SciELO - Scientific Electronic Library Online

vol.5 special edition author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Revista Brasileira de Ciências Sociais

Print version ISSN 0102-6909

Rev. bras. ciênc. soc. vol.5 São Paulo  2010


Marxism and elitism: two opposite social analysis models?


Marxismo e elitismo: dois modelos antagônicos de análise social?


Marxisme et élitisme: deux modèles antagoniques d'analyse sociale?



Adriano Codato; Renato M. Perissinotto

Translated by Thiago Gomide Nasser
Translation from Rev. bras. Ci. Soc., São Paulo, v. 24, n. 71, out. 2009.




The purpose of this article is to contrapose the propositions on power, class and political domination presented by a particular interpretation of Marxism - structuralist Marxism - through a critical dialogue with one of its most paradigmatic authors: Nicos Poulantzas. The article states, against Poulantzas suggestions, that the insertion of the concept of "élite" in theoretical Marxism may produce positive effects on it, specially making the classist analysis of politics scientifically manageable.

Keywords: Marxism; Élite theory; Social theory; Nicos Poulantzas; Class analysis.


Este artigo contrapõe-se às proposições sobre poder, classe e dominação política de classe elaboradas por uma vertente particular do marxismo - o marxismo estruturalista -, por meio de um diálogo crítico com um de seus autores paradigmáticos: Nicos Poulantzas. Defendemos que, ao contrário do que sugere Poulantzas, a introdução do conceito de "elite" no interior do marxismo teórico pode ser produtiva para o desenvolvimento dessa perspectiva de análise social, tornando a abordagem classista da política operacionalizável cientificamente.

Palavras-chave: Marxismo; Teoria das elites; Teoria social; Nicos Poulantzas; Análise de classe.


Cet article s'oppose aux propositions sur le pouvoir, la classe et la domination politique de la classe élaborés par un volet particulier du marxisme - le marxisme structuraliste -, au moyen d'un dialogue critique avec l'un de ses auteurs paradigmatiques: Nicos Poulantzas. Nous défendons que, à l'opposé de ce que suggère Poulantzas, l'introduction du concept d' "élite" au sein du marxisme théorique peut être productif pour le développement de cette perspective d'analyse sociale, de façon à permettre que l'abordage classiste de la polique soit scientifiquement opérationnalisable.

Mots-clés: Marxisme; Théorie des élites; Théorie sociale; Nicos Poulantzas; Analyse de classe.




When the issues of power, politics and the state regained the attention of Marxist sociologist in the late 1960s, in what can be aptly described as the first "institutionalist" uprising against the behavioral revolution and its culturalist offshoot, as well as against the theory of political systems and the ideological mirages of liberal pluralism, Nicos Poulantzas seized the occasion to defend the theoretical purity of theoretical Marxism. Both in Political Power and Social Classes, published in 1968, and in the polemic which thereafter he engaged in with Ralph Miliband in the pages of the New Left Review, Poulantzas criticized the analytical, political and ideological impertinence that was bringing in the problematic of political elites into Marxist theory (cf. Poulantzas, 1971, vol. II, p. 154ss; 1969).

Essentially, the arguments for this refusal were based on the following: the functioning of the capitalist state must be explained based on the objective (and not subjective, i.e., interpersonal) links between this political institution and class structure (Poulantzas, 1969); thus, whoever controls, manages and occupies the main nodes of power within the state apparatus (the "bureaucracy"), regardless of social origin, faith or specific motivations, has no choice but to reproduce the objective function of the state, which consists of maintaining the social cohesion of a given social formation (Poulantzas, 1971); this is equally valid for any type of political regime (bourgeois democracy, military dictatorship, fascism, authoritarianism) in which those in command of the political administration of the state are sensibly different (Poulantzas, 1970, 1975, 1978). It therefore follows that the central problem for any researcher of Marxist bent ought to be "what are the social relations of domination being reproduced by the state?", and not "who decides?" or "who governs?", being that these questions are lesser or less important compared to the first one.

Bearing in mind Poulantzas's critique of the theory of elites, our main goal in this article is to discuss these aforementioned propositions and advance an argumentation that can be used to verify to which extent it is possible to conduct social analysis in such a way that combines these two theoretical traditions, in spite of their remarkable ideological differences.

Pierre Birnbaum summed up rather paradigmatically this intellectual disposition which is also our theoretical parti pris. According to Birnbaum, only with "the study of French politico-administrative personnel" can one arrive at a "better understanding of the nature of the State in France" (1994, p. 11). The choice of this sort of object of investigation is based on a hypothesis (one that is, by definition, refutable) which would allow the social scientist to avoid two very common temptations of political sociology:

In order to avoid the traps of purely structural conceptions, which by means of metaphors spare themselves from using empirical procedures, but also without reducing the social system to the sum of individuals acting in a more or less voluntary manner, it is fundamental to recall that the action of the State, as an institution, depends greatly upon the people who run it (Idem. p. 11; italics added).

It is necessary, before moving on, to present the backdrop to this discussion concerning "elite" and "class." It essentially concerns how must we come to grips with the affinity (or divorce) between Marxism, as a theoretical system, and the social sciences, as a diverse ensemble of theories, methods and research techniques. There are, in this regard, three mutually excluding possibilities: Marxism as social science; Marxism against "bourgeois" social science; and Marxism as a parallel, revolutionary, and superior science/philosophy/ideology as compared to conventional social science1.

This third variant explicitly assumes that Marxism is at once a "correct" view of the world, a "privileged" point of view, etc., and a social science par excellence, although not only. It follows therefore that it is possible to diminish or simply ignore the debates, issues, methodological advances and conceptual innovations of non-Marxist sociology and political science and whatever else is not useful or does not contribute towards the overcoming of class-based society.

The second variant - Marxism contra social science - requires one to think of the former as a warranty for scientificity and objectiveness against the widespread diffusion of theories which, in the guise of "sociology" or "political science," are, in reality, more or less competent ideological rationalizations of partial points of view and/or undisclosed vested social interests. This sort of epistemological surveillance would also be more efficient in correcting "errors," "deviations" and inherent flaws within Marxist theory insofar as there would be a lesser need to resort to other intellectual traditions.

The first variant - Marxism as a "normal" social science - is the point of view which allows for a true dialogical relationship with other non-Marxist or explicitly anti-Marxist social theories. This implies several consequences, the least of which is the risk of sliding into the "ideological" terrain of the opponent. Understanding Marxism as a strand among many others within the social sciences implies submitting its own postulates to empirical testing, accepting certain premises of rival social theories and incorporating - with or without modifications - some concepts in order to advance scientific research.

We have divided this article into four sections. The first one is dedicated to translating the chief issues of the theory of elites for Marxism, or, to be more precise: to expound them in the official language of theoretical Marxism in order to outline not their dissimilarities (which are somewhat obvious), but, rather, the fundamental differences which set apart each of these theoretical models. In the second section, we will synthesize the main troubles pointed out by Poulantzas in how the elitists conceive of the relationship between the political and social worlds. The third section briefly presents the theoretical solutions Poulantzas himself provided in order to overcome such troubles. In the fourth and final section we suggest a few conditions if we are to resume (or in fact inaugurate) the dialogue between elitism and Marxism, pondering what should be preserved or discarded from the sociological critique the latter addressed to the former.2


Power, class (dominant) and bureaucracy

Nicos Poulantzas made the claim in Political Power and Social Classes that political problems, such as those traditionally laid out by the theory of elites (to spell them out: who wields power in a community? How many political groups are there? Where does their power com from? etc.), "can only be resolved within the scientific problematic proposed by Marxism" and that to this end it would be necessary to return to "scientific indications which Marx, Engels, Gramsci and Lenis provided us in this respect" (1971, vol. II, p. 155 and p. 154, respectively).

From this perspective, how should these very same problems be expressed in the language of Marxism, that is, according to its conceptual framework, and what was the theoretical solution Poulantzas devised for them?

Let us consider, first of all, the problem of the dominant class, which, on its turn, can be subdivided into two enigmas: i) is there in fact a politically dominant class, or is political life simply the result of the clash between countless interest and pressure groups who detain more or less equivalent portions of power?; and ii) is this class politically dominant class the same one which dominates economically?

The theory of elites is largely known as a critique of the Marxist theory of a dominant class and an attempt to refute the hypothesis according to which political power or, more appropriately, "the political resources of the dominant class" is derived from their economic power - or, to be precise, "the possession of economic resources" (Saes, 1994, p. 11). What is more, the neo-elitists will argue that the transformation of capitalist system since the mid twentieth century (the separation between property and the means of production, effective social mobility among groups, the decentralization of government functions, the transformation of individuals into a "mass," the new roles taken on by state bureaucracy and so on) rendered obsolete the idea of a class that is at once politically and economically dominant. This is the well-known stance taken by C. Wright Mills, for example (cf. Poulantzas, 1971, vol. II, p. 155-156).

Secondly, there is the problem of the state bureaucracy, one that implies an array of additional problems: i) what is the connection between the state bureaucracy and the dominant class?; ii) is this connection simply instrumental, in that the latter controls the former?; iii) or, quite the contrary, is this a reciprocally autonomous relationship, in which both stand independently?; and iv) if this is the case, do the bureaucracy and the other elites (military, political, technical etc.) who command the state's administrative apparatus have a political power of their own?

For classical elitism, political power, held and wielded by an autonomous bureaucracy (embodied by the "upper echelons" of the state and high-level cadres) is considered at any rate parallel to (political and economic) class domination and often independent from economic power. It could be said that these problems were either hidden or forgotten by the political sociology of the twentieth century as a result of the widespread use of the "political class" formula and whatever term later came as a substitute - power elites, governing class, ruling class - although never resolved.

The choice between the expressions "dominant class" and "political elite" is not, however, merely a matter of terminology. There are at least three questions to be dealt with in this regard. One of them is more theoretical in nature, which is the question of the foundation of political power. Where is power derived from? From the state itself, in that it is considered the exclusive source of political power (as in Weber and Michels)? Or can power be derived from parallel (and not more important) sources of power, such as economic might? There is a second, more empirical question which concerns the division of political power: is there a unity among elites (as Mosca, Michels, Mills and Meynaud, inter alia, argue) or rather a plurality of elites (as per Parsons, Aron, and Dahl)? And finally there is the question of the relationship of political and economic power: in Marxist terms, how to think of complex links between the political (level) and the economic (level)? This third question is, in sum the question of "representation": whether political, bureaucratic, and scientific elites represent themselves (their own interests) or social class interests.

Yet, before listing structural Marxism's theoretical solutions for this agenda of questions we shall briefly turn to Poulantzas's reproaches to elite theory.


The theoretical critique to elite theory

Poulantzas emphasized that the criticism the elitists inveighed against Marxist theory either refer to or are an outcome of "poor interpretations of Marxism." Whatever is the case, the specific problems these critiques raise - that of the dominant class, the state bureaucracy, the relationship between them and their sources of power - are far from being resolved by the "ideological perspectives" of classical elitism (1971, vol. II, p. 154-155). These problems would in fact result from a series of mistakes made by the elitists. What are these mistakes?

The first mistake made by the elitists was to suppose that Marxism assumes that there is an "empirical concentration of all the political functions in the hands of the politically-economically dominant class," being that power exercised, in practice, by "members of this same class" (Idem, p. 155).

This assumption does not, however, take into account the separation, postulated by classical Marxism, between state power (i.e. the social power exercised through the institutions of the capitalist state), effectively held by the dominant classes or fractions, and the state apparatus, which is where this power is exercised, and which can be occupied and operated by any other social category (the middle strata, the petit bourgeoisie etc.).

The second mistake is a sequel to the first one. There are two versions of the critique of the elitists to the Marxist conception of the dominant class. One argues there is a plurality of elites - being that these groups are defined according to the control positions they occupy in different realms of social life (hence, labor elite, party elite, religious elite etc.); the other claims that there is a unity of political elites.

"Elitist pluralism," represented by Dahl and Schumpeter, for example, is based on a factual assessment: the upper strata of different social groups (politicians, bureaucrats, union leaders, entrepreneurial leaders, etc.) barely have - and therefore do not represent - the same interests, neither do they have any political unity. The problem of this conception is that, in reality, it divides political power (Idem, p. 158), something that is, by definition, non-sharable. However, this conception admits and postulates certain foundations of power distinct from those assumed by Marxist and this, according to Poulantzas, is its main flaw. Another crucial shortcoming has to do with the fact it does not take into account the unity of political power and the centrality of the power of the state (and not of any other "powers") in capitalistic social formations.

The "elitist monism," a version of this theory which accepts and argues in favor of the unity of elites, is at any rate included in the original Marxist problematic of political domination, notwithstanding its rejection of the concept of the "dominant class." In its place, and as a result of the historical transformations of capitalism, it suggests the existence of a super elite. The cohesion of the social group who composes this new political group is conceived of distinctly (and erroneously) by Mosca, Michels, Meynaud, Wright Mills - either in terms of a unifying center, or in terms of the ascension of a new social group (the "administrators"), or, not least, in terms of the domination of one elite in particular over others. Whichever direction is taken, its power can derive both from the control one group has over relations of production, and as the control of the state apparatus itself (which can cumulatively assume control of economic power). These formulations, stresses Poulantzas, not only do not avoid escaping determinism - a common accusation aimed at Marxism - but also restore in its explanations economic overdeterminism (Idem, p. 158-159).

This is, in summary, the Poulantzian critique of elitism, from which it is already possible to deduce the premises, postulates and principles of this political science and which thwart any possible conceptual exchange with other traditions unwilling to accept this evidence. Let us take a closer look at this problem building upon the theoretical solutions proposed by Marxism in order to understand the relationship between social power and political power.


An alternative conceptual system

How does, on its turn, Marxist political theory, according to Poulantzas, conceive of the question of class dominance and also the question of the state bureaucracy?

The concept of "dominant class" is, Poulantzas reminds us, a lot more complex than the caricature outlined by Wright Mills. It is possible to read among the classics of Marxism countless analyses which point out to the dissimilarity and mismatch between the economically dominant class and the politically dominant class3.The actual ("empirical") concentration of politico-administrative functions in the hands of dominant classes and fractions is not mandatory (that is, an historical constant). What is more, its non-coincidence can only be explained by Marxism as a result of a thorough understanding of this problem based on the variations promoted by class struggle, by the forms of the state and by the forms of regimes in concrete social formations (Idem, p. 161-162).

Another inescapable topic is the question of belonging to the class of state bureaucracy. Raising the problem of "bureaucracy" only makes sense if the decisive difference between the state apparatus and state power are kept in mind (Idem, p. 164). Briefly put: the state apparatus is where power is exercised from; state power is power that is detained by dominant classes and fractions who benefit from decisions taken by the state.

Building upon a narrower, stricter definition, according to which the "ruling class" (or, more appropriately, the hegemonic fraction or class) is the one whose political interests is ensured to a great extent by state policies, Poulantzas advances two working notions in order to deal with this problem: the notion of a "ruling or hegemonic class" and a "class ‘in charge of' the state apparatus."4 The class in charge (which generally is inaccurately designated the dominant political class) is the social class which controls and administrates the centers of power within the state apparatus - and not the one which actually holds political power (Idem, p. 165). Who holds power is, by definition, the dominant classes. The class in charge of the state apparatus my or may not identify itself with the hegemonic fraction. The ruling class is the social class or fraction which is predominant in the political scene - and therefore assumes "the role of political representation" - as result of the political party game (Idem, p. 162).

All these differentiations are all the more important when we become aware of the misconceptions that can arise from not knowing them. For example: "if we place ourselves in the realm of the political scene with the intent of discovering class relations, reducing them to mere party relations, we are inevitably led to mistakes […]" (Idem, p. 73, authors' emphasis). This is because in the real political process there might be a wide range of available alternatives. As a general rule, the actions of the ruling class or fraction disguise its role as the hegemonic class or fraction in the political scene. However, there is the possibility, for instance, of a class of fraction in the bloc of power that does not necessarily have its own party-style organization or that does not make itself present in the political scene in this way; a class or fraction of a class can disappear from the political scene yet continue to exist in the power bloc; there can possibly be a class of hegemonic fraction in the political scene that is different from the class or hegemonic class in power bloc; "the ruling class or fraction […] [in the political scene] can not only not be [the class of hegemonic faction], but can even, at times, be absent from the power bloc" (Idem, p. 76).

On its turn, a "displacement of the index of hegemony from one class or fraction to another in the power bloc does not necessarily involve displacements of party representation in the political scene" (Idem, p. 74); the dislocation of the hegemony of a class or fraction to another power bloc does not "necessarily correspond […] to backdoor passages to the political scene (Idem, ibidem). The power bloc can ultimately express itself in the political scene through party alliances or even through direct confrontation between parties (Idem, p. 76).

Although this complicated equation, intelligently deduced by Poulantzas from Marx's analyses of European nineteenth-century politics, might correct the more simplistic views of the political phenomenon (and, by extension, many simplifying views of Marxism as whole), it does not nevertheless account for certain phenomena that are exclusively political or that can be reduced - or deduced from - class analysis.


A Defense of Dialogue in Social Research

Let us remember that the broader goal of this article is to (re)establish a dialogue between Marxism and elitism, which was interrupted after Nicos Poulantzas advanced his critique of elite theory (summarized above) and the reproaches to its uncritical incorporation by Ralph Miliband in The State in Capitalist Society (cf., in particular, Poulantzas, 1969). This aim, however, in no manner implies a simple refusal of these critiques. In fact, Poulantzas is correct regarding three important points:

There is no doubt that elite theorists, both classic and contemporary, criticize Marxism based on a caricature - a very crude one at that - of what this theory often portrayed as is made to be. Most often, Marxism is perceived as economicism, that is, a theory according to which political agents act at the behest of "economic interests" or, more appropriately, of economic agents. This deformation of Marxism, a maneuver which facilitates uncomplicated rejection can be found both in Gaetano Mosca (1939) and Raymond Aron (1991) or Pierre Birnbaum (1994). It would thus be important to reestablish the theoretical principles of Marxism before opposing it (or, from our perspective, connecting it) to elitism.

Secondly, Poulantzas was correct when he criticized the elitists for not providing a theory of the state - the ultimate center of political power. Overly preoccupied with the "subjects" of power, the elitists are incapable of thinking the state as an institutional structure (agents, apparatuses, roles, center of power, etc.) which serve the purpose of restraining decision-makers from acting according to their whims. Thus, we never know for sure the place and function of the state apparatus and its operators, the "state elite" in the reproduction of social and political domination.

Lastly, it is undeniable that elitism suffers from two interrelated limitations. On one hand, exaggerated weight is given to the autonomous power of the political elites, seen as a group responsible for the conducting of human communities. In this sense, the elitist approach is compromised by an excessive degree of voluntariness, being insufficiently able to account for the structural constraints which limit the actions of elite groups. On the other one, this theory is excessively focused upon the self-interests of "politically active minorities" and thus tend to shy away from choosing the possible (and in fact frequent) relationship between the behavior of the elites and certain outside interests as the object of analysis. Adding up these two shortcomings and pushing them to the limit, the "politically active minorities" seem to act in something of a social void. Thus, the elitists do not go beyond the analysis of either the elite-mass relationship (latter never being rigorously defined beyond conventional preconceptions), or intra-elite relationships. Social classes, as much as their empirical existence is acknowledged, are not taken into account in the explanation of political domination since they are considered aggregates that are overly broad and/or they do not ultimately produce politically important effects.

Are these flaws reason enough to suspend the dialogue between and elitism? We do not believe so and in order for this proposition to become accepted one must refute some other critiques Poulantzas applied to the theoretical problematic of political elites. We have three points particularly in mind, laid out and discussed below.

(i) The problematic of the agents of power and the source of power can only be resolved within the theoretical realm of Marxism. This observation comes with a caveat. These problems could only be resolved, according to Poulantzas, in a realm of theoretical Marxism, or at least what Poulantzas deemed theoretical Marxism to be. The fact that other Marxists incorporated some of the issues and concepts of elite theory - Miliband, Bottomore, Domhoff, for example - reveals that this statement is, to say the least, debatable. However, in addition, or yet, as a precondition, one must question whether Marxism can in fact account for some of these issues, regardless of one's understanding of this theory.

As discussed above, Poulantzas sought to resolve the problem of the relationship between agents of the state (the political and/or bureaucratic "elites") and the political domination of a certain class or fraction based on the conjugation of two concepts: "ruling class" and "hegemonic class (or fraction)." The concept of class or hegemonic fraction would supposedly identify the class that is systematically benefited by state policies, regardless of whether or not this class or fraction is a collective political agent able of organizing itself effectively (as the "ruling class" in the political scene, for example). In this sense, it is theoretically (and possibly empirically) different from the class "in charge of" a concept which describes the agents who directly control state positions (the "state elite", in Miliband's definition).

However, this conceptual framing only solves the problem by eliminating it arbitrarily. Given the fact that the class is a simple effect of the state structure and, thus, is condemned to carry out the imperatives of the objective function of the capitalist state (the reproduction of a capitalist social formation) it ultimately does not require analysis. Miliband (1970) was correct in claiming that, for Poulantzas, the state and its agents can only be seen as autonomous (confronted with the hegemonic fraction) under the condition they are mere automatons, that is, they completely lose their autonomy (confronted with the objective imperatives of the "capitalist" system) and, thus, lose, once and for all, their importance as an object of study. This lack of importance, however, is an arbitrary derivation of certain theoretical postulates, and not the product of historical, empirical analysis which prove them.

The studies on political elites overwhelmingly and convincingly demonstrate the scientific value of studying "politically active minorities" due to the (not necessarily intentional effects) of their actions and strategic options can have on the social system (cf. Guttsman, 1965; Keller, 1971; Carvalho, 1980; Czudnovski, 1982; Perissinotto; 2000; Hunt, 2007; inter alia). Let us recall the starting point of this article: if it is in any way possible to establish an interrelation between the nature of political/state elites (or the "class ‘in charge of'") and their decisions, and, on the other hand, between these decisions and the effects they produce upon the social system (are they reproducing effect or not; anticipating or nor; in favor or not of the hegemonic fraction), therefore one must acknowledge the importance of elites as an object of study, which includes its importance in the analysis (or "proof") of the reproduction/transformation of class domination relationships.


(ii) The concept of "elite" cannot effectively account for the problem of domination since it does not take into consideration the problem of society's class structure. The concept of elite (or "political class" or "oligarchy" and so on) was notably borne out of the explicit objective of refuting the concept of class as a little or non-workable theoretical notion. However, this justification need not be accepted and, consequently, Marxists do not need to reject "their" concept with no further ado.

We should not imagine that the concept of elite and its many specializations - political, economic, intellectual etc. - can have a role that is parallel or analogous to the concept of class within Marxism. This seems to be the case of Ralph Miliband (1972) and Tom Bottomore (1974). According to both, the concept of elite is useful insofar as it explains some social realities to which the concept of class cannot be applied to or cannot be adjusted to adequately. Even if this proposition is, for the sake of reasonability, fully acceptable, we must go beyond it. In reality, from our perspective, it is more reasonable to think that the concept of elite can be useful when empirically working out the class analysis of politics5

To this end, class analysis cannot, on its turn, be reduced to a principle which conceives of classes only as objective structures which produce "pertinent effects" at the political level, despite or even preceding their constitution as effective political agents. For this reason, this mode of analysis cannot be limited to identifying the morphology of the mode of production (and its stags or phases) with the aim derive, by theoretical deduction, the political effects the class structure supposedly produces. On the contrary, if the class perspective is to become an instrument of social analysis in an empirically-oriented social science, it is necessary to consider, first of all, if and how classes are constituted, in fact, as relevant political agents.

The accomplishment of this goal entails great obstacles, as it is no trivial thing to conceive of classes as voluntary collective actors, as Olson (1999) has demonstrated. It would be necessary to at least explain how "solidarity" among class members (a common way of thinking) becomes "cooperation" (a common way of acting) (cf. Kaplan e Lasswell, 1998, pp. 60-61).

The Marxists could refute these arguments by saying that neither they nor Marx defend the idea that social classes act directly in politics, as voluntary collective forces (Therborn, 1989). In fact, classes have always acted through "mouthpieces," that is, through parties, unions, civil associations and other institutions capable of speaking on behalf of classes. Once can easily notice that this move, instead of settling the score for good, just adds another term to the equation, as now we have an additional empirical problem: how then can it be proved that such institutions in fact represent or serve as a vehicle for the interests of the class in question?

It is exactly in relation to this problem - the problem of representation - that the concept of elite can be not only complementary to Marxism, but also important to render class analysis workable, in other words, to turn it into a useful tool in social science. "Class" can only be constituted as an analytically fruitful concept if we abandon for good the idea that it acts directly in politics. Stated otherwise, adequate use of this concept seems to require that we consider class as a collective entity that is "represented" in the political realm by a "politically active minority," as argued by Therborn (Idem, pp. 437-438). The problem then turns out to be how exactly to detect the relationship of representation of class in day-to-day political struggle, yet without resorting to the "key that opens all doors" of the "objective functions" of the state or the "intrinsic logic of the mode of production."

As we see it, class analysis of the political dynamic requires following three procedures, ordered in a hierarchy of importance, so as to prove the hypothesis of class political representation by a minority (or, an "elite"):6 a) the study of the actual behavior of this minority; b) the analysis of the content of the manifest discourse; and, lastly, c) the study of the social origins of their members. In sum, it is necessary to know whether the members of the minority at stake act in a way that is coordinated and convenient for the interests of the class they supposedly "represent"; whether they explicitly speak "on its behalf" and whether they belong to the class in question. The designation of a hierarchy to these three methodological procedures is fundamental, since they have varying impacts on the proof of the relationship of representation between the minority and the class it supposedly (and not by definition) represents. For example: a group can be recruited from a class (thus fulfilling the requirement of social origin), but can adopt a discourse and behavior that is guided by the ideology of another social group; in another scenario, the presence of action that is manifestly and consciously guided toward the accomplishment of class objectives would be enough to establish the relationship of representation, even if the members of the minority were not recruited by the benefited class and if they profess to uphold the ideology of a third social group. Evidently, all three dimensions together - action, "spiritual" affiliation and social precedence - render proof of the existence of "class representation" even more convincing.


(iii) The elitist perspective cannot identify the foundations of political power. There is little doubt that elite theory tends to be excessively voluntaristic in its analysis of the power of political elites since it tends to neglect elements external to politics as conditioning and limiting factors vis-à-vis the power of these special social groups. This is what has come to be called the sin of "formalism."7 However, concerning this point, we can initially observe that not only Marxism is deformed by its theoretical enemies, but that Marxists themselves tend to do the same to their ideological adversaries. The assumption that every investigator which chooses political elites as an object of study is doomed to commit the sin of formalism is not accurate. Hence, it is a mistake to argue that any elite theoretician cannot identity the "true" basis of political power. They do in fact identify it, however it does not lay in class structure, but rather in other social realms/domains. In this regard, a quick reading of the typology of political classes set forth by Mosca can help overcoming what can be politely called a misunderstanding (1939, p. 53-60).

Before moving on and in order to avoid any of the inconveniences typical of this kind of confrontation, it is necessary to shed light on the precise content of terms we are discussing. If we say that the political elites do not in fact exercise political power, it becomes necessary to clearly state what is meant. It seems evident that the concept of political power, in the case of structuralist Marxism, describes the production by the capitalist state of government policies capable of reproducing class structure (or "the structure of domination") of capitalist society. In this sense, the state serves the long-term interests of the dominant class, or, to be specific, the political interests of this class in particular, interests which basically consist of the reproduction of fundamental characteristics/interrelations that constitute the capitalist mode of production. This is power in the structural sense.

There is not much doubt concerning the fact that the structure of capitalist society creates several limits to the decisions, strategies and room for action of the political elites. However, what can we do and say about an entire range of political phenomena that has no relation to the reproduction of the social order? How do we explain them? Is it worthwhile to give up trying to understand and discuss a gamut of certain political events - which tend to be the majority - just because they do not fit in what is considered essential from the structural point of view (assuming that "structural" refers to all things that have to do with the reproduction of the mode of social production)?

Our perspective is that the answer to this question ought to be no. If political elites do not hold "political power" in the strict sense defined above, they certainly must possess, to some extent (to be empirically determined), authority, force, prestige, or "political influence" capable of producing effects worthwhile examining. In fact, several studies show (for example, Codato, 2008; Fausto and Devoto, 2004; Skocpol, 1984; Putnam, 1976; Perissinotto, 2000), that often the choices made by the political elites can help us understand the configuration and evolution of a certain political formation, as well as the processes of maintenance or destabilization of the social order.

If this is correct, we therefore must strive to elaborate concepts which will allow us to analyze, building upon a class perspective, "superficial" political interactions, that is, political phenomena that are not directly connected to the problem of "long-term" social reproduction. To this end, it is perhaps the case of resorting to a narrower yet more operable conception of power, such as the one elaborated in the Weberian theoretical tradition. In this sense, power would be no more than the ability to produce intended effects and ensuring that outcomes are achieved, despite the resistance of antagonistic groups. This is power in the strategic sense.

This kind of formulation lends itself well and preferably to the analysis of strategic actions of real political life. Based on it we can follow more closely the interactions among social and political agents, without allowing these interactions to become dissolved in the long duration timeframe of the "reproduction of the mode of production." This is, as it happens, the analytical strategy adopted by Marx in The Eighteenth Brumaire of Louis Napoleon. There we can observe him following the day-to-day decision strategies of several political agents, their calculation, hesitations, and positions in the face of concrete events. The central question of political research guided by the Marxist problematic would henceforth be: to which extent the strategies adopted by different political elites can be linked to a class base? After all, it is as dogmatic to believe that classes have no effect whatsoever on political life as supposing that, by definition, they indeed should.

With this said, there is no reason - other than those beyond the theoretical domain - to consider these conceptions of power mutually exclusive. If, on one hand, it is undeniable that elites act in a structural context which restricts their margin of actions/option and redefine the sense of their strategies despite their initial intentions and "projects," on the other hand, it is not less undeniable that these elites make choices, outline their tactics, redefine decisions and calculate the reach of their possibilities of power and thereby affect the concrete dynamic of the political and social worlds. Refraining from assuming them to be the demiurge of these two worlds does not require us to see elites as mere puppets of structural determinants.



ARON, Raymond. (1991), "Classe social, classe política, classe dirigente", in _________, Estudos Sociológicos, Rio de Janeiro, Bertrand Brasil.

BIRNBAUM, Pierre. (1994), Les sommets de l'État: essai sur l'élite du pouvoir em France. Paris, Seuil.

BOBBIO, Norberto. (2006), Nem com Marx, nem contra Marx. São Paulo, Editora da Unesp.

BOTTOMORE, Tom. (1974), As elites e a sociedade. Rio de Janeiro, Zahar.

CARVALHO, Jose Murilo de. (1980), A construção da ordem: a elite politica imperial. Rio de Janeiro, Campus.

CODATO, Adriano. (2008), Elites e instituições no Brasil: uma análise contextual do Estado Novo. Campinas, tese de doutorado em Ciência Política, Universidade Estadual de Campinas - Unicamp.

CZUDNOWSKI, Moshe M. (ed.). (1982), Does who governs matter? DeKalb, Northern Illinois University Press.

FAUSTO, Boris & DEVOTO, Fernando J. (2004), Brasil e Argentina: um ensaio de história comparada (1850-2002). São Paulo, Editora 34.

GUTTSMAN, W. L. (1965), The British political elite. Londres, MacGibbon & Kee.

HUNT, Lynn. (2007), Política, cultura e classe na Revolução Francesa. São Paulo, Companhia das Letras.

KAPLAN, Abraham & LASSWELL, Harold. (1998), Poder e sociedade. Brasília, Editora da UnB.

KELLER, Suzanne. (1971), Mas alla de la clase dirigente. Madri, Tecnos.

MILIBAND, Ralph. (1970), "The capitalist State: reply to N. Poulantzas". New Left Review, 59, jan.-fev.

_________. (1972), O Estado na sociedade capitalista. Rio de Janeiro, Zahar.

MOSCA, Gaetano. (1939), The ruling class: elementi di scienza politica. Nova York, McGraw-Hill.

OLSON, Mancur. (1999), A lógica da ação coletiva. São Paulo, Edusp.

PERISSINOTTO, Renato M. (2000), Estado e capital cafeeiro em São Paulo (1889-1930). São Paulo, Annablume/Fapesp, vol. 2.

_________. (2007), "O 18 brumário e a análise de classe contemporânea". Lua Nova, 71: 81-121.

PERISSINOTTO, Renato & CODATO, Adriano. (2009, no prelo). "Classe social, elite política e elite de classe: por uma análise societalista da política". Revista Brasileira de Ciência Política, vol. 1 (2).

POULANTZAS, Nicos. (1969), "The problem of the capitalist State". New Left Review, 58, nov.-dez.

_________. (1970), Fascisme et dictature: la Trosième Internationale face au fascisme. Paris, Maspero.

_________. (1971), Pouvoir politique et classes sociales. Paris, Maspero, 2 vols.

_________. (1975), La crise des dictatures: Portugal, Grèce, Espagne. Paris, Seuil.

_________. (1978), L'Etat, le pouvoir, le socialisme. Paris, PUF.

PRZEWORSKY, Adam. (1989), "A organização do proletariado em classe: o processo de formação de classes", in _________, Capitalismo e social-democracia, São Paulo, Companhia das Letras.

PUTNAM, Robert D. (1976), The comparative study of political elites. New Jersey, Prentice Hall.

RUBEL, Maximilien. (1960), Karl Marx devant le bonapartisme. Paris, Mouton.

SAES, Décio. (1994), "Uma contribuição à crítica da teoria das elites". Revista de Sociologia e Política, 3, nov.

SKOCPOL, Theda. (1984), Los Estados y las revoluciones sociales. México, Fondo de Cultura Económica.

THERBORN, Göran. (1983), "Why some classes are more successful than others?". New Left Review, 138: 37-55.

_________. (1989), "A análise de classe no mundo atual: o marxismo como ciência social", in E. Hobsbawn (org.), História do marxismo, Rio de Janeiro, Paz e Terra, vol. 11.



1 Bobbio discussed these oppositions in rather different terms in the essay titled "Marxism and Social Sciences" (2006, pp. 167ss).
2 Poulantzas was not, to be sure, the only one to deal with these theoretical problems in the Marxist camp. He was, however, the one to most explicitly and conscientiously step up to the task of dealing with this issue in theoretical realm. For this reason, this article has chosen him as a privileged interlocutor (see, especially, Poulantzas, 1971, vol. II, pp. 154ss).
3Take as examples Marx on the German Revolution (1848-1849) or on the political reality conveyed by the theoretical notion of "Bonapartism." Concerning the latter, see Rubel (1960).
4 Translator's note: The terms used by Poulantzas in his original discussion are classes régnantes and classes tenant de l'État and have been translated distinctly by commentators in different English language versions of Political Power and Social Classes.
5 The ideas presented next sum up a much broader discussion published in Perissinotto and Codato (2009, forthcoming) and in Perissinotto (2007).
6 Similar suggestions have been mad by Therborn (1983, 1989) and Przeworsky (1989).
7 Formalism is the outcome of the "internalist" perspective, adopted by some elite theory thinkers. They tend to explain political phenomena and the power of elites based only on factors internal to the political universe. See, in this regard, Saes (1994).