Wendt, Alexander (2003): Social Theory of International Politics. Cambridge: Cambridge University Press, s. 113 -- 138. (26 s.) The constitution of interests by ideas If an emphasis on the role of power is usually seen as one of the defining features of Realism, then an emphasis on egoistic national interest would be the other. Realists of all stripes believe that states do what they do because it is in their national interest, and that the national interest is self-regarding with respect to security. As with power, however, these cannot be uniquely Realist claims, since then almost every IR scholar would be a Realist. No one denies that states act on the basis of perceived interests, and few would deny that those interests are often egoistic. I certainly do not. To that extent I am Realist, but interests should not be seen as an exclusively "Realist" variable. What matters is how interests are thought to be contstituted. As I see it, the uniquely Realist hypothesis about national interests is that they have a material rather than social basis, being rooted in some combination of human nature, anarchy, and/or brute material capabilities. The argument in the preceding section was largely agnostic about this question. It acknowledged that material forces constrain and enable social forms at the margin, but its primary claim was that the distribution of interests helps constitute the meaning of power. Nevertheless it is widely thought in IR that power and interest are both ''material'' and therefore that the only way to challenge theories which emphasize them, like Realism, is to show that factors like ideas, norms, or institutions explain a lot of behavior. This has been the intuition behind Neoliberalism, which frames the explanatory problem as power and interest versus institutions versus norms, versus ideas. This framing has been fruitful, since there is much in international politics that power and interest cannot explain. On the other hand, this view implicitly suggests that power and interest are not themselves constituted by ideas. And since Realists have already claimed power and interest as ''their'' variables, this limits a priori the role of ideas - and thus non-Realist theories - to the superstructure, and thereby privileges Realist arguments about the base. Neoliberalism focuses on the ways in which ideas can have causal effects independent of other causes like power and interest. However, ideas also have constitutive effects, on power and interest themselves. Here I discuss how ideas constitute interests. If in some sense interests are ideas, then the causal, ''ideas-versus-interests'' model will be incomplete. This does not mean that all ideas are interests. Most are not. Nor does it mean that interests no longer have an independent explanatory role. They explain just as much as they did before, and exist independently of ideas that do not constitute them, as required by causal explanations. The claim is only that among the different kinds of ideas are some that constitute interests, and that the explanatory power of these ideas therefore cannot be compared to interests as interests should not be seen competing causal variables. To say interests are ideas brings us again to the definition of materialism. I argued above that meaningful power is constituted in important part through the distribution of interests. Here I argue that only a small part of what constitutes interests is actually material. The material force constituting interests is human nature. The rest is ideational: schemas and deliberations that are in turn constituted by shared ideas or culture. As in my discussion of power-explanations, in other words, my goal here is not to show that interests do not matter, but to show how little of them a properly specified materialism can explain, and to claim the rest for idealism. Rational choice theory is the conventional framework in main- stream IR for thinking about the relationship between ideas and interests. For that reason I shall organize my discussion with reference to it. The core of rationalist explanations is the view that preferences and expectations generate behavior. This is known in philosophical literature as the equation, ''desire plus belief equals action" It is not hard to see how this equation might encourage the interests ''versus'' ideas thinking that I'm arguing is problematic, and as such play into the materialist bias in IR theory. Rationalism treats desire (or preference or interest) and belief (or expectations or ideas) as distinct variables, which suggests that desires do not depend on beliefs and are therefore material. This connotation is further enabled by the fact that rationalists do not usually ask where interests come from. It is in this way that methodology can become tacit ontology. By the same token, however, strictly speaking the theory is agnostic about that question. Interests might be material or ideational; it simply does not say. Moreover, rationalism has a strong subjectivist aspect, which has led some people to emphasize its affinities to interpretive social science and thus, implicitly, an idealist ontology. These considerations suggest rational choice theory might be compatible with an idealist view of interests. Thus, in what follows I shall not be arguing ''against'' rational choice theory (nor, it might be noted, will I bring up some familiar, long-standing criticisms, such as about the theory's realism); on the contrary - I see it as part of my own understanding of agency (see chapter 7). But it is only part of the story and as such must be assimilated into a constructively framework. In what follows I first discuss the standard rationalist view of the relationship between interests and ideas, and then propose an alternative. The rationalist model of man Rationalism has both a macro- and micro-dimension. The macro- dimension is concerned with explaining broad patterns of behavior and aggregate outcomes rather than the behavior of individual agents. Often the patterns and outcomes arise via unintended consequences of behavior. What matters here are the structural constraints on choice rather than individual psychology, since the same aggregate outcome may be realizable under various psychological condition. While this might suggest that rational choice theory does not depend on assumptions about agents, in fact it does. Even if a macro-outcome is compatible with a variety of desires and beliefs, rationalist explanations presuppose that agents act at least ''as if '' they are maximizing certain desires and beliefs (see below). The macro level is important and it relates to arguments about the role of culture in constituting interests that I develop in chapter 4, but since my concern in this chapter is only with the nature of interests, I shall limit my discussion here to its micro-aspect, focusing on the logic of desire/belief explanation and the assumptions about human agency which it makes. To explain action as a product of desire and belief is to offer an '' intentional" esplanation. This is the kind of explanation most of us would intuitively give if asked to explain why we went to the grocery store: we had a desire for food and a belief that that desire could be satisfied there. This combination of desire and belief was the ''reason'' we went to the store, and in the intentionalist view reasons are causes of behavior. In effect, the intentional theory of action is a dressed up version of the folk psychology implicit in our everyday explanations of behavior. In the social sciences it has received its most systematic use in economics, however, and is now often seen as the core of an ''economic'' approach to human behavior, from where it has been colonizing other social sciences. Alexander Rosenberg offers a good summary: Economics is an intentional science. It holds that economic behavior is determined by tastes and beliefs, that is, by the desire to maximize preferences, subject to the constraint of expectations about available alternatives. Differences between the choices made by individual agents who face the same alternatives are due either to differences in preferences, to differences in expectations, or to both. Similarly, changes in the choices of an individual agent over time are due to changes in one or both of these causal determinants of his behavior. It is important to note that this explanatory logic says nothing about the content of desires and beliefs. This can be seen by distinguishing ''thin'' and ''thick'' versions of rational choice theory. The thin theory consists of propositions about the nature of desire and belief and their relationship - in short, intentional explanation as such. In the intentional theory of action the concept of desire refers to a motivation that moves the body in the direction of the object of desire. Desire is always for something, and as such plays an active explanatory role in the sense that it is the force or energy which moves the body. This force is activated only if an actor also believes the object of desire can be attained by acting, and so desire by itself is not sufficient to explain action, but given appropriate beliefs the energy for activity comes from desire. Belief plays a more passive explanatory role in the thin theory. Whereas desire is for things, belief is about them. Two kinds of beliefs are important: beliefs about states of the external world, and beliefs about the efficacy of different means to satisfy desires in that world. It does not matter whether these beliefs are accurate, only that actors take them to be true. A key assumption of the traditional rationalist model is that beliefs have no motivational force of their own; they merely describe the world. This creates within the model an explanatory bias in favor of desire/interest, which is deeply rooted in the intellectual history of rationalism, going back to Hobbes and Hume. Beliefs play an important enabling role in behavior by activating and facilitating the realization of desires, but the active, primary explanatory work is done by desire. Thick versions of rational choice theory add to this skeleton assumptions about the content of desires and beliefs. One of the most common thick theories is that actors are egoists with complete information about their environment, but thick rationalist theories could alternatively assume altruism and incomplete information. There is no one thick theory of rational choice, and so we need more than the the theory. Many disagreements in IR scholarship are rooted in different thick theories of human nature and/or the national interests Classical Realists offer varying permutations of fear, power, glory, and wealth as candidates. The debate in Neorealism about whether states are status quo or revisionist is in part about whether they are motivated more by fear or power. The debate between Neorealists and Neoliberals about the extent to which states seek relative or absolute gains is in part about whether states are more interested in security or wealth. The question of whether states are capable of collective security depends on whether they are necessarily selfish or capable of having collective interests. And so on. These are important disagreements, but all sides seem to accept the key rationalist premise that desire (the national interest) causes states to act in certain ways. The intentional equation is also a common baseline in recet IR work on beliefs.. One stream of scholarship has focused on the belief systems and perceptions of decision-makers. This work presents a challenge to thick rationalist theories that assume complete information, but it does not threaten the thin theory. And there is also recent rationalist work on the role of ideas in foreign policy. Goldstein and Keohane actually contrast this work to the ''rationalist'' concern with interested but it should be clear from the foregoing discussion that beliefs play an essential role in rationalist theory. In the past rationalist scholars may have neglected belief in favor of desire (usually by assuming that actors have complete information), which encouraged the view that rational choice theory is a materialist theory. Goldstein and Keohane have issued an important reminder that it need not be seen this way. But in itself a focus on ideas poses no inherent threat to rational choice theory's explanatory logic. Most of the recent mainstream IR scholarship on ideas is clearly based on an intentional theory of action: treating desire and belief as if they were distinct, with the latter relating to the former in instrumental rather than constitutive terms. Of course, to some extent desire and belief are distinct phenomena. Desire is ''for,'' belief ''about.'' The one is motivation, the other cognition. An interesting way to think about the difference is that they have different ''directions of fit'' with the world's Desire aims to fit the world to the mind, belief aims to fit the mind to the world. However, this difference does not rule out the possibility that desire may itself be a kind of belief - a belief not about the world, but a belief that something is desirable. I explore below the possibility that cognitive factors constitute desire. This raises the crucial question of ''what is desire (interest)". The received view, going back at least to Hume, is that desire is constitutionally unrelated to belief. Desire is a matter of passion, not cognition; and while beliefs activate and channel desires, they cannot be desires. Hume's view is ''dualistic'' in that it explains action by reference to two unrelated mechanisms. This view has two important theoretical consequences. First, if desires are not a function of belief, then it is natural to treat them in materialist fashion as material, and to treat ideas in rationalist fashion as a means for realizing exogenously given interests. Second, the Humean view also makes life difficult for the constructivist, because her point is that culture (a shared idea) constitutes interests. If interests and ideas are entirely different kinds of stuff, then it is not clear how they can mix and transmogrify one (mind) into another (body). Constructivism needs to overcome the Humean dualism of desire and belief. It can do so with an alternative, cognitive theory of desire. Simply put, we want what we want because of how we think about it. As we shall see, this need not vitiate intentional explanation, but it does suggest that there is more to the relationship between desire and belief than rationalism acknowledges. Beyond the rationalist model The Humean view that desire and belief are constitutionally unrelated is deeply embedded in rationalist discourse. It appeals to important intuitions in our everyday understandings of behavior, and the structure of intentional explanation (desire plus belief) tacitly connotes it. On the other hand, there is a growing body of scholarship in philosophy, cognitive psychology, anthropology and even economics which argues that desire is not separate from belief but constituted by it. This literature too appeals to important intuitions in everyday life. I discuss two different but related versions of this thesis, cognitive and deliberative. Judging from citations their advocates seem unaware of each other, and one seems to pose a deeper challenge to the traditional theory of intentional action than the other. But rather than assess their relationship, at this stage it seems more useful simply to present the two accounts and show how each links ideas to interests. An important premise of the argument I make here is that we should care about how preferences are constituted. The premise comes from scientific realism and many rational choice scholars might disagree with it. For them, as for the empiricist anti-realists I discussed in chapter 2, ''as if '' assumptions about preferences are sufficient for theorizing. A sophisticated version of this argument is advanced by Debra Satz and John Ferejohn and it merits a response. Satz and Ferejohn argue that rationalist explanations do not need to show that agents ''really'' are motivated by desires and beliefs just that they act ''as if '' they are. If this is right, then the issue of what desires are made of is without substantive import, beside the point. Satz and Ferejolm are expressing a consensus among contemporary economists on an old debate about whether their discipline needs robust psychological assumptions about ''utility.'' In the nineteenth century most economists thought it did. Systematized by Stanley Jevons, this view can be traced back to Bentham, who argued that utility was constituted by experiences, and before that to Hobbes and Hume who argued that ''passions'' were the source of desire. Beginning with seminal work by Paul Samuelson in the 1930s, however, economists have today largely abandoned this ''internalist'' view (''internal'' because it referred to states of consciousness), because of its intractability, unrealistic psychology, and, importantly, appeal to unobservable causes. Like behaviorists in psychology, rational choice theorists now take an ''externalist'' view, which treats desire in behavioral or operational terms as choice (revealed preferences) rather than as an unobservable cause of choice. This is legitimate, Satz and Ferejohn argue, because in rationalist theory what explains outcomes are the structural constraints in a system, which will often have the same effects regardless of individual motivations (back to the macro-level aspect of rational choice above). The result is an instrumentalist reading of rationalism, in which no assumptions are made about the ontological status of desire and belief. In a sense we are back to the epistemological anxiety discussed in chapter 2, which leads to a focus on what we can see and measure. In a response to Satz and Ferejolm, Daniel Hausman defends the necessity of an internalist view of action. On the grounds that even if the structure of a choice situation is highly constraining (as in a hotel fire), our explanation of the outcome (the occupants flee) depends on the accuracy of our assumptions about desires and beliefs. In the hotel fire example these assumptions are trivial (most people want to live and know that fire can kill them), and as such little will be gained by devoting much energy to refining them. But it remains the case that ''the correctness of the explanation depends on their truth. An adequate externalist story depends on an adequate internalist one. Otherwise it is a mystery why the occupants flee, and we should want to know why. One reason is practical: structural theories that make false motivational assumptions may sometimes successfully predict outcomes, but if we ignore their falsity we will not know when they might fail us or how to revise them most efficiently. From this standpoint, encouraging social scientists to ignore the truth of their assumptions is ''bad methodological advice". Another reason we should care about motivation is philosophical: unlike the instrumentalism espoused by Satz and Ferejohn, in which the goal of science should be merely to ''save the appearances,'' Hausman is a scientific realist who thinks that science should try to describe the causal mechanisms that generate appearances, and so we must care whether the psychological claims employed in rational-choice explanations are true. Scientific realists about rational-choice theory must be internalists. Social scientists do not always need to worry about the truth of their assumptions, but the question of how desire is constituted is not something that should be side-stepped completely. The cognitive basis of desire The first argument against a materialist view of interest is that interests are themselves cognitions or ideas. We find this thesis in two distinct bodies of scholarship, one in cultural anthropology, the other in philosophy. Drawing on cognitive psychology, anthropologist Roy D'Andrade argues that motivations, desires, or interests should be seen as ''schemas'' (or ''scripts '' ''frames" or ''representations'') which are knowledge structures that ''make possible the identification of objects and events.'' Many schemas are simply beliefs about the world that have no connection to desires. Other schemas are goals or desires that energize action. D'Andrade (p. 35) gives the example of a motivation for ''achievement.'' Achievement implies a social standard about what counts as a legitimate aspiration - and as such is a cultural rather than material fact. Individuals who have a desire to achieve have internalized this standard as a cognitive schema. Similarly, in capitalist societies some people have a desire to get rich on the stock market. This is a schema which includes beliefs about the external world (how the market works, where it is going, etc,), and also constitutes its holder with a particular motivation that drives her behavior in that world. Symbolic interactionists would argue that many of these goal-schemas or interests are constituted by identities, which are schemas about the Se1f. The identity or self- schema of professor, for example, constitutes an interest in teaching and publishing. Like other schemas, motivational schemas are organized hierarchically within the Self and so not all equally ''salient," which is important in trying to explain what someone will do in a particular situation. The important point is that none of these schemas is given by human nature. D'Andrade is careful to acknowledge that motivation is partly rooted in biological drives and as such is truly material. Sometimes, as in the example of fleeing the hotel fire, these are more important in explaining action than culturally constituted schemas. But biological drives explain few of the almost infinite goals human beings seem to be capable of pursuing. Most of these are learned through socialization. Those who would explain how desire is constituted, therefore, would do well to focus more on culture and its relationship to cognition than on biology. Much the same conclusion is reached without much reference to cognitive psychology by R. Howe, who uses recent philosophical discussions to articulate a cognitive theory of desire. Like DīAndrade Howe acknowledges a role for biological drives in the constitution of desire. Needs for food, water, reproduction and so on matter, and these are material. Yet Howe argues that even very primitive desires are mostly ''directionless", and depend on beliefs about what is desirable to give them content. Beliefs define and direct material needs. It is the perception of value in an object that constitutes the motive to pursue it, not some intrinsic biological imperative. Such perceptions are learned, partly through interaction with nature (fire hurts; dirt tastes bad), in which case they have a materialist explanation, but mostly they are learned through socialization to culture. Desires always involve a mixture of biological drives and beliefs, with the importance of beliefs varying along a continuum from low (a desire for water when thirsty) to high (a desire to do the right thing) These desire-constituting beliefs or cognition have a different ''direction of fit'' with the world than the beliefs-''about'' which figure on the belief side of the desire plus belief equation. To highlight their distinctiveness philosophers have dubbed them "desiderative belief". ''Goal-schema'' would do just as well. The arguments of D'Andrade, Howe, and others concerned with the relationship between desire and belief refer mostly to individuals rather than groups. I argue in chapter 5 that certain groups, including states, also have desires. This is an assumption of all state-centric IR theory, and one virtue of the cognitive approach to interests is that it is easier to defend this assumption than it is to defend it with a materialist approach, since states are not biological beings. Assuming for the moment that states have desires, let me illustrate the argument in this section with reference to the three state interests that figured in the earlier discussion of the distribution of power: status quo, revisionist, and collectivist. A status quo state is one that has no interest in conquering other states, redrawing boundaries, or changing the rules of the international system. It may attack another state to preempt a threat, but it has no intrinsic desire to infringe on other states' rights. How is this interest constituted? Undoubtedly part of the answer lies in basic material human needs for security and stability, but since all states are presumably subject to these needs and not all have status quo interests, this does not tell us enough. The cognitive theory of desire directs our attention to the schemas or representations through which status quo states define their interest. They may be hypothesized to have schemas as ''satisfied'' with their international position, as ''law-abiding'' as ''members of a society of states'' the rules of which are seen as ''legitimate'' and so on. These beliefs are not merely about an external world: they also constitute a certain identity and its relationship to that world, which in turn motivates action in certain directions. Status quo states have the interests they do, in other words, in virtue of their perceptions of the international order and their place within it as desirable, not because of brute material facts. Revisionist states, in turn, have the desire to conquer others, seize part of their territory, and/or change the rules of the game. Human nature helps constitute these desires too, most likely in the form of self-esteem needs, but again this explains little. More significant will be self-schemas like ''victim'' or ''master race," representations of Others as ''infidels'' or ''evil empires" of the system as ''illegitimate'' ''threatening''' war as '' glorious" or ''manly '' and so on. These schemas are a function of culturally constituted cognitives, not biology. Collectivist states have the desire to help those they identify with even when their own security is not directly threatened. Realist cynicism notwithstanding, biology surely plays a role here as well, since humans are social animals whose brains are hard-wired for ''team play,'' but this cannot explain why some states identify and some do not. The presence of certain schemas can: ''we-ness'' ''friend '' ''special relationship'' ''doing the right thing'' ''regional policeman,'' and so on. In foreign policy discourse these ''moral'' schemas are often juxtaposed to ''interests''' as in the debate about US intervention in the Bosnian civil war. One way to interpret President Clinton's speech to the American people justifying intervention is that it tried to define US ''interests'' in terms of the belief that Americans are the kind of people who do the right thing. In chapters 4 and 6 I will argue that these interest-constituting ideas are in turn constituted by the shared ideas or culture of the international system. Here I am arguing that ideas at that macro level get into the heads of states and become interests at this other, more micro level of international structure. The cognitive theory of desire violates the spirit but not the letter of the intentional theory of action. The traditional interpretation of intentionalism, following Hume, ruled out the hypothesis that beliefs could motivate, but nothing in the theory's propositional structure (the thin theory of rational choice) requires such an interpretation. It is perfectly consistent with the idea that beliefs and desires are distinct to hold that certain beliefs are about the external world and other beliefs constitute desires, and that the two play different explanatory roles. Desires are no less desires for being constituted by beliefs. As such, nothing said so far is inherently incompatible with rational choice theory, as long as rationalists concede that ideas play a larger role in explaining social action than is captured by the desire ''plus'' belief model. The resulting opening has been exploited by some rationalists in economics, who have modeled preferences as constituted by beliefs, and others in IR, who have argued that state interests are affected by expectations about the environment. Precisely because it is agnostic about what preferences are and where they come from, rational choice theory can be adapted to either an idealist or a materialist ontology. The deliberative basis of desire Cognitivism challenges the materialist view of desire, but it does not call into question the key assumption of the intentional theory, that desire and belief alone explain action. Desire still does all the motivational work, even if it has been reconceptualized as a kind of belief. An alternate argument for what explains action brings in reason or deliberation. Martin Hollis and G.F. Schueler, drawing from Kant, argue that Reason or deliberation should be considered a third factor in the model: desire plus belief plus reason equals action. The rationale for looking to a third factor stems from rational choice theory's paradoxically impoverished conception of ''rational choice". Rationality is normally defined in instrumental terms as nothing more than having consistent desires and beliefs, and choice involves nothing more profound than their automatic enactment in behavior that maximizes expected utility. Rationalists rarely ask whether preferences are rational in the sense of justifiable, and often specifically abjure such assessments. ''Rationality of action is always relative to the current desires of the agent," whatever their content. In this light humans differ from other animals only in the greater complexity of their desires and beliefs, not in their rationality. And indeed, experiments have shown that humans, rats, and pigeons are equally rational as defined by rational choice theory. What is missing from this conception of rationality is any sense of deliberation, which goes back to the Humean model of man. In that model deliberation involves nothing more complicated than weighing up one's desires on a '' grocerīs scale,'' or doing a ''vector analysis" of their relative strength. There is no sense in the Humean model of Reason as a distinct faculty of mind that decides what desires to have, which to act upon, or even whether to act at all. The perhaps surprising result, therefore, is that rational choice theory is highly deterministic. This is seen in the many metaphors which its critics have coined to describe it. Schueler calls it the ''blind forces'' model of intentionality, in which agents (now rather mixing metaphors) are pushed and pulled by desire ''rather in the way currents of air act on a falling leaf ''; Doris prefers the electronic imagery of agents as ''throughputs'' for desires and beliefs; Margaret Gilbert offers the mechanical metaphor of desire causing choice in ''hydraulic'' fashion; Harry Frankfurt calls people who do not reflect on their desires ''wantons''; Amartya Sen calls them ''rational fools.'' For rhetorical punch none tops Hume who argued that Reason ''alone can never be a motive to any action of the will and ''is and ought only to be the slave of the passions." But all point to the fact that his model of man lacks the free, deliberating agent which one intuitively associates with ''rational choice". Indeed, whereas rational choice seems to be nothing more than a formalization of folk psychology - and on one level it is - on a closer read it is also somewhat out of sync with our common sense under- standings of how and why people act. For example, the assumption that human beings do not reflect upon and choose their desires is hard to square with our intuitions about responsibility. If we are merely throughputs for desires and beliefs that we cannot control (since we are nothing but them), then how can we be held responsible for our actions? The reason we do not blame animals for their behavior is because we assume they lack the capacity for deliberation about their desires which would enable them to act differently than they do. Yet as we saw above in rational choice theory humans and animals are equally rational. Another problematic intuition is that people often engage in practices of delayed gratification, ''self-binding,'' and ''character planning'' which involve acting on behalf of desires they do not yet have. Rationalists may try to explain such behavior by introducing discounted future desires into the present, but this still raises the possibility of Reason shaping desire, which contradicts the Humean view. Finally, the desire/belief model ignores the sense in ordinary language that people can act against or in spite of their desires, that we can do something even though we ''wanted'' to do something else. Human beings are often deeply torn about whether to act on their desires, and sometimes restrain themselves because of Reason or morality. ''External'' rather than ''internal'' reasons sometimes prevail. Rationalists may try to explain such behavior as resolving a conflict between lower desires (e.g., be selfish) and higher desires (e.g., do the right thing), such that whatever an agent decides to do must have been what she really ''wanted'' to do: either lower or higher desires simply won out. But Schueler argues such an explanation conflates two senses of desire: ''proper desires" which are in the head and can be acted against, and ''pro attitudes" which are the actual choices agents make. The distinction matters because pro attitudes are known through choices, not before, and as such cannot enter into an agent's own calculus about what to do. Reducing all deliberation to a weighing of conflicting desires, in other words, is a non-falsifiable proposition that cannot explain behavior. The desires that can truly explain behavior are proper desires, and in order to know how proper desires affect choices we need to bring in deliberation. These intuitions all call into question the two-factor model of intentional action, but like the cognitivist argument, they can be made consistent with rational choice theory, if we detach it from its Humean moorings and view it as only a partial theory of action. In fact these intuitions suggest the fruitfulness of distinguishing two versions of intentional explanation, which Schueler calls the ''blind forces'' and ''reflective" models. The former, corresponding to the traditional Humean view, treats human agency as ''impulsive'' and lacking meaningful deliberation. The latter, corresponding to a Kantian view, treats Reason as a third factor that deliberates about and helps choose interests. While the blind forces model characterized rational choice scholarship for some time, rationalist social theory today is developing and strengthening its notions of deliberation and selfgovernance. Schueler sees an ''enormous difference'' between the two models (p. 186), but argues that the best description of a choice process in a given context, blind versus reflective, is always an empirical question. Moreover, since deliberation is a learned capacity, the balance between them for a given agent may change over time. The addition of Reason to rational choice theory seems particularly apposite for IR scholarship. The philosophical literature on deliberative rationality concentrates on individuals. A strong case exists even in that context against the traditional, two-factor model of intentionality. But an emphasis on the role of deliberation in constituting interests seems even more appropriate for decision-making in groups. Often one of the most difficult tasks facing foreign policy decision-makers is figuring out what their interests are. This process does not typically consist of weighing competing interests on a ''grocer's scale'' of intensity, or even of aggregating the exogenously given preferences of different individuals. It typically consists in a complex and highly contested process of discussion, persuasion, and framing of issues. In short, what goes on is collective deliberation about what their interests in a given situation should be. These deliberations do not take place in a vacuum, either domestic or international, but neither are they strictly determined by domestic or systemic structures. There are relatively few ''hotel fires'' in international politics. And sometimes deliberation can generate dramatic ''preference reversals'' even while structural conditions remain constant. Such was arguably the case with Soviet New Thinking under Gorbachev. Those wedded to the blind forces model of intentional action will say that the Soviet leadership had to change its policies because of its declining relative power position. Certainly the economic and military pressures on the Soviet state were a crucial impetus for change. However, a structural pressure theory alone cannot explain the form the Soviet response took (ending the Cold War rather than intensifying repression) or its timing (the material decline had been going on for some time). And it also ignores the role that the leadership's realization that its own policies were part of the problem played in conditioning that response. Structural conditions did not force self-awareness on the Soviets. Soviet behavior changed because they redefined their interests as a result of having looked at their existing desires and beliefs self-critically. The reflective model of intentional explanation captures this process more naturally than the blind forces model. This example also points to ways in which the cognitive and deliberative arguments may overlap. The principles informing Soviet ''Reason'' were not wholly independent of beliefs about the identity of the Soviet state, the feasibility of certain actions, and even about right and wrong. Deliberation about national interests takes place against the background of a shared national security discourse, in other words, which may substantially affect its content. This blurring of Reason and belief is also evident in the philosophical literature. Howe, who does not make the Kantian argument that Reason is a distinct factor in intentional explanations, treats morality as a belief or schema. Schueler, who does make the Kantian argument, places moral considerations under the heading of Reason. My own inclinations lie with Schueler because the cognitive theory alone, with its continued reliance on just desire and belief to explain action, does not escape the determinism of rational choice theory. But the relationship between the two idealist critiques of materialist theories of desire is complicated and need not concern us here. Toward a rump materialism II The overlap between the cognitive and deliberative critiques suggests a general proposition about the relationship between interests and ideas: ''interests are beliefs about how to meet needs.'' Since this depends on a distinction between interests and needs, let me first say a few words about the latter and then return to interests. As in my concluding remarks about power, having now taken the idealist line that interests are constituted mostly by ideas, in this section I turn around and defend the rump materialist view that they nevertheless must ultimately hook on to a material ground, human nature. Needs refer to the functional reproduction requirements of a particular kind of agent, what some would call ''objective interests.'' Two types of needs may be discerned: identity needs and material needs. Identity needs are as variable as the identities they sustain, which is to say practically infinite. To reproduce the identity of a state a group needs to sustain a monopoly on the legitimate use of violence in their territory. To reproduce the identity of a professor an individual needs to teach. In both cases these needs reflect the internal and external structures that constitute these actors as social kinds. There is no guarantee identity needs will be translated into appropriate beliefs about how to meet them, which is to say into (subjective) interests, but if they are not translated then the agents they constitute will not survive. Identity needs are ultimately a matter of individual and social coalitions rather than biology. They are still real and objective, but given that they are not material to focus on them here would do little to clarify the role of materialism. So let me turn to the material needs stemming from human nature and show just what exactly is a material basis for desire. Scientific realism assumes that human beings are self-organizing natural kinds with material reproduction requirements. All animals have such requirements. Material needs are no guarantee that individuals will try to meet them (people do commit suicide), but it seems likely that were humans not predisposed to meet their needs we would never have survived evolution. The content of this predisposition is ''human nature". Radical constructivists might deny the existence, or at least social significance, of biological needs. But despite its well-intentioned resistance to biological determinism, there is an anthropic exceptionalism or human chauvinism in the radical view that is hard to justify from the standpoint of evolutionary theory. It is impossible to explain social action without making at least implicit assumptions about human nature, since, without it, it is hard to explain why our bodies move at all, let alone their direction or resistance to societal pressures. If this is right, then even postmodernists have a theory of human nature. I shall not examine competing views of human nature here, but if all sociologies presuppose one there is not much point dodging the issue either. Let me therefore stipulate the following rump materialist ''theory'' of human nature. Unlike the open-ended list of identity needs, it posits just five material needs. These are needs of individuals, not groups. Groups also have needs, but since they do not have bodies these will be identity needs which cannot be reduced to the material needs of their members even though they help meet the latter (see chapter 5). Material needs may generate contradictory imperatives and thus practices, but they vary in importance and people will generally - though not always - try to meet their more fundamental needs first. In roughly descending order of importance. Physical security: human beings need food, water, and sleep to sustain their bodies, and protection from threats to their physical integrity. Fear of death comes under this heading. Ontological security: human beings need relatively stable expectations about the natural and especially social world around them. Along with the need for physical security, this pushes human beings in a conservative, homeostatic direction, and to seek out recognition of their standing from society. Sociation: human beings are social animals who need contact with each other. Needs for love and group membership are met through sociation Self-esteem: human beings need to feel good about themselves. This is achieved primarily through social relationships, and as such its content can vary hugely, including ''needs'' for honor, glory, achievement, recognition (again), power, group membership (again), and so on. Transcendence: human beings need to grow, develop, and improve their life condition. This is a source of creativity and innovation, and of efforts to remake their material circumstances. In the last analysis the energy that human beings expend in their lives stems from efforts to meet these material needs, and people will define their interests in ways that facilitate doing so in the material and cultural environments in which they find themselves. When needs are met people experience the emotion of satisfaction. When needs are not met we experience anxiety, fear, or frustration, which depending on the circumstances will motivate us to redouble our efforts, to change our interests, or to engage in aggression. Thus, in contrast to Classical Realists who would posit fear, insecurity or aggression as essential parts of human nature, I am suggesting these feelings are effects of unmet needs and therefore contingent. The effort to prevent the fear and anxiety associated with unmet needs is part of human nature, but fear and anxiety themselves are socially constructed. Regardless of the truth of this particular ''theory'' of human nature, rump materialism is an ontological argument that we need some such a theory to explain human behavior. Ironically, Neorealists seem as uncomfortable with this suggestion as radical constructivists, preferring to ground their theory on the ''structural'' materialism of power rather than the ''reductionist'' materialism of human nature. Human nature cannot be avoided, however, and the assumptions we make about it will condition our theorizing about world politics. Like power, interests are not ideas all the way down. This is a significant idealist concession to materialism, but the two are not contradictory. Biological realism is compatible with social construction. The question is to what extent does biology constitute interests? Perhaps thinking that it cannot or need not be answered, systemic IR scholars have largely avoided this question in recent decades, but with the emergence of sociobiology there is now the potential for a renewed and fruitful discussion. Sociobiologists would say biology matters quite a lot in the constitution of interests, as would perhaps most Classical Realists. Even Neorealists, when necessary to sustain their pessimism about anarchy, will fall back on the view that human nature is inherently selfish or power-seeking. In contrast, even though the kind of constructivism I favor is thin, in my view biology matters relatively little. Human nature does not tell us whether people are good or bad, aggressive or pacific, power-seeking or power-conferring, even selfish or altruistic. These are all socially contingent, not materially essential. Much more than other animals, human behavior is underdetermined by our nature, a fact attested to by the remarkable variety of cultural forms we have created. In developing this hypothesis we should not forget that human beings are animals whose material needs are a key constituting element of their interests, but in the end their interests are mostly a function of their ideas, not their genes. Let me conclude with three virtues of an idealist approach to the study of interests in IR. First, and most important, it suggests a program of empirical research for studying the content of real world state interests. Most traditions of IR theory rely on intentional explanations of action, and as such need a model of state interests. In practice mainstream IR scholars typically assume a model. This is perfectly legitimate for certain purposes, but it is nevertheless striking just how little empirical research has been done investigating what kinds of interests state actors actually have. Perhaps this is because everyone ''knows'' that states are egoists who want power (and wealth?, or security?), or because the influence of rationalism on the field has discouraged the empirical study of preferences, but it might also reflect the fact that materialist social theory offers little guidance about how exactly to find and study interests, especially in a corporate person like the state. By hypothesizing that interests are constituted by ideas, idealism suggests that schema theory and attention to deliberation processes - suitably adjusted for the fact that states have collective rather than individual cognitions - might prove to be fruitful approaches to this problem. Second, and by extension, an idealist approach to interests also suggests ways to operationalize the relationship between cognition (agency) and culture (structure). In social (and IR) theory it has become commonplace to describe action as culturally or discursively structured, but rarely is a mechanism supplied through which this effect might actually work. Somehow it is thought to be enough to point to the existence of cultural norms and corresponding behavior, without showing how norms get inside actors' heads to motivate actions. The materialist theory of interests may help explain this neglect, since it makes it difficult to see how an ideational phenomenon like culture could affect a material phenomenon like interests. Recognizing that interests are constituted by ideas removes the problem of mixing two kinds of ''stuff.'' In IR this points toward a potentially fruitful dialogue between cognitive theories of foreign policy and cultural theories of structure, perhaps organized around the concept of foreign policy ''role'' (see chapters 4 and 6). Finally, this approach suggests new possibilities for foreign policy and systemic change. In raising this issue it should be emphasized that saying that interests are made of ideas does not mean they easily can be changed in any given context. Idealism is not utopianism, and it is often harder to change someone's mind than their behavior. As such, ironically enough materialists may sometimes have a rosier view of the future than idealists, as in Waltz's view that controlled nuclear proliferation can cause system stability. However, to the extent that interests are constituted by beliefs we can have more hope of changing them than we could if they simply reflected human nature (short of genetic engineering). It may be difficult for an actor to change its interests if the beliefs that constitute them are part of a culture that simultaneously constitutes the interests of other actors. This helps explain why cultures tend to reproduce themselves once created. But the fact remains that if interests are made of ideas, then discursive processes of deliberation, learning, and negotiation are potential vehicles of foreign policy and even structural change that would be neglected by a materialist approach. Conclusion The argument of this chapter has been that the meaning of the distribution of power in international politics is constituted in important part by the distribution of interests, and that the content of interests are in turn constituted in important part by ideas. The constitutive as opposed to causal nature of this claim bears emphasis. The claim is not that ideas are more important than power and interest, or that they are autonomous from power and interest. Power and interest are just as important and determining as before. The claim is rather that power and interest have the effects they do in virtue of the ideas that make them up. Power and interest explanations presuppose ideas, and to that extent are not rivals to ideational explanations at all. My claim is therefore different than the Neoliberal argument that a substantial proportion of state action can be explained by ideas and institutions rather than power and interest. That treats ideas in causal terms which, while important, is not enough. The issue of ''how'' ideas matter is not limited to their causal effects. They also matter insofar as they constitute the ''material base'' in the first place, that is, insofar as it is ''ideas all the way down". An argument that power and interest are just as important as before, but constituted more by ideas than material forces, inevitably raises the question, ''so what?" If the balance of variables has not changed, what difference does this make to our understanding of international politics? Part II of this book is one answer to this question. But let me answer for now in programmatic terms by proposing a rule of thumb for idealists: when confronted by ostensibly ''material'' explanations, always inquire into the discursive conditions which make them work. When Neorealists offer multipolarity as an explanation for war, inquire into the discursive conditions that constitute the poles as enemies rather than friends. When Liberals offer economic interdependence as an explanation for peace, inquire into the discursive conditions that constitute states with identities that care about free trade and economic growth. When Marxists offer capitalism as an explanation for state forms, inquire into the discursive conditions that constitute capitalist relations of production. And so on. Enmity, interdependence, and capitalism are to a large extent cultural forms, and to that extent materialist explanations that presuppose those forms will be vulnerable to the kind of idealist critique featured in this chapter. This is not to say that we should never treat cultural contexts as given, within which materialist explanations may be compelling, but in doing so we should recognize that the latter acquire their causal powers only in virtue of the contexts of meaning which make them what they are. Nor, on the other hand, is this to say that material forces like human nature, technology, or geography play no role in state action whatsoever. However, the materialist explanations offered above go well beyond such factors, in effect ''cheating'' on the materialism-idealism test by building implicit cultural elements into their claims. Only after we have stripped the discursive conditions of possibility from those claims will we know what material forces can really do. This argument tries to change the terms of the materialism-idealism debate in social theory by reducing ''materialism'' from its traditional, expansive definition focusing on the mode of production (or destruction), to a stricter, rump definition focusing on materiality per se. This is not definitional sleight of hand, but an attempt to get at issues that are obscured in the traditional base-superstructure model. The key here is recognizing that materiality is not the same thing as objectivity. Cultural phenomena are just as objective, just as con- straining, just as real as power and interest. Idealist social theory is not about denying the existence of the real world. The point is that the real world consists of a lot more than material forces as such. Unlike a potentially more radical constructively position I do not deny the existence and independent causal powers of those forces, but I do think they are less important and interesting than the contexts of meaning that human beings construct around them. Finally, this reframing of the issue casts new light on the NeorealistNeoliberal debate. In my view, Neoliberals are caught in a Realist trap. It is the same trap that structural Marxists like Louis Althusser and Nicos Poulantzas were caught in when they tried to show, against orthodox Marxists, that the superstructure was ''relatively autonomous'' from its base. Structural Marxists conceded Primary explanatory importance to the mode of production (material base), but tried to show that institutional and ideological superstructures were important intervening variables. This theory ultimately failed, however, because of the inability to make coherent the argument that the superstructure was ''relatively autonomous'' while the material base remained still ''determinant in the last instance.'' (Interestingly, with the failure of structural Marxism many erstwhile adherents became post-structuralists, a move not unlike what happened in the 1980s in lR.) As in the case of structural Marxists, Neoliberals have done important work showing that by itself the material base (here, power and interest) cannot explain international outcomes by itself, but by conceding the base to Neorealists they have nevertheless exposed themselves to the same problem. This trap underlies Mearsheimer's argument that Neoliberals are tacit Realists; structural Marxists, after all, were still Marxists. From Mearsheimer's perspective and mine, in other words, Neoliberals face a hard choice: either acknowledge the ultimately Realist character of their theory (because it buys into the base-superstructure interpretation of materialism) and deal with the problems of sustaining an independent theoretical position using a ''relative autonomy'' thesis, or refuse the Realist trap by problematizing the ''materialist'' nature of power and interest explanations from the start. Either way, in the end there can only be two possibilities, materialist and idealist, because there are only two kinds of stuff in the world, material and ideational. Throughout this chapter I have used the language of ideas and the term idealism to make the case against materialist approaches to structure. This permitted economy of expression, but it might have suggested that I advocate a subjectivity approach to social theory in which all that matters is how individual agents perceive the world, or a voluntarist one in which agents are thought to be free to choose any ideas they wish. I advocate neither. How agents perceive the world is important in explaining their actions, and they always have an element of choice in defining their identities and interests. However, in addition to idealism, a key feature of constructivism is holism or structuralism the view that social structures have effects that cannot be reduced to agents and their interactions. Among these effects is the shaping of identities and interests, which are conditioned by discursive formations - by the distribution of ideas in the system - as well as by material forces, and as such are not formed in a vacuum. I have so far largely ignored the effects of this distribution, as well as the senses in which it might be structured. It is to these issues that I now turn.