avec le catalogue

En plein texte
avec Google

Recherche avancée

Tous les ouvrages
numérisés de cette
bibliothèque sont
disponibles en trois
formats de fichiers :
Word (.doc),

Pour une liste
complète des auteurs
de la bibliothèque,
en fichier Excel,
cliquer ici.

Collection « Les sciences sociales contemporaines »

Une édition électronique réalisée à partir de l'article de Patrick Pharo, “How is Sociological Realism Possible ? Sociology after Cognitive Science.” Un article publié dans la revue European Journal of Social Theory, vol. 10, no 3, August 2007, pp. 481-496. [Autorisation accordée par l'auteur le 19 décembre 2011 de publier cet article dans Les Classiques des sciences sociales.]


Patrick Pharo

How is Sociological Realism Possible ?
Sociology after Cognitive Science

Un article publié dans la revue European Journal of Social Theory, vol. 10, no 3, August 2007, pp. 481-496.

Realism of Action
Cognitive Realism
Moral Realism
Patrick Pharo


This article explores the limits of social constructionism and criticizes the ‘demiurgic conception of society’ associated with it. It contemplates the possibility of sociological realism by investigating the intrinsic and objective properties of action, cognition and morality. The incorporation of intrinsic meanings and intentions in social actions, the objective information supporting cognitive processes and human sensitivity to pleasure and pain as well as the normative rejection of undue suffering, delineate the objective core of social facts, which can be interpreted or influenced, but not arbitrarily or capriciously constructed or manipulated. This general argument is supported by various illustrations drawn from the semantics of social actions and classical puzzles of interpretative sociology such as the meaning of suicide or the morality of social sanctions.

Key words

• civil action • cognition • conceptual constraints • morality • natural causality • realism


After Social Constructionism
and Cultural Relativism

According to Berger and Luckmann (1966), what is real for a Tibetan monk is not real for an American businessman. Why ? Because what is real depends on social knowledge. And social knowledge itself depends on common typifications peculiar to various cultures. Consequently, reality cannot be the same in different cultures. According to Howard Becker (1963), deviance is not a quality of acts committed by a person. Why ? Because social groups create deviance by establishing norms which provide labels and sanctions to apply to certain acts. Therefore, deviance is a consequence of successful labelling and cannot be an intrinsic reality.

Phenomenological sociology and interactionism typically epitomize a very deep trend in the social sciences, which goes back at least to Marx and Durkheim, ascribing a very strong creative power to society. We could call this trend the [482] ‘demiurgic conception of society’. Philosophically, this conception is holist, and sometimes nominalist, regarding social phenomena as a ‘reality sui generis’ (Durkheim, 1951), distinct from physical or psychological entities. Abstract social phenomena, such as norms, values and qualities are considered derived from opinions, descriptions, narrations, interpretations, and not anchored in the proper nature of things. Sociologically, this conception is culturalist in so far as it considers different cultures as sharing neither the same knowledge, nor the same beliefs, nor the same norms, nor the same values. Politically, this conception is often critical, because social constructions are connected to arbitrary conventions and prejudices which have to be overcome when they harm people’s liberty or happiness.

What is true in this demiurgic conception of society ? As a matter of fact, a large proportion of its assumptions is plainly right. Social ideas and institutions are often grounded on arbitrary conventions and prejudices. Various cultures do not share the same knowledge, beliefs, norms or values. And society as a whole creates institutions, rules, symbols and other abstract or semi-abstract entities, let alone physical objects, which did not exist until the appearance of human societies in the natural evolution of life. The part constructive activity plays in law, in politics, in art, even in science, is so large that it could hardly be denied. For instance, any scientific theory has to be regarded as an intellectual construction, even when it has obtained incontrovertible evidence in its favour. And we could say the same about a text of law or a legal judgement. Moreover, many constructionist sociologists do not reject the objectivity of social facts : for example, ethnomethodology considers social accountability, which is plainly objective as practical accomplishments (Garfinkel, 1967), as the core of social facts. In the same way, Berger and Luckmann clearly recognize the objectivity and the externality of social facts, even if they are socially constructed, even if they are a ‘human product’.

Is there anything wrong, then, with the demiurgic conception of the society ? I think so. In my opinion, the flaw of the constructionist view consists in overlooking the following considerations : (1) human actions largely escape social decisions and they display causal processes and intrinsic features which do not depend on external descriptions, narrations, interpretations or evaluations ; (2) social cognition is at once a physical and an abstract process which is strongly and universally constrained by the flow of natural information ; and (3) social norms and morality rest on independent features of human sensibility which prefers pleasure to pain in the normal case and recommends the avoidance of undue suffering. If properly understood, these assumptions open the possibility of a sociological realism which appears as a genuine, exciting and fruitful alternative to sociological constructionism and the demiurgic conception of society. Actually, realism in the human and social sciences is not at all an exception, since we find some of its ingredients in economics, in psychology and even in sociology, for instance, in the theory of rational choice. But realism in the social sciences, particularly in sociology, is often simply naïve and does not address directly the question of the cognitive and logical status of collective or individual states and [483] events. On the contrary, the present approach, which is derived from Weber’s sociology of meaning, focuses on the steady constraints of intersubjective settings and the ordinary tools of shared understanding. [1] The rest of the article is aimed at some clarification of this approach.

Realism of Action

For a half-century now, cognitive sciences and philosophy of mind have brought radically new insights to bear on the place of human action and cognition in the natural processes of life through computational theories of mind, animal ethology, evolutionary psychology, cognitive neuroscience and cerebral imagery. One of the main contributions of these manifold works was to restore the links between social action and natural causality by revealing the neuronal and functional bases of many ordinary behavioural abilities : language learning, unconscious perceptions and calculation, structures of memory, recognition of faces, naïve physics, ‘theory of mind’, ecological rationality, reciprocal altruism, cheater detection, acceptance of norms, parental attachment, sympathetic or aggressive feelings, spontaneous motivation for food and sexuality, mechanisms of addiction, etc. [2] According to some proponents of this new social naturalism, there is an insurmountable gap between classical cultural theory (in short, the theory of the ‘blank slate’) and new cognitive, evolutionary and causal conception of human behaviour. The former wrongly emphasizes the paramount power of culture in human training and the social adaptation of individuals, while the latter rightly underlines the natural determination of human behaviours by neuronal systems and specialized cognitive modules which developed during the natural evolution of the human species.

Actually, this contrast between the two approaches seems rather exaggerated. As I admitted in the previous section, the constructionist or culturalist theses remain true for a large class of human behaviours. Moreover, neuronal systems are probably quite flexible and it is generally admitted that they develop very differently according to the course of individual lives, the organic propensities usually interacting with cultural and contextual settings. On the other hand, most of the works which I have just alluded to are highly speculative : they suppose the existence of causal mechanisms underlying cognitive processes, but do not specify exactly how these mechanisms act on complex social behaviours. But human decision, either individual or collective, does plays a role in the accomplishment of some social behaviour. And in the current state of knowledge, we are unaware of its exact relation with the causal mechanisms characterizing the operation of the brain. We do not have sufficient reason, then, to want to naturalize the whole of social behaviour.

Nevertheless, the contemporary cognitive sciences conspicuously restate the possibility of intrinsic objectivity of social actions and events by providing an important support for the sociological realist perspective on the causal processes of action, whether they are decided or not. The cognitive sciences can thus help [484] specify the limits of social constructions and to draw a clearer boundary between social realities which can be ‘constructed’ and those which escape any social construction by virtue of their intrinsic properties.

In the philosophy of the social sciences, this boundary is usually located between interpreted and not-interpreted facts. For example, from an epistemological perspective, Ian Hacking (1999) has reviewed the various objects, ideas and ‘elevator words’ (i.e. abstract or semantic objects) which have been considered ‘socially constructed’ in social science literature. The list is very long, stretching from authorship to Zulu nationalism by way of gender, illness, serial homicide or urban schooling. But, according to Hacking, the main characteristic of social objects is eventually interactive, in contrast with natural objects which are indifferent and are not so easily ‘constructible’. In a related but more genealogical perspective, John Searle (1995) wondered how brute facts could become social or symbolic ones. According to Searle, the reality of social facts is not at all intrinsic but depends on human representations and collective intentions which bestow conventional values on them. In the two analyses, the construction of social facts mainly results from social interactions or conventions which add extrinsic properties to indifferent objects or brute facts. Social interactions or conventions thus attribute an external meaning to natural facts, either human, such as death, birth, sexual intercourse, fights or interpersonal encounters, or nonhuman such as animals, plants, territories or physical objects, and this extrinsic meaning organizes the course of social life in a ‘constructionist’ manner through narrations, interpretations, evaluations, rules of law, and so forth.

As we can see, the two analyses are mainly interested in synthetic, complex or intricate social facts which combine arbitrary symbols or conventions with real objects or events, such as child abuse or madness (Hacking), money and other institutional facts (Searle). But we can also wonder, in a more analytic perspective, up to which point the social objects could have intrinsic properties which cannot be ‘constructed’ or could resist any attempt of interpretative construction. Now, according to modern cognitive science, the answer to this question is almost obvious : the social properties which cannot be ‘constructed’ are merely cognitive, that is, mental, intentional or reflexive properties which go together with functional and physical behaviours, such as practical desires, practical beliefs, practical norms and values, or, more generally, what we can call ‘intrinsic reasons of action’. Of course, a subject or a group can shape, construct or contribute to the construction of its own desires, beliefs and values through reflection, arguments, personal exertion, training, influences, fashions, social trends, rules of law, and the rest. But when such desires, beliefs or values are incorporated in a particular action, they become an intrinsic part of it, which cannot be constructed or reconstructed at pleasure. This point was emphasized by the philosopher Donald Davidson (1980) who contends, against the Wittgensteinian tradition, that desires and beliefs can be genuine causes of action, even if they are also reasons and justifications. But Weber had already stated this in Verstehende Soziologie (1921) with his notion of subjective meaning (gemeint Sinn) of action which is primarily intrinsic to the action, even if we can interpret it or reconstruct it [485] through ideal-types. And, to a certain extent, we find similar approaches in Bourdieu’s theory of incorporated habitus (1980) or in the theory of rational choice (Coleman, 1990), since unconscious habitus as well as self-interests or conscious preferences intrinsically belong to actions. Unfortunately, these sociological approaches remain very vague about the logical status of the intrinsic reasons of action.

For a better understanding of the exact role of intrinsic practical reasons in regard to ex-post justifications, interpretations or reconstruction of the meaning of action, let us consider the semantics of the usual terms of actions which can be called ‘civil’, either in a positive sense, such as, for instance, congratulations, celebration, invitation, reconciliation, care, or in a negative (or uncivil) one, such as murder, robbery, rape or adultery. Civil actions can be defined at the outset as Weberian ‘social actions’ (1921), that is to say, actions endowed with a subjective meaning which relate to behaviours of other people and are directed according to them. But the definition of civil actions includes also the agent’s expectation of the good or bad consequences for the addressee or recipient. Accordingly, a civil action is positive when its predictable consequences promote the good or the liberty of the recipient (or related others), negative (or uncivil) in the inverse case, and indifferent if the predictable effects are themselves indifferent. Now, in a sense, civil actions are not accomplished if the agent does not aim, expect or predict some consequences at the moment when he acts. For instance, congratulations or murder are not accomplished if the agent does not aim, expect or predict the satisfaction or the death of the addressee in consequence of his action. [3] More generally, it is easy to see that the usual semantics of civil actions requires the fulfilment of various cognitive, intentional, but also normative or juridical conditions so that the action could be accomplished according to its own intrinsic meaning. For example, if the congratulations are ironic or if the death is accidental, we cannot say salva veritate that the action consists of congratulations or a murder. Likewise, we cannot describe an action as an invitation if the host did not ask for the presence of the intruder, as a reconciliation if the opponents do not expect peace between them, as care if the agent does not aim at the recipient’s well-being, as robbery if the agent was the owner of the stolen goods, as rape if the sexual intercourse was consensual, as adultery if the sexual partners are not married. The ‘subjective meaning’ or the ‘intrinsic reason’ of a civil action must then be regarded as a part of its intrinsic reality, because eventually every true sentence about the action is constrained by it.

It is very useful to keep in mind this logical structure of civil actions in taking up the question of the intrinsic properties of more complex or synthetic social facts, such as historical events, scientific discoveries, public ceremonies, private relations, and so forth. For instance, the event of the death of Louis XVI was interpreted, ‘constructed’ and ‘reconstructed’ in numerous ways by different interpreters and historical agents, and these constructions constituted a large part of the effective story of the French Republic. But the event itself does not entirely depend on its narratives, because all the agents involved in Louis XVI’s death, whether the Convention’s representatives voting pro or contra, the executioner [486] of the sentence, the Parisian witnesses and the king himself, knew something, wanted something or advanced normative judgments about what really occurred. In this sense, the king’s death, which was decided or witnessed by the French people, and all the thoughts and cognitive contents making possible this event, were not at all a social construction, but an objective fact endowed with intrinsic meanings. We could apply the same kind of analysis, for example, to the discovery of HIV by two research teams in France and the USA, or the refusal of the Nobel Prize by the French philosopher Jean-Paul Sartre in 1964, or the second wedding of Prince Charles in England, and eventually to any historical fact or any ordinary social action, including child abuse, madness or economic transaction. The intrinsic properties of social actions merely result from the occurrence of one or several physical events in relation to defined meanings, allowing an observer the logical and empirical discrimination of the action from any other. The conventional or interactive narratives do not then take in or abolish these properties but, on the contrary, must reckon with them, even when it is impossible to confirm all the descriptions of the action.

We could readily extend these semantic observations to other social terms such as roles, relations, feelings and qualities, the referents of which are in fact tightly associated with cognitive and practical stances of social agents. In the sociological tradition, more specifically, there was a controversy between two positions, namely institutional versus interpretative conceptions of the social roles (Parsons and Shils, 1951 ; Cicourel, 1972), because the practical scope of social roles cannot be completely defined and gives rise to some discrepancy between scientific and ordinary typifications. But, in any case, it would be difficult to deny that, in some defined social settings, a policeman, a priest, a judge or a manual worker abide by intrinsic rules of action and expression which allow discrimination between the various social roles and are known well enough by the agents to regulate institutional and ordinary relations in society. Likewise, social relations such as friendship, contracts, weddings, associations, work subordination, or social feelings such as indignation, shame, love, envy, or social qualities such as modesty, temperance, courage, kindness, which are usually built up or acquired through individual and collective stories, become intrinsic to social facts when they are incorporated into social actions and individual experience.

More generally, we can assume that the cultural or social qualities of persons and relations, whether they are consciously elaborated or more spontaneous, do not primarily become objective because they are judged as such by everyone, but because they are intrinsically incorporated into the cognitive component of social actions. Now, this assumption has important consequences for the question of human responsibility. Indeed, contrary to the demiurgic conception of society, this point makes it possible to consider human responsibility not as a social construction, but rather as an objective fact which is carried out through the causality of action and within the limits of the natural causality of the world. The Durkheimian sociologist Paul Fauconnet (1920) regarded individual responsibility as a social decision which serves collective aims and could be entirely disconnected from the real causality of, for instance, damaging actions. And the [487] ethnomethodologist Harold Garfinkel (1967) regarded the accountability of action as a continuous reconstruction and a permanent indexation to instantaneous properties of social interaction. But these two approaches overlook the effective causal role of social actions which bring about individual and collective stories, not only because of their social interpretation, but also because of their own cognitive structure and their moral (or immoral) orientation. Consider, for example, the case of racist stigmatization. Of course, the association of migrants and foreigners with wicked and dreadful characteristics is clearly a type of social construction that we find in many societies. But, as I have tried to show elsewhere (Pharo, 2001a, 2001b), the reciprocal acts of rejection or aggression, the political instrumentalization of foreigners, the numerous ordinary decisions associated with in-group/out-group perceptions and relations, the secessionist consequences of exclusive forms of self-esteem, are not at all social constructions.

Finally, the causal role of social action is a part of natural causality, which involves the occurrence and the succession of social facts, and contributes to their irrevocability and indefeasibility. However, the nature of what is really incorporated in actions remains the principal question of cognitive realism, to which I now turn.

Cognitive Realism

The cognitive sciences and evolutionary psychology by and large assume a sort of cognitive realism, since natural cognitive processes are physically realized and can be supposed to be approximately similar in the same natural species. But from a sociological perspective, cognitive realism immediately raises two intertwined difficulties : one is observational and the other ontological. The observational difficulty is related to the under-determination of meaning by visible and conspicuous behaviour. The ontological one is related to the abstract part of social phenomena in general, particularly intentional and reflexive states. To illustrate these difficulties, let us take the examples of respect and promise. If someone behaves with ostensibly polite, careful, deferential or refrained manners when she meets her colleagues or her superiors, we would probably judge her posture respectful. But, if we learn that in other settings the same person makes fun of the same colleagues and superiors, utters insults about them or plans to harm them, we would conclude that, in fact, the conspicuous behaviours are not at all respectful, but rather deceptive, hypocritical, fraudulent, and so forth. [4] Likewise, if someone promises to bring back something, to do a favour or the like, but we know from her diary or a public confession that the ostensible statement went along with a mental reservation, we would probably hesitate to call such an act a promise and instead use notions of false promise or deceit. [5]

As we can see from these examples, the observational difficulty is a direct consequence of the ontological one. It is indeed the abstract aspect of practical postures such as respect and promise, but also admiration, love, trust, marriage and a large number of other social phenomena, which makes them so difficult [488] to observe. In fact, most social phenomena can be divided into a visible part such as the figure, the outfit or the physical movements which can be described by way of empirical observation, and an invisible, mentally embedded part which can only be inferred or conjectured from contextual data. What we call ‘folk psychology’, that is, the external attribution of motives to behaviours, is the usual way of dealing with this difficulty. Folk psychology is a colloquial method which seems to be used universally to grasp the logical structure of intentional or reflexive behaviours. And, as a matter of act, folk psychology proved to be very useful in the interpretation of a large part of social action : for example, to infer the motives of people standing on a station platform (waiting for the train !) or the motives of students writing in an examination room (sitting the test). Without folk psychology, no sociological interpretation of action would be possible, whatever the theoretical framework : rational choice, interpretative sociology or interactionism.

But what is the proper ‘matter’ of mental states and folk psychology ? According to numerous theorists in the cognitive sciences, the answer is merely : nothing, because only neuronal structures really exist and there is only the brain under the skull. It is thus very easy to be realist in the cognitive sciences when we assume either an eliminativist approach to cognition which completely rejects ‘folk psychology’ (Churchland, 1986), or even an ‘intentional stance’ which does not grant any proper reality to intentional properties (Dennett, 1987). In this case, realism is simply equivalent to radical materialism, and by comparison with the neurosciences folk psychology is only a very primitive theory of the human mind.

However, the case is much more difficult if we also admit an abstract component in cognitive abilities. Such a hypothesis was assumed by the functionalist theory of mental states (Putnam, 1967 ; Fodor, 1975) which took human cognition as abstract causal relations, defined by their functional role and not by the material structure in which they are instantiated. But the majority of functionalist theorists eventually jettisoned this hypothesis, because it was linked to a wholly computational conception of mental processes which proved implausible : no algorithm can be strong enough to represent the whole of mental calculations (Putnam, 1988), and the brain is probably not a gigantic computer (Edelman and Tononi, 2000). Moreover, the functionalist story propounds a very impoverished picture of the relations between mind and world which actually are not always well fitting or ‘functional’, unless we take the notion of function in a greatly enlarged sense, including cognitive errors and shortcomings. Nonetheless, the functionalist approach of mental states opened new ways to investigate the seminal question of ‘subjective meaning’ or ‘intrinsic reasons’, discussed in the previous section, by recognizing the abstract aspect of mental states and the logical gap between physical processes in the brain and the intentionality of action and cognition. And even if all mental states do not have a clearly definite functional role, they all could probably have a conceptual one.

Actually, the basic reasons why it is usually possible to grasp the abstract aspect of cognitive and social behaviours stem from the deep-seated common condition of human beings : (1) who are members of the same animal species ; (2) who live [489] in the same natural world ; and (3) who use the same ordinary logic. These three kinds of human kinship make highly probable an approximately similar treatment of approximately similar information by any human being. As Davidson (1984) put it, ‘radical interpretation’ of the sentences and actions performed by members of foreign cultures is possible because we impute to them rational abilities and true beliefs in relation to the same surrounding world. Surely, constructionist and culturalist theories are right in emphasizing the differences in beliefs between cultures, but these differences must rely on a large core of obvious data for every member of human species.

In order to clarify this point, let us take some examples. Consider the case of the widows of princely families in Bali who had to be burned alive with their husbands when the latter died (Geertz, 1983). Obviously, we do not know what exactly they thought at the time of jumping into the blazing inferno, and the difficult question of genuine consent or internal revolt cannot be answered by mere observation. But we can at least suppose that the young spouses at the outset had an idea of their impending death under atrocious conditions because of a dreadful misfortune, and we understand that some could hesitate when they were led to the torment.

Or, to be still more specific, let us consider the classical sociological problem of suicide. According to Durkheim (1897), it would be easier to detect from the outside the agent’s knowledge of the consequences of his action than his actual intentions, and consequently the former could be empirically ‘observed’, but not the latter. Perhaps it is sometimes the case, but it is probable that when we know what the agent knows about the consequences, we can also imagine his or her intentions. Or, according to Garfinkel (1967), the so-called ‘coroner’s question’ raised by an equivocal mode of death consists in deciding between four elementary possibilities : natural death, accident, suicide and homicide. In order to know ‘what really happened’ and assign an adequate title to each case, the coroner looks for compatibility between the remains on the slab and the possible courses of action. In his comment, Garfinkel emphasizes the practical decidability of the case : ‘What the inquiry can come to is what the death came to’ (1967 : 18). So be it. But practical decidability is semantically constrained by the objective data which are simultaneously physical and ‘abstract’ or intentional and display something relevant (Sperber and Wilson, 1986) about the actual conditions of death. In my own research involving the Centre medico-légal de Paris, it was confirmed that the actual scientific methods of investigation allowed the acquisition of certainty in almost all cases. Of course, we cannot know precisely the deepest motives or desperate thoughts of persons who committed suicide, but we are generally able to discern their practical intentions.

The previous point could be generalized with reference to an important contribution of cognitive sciences that Jerry Fodor (1995) subsumes under the title of ‘informational semantics’. The main idea of informational semantics is that the reference of thoughts and representations must be fixed by information coming from the external world and not by the thinker’s mode of presentation of objects. At the origin of this idea, one finds Kripke’s notion (1972) of ‘rigid [490] designators’ applied to natural phenomena and Putnam’s demonstration (1988) according to which mental representations cannot depend at the same time on an internal construction and an external determination. In Fodor’s work, the thesis consists in saying that mental representations are causally, and even nomologically, determined by the objects of the external world. In other words, if you have the concept of something and you recognize a being as something, for example, a cow or a tree, it is because your representations depend directly on the information coming from the external world – with, however, certain possibilities of error or variation from one subject to another because the laws in question are not strict laws. And this conclusion would not apply only to natural phenomena, but it would be valid also for social phenomena such as actions, feelings, social roles and relations, and even qualities and virtues which can be described with the help of ‘thick’ concepts (Williams, 1985).

Informational semantics raises numerous questions, in particular the question of the complex relations between analytic social knowledge and broad cultural syntheses. But if we seek out the analytic components of large institutions such as states, churches, social rituals, popular trends, fashionable categories, and so forth, this question is probably not an irresolvable one for sociological analysis. [6] We have seen in the previous section that it is not impossible to fix the boundary between highly aggregated social or historical facts and their elementary practical and intentional components. Moreover, if we acknowledge informational semantics in sociological research, we can also assume a kind of conceptual universalism that in principle excludes cognitive relativism and limits the scope of cultural relativity to the epistemic consequences of local data which are more salient in the various cultural domains and impact on the knowledge of local thinkers. In this sense, cultural differences would not be such because of differences in social knowledge, but social knowledge would be different because of actual cultural and local differences. It is indeed more economical to impute cultural differences to inescapable, contingent, ecological conditions and social history, which logically entail the adjusted contents of knowledge of human beings of a common type, than to suppose infinite, obscure differences in human structures of knowledge which allow the capricious construction of cultural differences.

Moral Realism

In fact, the ultimate and genuine stake of sociological realism is moral. The immediate objection to practical and cognitive realism is indeed that people do not follow the same norms and do not value the same things in different cultural settings. Maybe Balinese widows attribute moral value to the fact of being burned alive ? Maybe the North-east African women ascribe moral value to excision ? Maybe Sicilian mafia members attach moral value to the rules of omerta and the skills of murderers ? Maybe Eskimo men attribute moral value to the sexual sharing of their spouses with foreign guests ? Maybe Chinese citizens ascribe moral value to strict obedience to dictatorship ? Now, if moral meanings are so [491] culturally variable and erratic, then moral realism seems perfectly impossible. And if moral realism is impossible, practical or cognitive realism would be strongly threatened, for morality is an important component of the meaning of action and cognition. However, it is not impossible to overcome this threat if we manage to show that moral relativism itself rests on a confusion about the meaning of the moral concepts : no concept can mean both one thing and its opposite, while moral concepts and statements should also be able to be satisfied by their objective referents, in spite of their tricky normative content. Such is the basic idea of moral realism.

The question of the ‘reality’ of moral facts is a very old philosophical question whose basis is to be found in Plato’s treatment of the reality of the ideas of the Good, Beauty and Truth. In modern philosophy, the question was renewed by authors such as Max Scheler (1916), who proposed treating moral values as ideal entities accessible to intuition, or G.E. Moore (1902), who saw in the notion of intrinsic moral value a possible way between Platonic dualism and reductive naturalism. More recently, there has been an important renewal of reflection on moral realism when philosophers became interested in the conditions of the truth of moral sentences and in the putative independence of moral properties which emerged from evolutionary cognitive processes (Sayre-McCord, 1988). For some later modern philosophers, moral facts or moral properties should be considered not as primary qualities pertaining to objects, like mass or electrical charge, but rather as secondary qualities, like colours and other apparent features, the perception of which depends on objects’ dispositions to be perceived by particular organisms (McDowell, 1978). At any rate, the examination of topics such as truth conditions, intrinsic value or perceptual qualities opened the theoretical possibility of finding an objective basis for morality and to overcome moral relativism. Nevertheless, moral realism appears today as a very appealing yet still conjectural perspective of research – as long as we don’t have more precise criteria of moral objectivity. And, as everyone knows, it is not very easy to find such criteria.

At this juncture, the best way to circumscribe the normative root of the problem is probably to go back to the Durkheimian conception of ‘the determination of the moral fact’ (Durkheim, 1906). Durkheim was by no means relativistic in morals. He thought, on the contrary, that the ‘science des moeurs’ towards which he was working, which started from the observation of the social facts, would be the only means of providing morality with an objective and non-metaphysical foundation. He maintained indeed that society is the true foundation of moral life and that this everywhere requires unselfishness and the devotion of citizens to the collective interest. But he also thought that every society has its morality and that to know what morality is in a given society, it would be enough to observe the positive or negative social sanctions, that is to say the praise or blame, which every society awards to different behaviours. The criterion of sanction was thus considered a kind of ‘reaction’ which allowed the drawing of a distinction between the ‘synthetic’ rules of morality and the simple ‘analytical’ rules of technique. As a matter of fact, however, all synthetic social rules, that is to say those which according to Durkheim involve praise or blame, are not necessarily moral.


In social life, there are indeed all kinds of aesthetic, legal, hedonistic, play and other rules the observance or transgression of which can give rise to positive or negative sanctions, but which, fortunately, are not moral. For instance, a master of music or cabinet making could congratulate a pupil for having achieved a difficult gesture, or we could blame a football team for having played a poor match without giving a moral sense to such approbation. Consequently, we need not confuse moral norms and values with other social rules which are not strictly technical yet pertain to other normative and value spheres. Besides, if all social norms and values were presumed moral, social life would be overwhelmingly oppressive.

In spite of its shortcomings, however, Durkheim’s interrogation about the determination of moral facts probably still provides the best means to make clear the actual place of morality in social life. For this apparent philosophical question must definitely be addressed if we were to tackle the socio-anthropological question of moral relativism and to distinguish between moral, immoral and indifferent facts. Otherwise, morality should be considered a mere ideology, invested perhaps with a practical priority, but without any normative specificity regarding different kinds of good or evil in human life. The classical criteria of moral philosophy, such as practical virtues (Aristotle), categorical duties (Kant) or the sum of happiness (Bentham, Mill), were aimed precisely at specifying moral good and evil. And Durkheim himself thought he was able to synthesize utilitarian and deontological morals by means of a strictly sociological and empirical approach of morality according to which sanction, imposed first from outside, becomes desirable thanks to moral education. But the reason for the failure of Durkheim’s ‘reactive’ determination of moral facts is merely that a social sanction is a possible consequence of some ‘immorality’, but not a criterion of morality as such. As a matter of fact, the point was established by many post-Durkheimian authors, including, for example, Erving Goffman or Michel Foucault, who demonstrated by and large the frequent unfairness of social sanctions, stigmatization and punishment. It is thus an all-together sociological and normative reason which makes it impossible for us to rely on sanction to determine moral facts.

What then could the intrinsic properties of moral or immoral facts be ? In the present state of the art, it would probably be difficult to propound a positive criterion of what a moral fact really is – one which would allow decision-making not only in ordinary settings, but also in hard cases of medicine, economic welfare, political choice, educational institutions, religious affairs, and so forth. However, the difficulty seems much less critical in the case of ‘immoral facts’ which could be commonly discerned and criticized in our own culture as well as in foreign or past cultures. And from a sociological perspective, it seems sensible to use an inductive step by trying to determine what, in contemporary moral conscience, appears as the prior characteristics of moral ideas or, more exactly, of moral aversion. [7] Now, such a survey, however general it may be, easily allows us to gather some important observations, such as : (1) the consensus expressed in international organizations, starting with the Universal Declaration of Human [493] Rights of 1948 and the Charter of the United Nations, accepted by all the member states in spite of its clearly Western philosophical origin, and including future perspectives such as ‘Millennium Development Goals’ ; (2) the usual rules of civil life, including the protection of private life and interests, prohibition of murder, theft, rape and antisocial acts, which are recognized by most of the contemporary states, even when they are not democratic ; and (3) the ethical requirements expressed in many spheres of social life such as, for example, in medicine with regard to respect for the person or the topic of free consent, in international relations with regard to the protection of civil populations, or in economic life with regard to the universally allowed requirements of protection and assistance to populations touched by economic crisis.

If, moreover, we take account of current ethical debates in modern liberal societies, for example, about basic individual rights, biomedical ethics, sexual liberty or interpersonal care, then it seems possible to hypothesize that today immoral facts are most often defined by the very simple but powerful criterion of undue suffering – that is to say, suffering that results from unfair action or, if this is not the case, that could be avoided by just interventions so as to exclude or mitigate unintended and unavoidable events, consequences and side-effects. Social agents are indeed sensitive beings who usually prefer pleasure and wellbeing to pain (except when pleasure comes through pain) and, as I have already stated, can also predict the good or bad consequences of their actions for others, as well as for themselves [8], in the present and in the future. Moral and immoral facts, irrespective of their direction, would thus be a compound of intentional aims and effective consequences of the related actions. Accordingly, the empirical and normative maxim sustaining the criterion in question could be something like the following : ‘if suffering did not exist, everything would be permitted ; but, as suffering exists, there are also some moral limits in human societies’. In this sense, the Durkheimian criterion of sanction would be only a consequence of the more primary criterion of undue suffering, the importance of which is also emphasized today by many works in women studies (Gilligan, 1982), political philosophy (Honneth, 1992 ; Margalit, 1996) or evolutionary psychology (Barret et al., 2002) about care, sympathy, attachment, reciprocal consideration and naïve morals.


In this article, I have tried to contemplate the possibility of sociological realism from the point of view of the intrinsic and objective properties of action, cognition and morality. This perspective of research is both theoretical and empirical, since the intrinsic practical, cognitive and moral properties of social facts could explain a part, if not the whole, of their effective development. And one could guess that the cultural and social constructions – that is, the demiurgic power of society as a whole – which largely explain the other part, depend themselves on the intrinsic reality of action, cognition and morality.



Abramson, Paul R. and Pinkerton, Steven D., eds (1995) Sexual Nature, Sexual Culture. Chicago : University of Chicago Press.

Axelrod, Robert (1984) The Evolution of Cooperation. New York : Basic Books.

Barkow, Jerome H., Cosmides, Leda and Tooby, John, eds (1992) The Adapted Mind :

Evolutionary Psychology and the Generation of Culture. Oxford : Oxford University Press.

Barrett, Louise, Dunbar, Robin and Lycett, John (2002) Human Evolutionary Psychology. New York : Palgrave.

Becker, Howard S. ([1963] 1973) Outsiders : Studies in the Sociology of Deviance. New York : Free Press.


Berger, Peter L. and Luckmann, Thomas (1966) The Social Construction of Reality. New York : Doubleday.

Bourdieu, Pierre (1980) Le sens pratique. Paris : Minuit.

Carruters, Peter and Chamberlain, Andrew, eds (2000) Evolution and the Human Mind : Modularity, Language and Meta-Cognition. Cambridge : Cambridge University Press. Cavalli-Sforza, Luca (1996) Genes, People and Language. New York : North Point Press.

Churchland, Patricia (1986) Neurophilosophy. Cambridge, MA : MIT Press.

Cicourel, Aaron V. (1972) Cognitive Sociology. Harmondsworth : Penguin.

Coleman, James S. (1990) Foundations of Social Theory. Cambridge, MA : Harvard University Press.

De Waal, Frans (1996) Good Natured : The Origins of Right and Wrong in Humans and Other Animals. Cambridge, MA : Harvard University Press.

Davidson, Donald (1980) Essays on Actions and Events. Oxford : Oxford University Press.

—— (1984) Inquiries into Truth and Interpretation. Oxford : Oxford University Press.

Dennett, Daniel C. (1987) The Intentional Stance. Cambridge, MA : MIT Press.

Durkheim, Emile ([1897] 1930) Le suicide. Paris : PUF.

—— ([1906] 1951) ‘La determination du fait moral’, in Sociologie et philosophie. Paris : PUF.

——, ed. (1951) Sociologie et philosophie. Paris : PUF.

Edelman, Gerald M. and Tononi, Giulio (2000) A Universe of Consciousness : How Matter Becomes Imagination. New York : Basic Books.

Elster, Jon and Skog, O-J, eds (1999) Getting Hooked : Rationality and Addiction. Cambridge : Cambridge University Press.

Fauconnet, Paul ([1920] 1928) La responsabilité : Etude de sociologie. Paris : Alcan.

Fodor, Jerry A. (1975) The Language of Thought. New York : Crowell.

—— (1995) The Elm and the Expert. Paris : CNRS éditions.

Garfinkel, Harold ([1967] 1984) Studies in Ethnomethodology. Cambridge : Polity Press.

Geertz, Clifford (1983) Local Knowledge : Further Essays in Interpretive Anthropology. New York : Basic Books.

Gibbard, Alan (1990) Wise Choices, Apt Feelings. Cambridge, MA : Harvard University Press.

Gigerenzer, Gerd and Selten, Reinhard, eds, (2002) Bounded Rationality : The Adaptive Tool-Box. Cambridge, MA : MIT Press.

Gilligan, C. (1982) In a Different Voice : Psychological Development and Women’s Development. Cambridge, MA : Harvard University Press.

Hacking Ian (1999) The Social Construction of What ? Cambridge, MA : Harvard University Press.

Honneth, Axel (1992) Kampf um Anerkennung. Frankfurt-am-Main : Surhrkamp.

Kripke, Saul A. (1972) Naming and Necessity. Oxford : Basil Blackwell.

Margalit, Avishaï (1996) The Decent Society. Cambridge, MA : Harvard University Press.

McDowell, John (1978) ‘Values and Secondary Qualities’, in G. Sayre-McCord (ed.)

(1988) Essays on Moral Realism. Ithaca, NY : Cornell University Press.

Mill, John Stuart ([1859] 1972) On Liberty. London : Everyman’s Library.

Moore, G.E. (1902) Principia Ethica. Buffalo, NY : Prometheus Books.

Parsons, Talcott and Shils, Edward A., eds (1951) Toward a General Theory of Action : New York : Harper & Row.

Pharo, Patrick (1985a) Savoirs paysans et ordre social, Paris : C.E.R.E.Q.

—— (1985b) Le civisme ordinaire. Paris : Méridiens-Klincksieck.

—— (1992) Phénoménologie du lien civil : Sens et légitimité. Paris : L’Harmattan.


—— (1996) L’injustice et le mal. Paris : L’Harmattan.

—— (1997) Sociologie de l’esprit, conceptualisation et vie sociale. Paris : PUF.

—— (2001a) La logique du respect. Paris : Cerf, Collection Humanités.

—— (2001b) Le sens de la justice : Essais de sémantique sociologique. Paris : PUF.

—— (2004) Morale et Sociologie : Le sens et les valeurs entre nature et culture. Paris : Gallimard.

—— (2006) Raison et civilisation : Essai sur les chances de rationalisation morale de la société. Paris : Cerf.

Pinker, Steven (2002) The Blank Slate : The Modern Denial of Human Nature. New York : Penguin.

Putnam, Hilary (1967) ‘The Nature of Mental States’, in H. Caapitan and D.D. Merril (eds) Art, Mind and Religion. Pittsburgh, PA : Pittsburgh University Press.

—— (1988) Representation and Reality. Cambridge, MA : MIT Press.

Sayre-Mccord, Geoffroy, ed. (1988) Essays on Moral Realism. Ithaca, NY : Cornell University Press.

Searle, J. (1995) The Construction of Social Reality. London : Penguin.

Scheler Max ([1916] 1982) Das Formalismus in der Ethik und die materiale Wertethiks, in Gesammelte Werke, 2, edn. München : Francke.

Sperber, Dan, ed. (2000) Metarepresentations ; A Multidisciplinarity Perspective. Oxford : Oxford University Press.

Sperber, D. and Wilson D. (1986) Relevance, Communication and Cognition. Oxford : Basil Blackwell.

Weber, Max ([1921] 1956) Wirtschaft und Gesellschaft. Tübingen : Mohr.

Williams, Bernard (1985) Ethics and the Limits of Philosophy. London : Collins.

Patrick Pharo is Director of Research at the French CNRS, member of the CERSES (Centre de Recherche Sens Ethique Société) and teaches at the Université Paris V (Ecole doctorale de sciences sociales). His publications include Le civisme ordinaire (1985), Le sens de l’action et la compréhension d’autrui (1993), Sociologie de l’esprit (1997), La logique du respect (2001), Le sens de la justice (2001), Morale et sociologie (2004) and Raison et civilisation (2006), and with Anni Borzeix and Alban Bouvier, he edited Sociologie et connaissance : Nouvelles approches cognitives (1998). Address : University of Paris V – René Descartes, 12 rue de l’École de Medecine, Paris Cedex 6 F-75270 Paris, France. [email : ]

[1] My own sociological background consisted of three types of successive research : initially, realization of monographic work in the working class and rural workers to study logical structures and morality of social representations and ordinary civility (Pharo, 1985a, 1985b, 1992) ; then research on the formal structures of the civil acts, such as compliments, reproaches, promises, orders, obedience... implemented by computer tools – namely an expert-system called “Civilité”, which was elaborated in 1993 for the French CNRS ; and finally, application of these conceptual tools to concrete cases such as the pressures of the truth in public debates, reception and rejection of foreigners, reactions to the requests of bums and beggars, acts of cruelty, moral feelings... (Pharo, 1996, 1997, 2001a, 2001b), and today ongoing research on dependency on pleasure and addictive practices for the INSERM (Institut National de Recherches sur la Santé) and the MILDT (Mission Interministérielle de Lutte contre la Toxicomanie).

[2] It would obviously be impossible to give even an outline of the enormous quantity of publications on these topics. But we can indicate some useful references for the social sciences such as Axelrod (1984), Gibbard (1990), Barkow et al. (1992), Abramson and Pinkerton (1995), De Waal (1996), Cavalli-Sforza (1996), Elster and Skog (1999), Carruters and Chamberlain (2000), Sperber (2000), Gigerenzer and Selter (2002), Pinker (2002).

[3] For more details on this topic, see Pharo (1997, Ch. 6).

[4] For more details, see Pharo (2001a), which explores the semantics of respect and applies the conceptual model to respectful or disrespectful relations in various social settings.

[5] These conceptual constraints play an important role in social contracts and they are a crucial component of the ordinary sense of justice, as I try to show in Pharo (2001b).

[6] In my most recent book, Raison et civilisation (2006), I have tried to deal with the question of moral rationalisation or moral progress by way of an analysis of the rational and reflexive properties in the main spheres of social values : knowledge, aesthetics, religion, economy, politics and pleasure.

[7] I present the prospect of moral sociology in Morale et sociologie (2004), with its main inductive and conceptual criteria : avoidance of undue suffering, politics of others’ justice, practical virtues.

[8] The criterion of undue suffering is then a little larger than the ‘harm to others principle’ stemming from Mill (1859).

Retour au texte de l'auteur: Jean-Marc Fontan, sociologue, UQAM Dernière mise à jour de cette page le mercredi 21 décembre 2011 9:52
Par Jean-Marie Tremblay, sociologue
professeur de sociologie retraité du Cégep de Chicoutimi.

Saguenay - Lac-Saint-Jean, Québec
La vie des Classiques des sciences sociales
dans Facebook.
Membre Crossref