Sociology without reductionism

With abundant media coverage, quite rare at this level for a work in the social sciences, the latest book by Gérald Bronner and Étienne Géhin is fueling lively controversies this fall, which the authors cannot be surprised by. The book, which alternately discusses a science “in danger” and a “dangerous science,” was clearly partly written for this purpose. In this respect, the work, touted by its publisher’s banner as “the event book” of this new season, undoubtedly renders an ambiguous service to the cause it claims to defend. Many of the arguments presented therein are rendered barely audible by the passionate reactions provoked by its rather sensationalist title and its slightly arrogant marketing. However, these arguments deserve to be heard and discussed for what they are.

For an analytical sociology

So what is this danger that the title of the book refers to? G. Bronner and É. Géhin first attack the prevalence among their French colleagues of the holistic paradigm inherited from Durkheim (“considering social facts as things” and “explaining the social by the social”) and strongly associated, according to them, with the sociology of Pierre Bourdieu and its various contemporary avatars. In their eyes, this paradigm supports a type of sociological reasoning affected by an “agency bias” consisting of attributing to collective entities (the State, class, power, society, etc.) an intentionality that they cannot intrinsically be endowed with, and a causal role that cannot be theirs. According to them, this results in a form of sociological determinism that leads to considering individuals as agents of a destiny driven by social structures that escape their control. To this paradigm, G. Bronner and É. Géhin opposes that of an “analytical” sociology which, in line with the methodological individualism of Raymond Boudon and the comprehensive sociology of Max Weber, places the intentionality of the actor and the understanding of the reasons he gives himself to act at the starting point of the analysis of social facts, conceived as essentially unstable, contingent and unpredictable emerging effects of interactions between individuals.

It is perfectly legitimate for the authors to defend the virtues of this alternative epistemology. It should be noted, moreover, that outside the French context, the actionalist paradigm underlying this analytical sociology is not necessarily irreducibly opposed to structural approaches. Various contemporary authors have long been working to build bridges between the micro and macro levels of sociological analysis, as can be seen, for example, in North America in the work of Douglas Massey.

The argument of G. Bronner and É. Géhin nevertheless raises some difficulties on this point. If it is always possible to identify, in contemporary sociological discourse, as in the prose of the founding fathers, the type of deterministic and intentionalist drift that they intend to attack, the picture that they paint of the ordinary practice of sociology seems nevertheless quite largely exaggerated. So that the work sometimes falls into this classic rhetorical flaw which consists of exaggerating the line of an argument that one then easily gives oneself the means to disqualify. Even when they do not fit into the individualist or analytical framework defended in this book, sociologists who mobilize the role of norms, structures, classes, collective actors and other conceptual entities rarely do so within the ultra-determinist (or hyper-culturalist) framework decried by the authors, who incidentally lump together theoretical traditions that have little in common, such as culturalist anthropology and Bourdieusian-inspired “critical” sociology. Contemporary sociological reasoning is clearly most often of the probabilistic type, and while a certain vigilance must be exercised with regard to the forms of uncontrolled extrapolations that it sometimes allows itself, there is no point in caricaturing it. At the risk of opposing the excesses of deterministic thinking with the truism of the irreducible contingency of individual behaviors and social facts, which authors do not always escape, as in the following passage, on the subject of the deterministic tendency attributed to certain currents of the sociology of delinquency:

Since we are talking about crime, let us recall here that in France, in the “sensitive neighborhoods” considered to be criminogenic by a part of the sociological community that likes to say that criminal acts are the effects of social causes, crime obviously poses a problem, but the fact is that it only concerns a minority of individuals. Thus, in Saint-Denis, the most criminogenic city in France, there were approximately 166 crimes per 1000 inhabitants in 2015. If we compare it to another city not known for its delinquency like Annecy (both belong to the cities of 50,000 to 100,000 inhabitants), we note, for the same period, 79 crimes per 1000 inhabitants. The difference remains significant, but if we assume that each crime was committed by a different person (maximalist hypothesis), we see that in Annecy, 92% of individuals resist the temptation to commit a criminal act, while they are 83% in Saint-Denis. How can we explain by deterministic models that between the crime rates of two cities, one of which is twice as criminogenic as the other, the difference, although real, is nevertheless “only” what it is? (p. 172-173)

The pursuit of the deterministic narrative, to which few sociologists specializing in these issues actually give such credence, leads to treating as secondary the very great significance of the difference in crime rates between the two cities, which logically attracts the attention of the sociologist, and to insisting on the trivial observation that, in both cities, criminal behavior is largely in the minority, which defines it as statistically deviant. On this basis, one would be justified in incriminating the “deterministic” inclination of modern oncology: if smokers die more often than non-smokers from lung cancer, the majority of them die from other causes, like non-smokers, which does not, however, weaken the scientific consensus on the carcinogenic effect of tobacco.

The promise of neuroscience

The second danger pointed out by G. Bronner and É. Géhin concerns the risk of marginalization that sociology would run by ignoring the contribution of disciplines outside the field of social sciences, and in particular that of cognitive sciences and neurosciences. This is undoubtedly the richest aspect of the work. The most measured too. G. Bronner and É. Géhin have expertise on these issues that protects them from some of the scientistic illusions that specialists in the human and social sciences sometimes indulge in when in contact with the “hard” sciences. The criticism they address to the neuronal reductionism of orthodox cognitivists is here unequivocal (pp. 178-179), and the hybridization of sociology and cognitive sciences that they call for intends not to separate the analysis of the processes of arbitration and retro-judgments that have their seat in the human brain from that of the social variables of action.

Moreover, although brain sciences are mainly called upon here to support the actionalist paradigm, a large part of the arguments presented describe a scientific perspective that could achieve consensus outside this framework. G. Bronner and É. Géhin also suggest in passing that some of the concepts of Bourdieu’s sociology, such as that ofhabitusare not at all incompatible with the approach of these disciplines, recalling the interest that this notion was able to arouse in the neurobiologist Jean-Pierre Changeux, who saw it as a bridge between contemporary discoveries on the plasticity of the brain and the emphasis placed by sociologists on the incorporation of dispositions to act through experience and repetition. On all these questions, the idea defended by the authors that there exists today in France, in research in social sciences, a somewhat irrational fear of cognitive sciences that should be dispelled is in fact rather convincing.

An irresponsible science?

The third danger to which the title of the book refers attracts particular attention in the comments to which the work gives rise in the media and echoes certain recurring trials to which sociology is the object. The latter would in substance be the bearer of deterministic narratives, the dissemination of which would produce deleterious effects on behavior, by devitalizing “the notions of merit, responsibility or morality” (p. 211). We think here of the “destructive teachings” of this “stammering science” and “without clearly defined utility” that Georges Pompidou mentioned about the student leaders of May 68 …

G. Bronner and É. Géhin are particularly interested in the academic failure of students from modest backgrounds, particularly children of immigrant origin, which would be partly attributable, according to them, to the dissemination of a sociological vulgate inciting the most deprived to a form of self-complacency functioning as a self-fulfilling prophecy: exposed to deterministic accounts of the social causes of academic failure, the most deprived students would come to adhere to them and would give up making the necessary efforts to overcome the obstacles that stand in their way. The hypothesis is not necessarily uninteresting. It would nevertheless be worth testing rigorously, which the authors do not do. At the very least, it would be necessary to be able to measure somewhat the degree of exposure of children of about ten years old to this “irresponsible” sociological vulgate, in the absence of which we condemn ourselves to conjectures that are based only on vague impressions about the prevalence, in the mass media in particular, of discourses based on deterministic sociology, to which we could oppose equally vague impressions about the prevalence of opposing discourses. Presented as a result, the assertion of the deleterious effects of sociological discourse is thus not supported in the book by any empirical basis. The fact that exposure, in the family environment, to positive attitudes towards school and merit can promote academic success, as the authors point out, in no way proves, on the contrarythe harmful effects of the dissemination of sociological discourse. The fact that a parent does not show his children any particular adherence to the virtues of school meritocracy does not imply ipso facto let him fill them with deterministic stories inspired by Bourdieu and Passeron!

Here is a somewhat disconcerting book, which is much better than its title and some of the argumentative simplifications to which its authors sometimes indulge. One can thus fear that the interest of some of the theses put forward in the book, on the merits of the actionalist paradigm in the analysis of social processes as on the interest of the meeting of the social sciences and the cognitive sciences, in particular, is polluted by a polemical spirit which deviates quite widely from the canons of science to which its authors nevertheless claim to be very attached.