Durkheim’s Science of Morality

December 5, 2012

[Attachment saved to "C:\durkheim2.JPG"]

In the introduction to the first edition of his classic The Division of Labour in Society (1893), Durkheim inquires into the possibility of a science of morality. He writes that it is untenable to begin one’s scientific investigation of morality with the underlying rules that guide ethical behaviour, because there is considerable disagreement as to what those rules might be:

each moralist has his own particular doctrine, and the diversity of doctrines proves the flimsiness of the so-called objective value. (Emile Durkheim, The Division of Labour in Society, 1964, The Free Press: Toronto, p. 411)

He discusses some of the most prominent attempts to prove the objectivity of a given universal moral law (most notably Kant’s ‘derivation’ of the categorical imperative), and argues that they fail to establish the objectivity of the laws they advocate. If we begin with a dogmatic conviction that a given moral law is the right one, Durkheim argues, then our investigation cannot be a scientific one, for this conviction cannot be scientifically established.

Possibly, there is an eternal law of morality, written by some transcendental power, or perhaps immanent in the nature of things, and perhaps historical morality is only a series of successive approximations; but this is a metaphysical hypothesis that we do not have to discuss. (423)

How then should a scientific study of morality begin? It should look at observable behaviour, and make this the starting point of its analysis. Specifically, it should look at social sanctions.

This predetermined reaction, exercised by society on the agent who has violated the rule, constitutes what is called a sanction. (425)

By starting with sanctions – the manifestation of the moral rules by which agents live their lives – rather than with rules themselves, our investigation can be empirically grounded. Durkheim stresses that with this methodological principle he is not reducing moral rules to empirical sanctions:

If… we define the moral rule by the sanction which is attached to it, it is not as if we were considering the sentiment of obligation as a product of the sanction. On the contrary, it is because the latter derives from the former that it can symbolize it; and as this symbol has the great advantage of being objective, accessible to observation and even to measurement, it is a good method to prefer it to the thing it represents. (425-6)

Durkheim proposes, therefore, that we scientifically study morality by studying the social sanctions that are applied to transgressions of the presumed moral code of a given milieu. He notes that sanctions need not be physical – they can themselves be ‘moral’:

there are, and there have always been, legal punishments which are purely moral; such as those which consist in the deprivation of certain rights, as infamy among the Romans, dishonour among the Greeks, civic degradation, etc. (427)

However, Durkheim notes that the distinction between ‘moral’ and ‘material’ sanctions “is not exact, for all moral punishment necessarily assumes a material form.” (427)

Durkheim continues:

This definition alone proves that the positive science of morality is a branch of sociology, for every sanction is principally a social thing. The duties comprising that part of ethics called individual morality are sanctioned in the same manner as the others. That is to say, they are individual only in appearance, for they, too, depend on social conditions. (428)

However, Durkheim writes: “Our definition is still faulty” (431). A wide variety of sanctions are applied in any given social milieu – many of them seemingly incompatible with one another. How are we to select which sanctions, in any given milieu, count as the implementation of the milieu’s moral code, and which are deviations from that code?

[For the record: up until this point I myself have more or less been on board with Durkheim’s analysis. As any regular readers will know, the Brandomian metatheoretical apparatus I endorse is within the Durkheimian tradition, in that it also understands ethical rules in terms of sanctions – indeed, is more openly reductionist than Durkheim is willing to present himself as being. The main reason I’m going through this discussion of Durkheim is to illustrate how closely Brandom’s core commitments align, in many respects, with this foundational work in social theory – and then to show how they diverge. It is at the following point in the argument that I start disagreeing with Durkheim.]

Durkheim is asking himself how we recognise “the moral facts” through sanctions that express them, when there are also sanctions that contradict those “moral facts” (424). Here Durkheim draws an analogy between the sociologist and the biologist.

The question does not differ essentially from the one the biologist asks when he seeks to separate the sphere of normal physiology from that of pathological physiology (432)

For Durkheim, the scientific investigation of moral facts is the investigation of the normal and the pathological in social life. What is normal in a given social milieu and what is pathological? Biologists

call a biological phenomenon normal in a determined species when it is found in the average specimen of that species, when it is a part of the average type; and, contrariwise, it is pathological when it is not within the average, whether it be above or below it. (432)

A biologist can study a wide range of individuals within a species, and derive the normal characteristics of the species by finding the most typical or average attributes.

If, for example, in a given society, the heights of all individuals are taken and if one puts in columns the figures thus obtained, beginning with the highest, one observes that the most numerous and most closely related statistics are massed in the centre…. It is this central dense mass which constitutes the average (432)

(It is worth remembering that Durkheim is writing at a time when statistical science is in its infancy, and is pioneering the practices of applying statistical methods to ‘sociological’ questions, just as he is carving out a field of questions that can be referred to as ‘sociological’. So the simplicity of the statistical concepts here shouldn’t really be held against him.)

Durkheim writes that:

The same method must be followed in ethics. A moral fact is normal for a determined social type when it is observed in the average of that species; it is pathological in antithetical circumstances. (432)

It is not enough, however, Durkheim continues, to determine whether a moral fact is normal for a given society by analysing that society at a particular point in time. For just as organisms grow and change, so do societies. What is normal for an organism in its infancy may not be normal in its maturity – and we should not judge a mature creature pathological because we are judging it by standards derived from the analysis of infants.

Consequently, to know if a moral fact is normal for a society, we must take into account the age of the society and determine the normal type which serves as landmark. (433)

Durkheim acknowledges that this indicates a limitation of his analytic approach, as it currently stands – for the scientific analysis of societies is not yet sufficiently advanced to allow us to determine their degree of development.

Only according to certain characteristics of the structure and functions is it possible scientifically to distinguish old age from infancy, or maturity, and these have not yet been determined with sufficient precision…. Ultimately, the progress of science will make this determination more exact. (433)

Furthermore, Durkheim writes, we may not always be able to observe a mature society of a given type – for example “when the moral conscience of nations is not yet adapted to the changes which have been produced in the milieu” (434). In this scenario:

the method remains the same…. We can determine the new conditions of the state of health only in the functions of the old, for we have no other point of comparison. To know if such and such a precept has moral value, we must compare it with others whose intrinsic morality is established. (434)

After these considerations are taken into account, Durkheim arrives at the following definition:

One considers as a normal moral fact for a given social type, at a determinate phase of its development, every rule of conduct to which a repressive diffuse sanction is attached in the average society of this type, considered at the same period of evolution; secondly, the same qualification applies to every rule, which, without precisely presenting this criterion, is, however, analogous to certain of the preceding rules; that is to say, serves the same ends, and depends upon the same causes. (435)

This is the starting point for Durkheim’s ‘science of morality’ – and for his overall intellectual project, that he would pursue through his subsequent major works.


What are we to make of Durkheim’s framing of his project here? Various criticisms could be made – I’ll highlight a few.

The first broad charge that could be leveled at Durkheim’s project here, is that he is guilty of what Brandom calls ‘regularism’. He is equating moral obligations with regularities of practice, which is to say he is equating what ought to be done with what is done. There are various objections to this move:

– that it is morally bankrupt, because it fails to differentiate morality from actual practice in a manner that would allow us coherently to criticise the practices of our social milieu;
– that is conceptually incoherent, because a defining feature of morality is its capacity to be differentiated from what is actually done;
– that is is not empirically adequate to even the social practices it describes, because it may be the case that a society accepts as immoral practices that it in fact does not adequately sanction, and vice versa
– that it is impossible to adequately delimit the relevant regularities without importing one’s own normative judgements to the analysis;
– that even if the relevant regularities could be adequately delimited, there are always multiple moral ‘rules’ compatible with any given empirical regularity of practice (Wittgenstein’s argument about following a rule).

In response to at least some of these objections Durkheim (or a neo-Durkheimian like David Bloor) could respond that we are not here interested in whether a set of moral rules is correct or not – we are merely interested an objective sociological analysis of the ethical behaviours of a given society. This is (to my mind) a legitimate response. But, of course, the sociological analysis of societies itself must follow a set of normative and ethical criteria, and if it is to be scientifically reflexive the account should be applicable to our own rules as well. And this raises the problem of how and whether it is possible to justify a set of ethical criteria (our own) that are analysed in these terms. So a Durkheimian still needs to address the question of why any ethical rules should be regarded as legitimate, once they have been analysed in these terms.

As readers of the blog will know, I regard Brandom’s work as having addressed and basically settled this issues. I think Brandom shows how ethical (and/or normative) behaviour (including that of the researcher/theorist) can be analysed in terms of regularities of practice (indeed, regularities of sanctions) as Durkheim advises, without becoming vulnerable to these criticisms. A more sophisticated apparatus than Durkheim’s is required to get this to work – but Durkheim is a trailblazer, here, in the foundation of sociology and theories of practice, and I think a theoretical framework operating in a broadly Durkheimian tradition can comfortably handle this set of objections. All this has been exhaustively discussed on the blog previously, and I won’t rehearse it again now.

In this area, then, I think that (a modified; more sophisticated version of) Durkheim’s approach can be made to work.

What other objections are there to Durkheim’s approach?

The second broad charge that could be leveled is that Durkheim’s position is a relativist one. If moral facts are to be understood relative to a given social type, then there would appear to be different moral facts for different social types. How is one to decide between them?

This objection can be folded into the previous one, in the sense that if we have the resources to explain the warrant for our own normative commitments, without giving up on the analytic machinery of a sanctions-based scientific analysis of ethical social practice, then we’re pretty much home dry re: relativism. I’ve just said, though, that Durkheim’s own apparatus isn’t able to pull this off. And of course Durkheim does not take the Brandomian approach to this problem (as I would advise). Instead, Durkheim embeds his different social types within a tacitly teleological evolutionary narrative: there is a hierarchy of more or less developed societies, where the scientific social type – the social type of the researcher or theorist – is the most advanced.

This brings us to the third broad objection to Durkheim’s theory: that the analogy between social analysis and biological analysis that Durkheim draws does not work. I agree with this; or, rather, I think that the model of biological analysis that Durkheim is tacitly relying on is faulty, and that even if it were not, the analogy would still be an unhelpful one. But this needs to be broken down into several sub-points.

– Durkheim draws an analogy between a society and a species. There are a couple things wrong with this. In the first place, a society is a more fluid and permeable thing than a species. While, of course, species change over time, and the boundaries of what counts as a species are by their very nature unclear, there is a point at which any two given individuals are sufficiently genetically differentiated that you’re not going to get sexual reproduction. This analogy doesn’t hold with cultures: there’s no reason why you can’t get cultural transmission between two very ‘distant’ cultural milieus. So Durkheim is using the analogy with biology to reify social spaces. At the same time (secondly), it’s important to note that while one can characterise a biological species by finding its ‘average’ qualities, the motor of evolution is deviation from the average: one perspective’s ‘pathology’ is another perspective’s ‘development’ (in biology as in culture…) and there is no theoretical way to differentiate between the two.

So Durkheim’s (poor) analogy between societies and species enables him to reify the former, and not deal with the extensive mutual permeation of ostensibly discrete societies’ social practices. At the same time, this analogy allows Durkheim (like many) to write as if it is unproblematic to refer to a single set of ethical rules for a given social milieu: Durkheim’s social spaces are tacitly lacking in social contradictions – that is, contradictions where both sides of the contradiction are core to the constitution of the space. Instead the social contradictions Durkheim’s theory can accommodate are those between central norms and deviations from those norms. This greatly restricts the kinds of social spaces that Durkheim’s apparatus can adequately theorize.

Durkheim’s approach errs, then, in imagining or implying discrete and largely non-permeable social spaces with a single coherent set of ethical rules associated with each space. This is the first thing wrong with Durkheim’s analogy with biology. This is not all, though.

– Durkheim also embeds his discrete societies within an evolutionary continuum. The dynamics of this evolutionary process are imperfectly known to social science – but Durkheim has no doubt that they will, over time (as our own social milieu evolves and develops) become better known. This evolutionary dynamic within which the social spaces analysed by social science can be placed, is a social reality which transcends the social spaces that are otherwise, for Durkheim, generative of our ethics. The scientist has access to a transcendent reality that informs and grounds the analysis. This knowledge does not transcend history, but it transcends any given historical moment – it is the overarching logic of the development of social spaces. Our sense of this dynamic is not (unlike the analysis of the sanctions of a given milieu) historically informed. It is instead (I believe) a quasi-mystical belief. The Durkheimian social scientist thus exempts their own analysis from the empirical historicisation to which they subject other social practices. And this evolutionary movement is not a Darwinian evolutionary movement, of contingent change, but rather a teleological movement of historical development, as a given social space moves from its infancy to its maturity to its decline.

So: to sum up. From our perspective – or at least from the perspective that I’m advocating here – Durkheim’s most general theoretical commitments, as expressed here, can be separated into two strands.

1) A rationalist, empiricist, or positivist commitment to analysing social actors’ ethical commitments entirely in terms of empirically observable social practice, specifically social sanctions. I endorse this, and believe that Brandom’s work shows how sophisticated and rich a theoretical apparatus can be developed within this tradition.

2) A quasi-mystical analysis of societies in terms of non-Darwinian biological theory, in particular:
– the treatment of social spaces as characterised by discrete and unitary sets of normal ethical commitments accompanied by pathological deviations from those commitments. (This is not always and necessarily a bad way to analyse a social space, but we should never assume it is an appropriate framework for any given analysis.)
– the embedding of these discrete social spaces within an overall historical developmental dynamic, which is non-empirically known and teleological in its movement.
I think this set of commitments is wrong and analytically unhelpful.

Obviously in the ‘meta theoretical’ work I’m doing on the blog these last few years I’m interested in working within the tradition of, and extending, (1). In the future I’d also be interested to spend more time looking at (2), and the intellectual history of the social-theoretic use of bad biological analogies (lebensphilosophie, etc.).


2 Responses to “Durkheim’s Science of Morality”

  1. duncan Says:

    I’m not entirely happy with my characterisation of Durkheim’s sense of social/historical teleology, towards the end of this post – feel I’m missing nuances / detail there. I need to spend more time on the social-theoretic use of teleological biological theory. It’s very prevalent (not just in older figures like Spencer or Durkheim but even in much more recent canonical social theorists like Habermas) – but I don’t feel I have a nuanced handle on the different ways this biologism can be used or understood.

  2. duncan Says:

    So – I need a clearer sense of the specific way Durkheim is deploying biological analogy and teleological understanding of social development: I think my characterisation of Durkheim’s treatment of these things above is a bit simplistic. Also: although I think I’m right about the assumption of unity, above, Durkheim also of course centrally thematises the unity – or not – of social spaces, so it’s not as if this is a tacit, unconsidered assumption. So I need to do more work on Durkheim’s treatment of social unity (if I want to keep pursuing this line of thought / research).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: