[Taylor, Charles (1980), Formal theory in the social sciences, Inquiry 23:139-44]

[ Books | Articles | Reviews | Index | The main menu ]

Formal Theory in Social Science

[start of page 139]

Charles Taylor

McGill University and All Souls College, Oxford

Contemporary social science tends to suffer from too many misplaced attempts at mathematical or game-theoretical formulation, and much effort is 'wasted in either propounding such formulations, or in showing their inanity. Jon Elster does not entirely escape this himself, but Logic and Society is truly remarkable in pointing the way to some possibly very relevant formalizations. These are particularly to be found in the chapter on 'contradictions of society'. There Elster attempts to delineate the properties of certain self-frustrating predicaments of action, and to relate them as well to Hegelian-Marxist conceptions of contradiction. It may be, however, that the relevance of this analysis is restricted to societies whose form of life has become atomistic, and whose members thus function in an individual-calculative way in many spheres of life. But even so, formalizations of the kind Elster provides would be useful in defining one kind of historically evolved society among others.

I have to confess that I started reading Jon Elster's Logic and Society (LS) with great scepticism. I believe that we suffer from a veritable pandemic of misplaced mathematization in the contemporary sciences of man. There is a massive violation of Aristotle's injunction not to try to treat a subject with a degree of exactness it will not admit of.

The disease is in fact even more insidious than it might at first appear. For the rash of inappropriate formalizations encourages in its turn a spate of refutations. For every new formal gimmick advanced for bad reasons, we have an elegant demonstration of its inaccuracy or uselessness proffered for the best of reasons. But after even the best of refutations, one is left with a profound sadness: was all the effort worth while, just to bring us back to point zero of insight into man and society? As one who devoted half a middle-size book to an attempt to refute behaviourism, I can speak with feeling on this matter.

I think that Jon Elster's book contains its share both of sterile new formalizations, and of refutations of past inanities. The proposed machinery of states politically possible with respect to other states strikes me as

[end of page 139, start of page 140]

an example of the former. How do we even begin to develop some rigorous mode of identification of a political state? And what possible way do we have of determining political possibility? Of course, we all do just this when we determine what to do as political agents. Social Democrats may have decided that a more radical transformation of society to an integrally socialist one is impossible; or impossible without violence; and therefore have opted for the brand of reformism they in fact espouse. Their liberal opponents may be such because they believe that even this programme of reform will be unrealizable without imposing unacceptable limits of inflation or state intervention (therefore unrealizable without coercion). Their radical critics probably hold that the system is heading for a crisis which it will not be able to resolve, and therefore that non-change is impossible.

But each of these positions involves different identifications of the current state, different notions of what is constant and what is changeable in human life; ultimately different notions of man and the bases of human life, when we get to the major divergences. The conflict between these views is the very stuff of political controversy. But this involves argument over the very terms of our thought about man and society. There is no possibility of our being able to couch our differences in terms of some agreed vocabulary of political possibility with respect to given states. Even the field of political possibility is very differently structured for people with different outlooks. For those committed to the liberal-democratic-representative institutions common in the West, there is a circumscribed field of outcomes which are possible within this framework (though even they cannot agree among themselves on how to categorize what falls within this framework); this contrasts with changes possible through revolutionary transformation. These are seen as quite different in kind. But this is not the case with certain revolutionaries, e.g. those for whom the present regime is simply a form of institutionalized violence. Nor can this be dismissed as simply a difference in 'values'. It involves a fundamental difference in the understanding of society.

But leaving the criticism of details aside, I find that my prejudices are somewhat upset by this book. Elster is obviously as adverse in principle as I could ever be to useless formalizations. What impresses the reader from the very beginning is Elster's determination that formalizations should serve the purpose of opening real insight into history and society. 'In the writings of people who are good at some or other formalized science, logic, mathematics, choice theory, one senses a joy like that of a child with a new toy. This is extremely engaging, and sometimes exhilarating, because it is

[end of page 140, start of page 141]

the pleasure inseparable from the exercise of the highest intellectual virtues. But for experienced readers concerned with social science, such ludic display brings on that sinking feeling that we are following another dervish dance into a dead end.

With Elster, however, this joy is severely controlled by a sense of higher purpose, which he expresses very tersely on p. 158: 'The basic postulate from which I start is that the goal of the social sciences is the liberation of man.' Not surprisingly, therefore, the most interesting part of his book is Chapter 5, on the 'contradictions of society'. Elster explores there at length the kinds of self-frustrating predicaments which he designates by the terms 'counterfinality' and 'suboptimality'. Both of these arise in circumstances where agents in a society have not yet achieved, or have shunned, explicitly collective action, where each pursues an individual policy which is bound to be frustrated by their joint predicament. In these situations, a negative variant of the invisible-hand mechanism holds: behind their backs the agents are brought to ruin (or at least unhappy consequences) through their own efforts.

Elster tries to show the close link between the mechanisms at work in this kind of situation and (a certain definition of) the fallacy of composition. In cases where it is rational for a single agent to follow a given policy, provided he does so alone, all may come to ruin in doing so. E.g. it is rational for me as a small farmer among many to raise production when the price drops, in order to maintain my income; but when all act similarly the price may drop to the point where we fail to recover our costs of production.

Elster is interested in how contradictions of this kind (and he has a lengthy defence of the use of the term 'contradiction' in this context, which does not open it to the vacuous indiscriminate application so common among Marxists) can illuminate social change. And that in two ways: certain contradictions can generate unintended change. Others can induce us to take effective action to overcome them.

In this latter context, he puts forward (on p. 135) an interesting hierarchy of levels of social action. At the lowest level, we have simply response of the individual to his environmental situation, others being taken just as part of this environment. At the first level of learning, the agent realizes that others too are responding to their environment, and adapts his action to take advantage of this. At a still higher level, the agent can take account of the fact that others are reasoning, like him, about everyone else's reasoning, including his. We are now on the level of

[end of page 141, start of page 142]

game-theoretic or strategic thinking. The contradictions at this level can only be overcome by moving to the highest one, viz. collective action.

I think this is very interesting and relevant because modern society has developed in such a way that these different types do really correspond to widely-experienced predicaments of social action. That is because modern society has developed institutions, and a mode of identity and action, in which people do act more or less as atomic individuals in some sense calculating their advantage. This has opened them to the kind of self-frustrating predicament that Elster explores in different variants, and which Marx tried to make central to his analysis of capitalism.

One of the most powerful criticisms that can be levelled at modern Western society lies in the claim that the emancipation of the atomic individual leads inevitably to such self-defeating predicaments. A critique of this kind, if valid, would justify the attempt at some variant or other of socialist transformation, i.e. one to a society where collective control over our social and economic life was much greater. Elster has considerably refined the tools by which a critique of this kind can be made. In this he is following a direction pioneered by Marx, but in a way which can learn from Marx's mistakes, and indeed, in many ways leaves Marx far behind.

It is not surprising, therefore, that in the second part of this fifth chapter, Elster tries to discuss some of the conditions that might favour or retard such an extension of collective action. But his discussion here seems to me to be partial (Is this a criticism? Whose discussion of such an issue could be comprehensive?). He looks at the various obstacles to the formation of a common disposition to collective action. But he doesn't take on the issue of the possible limits to such effective action.

These may spring either from the undesirable side-effects of collective action, in a domain where it is indispensable, on individual autonomy in other domains. It may be that the requirements of completely overcoming counterfinality in the economic domain might involve such a bureaucratization of life, or such a powerful state apparatus, or such a degree of regulation of other activities, that we might find the price too high to pay. Or else, the obstacles might spring from the very complexity of a modern technological-industrial economy. Planning control over such an economy may require such a sophisticated and many-sided style of intervention that it will either defeat the planners, in the sense of failing to produce the consequences they aim at, or only succeed at the cost of elite control by trained experts.

Obstacles of both kinds have been explored by right-leaning thinkers,

[end of page 142, start of page 143]

but, strangely enough, with a lack of depth and realism which is astounding. This is because the intellectual right in our civilization is generally under the spell of an absurdly unrealistic atomist view of society. From this position it can denounce in stark terms the dangers of serfdom under the social-bureaucratic state, but it is not very good at understanding how this atomistic society would actually work as a collective form of life. This is because the intellectuals of the right have trouble understanding how any society works, as one must if one thinks on atomist premisses.

But the kind of analysis which Elster proposes is in some danger of falling into the same trap. This is because the self-defeating predicaments explored in Chapter 5 are those of a society many of whose practices have been marked by atomism. I mean by that a society within which institutions and practices presupposing an atomistic identity have come to assume an important place. It is not an accident that the perspicuous examples of such predicaments are in a great majority of cases drawn from the behaviour of agents in a modern economy. This has been the prime locus in which the atomistic-calculative side of the modern identity has been institutionalized. Even when we operate in this way in other areas of life, we tend to talk in economic metaphors about what we are doing.

We need to put this kind of analysis in historical perspective. There are other ways of being together in society than either atomistic calculation of individuals on the one hand or consciously collective formation of policy on the other. We can also be related to our compatriots through a common formula which is seen as giving the pattern of our common existence, and of the respective place each one of us holds in it. Some variant or other of this seems to have been dominant in most pre-modern societies. Indeed, it would perhaps be better to say that some variant of this common formula exists in all societies, and define modem society as a special case in which the common formula calls for members to act as atomic agents through important ranges of activity.

We need a theory of different types of social cohesion if we are to identify properly the contexts in which the kinds of self-defeating predicaments analysed by Elster can arise. This is not to say that self-defeating predicaments may not arise in any society. There are clear examples of pre-modern societies following ecologically disastrous agricultural practices, for instance, which Elster also evokes; and there are others where firmly entrenched criteria of honour and success were economically disastrous.

But the peculiarity of the predicaments singled out by Elster in this

[end of page 143, start of page 144]

chapter is that they permit of analysis in game-theoretic language; and this is not true of the hidalgos, for instance, who couldn't take advantage of Spain's sixteenth-century windfall, or of tribesmen practising slash-and- burn agriculture in the face of an advancing desert. In each case there is a self-frustration: the attempt to remain superior led to inferiority; the attempt to survive leads to extinction. But there is not the element of calculation by agents who conceive of themselves as individuals. It is where this latter is true that Elster's applied formalisms can be illuminating.

The understanding of the pre-modern may help us to attain the post-modern as well. It may be that we will be able more successfully to overcome some of the counterfinalities of our modern economies if we are able to achieve some degree of self-understanding as communities. And this in turn may require just as much a recognition of a common formula which we see as binding on us all, as it will a common resolution to adopt some agreed policy. I sense in Elster's way of posing the problem a highly voluntaristic perspective. And this is borne out by his espousing what he calls 'the Kantian idea of freedom', which he sees as requiring 'that man should somehow be able to choose himself' (LS, p. 16).

I confess that I believe that this is a wrong, and ultimately incoherent, notion of freedom; that conceptions of socialism based on it are unrealizable, and fraught with destructive consequences in their attempted implementation. I think one of our biggest intellectual tasks these days is to find an adequate conception of situated freedom. I think that this turns out to be closely related to the task of situating the contexts in which our formal theories can be fruitful. It may be here that I disagree with Elster.

Wherever we draw the boundaries of their application, however, we are all in his debt for the categories he has developed, which have the makings of an insightful new social theory. In one bound, he has taken the discussion of formal theory in social science several leagues ahead.

[end of page 144]

[Taylor, Charles (1980), Formal theory in the social sciences, Inquiry 23:139-44]

[ Books | Articles | Reviews | Index | The main menu ]

1 1