versão impressa ISSN 0011-5258
Dados v.4 n.se Rio de Janeiro 2008
The social conditions of instrumental action: Problems in the sociological understanding of rational choice theory*
As condições sociais da ação instrumental: Problemas na concepção sociológica da teoria da escolha racional
Conditions sociales de l'action instrumentale: Problèmes dans la conception sociologique de la théorie du choix rationnel
Bruno Sciberras de Carvalho
Translated by Paul Charles Freston
Translation from Dados - Revista de Ciências Sociais, v.50, n.4, p. 827-862, 2007.
This article critically analyzes new sociological approaches to rational choice theory which beyond examining political or economic practices link the notion of instrumental rationality to social issues and themes. The article begins by highlighting the issue of trust, indicating the functionality of certain social arrangements in collective problem-solving. The paper goes on to demonstrate that problems emerge with the theory when it attempts to explain the feasibility of social norms in impersonal, comprehensive contexts. Thus, the fundamental point that appears to be missing from rational choice theory is the perception that individual decisions and instrumental conduct themselves incorporate dispositions that in a sense are beyond the actors' control.
Key words: rational choice; social theory; individualism; rationality
Dans cet article, on fait la critique des nouvelles approches sociologiques de la théorie du choix rationnel lesquelles, au-delà de l'examen des pratiques politiques ou économiques, relient la notion de rationnalité instrumentale à des questions et des thèmes sociaux. On remarque, d'abord, le thème de la confiance, qui montre la fonctionnalité de certains arrangements sociaux dans leur solution des problèmes de l'action collective. On montre, pourtant, que les problèmes de la théorie surgissent lorsqu'on essaie d'expliquer l'applicabilité des normes sociales dans des contextes impersonnels et plus larges. Ainsi, le point fondamental qui semble absent de la théorie sociale du choix rationnel est la perception que les décisions individuelles, ainsi que la conduite instrumentale, assimilent des dispositions qui sont, d'une certaine façon, hors du contrôle des agents.
Mots-clé: choix rationnel; théorie sociale; individualisme; rationnalité
In general, the propositions of rational choice theory are linked to political themes brought up, above all, by precursors like Kenneth Arrow, Anthony Downs, James Buchanan, and Mancur Olson. Thus the central tenets of the theory, deriving from the tradition of economic thought, serve to demonstrate that the normal possibilities of political interaction are undergirded by interest groups that seek particularistic benefits and by parties that maximize votes. On the other hand, the last few years have seen the growing relevance, in the field of rational choice, of a literature linking the notion of instrumental rationality to the examination of social problems and issues that exceed the limits of political practice. The theme of social norms inaugurates a theory of action which relates instrumental orientation with the idea of a social structure controlled by its agents, leading to an understanding of society as a strategic game. The analytical object of the sociological understanding of rational choice refers to goods or human activities which are not comparable and cannot be studied only through traditional economic science which sees social situations as generalizations of decisions taken by isolated subjects. A new approach to behavior that is both rational and non-economic is thus assumed to be necessary (Elster, 1984:127).
The original point emphasized is the fact that, unlike in standard social thought, there is no resort to cultural restrictions, tradition, chance, or imaginary structures. Social phenomena are analyzed through the perception of norms and values within complex mechanisms of exchange directly linked to agents' interests.
The essential project of the sociological theory of rational choice is to establish an analysis that coherently examines the movements from the macro-level of the social game to the micro-level of economic actions, and vice-versa. The objective is to understand social relations as a game dependent on rational practices, based on the understanding that individuals freely and actively participate in the structural construction of society, which is seen as a set of rules that restrict, but do not constitute, interactions. Therefore, the sociological proposition is to study the contractually established rules that organize actions and are subsequently modified and developed in the strategic movement of the social game. The need to analyze these rules becomes evident when interpersonal exchanges reveal great differences of resources, deriving from the diverse social positions of individuals. The admission of social life as a game demonstrates the postulate of a space for needs and struggles over scarce goods. Thus, society is seen as synonymous with an organization of interests in conflict. More specifically, social structure relates to the exchange mechanism of the market in the following way:
"[...] social exchange often occurs not in isolated two-person transactions but within the context of systems of exchange in which there is competition for scarce resources. These social markets sometimes resemble economic markets, although they often show major differences. An idea of both the similarities and differences can be gained by examining the role of money in economic systems for as much as any other single difference, it is the absence of money that sets noneconomic exchanges from economic ones (Coleman, 1990:119)"
In other words, the separation of the social and economic systems occurs through the mere absence of an objective means. Essentially, this quote reveals the social understanding adopted by rational choice theory.
This theory connects the theme of scarcity, a natural fact that obligates individuals to relate to one another and to create a society, to the theme of exchange, which constitutes its central mechanism of interaction. The social system is seen as a system of exchanges, in a way that norms and values become contingent on individual preferences. Instrumental practices are the micro-groundwork of this sociological theory, and are in opposition to the perspectives that do not stress the autonomous interests of individuals in all situations. Meanwhile, structure is not a proper dimension of social life, but an intentionally constructed necessity, a posteriori, by agents' interests. Norms and rules spring from conflict and scarcity, and create conditions for the emergence of equilibrium. Thus, Russell Hardin (1995:26-27) points out the need to study not only the contexts of what he calls "cooperation" which represents the regular model of the exchange of goods , but also "coordination", which is regulated by social norms and matches situations in which a party involved in the interaction maximizes something only if others do also. The important thing is to point out that the strategies elaborated by individuals are based on conventions that aim to promote more transactions. Subsequently, scarcity may be largely mitigated if related with a context that efficiently regulates free exchange. The sociological understanding is, therefore, clearly functional and circular: the logic of collective action and the opportunism of individuals cause problems of interaction which must be resolved by comprehensive instances and norms.
Social rules are connected to an individual agency prior to its own regulations. Commonly, the idea of norm in the sociological theory of rational choice is based on the concept of information. It is this category that reiterates the perspective of a control of social facts by individual cognition, since there is the assumption that social rules are informational instruments that give predictability to agents involved in the domains of exchange. Hence, norms are principles that regulate and inform the interactions, making them more or less predictable, but which, on the other hand, may be manipulated by the instrumental orientation of one of the parties involved. According to Elster (1983:16), the issue of social action refers precisely to the efficient relation between beliefs and evidence available to an individual. The notion of "judgment" asserts the possibility of agents reaching a "real belief", based on the total control of information or norms in a social context.
The problem is that the search for information is limited by rationality itself, due to the costs of acquiring new forms of knowledge of reality. (Hardin, 1995:15).1
The fundamental point of this sociological perspective is that norms are always circumstantial and do not acquire an autonomous social statute, insofar as they are only "used" if they are in accordance with the strategic acquisition of goods or services. One of the consequences of such an understanding is to see norms as "restrictions" on actions. Disconnected from a prescriptive dimension, they are conceived as instances of individual decisions, related to personal cost-benefit calculations. The understanding that norms and values are constituted in an ahistorical and contingent manner is reiterated. Actually, the value of social facts may be noticed in the economic domain, that is, in the exchange value they possess. Social norms and prices become equal, because the former necessarily depends on the interest of many agents.
One notices that this perspective objects to classical theories of sociology that indicate a complex relation between the domains of social structure and individual agency. Consequently, a series of social data are deprived of their characteristics and become irrelevant, especially the relations of conditioning between culture and norms and the explanation of differences between societies. Such issues are set aside in favor of an understanding that makes social spheres homogeneous and sees society as an aggregation of independent exchanges. In this way, the explanation of the coordination of agents through norms or conventions becomes empty. As Hardin claims, coordination "can just happen. And if ti just happens the same way a few times the result may be a forceful convention that then governs future behavior by giving us specific incentives for action" (idem:45). The main target of social relations is an isolated individual, free from imperatives, and structure is examined as a result of the constant volition that comes from his self-interest. Instead of being a decisive component of the constitution of individuals, social structure appears as a controlled fact, formed by the contingencies of bribes, threats, and promises, and may not be met if contrary to the conditions imposed by instrumental conduct.
The uniqueness of the social theory of rational choice may be perceived in the problem exposed by Parsons regarding the possibility of the constitution of a social order based on instrumental conduct. Parsons (1968:89-94) points out that there are empirical limits to the perspective of conceiving social integration through the aspect of maximization of personal means and ends. According to the author, such an ideal means either social disorganization itself, since individuals will attempt to use others as means for private objectives, thus creating social chaos, or metaphysical solutions, as exposed in the fictitious social contract of Hobbes and in the optimistic liberalism of a natural identity of interests. Thus, a sociological perspective that emphasizes instrumental behavior, such as rational choice, could not adequately express the normative elements of action that support social integration and condition individual choices and decisions. Such problems noted by Parsons, which in fact reflect some central tenets of the sociological tradition, did not impede rational choice from attempting to develop a social theory and the possibility of order through utilitarian conduct. However, as we shall see, the theory often reveals difficulties in the analysis of more complex social groups or realities, and generally seems unable to create a consistent argument on the role of social structure in relation to instrumental attitude.
The objective here is to describe some problems with this social theory, which means restricting the analysis to new and uncommon understandings of rational choice regarding social norms. Thus, we have no intention of dealing with the ontological bases and the already classical perspectives of the theoretical tradition which starts out from the concept of instrumental rationality, about which there is already an extensive literature especially in the field of political theory. To consider the singularities of the sociology of rational choice, I will set out the questions of trust and availability of information for agents, which reflect some of its essential themes. The theme of trust shows the functionality of certain social arrangements (generally of a reduced size) for resolving problems of collective action arising from instrumental behavior. Later on, I will show that this type of trust is far from representing the normative character of a social belief. I try to show, however, that the more obvious problems of the theory arise when it tries to explain the viability of trust and social norms in more extensive social contexts, since it cannot coherently demonstrate the relation between individual agency and social structure. I then resort to the debate found in recent works by the sociologist Raymond Boudon, demonstrating differences between an ontology of the social sciences and the economic ontology that is the basis of rational choice analysis. Finally, articulating criticisms of the concept of instrumental rationality by some authors from outside the field of rational choice theory, I argue that it does not seem sociologically consistent to evaluate social norms and conducts through the maximization of personal needs, interests, and benefits. I suggest that it is the characterization of these very needs and interests that must be explained in its genesis, which may only be done through the analysis of the specificity of instrumental behavior in modernity.
The calculation of trust and the evolution of cooperation
Insofar as a system based on unpredictable relations of competition is pressuposed, trust emerges as a parameter that shows the functionality of certain environments in the resolution of collective action problems. Giving others credit is analyzed through its potential in allowing predictable exchanges in egotistic contexts. In this case, the sociological question is to know when an agent should cooperate and in what conditions he should follow his interests in social relations (Axelrod, 1984:vii)2.
Trust is connected to the risk that individuals expose themselves to in their transactions. Risk presupposes the time that exchanges take to be concluded, which implies uncertainty for the agent who invests certain resources before receiving any return. The theory of rational choice claims that, in social spaces of restricted size, uncertainty may be mitigated in two ways. The first one is through the creation of formal mechanisms, like contracts and legal systems. However, what is of interest to the sociology of rational choice is the second way, based on factors of trust that stimulate low-risk exchange in an informal manner. These factors reveal and provide a predictable environment that dispenses with the costs of organization of a legal or contractual system. Nevertheless, the sociological perspective of rational choice establishes a union between the idea of a social dimension that imposes itself on personal preferences and a cost-benefit calculation that the agents autonomously elaborate. Therefore, trust and risk are seen as parallel phenomena, because even if social principles may acquire some independence in relation to the contingency of the transactions, they may also be instrumentalized by personal interests. Trust is subordinated to a sort of individual gamble, mathematically calculated to the smallest details, insofar as
"[...] the actor knows how much may be lost (the size of the bet), how much may be gained (the amount that may be won), and the chance of winning. These and only these are the relevant elements [...] this simple expression is based on the postulate of maximization of utility under risk. The potential trustor must decide between not placing trust, in which case there is no change in his utility, and placing trust, in which case the expected utility relative to his current status is the potential gain times the chance of gain minus the potential loss times the chance of loss (Coleman, 1990:99)"
So, the norm of trust does not have an element of behavioral prescription. Social relations are submitted to a cost-benefit calculation of cooperating or not, so that the trustor must possess appropriate mechanisms to dissuade the other party from breaching the agreement. Therefore, fulfilling the promises is linked to the beginning of frequent exchanges between the same individuals. That would be a natural feature of social relations, because rational agents try to interact with the same people. The main incentive for fulfilling the agreements is that the partner of the current interaction must be useful for future undertakings. Therefore, transactions tend to be like repeated games, such that the agents involved possess a reciprocal interest in reinforcing the promises made. The trustee of a benefit expects to lose a lot if the relationship established has temporal continuity, unlike a transaction or a straightforward and contingent "bet" with unknown subjects. Information gains special status in this perspective, because there is a direct correlation between the acts of an individual and the next move to be made by another agent (Smith, 1982:149). Since social relations take the form of risk situations, it is information that specifies the possibility of a context of minimum trust. At the same time, as more information regarding moves made is spread throughout a communicative structure, the greater becomes the element of trust in the interactions.
As long as games are repeated, it becomes possible to guide expectations and establish punishments according to past actions. In addition, reputation becomes a variable in the game (Axelrod, 1984; Heap and Varoufakis, 1995:167-194). The main point is that repetition allows for the development of cooperation without the need for external agencies to stimulate it. The prisoner's dilemma is usually resolved, and the possibility of joint moves for mutual gains is no longer seen as irrational. Players may adopt the tit-for-tat strategy a repetition of the adversary's previous move, in a way that will threaten defection if he tries to do the same to resolve the game satisfactorily. This perspective is called "evolutionary", based on the assumption that cooperation and reciprocity of behavior come spontaneously, even in an environment composed of selfish individuals (Axelrod, 1984:viii). Therefore, besides its spatial composition, a central aspect of cooperation is the temporal dimension. The calculation made by agents shows the ratio between the immediate gains of opportunistic conduct in relation to cooperation based on a more or less stable evolutionary network of future gains, although not as certain as present benefits.
One should observe that the understanding of the theory of repeated and evolutionary games starts from zero with a society without social norms. The essential issue regards the emergence of cooperation in a world filled with selfish people and no central authority (idem:3). One assumes that morality, in its normal sense, is not important, since only the agents' self-interest is required in the interactions. Therefore it is the sphere of information that is valued, instead of social rules or conventions. When the authors reveal a concept of culture for the study of cooperation, a simplified model is suggested, in which individuals acquire their behavior through the imitation of successful actors, which means a "stabilized evolution" (Smith, 1982:172).
What seems essential is that, even though some mistrust is inevitable, it can be partially contained by the classification of information regarding the parties involved in the interactions. This partly solves the problem, as shown by Anatol Rapoport (1980:178-179), with game theories that discard the importance of communicative acts and individual beliefs, in such a way that a sphere of information and social communication becomes essential for the analyses of social exchanges.3 The importance of information to deter unpredictable behaviors is not restricted to impersonal interactions, but is assumed even in the private spheres of family or amorous relationships. According to Coleman, "it is principally for this reason that parents of high-school girls want to meet and know as much as possible about the boys their daughters go out with" (1990:103). It's interesting to stress this commonplace example, because it clearly expresses sociological formulation regarding the relation between trust, risk, and social norms. The conduct of fathers and boys in relation to girls is compared to calculations made, for example, in exchanges between exporters and future importers. Boys proceed to calculations on gains and losses of girls' sexual aggression, which is, according to Coleman, the most desired attitude were it not within a relation of strategic reciprocity. On the other hand, for girls, "that gain consists of attention from the boy and being able to go out on dates in the future, rather than staying at home as she has in the past" (idem:102). Therein, building friendship ties expresses the partly unstable expectation of placing trust in certain individuals and not in others and establishing transactions based on self-interest. Rupture or transgression of a friendship, for example, is not linked to certain shared social duties, but to a personal evaluation that shows that
"[...] the potential gain from a close relationship may also be great, but since there are many other potential friends who might provide a nearly equivalent gain, the relevant comparison is not between the absolute potential loss and the absolute potential gain, but between the absolute potential loss (such as a violation of one's confidence or abuse of one's body) and the difference in gain expected from this friend and that expected from another (idem:104-105)"
The radicalism of such propositions implies the contingency of trust and social norms, which appear as factors to be used or discarded depending on the various circumstances of personal interests. The concept of instrumental rationality in these examples leads to extreme situations, taking into account the expressive source that the dimensions of love or friendship, through complex normative schemes, possess in modern societies.
However, the theory does not satisfactorily answer the logical question of how actors may generate intense relationships and easily escape the prisoner's dilemma situation, if they do not truly trust their fellow creatures. On the other hand, one might question if trust, in the rational choice analyses, is considered a social norm, since there is no reference to an unquestionable prescriptive sphere. What one can see is the centrality of the concept of placement of trust (idem:97). That is, trust is not a structuring dimension, independent of the exchanges being made, but a result of the instrumental mathematics of a specific situation. Thus certain authors show the need for social relations, especially in impersonal contexts, to be protected by formal mechanisms or regular contracts guaranteed by the state (idem:98; Hardin, 1988:46-47).
Essentially, the sociological propositions of rational choice suggest the irrelevance of the structural and cultural dimension for the explanation of social relations. On the other hand, as we shall see, the most obvious problems of the theory can be seen when it leaves the sphere of the contexts of interaction of a reduced size and tries to explain the viability of trust and norms in major communities. Demonstrating analytical indecision between the spheres of individual calculation and social structure, the theory seems unable to explain events that have multiple agents. It remains undecided, above all, among propositions that claim the autonomous existence of norms and others that suggest the inexorable need for sanctions in the expression of non-selfish orientations.
Norms and individual autonomy in complex societies: theoretical problems
Leaving the sphere of direct exchanges between a limited group of individuals, or what might be called microsociology, part of the rational choice literature also tries to explain the emergence of norms in comprehensive social contexts. Analyses demonstrate a greater preoccupation in trying to understand collective phenomena, especially those linked to a sphere characterized by impersonal and sparse transactions. Studies attempt, above all, to understand the production of social norms that are functionally adequate to the problems stemming from strategic interaction, which may generate a more appropriate functioning of the socioeconomic context.
Part of the sociological theory of rational choice claims the centrality of "externalities" as an essential parameter for the analysis of social interdependence in complex environments (Coleman, 1990:20-21; Opp, 2001:15-16). The existence of problems related to actors that are external to a transaction, over which they do not have direct control, leads to a unique and functional preoccupation in relation to social norms. There are two types of externalities: a "positive" one, when an action benefits other individuals situated outside the transaction process; and a "negative" one, when an action leads to bad consequences for the agents. For the authors of rational choice, it is the latter that generates larger interest, because research may lead to impeding certain actions that cause collective harm.
One of the ways of deterring negative externalities is by instituting a system of compensations for the individuals affected by certain actions, producing a sort of "market of externalities". Thus, the agents interested in actions that lead to harm establish direct contact with those affected, with the intention of purchasing the rights to carry out such actions. Therefore, a social structure is generated in the moment when individuals concede to others the rights to control over certain practices, which generates benefits for all parties involved in the process. According to Coleman (1990:145), it is the foundation of this relationship that creates a transition from the micro plane to a fundamentally social dimension. For such a context to be predictable, norms emerge as mechanisms that guide attitudes in a certain direction. The direction of such actions does not need a legal definition, as it may occur through a situation of informal and voluntary exchange between the interests at play. However, certain authors of the rational choice field point out that a market of externalities is out of the question in big societies, due to the impossibility of establishing agreements between a vast array of agents. In such contexts, individuals who impose externalities find themselves far away from the people who suffer them. The origin of social norms must be found, then, in other mechanisms of control over actions that cause collective harm.
As in the circumstances in which the problem of externalities imposes itself, the sociological theory of rational choice lays claim to an "instrumentality proposition" (Opp, 2001:15), because social norms must emerge as tools for obtaining collective objectives. Besides the conditions that generate externalities, norms of social structure emerge as mechanisms to solve prisoner's dilemma situations and those that need coordination for achieving a better equilibrium. Thus, better social situations may be obtained through an explicit or tacit agreement, regarding problems of coordination, or through sanctions that make individuals fulfill promises made in situations characterized as the prisoner's dilemma. One can say that there is a "utilitarian use" of norms, insofar as subjects have advantages if they act according to them. In this way, "the reward for conformity to a regularity in these circumstances consists in the very act of conforming, since it guarantees what is desired by all the achievement of a coordinated equilibrium" (Ullmann-Margalit, 1977:85).
Coleman (1990:253) points out the difference between this social concept and the norm-free environment defined by the traditional model of the prisoner's dilemma. While in the latter the capacity to establish a predictable cooperation is excluded since agents cannot communicate with one another (and even if they could, they would be unable to trust each other), the former perspective presumes a context of free communication, which involves the possibility of the constitution of self-fulfilling rules. The study of social norms becomes relevant in paradoxical contexts in which game theory is indifferent regarding the solutions presented, or in the cases in which the result is sub-optimal and socially unacceptable (Ullmann-Margalit, 1977:15). In this way,
"[...] when a large group or society faces a coordination problem, it may not simply be able to use a coordination promise. But, as Huma argues, it may readily resolve its problems by convention. If the problem recurs often enough, any successful coordination by much of the relevant population may signal how to coordinate again at the next recurrence of the problem. The strategy onto which the group happens may become a convention in the sense that it is the obvious choice of virtually everyone whenever the standard problem occurs (Hardin, 1988:49)."
However, these analyses do not satisfactorily explain which specific social mechanisms could impede the emergence of opportunistic behavior, that is, there needs to be an explanation of the resolution of the prisoner's dilemma.
The question is how to protect a context of collective benefits that is unstable due to the temptations of the participants involved. The principles of individual rationality clearly clash with social propositions, creating a contradiction that authors seem incapable of resolving. The problem is that along with the principle which states that agents attempt to minimize the risks and the lack of predictability of interactions, one must add the idea that they spontaneously determine the most efficient social domain for the attainment of such an objective. The sociological question to be faced, therefore, is the explanation of how this domain is chosen, because this depends on a theory which observes the cognitive and social views of individuals, that is, of a more extensive scrutiny than the model of rationality used by rational choice.
Due to the possibility of opportunistic behavior especially in situations in which there is an element of individual sacrifice in relation to more advantageous alternatives -, most analyses consider social norms as dependent on specific sanctions, matching the need for control to prevent individuals acting only in their own interests. The theme of sanctions attempts to unite rationality and social norms, demonstrating that deviant actions may directly affect personal expedience. Therefore, norms may only be institutionalized through vigilance, so that those who do not participate in the collective action become seen as losers or exploiters (Wolfelsperger, 2001:85). It is the desire not to be exploited that establishes a link between social sanction and the value of the elements at the rational level of the personal preferences at stake. Rewards and punishments become essential mechanisms in this model of society, in which interests and norms relate to each other in an unstable way.
However, issues become more complex in the contexts in which the establishment of sanctions produces both beneficiaries and losers who cease enjoying the right to perform more advantageous actions. Since there is no way to institute a system of complete or uniform vigilance, there is no explanation why agents should accept personal losses in favor of the abstract general well-being. On the other hand, it is essential to realize the substantive problem of sanctions in relation to the concept of efficiency defined in the parameters of Paretian optimization, since one of the parties involved in the transaction may lose something. To try and avoid certain contradictions, Ullmann-Margalit (1977:117) makes only vague references to internal sanctions in relation to non-fulfillment of social norms, such as personal feelings of guilt, remorse, and shame. Yet Coleman (1990:260-262) points out the inevitability of individual losses, which means a sub-optimal situation contrary to his presuppositions. One must also note that the domain of social norms necessarily presupposes a dimension of power, which eliminates egalitarian and consensual distribution of rights among individuals. Thus, norms and sanctions become weaker if they oppose the interests of agents with more resources. Even though Coleman (idem:270) notes that impeding externalities depends on the entirety of the resources of power that individuals possess, the author does not expand on the subject, preferring to point to agents' capacity to establish sanctions through dispersed channels of communication and the possibility of general classification of relations of trust.
These problems reflect the lack of resolution in the sociological theory of rational choice in general, concerning the dilemma between the notion of opportunism and behavior subjected to social norms. Analyses are generally limited to pointing out that norms sometimes constitute "second-order" factors, which are instrumental in impeding certain conducts in the attempt to generate first-order public goods (Opp, 2001:15-16). In this way it becomes difficult to understand efficiency and general well-being, because the theory contradictorily claims the autonomy of the pursuit of self-interest while it presupposes processes of socialization that weaken its concept of action. In addition, it does not explain why rational individuals, with their collective action problems, would join up in impersonal environments in an attempt to overcome certain externalities or cause a socially beneficial result.
One of the central problems with analyses of rational choice is that they slide into tautological explanations. In an attempt to consider the sources of social coordination (language, religion, community, and local customs, for example) as factors in personal interests, the authors allow in events which distort the concept of rational action. Values or beliefs are not defined as discursive frameworks that are a major part of social reality, but are restricted to contingent instruments to be used in the maximization of exclusive interests. Even if studies accept that many properties of culture are collectively produced, they try to argue that the satisfaction and benefits of these aspects are always linked to a dimension of rational consumption by certain agents (Hardin, 1995:68). In this clearly circular logic any social behavior may be described as rational, simply because the individual wants and acts that way. Therefore, whenever the theory accepts the existence of normative conducts that do not demonstrate any instrumental reference, the tendency is to call them irrational or place them with the idea of maximization of epistemological comfort (idem:61-63; 139), in such a way that any practice becomes a question of interests. The standard conclusion is limited to claiming that social norms tend to be stronger when they serve certain personal preferences.
Essentially, they do not explain satisfactorily the constitution of wide-ranging social coordination. In addition, the issue of differences from one community to another is eliminated, because it is presumed that they are all subordinated to the same type of instrumental behavior. What is also beyond a consistent understanding is the expression of groups which provide very little benefit to their members, such as peripheral religious communities, marginalized ethnic identities with no expectation of improvement, or even failed sporting associations. It is interesting to observe that, while theoretical propositions on the "macro" level are conditioned by the outcome of the "social game", no one ever questions the "micro" dimension in the social theory of rational choice. The latter is always classified through the idea of a nature which cannot be modified or developed through agents' exchanges.
One should also observe how the sociological theory of rational choice studies social norms and values which in principle do not have a specific functionality. In this case, the tendency of the theory is to emphasize the unpredictable and casual social direction of rational behaviors. Through a simple and non-evaluating process, certain norms appear when actors that have similar interests generate, through their uncoordinated but regular actions, the emergence of a new rule. Therefore, social norm emerges as a non-expected aggregate result (Opp, 2001:23-26). An example of this type of understanding may be given in female genital mutilation:
"[...]I will illustrate this model of the emergence of unplanned norms with Mackie's (1996) explanation of female genital mutilation (FMG). His basic assumptions can be summarized as follows. (1) Men as well as women strongly desire to successfully raise their biological children' (p. 1007). (2) Men have an interest to control women's sexual behavior in order to assure that they don't raise other men's offspring. (3) In polygynous societies with high resource inequality the costs of monitoring are high for men who have several women. (4) Parents and their daughters have an interest that the latter marry wealthy men. (5) FMG serves this interest: it is a signal for fidelity and constrains sexual promiscuity. (6) Circumcised women are preferred by men as wives" (idem:27).
This example exhibits the structural issues that the theory does not explain. The first one refers to the explanation of the emergence of the specific social fact of circumcision, and not of other mechanisms of "monitoring". Another sociological issue that is not analyzed is the way women's submission, through suffering violence towards their bodies, becomes male self-interest. Facts that seem to emerge exclusively from the social structure are also not explained, such as "men are interested in controlling women's sexual behavior", or "parents and women are interested in having their daughters marry rich men". These facts and values are not natural; other societies do not value them. The author seems to use, ethnocentrically, relevant cultural features of his own society to give ad hoc explanations. However, it is the characterization itself of these features, which agents internalize and which vary from society to society, that needs to be explained in its origin and specificity.
It is important to note that no historical reference is mentioned by the theory, even if conducts possess a traditional aspect that individuals do not question. Prescriptive dimensions such as the caste system in India, reprimanding for dirtying the streets of a big city, restrictions for sexual relations before marriage in religious communities, are seen as examples of results of consensual exchanges, and not as cultural models that exert some autonomous form of social conditioning. In the attempt to understand behavior that is not linked to the concept of instrumental rationality (such as strongly emotional actions in following a leader; a hunger strike for abstract and distant interests; the expectation of assuming a position in the frontline of a war; terrorists that blow themselves up with bombs) Coleman (1990:273-282) points to certain personal satisfactions based on incentives and rewards resulting from the union of individuals who seek the same objective.
However, it is not clear that individuals who follow a certain leader do so by encouragement from other individuals. In modern society, for example, this action may perfectly well occur through impersonal bases. Besides, the fact is that being on a hunger strike to help others does not necessarily relate to socially disseminated rewards or incentives. It would be even more complex to treat this behavior as the bearer of personal satisfaction. That is also the case of the person who autonomously goes to the frontline of a war. On the other hand, to treat as rational a group of diverse conducts implies tautological analyses, making it impossible to imagine any attitude that would not be considered rational. The concept of rationality becomes related to many different elements, hence losing its explanatory capacity.
To try and solve this impasse, some authors point to the emergence of informal norms that lead to processes of internal policing. According to Coleman, "it is the installation in the individual of something which may be called a conscience or a superego: I will call it an internal sanctioning system" (idem:294). Sometimes, the process of socialization is so influential that it may create another self (idem:517-518). Rationality and internal sanctions emerge, therefore, as two parallel mechanisms. The agent internalizes certain rules and norms that restrict his behavior, and is capable of lying to himself to deny their reality as sanctions. Thus, the satisfaction of interests may be reached through the transformation of the internal structure of the individual, adapting it to the imperatives of the external social order. This expansion of the self unites the individual to his context without questioning his rationality. This could explain, for example, processes of social identification such as individual orientations that lead to general mental benefits, sharing of experiences (even bad ones) or processes of transfer of power such as the identification with political or religious leaders.
Analyses often highlight mental facts that are not duly explained. Even if the denial of instrumental cognition is possible through processes of socialization, rational choice research does not comprehend such social mechanisms. On the one hand, norms seem to be internally produced by the individual. On the other, they seem to be strategically stimulated by certain agents, such as religious organizations or the national state.
Sometimes, there is the assumption that individuals have two useful functions, one based on the moral point of view and another one based exclusively on instrumental rationality (Wolfelsperger, 2001:74-75). Issues become more complex through the idea that the general norms of a society tend to be weaker when disconnected from a context of repeated interactions and personal incentives (Hardin, 1995:107). Essentially, the problems reflect the fact that the internalization of social norms emerges in opposition to the rational nature of individuals, generating competition between two dimensions that, at first, are considered to be incompatible in rational choice thought.
Voting is paradigmatic of these problems, especially of the relations that the sociological theory of rational choice causes between instrumental reason and irrationality, or between an agency free of determinants and a type of social structuring of behaviors. In fact, the issue is automatically attached to the expression "the paradox of voting". According to the propositions of the theory, a voter has no benefit in voting, given the calculation between the costs of the act in time and effort, and the relative impossibility of the act of voting making any difference to the final result. The rational voter, even though interested in the election result and in personal gains if a specific candidate wins, clearly knows that his own vote is not decisive a fact which points him toward abstention. Elections reflect the problem of collective action in which no citizen may be excluded from the benefits of a public good even if he has not contributed to its production. Nevertheless, most people do vote, which leads to a paradox. Occasions when empirical fact contradicts the theory's predictions become one of the most debated issues in the field of rational choice.
Studies contain different assumptions on the act of voting. Some contend that agents do not correctly judge the probability of their vote influencing the result, causing a reflection that increases expectations of their private decisions in relation to the social context. Other analyses resort to facts like civic duty, classified as irrational. Thus, it is suggested that agents go to the voting booths when they believe they have an obligation to do so or when they try to assert their partisan identity. Some research is based on the utility resulting from the contribution to a successful collective effort. On the other hand, there are studies that continue to propose the rationality of the vote calculation and try to show the fear that the individual has of losing the opportunity to decide an election, in the case of a probable tie. Others are based on game theory and claim that, from a strategic point of view, it is coherent to vote due to the expectation that other rational agents will not, which would make a personal vote decisive for the final result. As an unintended consequence, these individual expectations cause a massive turnout.
However, empirical evidence does not prove these expectations, and the tendency of studies on electoral conduct is to transform data and make them consistent with the theory of rational choice (Green e Shapiro, 1994:55). Analyses are restricted to concluding, with no consistent arguments, that the act of voting is generally more gratifying than its costs; therefore, it is rational. However, the notion that electoral abstention introduces some sort of personal loss is clearly in contradiction to the theory's concept of rationality, especially considering that the costs and benefits of such an attitude are explicit for any instrumental reflexivity. In general, this contradictory character seems to be linked to the fact that the theory does not refer to the sociopolitical structure in which actors participate. In the case of voting, for example, the lack of structural analysis discards the essential study of circumstantial modification of civic and political behavior from one election to the next, or indeed among different societies.
One must point out that some authors consciously accept that their analytical direction is restricted to the individual dimension and assert the advantages of a "positive" and simple methodology to show the essential phenomena of actions. Theoretical construction tries to perform a certain purification of empirical facts, and is more interested in what could happen in concrete situations if they were guided by rationality than in what actually happens (Ullmann-Margalit, 1977:1-2). The social interaction model designed by Axelrod (1984:17-18), for example, expressly avoids certain issues, such as behavior based on social habits or rules and unconscious or non-strategic choices. Also, when trying to explain the origin of agents' preferences, which could generate a decidedly sociological analysis, Coleman (1990:515-516) prefers to follow strictly psychological explanations. But Elster (2000:692) embraces the limits of the theory of rational choice and accepts the social importance of quasi-rational conducts or those based on emotional instances that go against agents' strategic and instrumental reasoning.
According to the author, the way out of this is to accept the modest explanatory capacity of the rational choice model, which does not mean its capacity to explain and predict the essential facts of human life should be questioned.
The absence of a coherent analysis of the relation between social structure and rational agency results in rational choice theory's tendency to consider phenomena that do not fit into its theoretical corpus as irrational or unintelligible occurrences. This analytical direction is reminiscent of the methodological division elaborated by Pareto (1968) between logical and non-logical behaviors. Thus, many actions are considered facts that must be disregarded, insofar as they belong to a phase to be overcome by the systematizing of economic rationality. As a result, authors do not pay attention to the primordial fact that the concept of rationality which they support depends on specific passions and values formed by complex social mechanisms, which subverts a model that objectifies instrumental and solipsistic orientations.
Rationality and methodological individualism
The main point that seems absent from the analyses of rational action theories is that individual decisions incorporate aspects that are detached from strategic situations and from personal interest. Aspects connected to social dimensions that are, in a way, out of the agents' control. The criticisms that Raymond Boudon (2002) makes of the paradigm of rational choice are important because they show this social sphere. Also, his criticisms are fertile ground for showing the differences between the theoretical fields of methodological individualism and rational choice, which although often confused are actually delimited by borders that separate the ontology of economic theory from social perspectives identified with a more structural analysis.
It should be noted, in the first place, that Boudon himself tries to separate his theory of methodological individualism from a normative evaluation of modern individualism, as is the case with analyses based on the concept of economic rationality. One might say that methodological individualism expresses an epistemological orientation which tries not to become the basis for an ontological position. When explaining a social phenomenon, Boudon prescribes the need to reconstruct agents' motivations and to consider the social fact in question as a result of individual actions. This methodology can be distinguished from holistic analyses that do not observe the necessary relation of collective phenomena with individual reasons and treat certain collective institutions, such as parties or religious organizations, as possessing knowledge or self-will (1991:50). Thus, the methodological plane does not presuppose that one should analyze societies as sets of independent atoms, but that the relation between attitudes and the social domain should be understood, because only this relation allows one to identify and justify an action as rational.
To clearly display his questioning of rational choice theory, and its difference from methodological individualism, Boudon (2003:19-22) points out the series of postulates that summarize the epistemology of the former: (1) every social phenomenon is a product of individuals' actions, behaviors, and beliefs; (2) all individual action can be understood; (3) with the exception of automatic actions (the organic action of winking, for example), one assumes that individual actions or beliefs are the product of reason; (4) the reasons considered by the agent always take into account the consequences of their actions; (5) individuals are exclusively or primordially interested in their own personal interests; (6) one assumes that the individual knows and evaluates the advantages and disadvantages of an action or belief, choosing the one which brings him most benefits. Boudon (idem:49-50) argues that the concept of rationality proposed by methodological individualism is restricted to premises 1, 2, and 3. Rational choice theory has added prescriptions 4, 5, and 6. The non-incorporation of the last three postulates by methodological individualism causes the study of specific and contingent cases to be connected to the situations in which agents display a behavior that corresponds to those premises.
Even though Boudon (idem:37) values parts of rational choice for revealing the reasons through which individuals act in a certain way, he criticizes the fact that a vast array of social phenomena are not explained by the instrumental assumptions of the theory. Based on his singular model of rationality, Boudon elaborates two essential criticisms of the paradigm of rational choice.
First of all, Boudon (idem:42-43; 2002:19-20) asserts the existence of many social situations in which agents' orientations are based on non-trivial beliefs. The important thing is the perception that personal acceptance of certain beliefs reveals a conduct of a cognitive nature. Therefore, the problem of rational choice theory is its incapacity to clarify phenomena in which the behavior of actors reflects socially coherent and non-consequential beliefs. Boudon (2003:45) gives examples of actions that make the idea of a careful and self-referenced evaluation a problem: the voter votes; two players share a sum of money instead of the first one getting more for himself; individuals commit suicide. Only in certain specific circumstances are actions and beliefs directed to the consequences and to a cost-benefit calculation. Boudon (idem:44-45) criticizes, above all, the manner in which rational choice tries to explain the phenomena that reveal these beliefs which are non-trivial and contrary to economic rationality. The capital mistake is to assume that individuals sometimes adhere to a sort of false consciousness, operating within "frameworks" that express inauthentic emotional factors. In this way, the voter who does not realize that his vote does not influence the final result is operating within a mental framework that would make him believe the contrary due to false estimation of his behavior. According to Boudon, these assumptions are always insatisfactory, insofar as they suggest that the individual obeys mysterious and irrational forces that cannot be explained.
Boudon's second criticism is that rational choice remains impotent in relation to social values that reveal attitudes and capacities of justification that cannot be related to egotistical conducts. The author (idem:46) reminds us that every spectator of Antigone condemns Creon and approves of the protagonist without hesitation, no matter where Sophocles' tragedy is put on. The universal character of such an attitude is linked to the right all people have to the dignity emphasized by Antigone. The attitude in relation to the play represents a series of phenomena in which the agents evaluate situations in which they have not been implicated. The weakness of rational choice in such circumstances is due to the fact that, instead of remaining indifferent, most people give a strong opinion on an issue that does not involve them directly. The exemplary fact that many individuals, who will probably never be condemned by the law, have a normative tendency regarding the death penalty shows that there are many instances of personal non-interested involvement that cannot be explained in the theoretical framework of instrumental rationality. Similarly, the plagiarist always causes a feeling of repulsion, even when he does no harm to anyone in particular and in fact contributes to the fame of the person plagiarized. The impostor is also seen negatively, even if his acts are not inconvenient to anyone but himself. The general notion of "might is right" is shocking, and its practice, unacceptable. The common citizen vehemently disapproves of corruption, even if it does not affect him directly, and the same attitude of indignation is felt in countries with remote possibilities of being affected by major levels of this behavior. Essentially, the sociological perception shows that, independently of social context, these facts are serious issues for individuals and are treated in detail, with personal expenditure of time and information so that they should not occur again. Insofar as they are phenomena of insignificant personal consequences, the disapproval can only be explained because it goes against fundamental rules that are the basis of a social pact.
Therefore, the questionable issue of rational choice is its analytical limitation to instrumental rationality, which makes it neglect cognitive or axiological reasoning based on social principles or theories. According to Boudon, the agents' reasons are of cognitive nature when he accepts an ordinary theory for the simple reason that it seems fair to him, even if it has no relation to his personal interests. On the other hand, rationality has an axiological quality when the agent normatively obeys prescriptive principles. Boudon tries to show that methodological individualism coherently expresses the rationality of actors with the more or less conscious incorporation of beliefs and attitudes that present "good circumstantial reasons" for existence. Social institutions or norms, if consistently based on reality, are considered good, legitimate, and acceptable, as well as disconnected from individuals' direct interests. Therefore,
"[...] many discussions would be, without a doubt, avoided if we remembered that belief is only a noun stemming from believe, or rationality a noun based on the expression have reason to. The verb has over the noun the advantage of not inducing these substantialist views that lead to questions on what is actually rationality or what is actually a belief" (idem:43).
Thus, adherence to a social belief or value does not occur through a cost-benefit calculation oriented to minimizing means and maximizing ends. The incorporation of a regular or practical theory by the subject does not stem from the fact that these seem to him to have certain consequences. In methodological individualism, rationality consists in adopting a belief that, based on a theory, allows one to observe the data of a certain social context in a satisfactory way, simply because the belief has coherent reasons. Rational action requires the acceptance of a logical group of propositions, when there is no other more acceptable alternative theory. Boudon reminds us that, if taken literally by the idea of an exclusively instrumental cognition, most normal actions should be understood as bearers of absurd and irrational results. Through the propositions of cognitive and axiological rationality, Boudon (1999:131) proposes the end of the "black boxes" present in the neoclassical notion of rationality, producing a sociological study that allows the analysis of new social data without resorting to a priori conceptual references. Boudon (2003:57) suggests that the process of regular knowledge is composed of the same parameters of beliefs that create scientific knowledge. Therefore, the effectively social aspect of rationality is that the reasons become stronger whenever they present themselves to the agents participating in the same social environments, or "contextual parameters", because "a reason, a system of reasons, are only identified as such if the actor has the impression that this reason or system is shared with others" (idem:160).
Boudon's analysis of rational choice theory seems to reveal the actual difference between the epistemology of social science and that of economics. Even though it is a school of thought that stems from a method centered on the individual, methodological individualism does not discard a structuring social dimension. Boudon (idem:11) makes it clear that his perspective does not suggest that socialization, when understood as a group of structural effects on agents' behavior, is a factor without sociological interest. On the contrary, the theory of methodological individualism reveals the centrality of the description of the ways through which subjects incorporate social effects. Without delving into his specific arguments or conclusions, the important thing is that the theory tries to analyze how individuals interact with structural conditions if they accept them, reject them, or politically claim them. Subjectivity, the central fact of all sociological projects, is connected to certain structural determinants. It is this connection that remains absent from the social theory of rational choice, in which subjectivity is determined a priori and structure is seen as a mere aggregate of antagonistic interests. There is no exchange between agency and structure, which implies an objectification of the circumstantial environment based on the economic equilibrium of the market.
Final observations: the social conditions of the maximizing AGENT
From the sociological point of view of the relationship between individual and social structure, the most questionable element of the rational choice theory of action is its incapacity to link the manifestation of instrumental reason to the environment of modernity. The fact that agents act according to their decisions does not mean they control the conditions and social context in which they are inscribed. As Marshall Sahlins argues regarding the analytical perils of theories that do not observe the structural and symbolic reality: "the final alienation is a sort of de-structuration" (1976:220). It is only through the structural conditions that we can reveal the singular techniques and actions available to individuals, because their practices, whether economic or otherwise, depend on a cultural system which is propagated through diffused or explicit processes. Thus, what is problematic in rational choice theory is the incomprehension of the interdependence between personal ends and social means, which lends a specific normative meaning to individual actions.
Weber's thesis (1996) on the connection between the Protestant ethic and the spirit of capitalism expresses the type of understanding that is absent from the theory of rational choice. Inscribing rational behavior in a certain social framework, Weber tries to demonstrate that both the values of Protestantism and their manifestation in calculated and frugal economic action belong to a macrosocial and historical level. It is important to point out that the relevant issue of Weberian analysis is not the discussion of the historical relation between Protestantism and capitalism as a real and causal explanation, but the exhibition of interdependence between instrumental attitude and the orientation of values, between agency and social structure, insofar as the rationality that characterizes the capitalist spirit reflects, in its particular meaning, singular values of a religious ethic. Thus, actions appear to be structured by a symbolic form that is incapable of being directly controlled by individuals, as is demonstrated by the thesis of the creation of an economic arrangement by a religion that is its opposite. Action is not defined by its personal consequences, but by extensive social results produced by historical processes.
The Weberian concept of social action, which characterizes three other types of behavioral orientation, as well as the instrumental form developed by the "capitalist spirit", expresses the problem of restricting the concept of rationality to instrumental parameters. According to Weber (1968:24-26), social action may be characterized by four types of orientation: goal-oriented rationality; value-oriented rationality; affective motivations; and traditional action. Thus, even in the social environment of modernity, it is important to understand the centrality of the orientation which, in opposition to economic reason, is determined by values or beliefs that do not require correlation to future results. The meaning of this orientation reflects practices that, conditioned by some type of duty or "cause", do not assume the cost-benefit calculation of means and ends of instrumental action. The essential point of the argument is its sociological bias, which reveals the necessary relation of the production of personal preferences to specific values of the context in which individuals participate. Inserting these practices in a particular social context, this type of analytical procedure shows that rationality is not a transcendent datum and can be connected to a type of reflection that is disconnected from self-interest. This analysis also identifies the potential of a non-instrumental sociopolitical system, created by a reflexive sociology regarding diversified beliefs or norms.
Therefore, differently to what utilitarian moral philosophy in general proposes, there seems to be no sense in considering individual ends or "pleasures", save for an investigation of the social logic which they acquire.
All practices, including the rational behavior of maximization, are connected to a reality that partially escapes the individual himself. It becomes necessary to analyze personal interests in a perspective that does not define it a priori, reinforcing the historically engendered interactive and symbolic processes that establish certain "dispositions", that is, specific inclinations for action based on collective norms or routines stemming from the properties of a social system (Domingues, 1995:141). The theory of rational choice does not observe such social configuration, preferring to explain the emergence of norms and attitudes as so independent that they do not even come to form a structure (Cherkaoui, 2003:250-251). However, norms and beliefs are linked to one another, because that is the only way they can display a coherent body of institutions and functioning of societies. As Gellner reminds us regarding individual objectives,
"[...] if we [...] think of the states of affairs satisfying these ends in complete isolation, as if there were nothing else in the world, we end up with something which is absurd in itself, let alone as an object of desire. Standing on a peak in a universe containing naught else, or possessing gold ingots in an otherwise empty world: each has a somewhat surrealist quality" (Gellner, 1991:34).
A suggestion of the social dimension that I point out is presented by Bourdieu (1977). The author captures the temporal dispositions of modern rationality, analyzing the participation of agents in a system that structurally stimulates a specific type of behavior. Bourdieu claims that the central issue of the conduct of homo oeconomicus is his perception of time - the axis which determines original knowledge of predictability and calculation of an imaginary future. Thus, instrumental action is determined by a unique symbolic network, based on the accumulation of indirect goods that contradict the economic notion of direct exchange of objects in pre-capitalist societies. The social representation of the system of indirect goods is based on money, which as a symbol is not in itself a source of any satisfaction. Its use is linked to a calculated future, defined by the availability of goods that are proposed by the hierarchy of personal ends. Functioning as a universal mediator, money suits a new ordering of relations between individual and society, satisfying a disenchanted temporal understanding that does not consider the future to be controlled by forces beyond individual actions.
Bourdieu uses the differences in practices of capitalist rationality in relation to traditional conduct present in Algerian society as as example of this thesis. In this way, the modern spirit of calculation is countered by the cultural system based on honor. Bourdieu demonstrates that work has a social function in Algerian communities that is mostly disconnected from economic ends. Laborious activities are valued in themselves, and occupations have a character that restricts the distinctions between profitable or non-profitable, productive or non-productive. The main point is the thesis that modern rational condition is not an objective "choice", but a historical and complex phenomenon. On the other hand, modern agents are not constituted through a simple passive and forced accommodation to "real" economic mechanisms. Historical movement displays many processes through which traditional habits are maintained in parallel to capitalist ordering, which reveals a "creative transformation" and the interdependence of individuals and socioeconomic structures (idem:11-18).
The restricted character of instrumental practices may be verified also by their contrast with the characteristics of the most important political meetings of the Douala African community, which occur with the consumption of champagne, or with the fact that workers of the Kigali community gather their savings to spend it in beer parties (Latouche, 2001:28-29). Similarly, in most Indian tribes work is done not only as a rational mechanism for feeding, but above all as a moment of religious activity, partying, or dancing (Castoriadis, 1982:40). Work in these contexts seems subordinated to symbolic goods that are not liable to being quantified or calculated. Thus, there is no "exchange" without prior identification and recognition of the participants of the interactions (often based on family ties). Far from being "irrational", these behaviors reflect cultural dimensions that reinforce solidarity, while the accumulation of material goods is seen as innocuous and a nuisance to individuals (Latouche, 2001:19-21). These orientations do not represent some sort of illusion or mystification, and their explanation rests on the sensibleness of it within its own context.
Considering the insertion of the notion of rationality in a social context that gives it some meaning, the arguments that objectify instrumental behaviors may be questioned. On the other hand, the type of instrumental social configuration expresses the relations in which modern agents' practices are based. The context of production, essential to the economic field, receives a meaning in modernity that leads to unique experiences, characterized above all by maximizing behavior. It is essential to point out that instrumental action and the modern economy depend on certain institutions, especially on the generalization of markets in which goods and services, including work and capital, are defined by prices and by free supply and demand (Polanyi, 1957:247). Only in this environment does the idea of scarcity of means acquire relevance, in this case represented by the universal and quantifiable mechanism of money, which sums up all personal ends.
Therefore, a more promising sociological outlook is to understand agents' interests through collective routines, habits, and rituals. The production of social institutions is now seen as a complex process, and disconnected from the assumptions that consider it a result of individual objectives. In this way, social norms reflect a structuring domain, influencing individuals' identity and meaning, as well as their respective activities. The consequence of this influence is that institutions sometimes acquire such power that they begin to be seen as laws, whether defined by formal rules or by customs and beliefs. Therefore, it becomes clear that the cultural power of instrumental behavior comes from a social dimension that induces, and gives legitimacy to, the discourse of an agent that is a natural bearer of fundamental rights and has the competence to make autonomous choices.
This type of analysis shifts rationality from the individual sphere to the extensive rules that show that modern individualism has an institutional history produced by theories that assume liberal democracy in the political field, free markets in the economic field, and a model of self in the cultural and religious fields (Meyer, 1987). One then realizes that modern social systems prescribe, through economic, religious, and political elements, a basis for human beings to work their subjectivity in an apparently independent way. In this model of analysis, the individual does not appear only as a subject, but also as an object of the project of rationalization that organizes a compulsory subjectivity composed of a legitimate type of private life. It is these elements that give meaning to instrumental rationality, emphasizing personal reasons and aspirations, among which: that the actor must find self-esteem in himself, staying away from moral standards; that he must have a sense of efficiency and initiative; that he must relate to the environment in a way that preserves his freedom and independence (idem:253).
Hence, human actions express subordination to a social dimension where the concept of economic action as an independent sphere loses its meaning. One way of realizing this is to pay attention to the "institutionalization of the economic process" (Polanyi, 1957:249-250), which inserts the economy into the most diverse social and symbolic forms and dependent on the historical context. The issue becomes more complex when instrumental practice acquires relative freedom in relation to other social phenomena in modernity, which is reflected in the institutionalization of private property as a central mechanism of social regulation. While exchanges and circulation of material goods occur through a complex classification in pre-modern economies, in the modern economy market functions determine decentralized forms of behavior (Godelier, 1974:174). The classical studies in sociology show that when there is more social division of labor, economic functions tend to gain more preponderance, which is reflected in the paroxysm of a production that is not controlled by its direct producers. It is only in this particular historical context, in which goods are individually appropriated, that the cultural conduct of maximization and productive profit makes sense. However, this must be explained and not reified into categories that suggest a concept of human nature. To take economic activity as an independent and universal sphere requires substitution of the analytical emphasis in social relations to an objectification of the material exchange of man with nature. It is in this sense that critical theory questions instrumental action, which supposedly absorbs the liberalizing or "negative" characteristics of thought. As has been generally said in studies by Adorno, Horkheimer, Marcuse, and Habermas, the conducts of maximization tend to transform rationality into technical reason. The result is a reflection connected to mechanisms of control and perpetuation of social relations directed to profitable efficiency and to monopolistic standardization. The dichotomous concepts of "critical reason" and "instrumental reason", "subjectivity" and "apparatus", "administration" and "negative thought", "life-world" and "system", express the dialectic of rationalization, reflecting the contradiction between human control of nature, or the potential for autonomous action, and the subordination of the internal and creative condition of individuals.
On the other hand, one must observe that, even with the stimulus of individualization and the relative independence of economic action brought about by the modern condition, there is no way to determine the borders of the symbolic dimension, as if it were possible to establish a clear difference between essential and secondary instances. That is, a separation between instrumental orientation, the true substance of society, and phenomena that are considered irrational and dispensable. The "utility" that individuals try to satisfy in their actions is related to a cultural code of concrete properties that governs both production and the dynamic of goods displayed in the market (Sahlins, 1976:166). This code demystifies the rational choice idea that reality is the aggregate result of exchanges occurring through autonomous interests, in which prices and supply and demand are the only elements that structure those transactions. Even if no society can live without the production of material life and its own reproduction, these are not dictated by a natural or rational law. There is always a margin of indetermination in which the field of history is inserted. Social movement is produced through meanings that manifest "correlative distinctions of what has worth and what does not (in all meanings of the word "worth", from the most economic to the most speculative), of what should be done and what should not" (Castoriadis, 1982:176).
Even though it is different and original, individual decisions and the construction of functions of utility in the modern social context depend on the social significance of consumption and of goods displayed in the market. This significance establishes contrasts that represent, and give legitimacy to, the inequalities of power present in society. Therefore, there is no free instrumental behavior, because even the sphere of consumption, which the theory of rational choice tries to classify in an attempt to predict actions, does not appear as a subjective option, since it is a process structured by social values and by a symbolic system that determines attitudes and forbids certain objects. It can be said that modernity is characterized by a new sort of totemism (Sahlins, 1976:178), composed by a predetermined system of objects that expresses a discourse of differences. Cultural order is defined through the differential characteristics of the goods on display, which are classified in a way that determines the type of consumption of each group or class.
In this way, we can question the basic reasoning of rational choice theory, which has a minimal concept of agency based on the idea of an independent capacity to produce decisions. This understanding asserts a separation between individual and structure that restricts the perception of whatever is outside personal cognition. However, the theory must explain the existence of organizations that are not reducible to individual practices and that reflect a subjectivity composed of many actors (Hindess, 1988:103-104). In many social instances, such as a productive corporation or a state apparatus, there is no possibility of understanding decisions through evaluation of objectives that are individually defined. In these cases, choices are a product of decisions made in many places and the results subsist on their own. Choices tend to express institutionalized forms of evaluation and information gathering, causing routine and habitual orientations that are distant from instrumental reflection.
Therefore, the biggest problem in the sociology of rational choice is the assumption that action follows a transparent path traced by individual objectives. In opposition to this ideal, it is suggested that agents usually base themselves on beliefs or habits that relate, in a specific way, to personal desires and social structure. In this way, the means of reaching a decision should not be prejudged in the way the models of economic rationality suggest. Even the type of questions that an individual poses to himself and the conclusions he himself reaches display mechanisms or forms of specific thought present in the community of which the individual is a part. The essential point is to identify the existence of something outside the individual sphere which establishes unique relations between agency and structure. It is exactly the understanding of these relations that may reveal the forms of thinking and the available, legitimate, and justified means for people's actions.
AXELROD, Robert. (1984), The Evolution of Cooperation. New York, Basic Books.
BOUDON, Raymond. (1991), "Individualisme et Holisme dans les Sciences Sociales", in P. Birnbaum and J. Leca (eds.), Sur L'Individualisme. Paris, Presses de la Fondation Nationale des Sciences Politiques.
___. (1999), Le Sens des Valeurs. Paris, PUF.
___. (2002), "Théorie du Choix Rationnel ou Individualisme Méthodologique?". Sociologie et Sociétés, vol. XXXIV, nº 1, pp. 9-32.
___. (2003), Raison, Bonnes Raisons. Paris, PUF.
BOURDIEU, Pierre. (1977), Algérie 60: Structures Économiques et Structures Temporelles. Paris, Les Éditions de Minuit.
CASTORIADIS, Cornelius. (1982) , A Instituição Imaginária da Sociedade. Rio de Janeiro, Paz e Terra.
CHERKAOUI, Mohamed. (2003), "Les Transitions Micro-Macro. Limites de la Théorie du Choix Rationnel dans le Foundations of Social Theory". Revue Française de Sociologie, vol. 44, nº 2, pp. 231-255.
COLEMAN, James S. (1990), Foundations of Social Theory. Cambridge, Harvard University Press.
DOMINGUES, José Maurício. (1995), Sociological Theory and Collective Subjectivity. Hampshire, Palgrave.
ELSTER, Jon. (1983), Sour Grapes. Studies in the Subversion of Rationality. Cambridge, Cambridge University Press.
___. (1984), Ulysses and the Sirens. Studies in Rationality and Irrationality (revised edition). Cambridge/Paris, Cambridge University Press/Éditions de la Maison des Sciences de L'Homme.
___. (2000), "Rational Choice History: A Case of Excessive Ambition". The American Political Science Review, vol. 94, nº 3, pp. 685-695.
GELLNER, Ernest. (1991), "L'Animal qui Évite les Gaffes, ou un Faisceau d'Hypothèses", in P. Birnbaum and J. Leca (eds.), Sur L'Individualisme. Paris, Presses de la Fon- dation Nationale des Sciences Politiques.
GODELIER, Maurice. (1974), Rationalité et Irrationalité en Économie - II. Paris, François Maspero.
GREEN, Donald and SHAPIRO, Ian. (1994), Pathologies of Rational Choice: A Critique of Applications in Political Science. New Haven, Yale University Press.
HARDIN, Russell. (1988), Morality within the Limits of Reason. Chicago, The University of Chicago Press.
___. (1995), One for All. The Logic of Group Conflict. Princeton, Princeton University Press.
HEAP, Shaun Hargreaves and VAROUFAKIS, Yanis. (1995), Game Theory: A Critical Introduction. London, Routledge.
HINDESS, Barry. (1988), Choice, Rationality, and Social Theory. London, Unwin Hyman.
LATOUCHE, Serge. (2001), La Déraison de la Raison Économique. Du Délire d'Efficacité au Principe de Précaution. Paris, Éditions Albin Michel.
MEYER, John. (1987), "Self and Life Course: Institutionalization and Its Effects", in J. Meyer, J. Boli, F. Ramirez and G. Thomas (eds.), Institutional Structure: Constituting State, Society, and the Individual. Newbury Park, Sage Publications.
OPP, Karl-Dieter. (2001), "How do Norms Emerge? An Outline of a Theory", in R. Boudon, P. Demeulenaere, R. Viale (eds.), L'Explication des Normes Sociales. Paris, PUF.
PARETO, Vilfredo. (1968) , Traité de Sociologie Générale. Genève, Librairie Droz.
PARSONS, Talcott. (1968) , The Structure of Social Action: A Study in Social Theory with Special Reference to a Group of Recent Writers. New York, Free Press.
POLANYI, Karl. (1957), "The Economy as Instituted Process", in K. Polanyi, C. M. Arenserg and H. W. Pearson (eds.), Trade and Market in the Early Empires. Economies in History and Theory. New York/London, The Free Press/Collier-Macmillan Limited.
RAPOPORT, Anatol. (1980) , Lutas, Jogos e Debates. Brasília, Editora UnB.
SAHLINS, Marshall. (1976), Culture and Practical Reason. Chicago, The University of Chicago Press.
SIMON, Herbert A. (1967), Models of Man. New York, John Wiley & Sons.
SMITH, John Maynard. (1982), Evolution and the Theory of Games. Cambridge, Cambridge University Press.
TAYLOR, Michael. (1987), The Possibility of Cooperation. Cambridge, Cambridge University Press.
ULLMANN-MARGALIT, Edna. (1977), The Emergence of Norms. Oxford, Oxford University Press.
WEBER, Max. (1968), Economy and Society. New York, Bedminster Press.
___. (1996) , A Ética Protestante e o Espírito do Capitalismo. São Paulo, Livraria Pioneira Editora.
WOLFELSPERGER, Alain. (2001), "La Modélisation Économique de la Rationalité Axiologique. Des Sentiments Moraux aux Mécanismes Sociaux de la Moralité", in R. Boudon, P. Demeulenaere e R. Viale (orgs.), L'Explication des Normes Sociales. Paris, PUF.
(Received for publication in September 2006)
(Definitive version in October 2007)
* This article is a revised version of chapter III of my doctoral thesis "Rational Choice as Social and Political Theory: A Critical Interpretation". The research enjoyed the support of FAPERJ (Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro) and CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior).
1. Part of this debate occurs through Herbert Simon's (1967) concept of bounded rationality, according to which individual choices are characterized by the limitation of cognitive capacity, based on lack of knowledge and information. In these conditions, agents adopt a conduct of "satisfaction", instead of an unbounded maximization. Simon (idem:271) describes satisfaction as a mechanism based on short sequences of decisions, which presuppose subjects' inertia in uncertain conditions. Taking into account psychological aspects, it is suggested that there are internal restrictions to instrumental conduct, way beyond external restrictions such as prices, supply, and demand. However, one must point out that the behavior of satisfaction is a premeditated reflection of how to construct optimal decisions in environments of limited information. In spite of the major influence of Simon's works in the field of rational choice, I do not delve deeper into his theses in this article due to the fact that it reinforces, above all, non-sociological aspects of a psychology of action and of individual construction of mental models.
2. Another essential study on trust and the possibilities of cooperation in social contexts that may dispense with formal state mechanisms is that of Michael Taylor (1987).
3. Rapoport is the formulator of the strategy of tit for tat, applied by Axelrod. The author is decisive in the debate on some limits of the traditional concept of rationality used by game theory. However, his critical assertions refer, substantively, to psychological factors, accentuating the sensorial aspects of perception, of individual perspective, and of mental persuasion.