1979). Discipline and Punish. New York: Vintage. Foucault, Michel. (1980). The History of Sexuality, Volume 1: An Introduction. New York: Vintage. Foucault, Michel. (1983). The Subject and Power. In Hubert L. Dreyfus and Paul Rabinow, Michel Foucault: Beyond Structuralism and Hermeneutics. 2nd ed. Chicago: University of Chicago Press. Garst, Daniel. (1984). Thucydides and Political Realism. Paper prepared for the American Political Science Association meeting in Washington, DC. Gilpin, Robert. (1975). U.S. Power and the Multinational Corporation. New York: Basic Books. Gilpin, Robert. (1981). War and Change in World Politics. Cambridge: Cambridge University Press. Gilpin, Robert. (1984). The Richness of the Tradition of Political Realism. International Organization 38 (Spring): 287–304. Gombrich, Ernst H. (1969). Art and Illusion. Princeton: Princeton University Press. Greybeal, Jean. (1990). Language and the “Feminine” in Nietzsche and Heidegger. Bloomington: Indiana University Press. Habermas, J¨urgen. (1971). Knowledge and Human Interests. Boston: Beacon Press. Hacking, Ian. (1983). Representing and Intervening. New York: Cambridge University Press. Hanson, Norwood Russell. (1965). Patterns of Discovery. Cambridge: Cambridge University Press. Hartz, Louis. (1955). The Liberal Tradition in America. New York: Harcourt Brace Jovanovich. Hesse, Mary. (1980). Revolutions and Reconstructions in the Philosophy of Science. Bloomington and London: Indiana University Press. Hoffman, Stanley. (1978). Primacy or World Order. New York: McGraw Hill. Holsti, Ole, Randolph Siverson, and Alexander L. George (Eds.). (1980). Change in the International System. Boulder, Colo.: Westview Press.

AN ANALYSIS OF NEOREALIST THEORY

185

Irigaray, Luce. (1991). Marine Lover of Friedrich Nietzsche. New York: Columbia University Press. Kant, Immanuel. (1797/1964). Groundwork of the Metaphysic of Morals. New York: Harper and Row. Kaplan, Morton. (1979). Towards Professionalism in International Theory. New York: Macmillan. Keohane, Robert. (1984). Beyond Hegemony. Princeton: Princeton University Press. Keohane, Robert O. (1986). Theory of World Politics: Structural Realism and Beyond. In Robert O. Keohane (Ed.), Neorealism and Its Critics. New York: Columbia University Press. Krasner, Stephen. (1978). Defending the National Interest. Princeton: Princeton University Press. Kratochwil, Friedrich. (1984). Errors Have Their Advantage. International Organization 38 (Spring): 305–320. Kuhn, Thomas. (1970). The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lake, David. (1983). International Economic Structures and American Foreign Policy, 1887– 1934. World Politics 34 (July): 517–543. Lake, David. (1984). Beneath the Commerce of Nations: A Theory of International Economic Structures. International Studies Quarterly 28 (June): 141–168. Machiavelli, Niccol`o. (1532/1950). The Prince and the Discourses. New York: Modern Library. Mill, John Stuart. (1872/1988). The Logic of the Moral Sciences. La Salle, IL: Open Court Press. Mulkay, Michael. (1979). Science and the Sociology of Knowledge. London: George Allen and Unwin. Nietzsche, Friedrich. (1886/1966). Beyond Good and Evil. New York: Vintage. Patton, Paul, (Ed). (1993). Nietzsche, Feminism, and Political Theory. New York: Routledge. Pitkin, Hanna Fenichel. (1984). Fortune Is a Woman. Berkeley and Los Angeles: University of California Press. Popper, Karl. (1968). The Logic of Scientific Discovery. New York: Harper Torchbook. Ricoeur, Paul. (1974). Structure, Word, Event. In Paul Ricoeur, The Conflict of Interpretations. Evanston, IL: Northwestern University Press. Rorty, Richard. (1979). Philosophy and the Mirror of Nature. Princeton: Princeton University Press. Rorty, Richard. (1980). A Reply to Dreyfus and Taylor. Review of Metaphysics 34 (September): 39–46. Rorty, Richard. (1982). Method, Social Science, and Social Hope. In Richard Rorty, Consequences of Pragmatism. Minneapolis: University of Minnesota Press. Rorty, Richard. (1988). Is Natural Science a Natural Kind? In Ernan McMullin (Ed.), Construction and Constraint. Notre Dame, IN: University of Notre Dame Press. Rosecrance, Richard. (1981). International Theory Revisited. International Organization 35 (Autumn): 691–714. Ruggie, John. (1980). Review: Defending the National Interest by Stephen Krasner. American Political Science Review 74 (March): 298. Ruggie, John. (1984). Continuity and Transformation in the World Polity: Toward a Neorealist Synthesis. World Politics 35 (Spring): 261–285. Runciman, William. (1978). What is Structuralism? In Alan Ryan (Ed.), The Philosophy of Social Explanation. Oxford: Oxford University Press. Sahlins, Marshall. (1981). Historical Metaphors and Mythical Realities. Ann Arbor: University of Michigan Press. Sahlins, Marshall. (1982). The Apotheosis of Captain Cook. In Michel Izard and Pierre Smith (Eds.), Between Belief and Transgression. Chicago: University of Chicago Press. Searle, John. (1984). Minds, Brains, and Science. Cambridge, MA.: Harvard University Press.

186

KEITH TOPPER

Shapin, Steven. (1982). The History of Science and Its Rational Reconstruction. History of Science 20 (September): 157–211. Symposium. (1994). Is Russian Democracy Doomed? Journal of Democracy 5 (April): 3–41. Taylor, Charles. (1979). Interpretation and the Sciences of Man. In Paul Rabinow and William Sullivan (Eds.), Interpretive Social Science: A Reader. Berkeley and Los Angeles: University of California Press. Topper, Keith. (1994). Sciences of Uncertainty: Perspectives on Naturalism, Politics and Power. Ph.D. diss. University of California, Los Angeles. Vico, Giambattista. (1725/1961). The New Science of Giambattista Vico. Garden City, N.Y.: Anchor Books. Waltz, Kenneth. (1979). Theory of International Politics. Reading, Massachussetts: AddisonWesley. Waltz, Kenneth. (1986). Reflections on Theory of International Politics: A Response to My Critics. In Robert O. Keohane (Ed.), Neorealism and Its Critics. New York: Columbia University Press.