Sveinbjorn Thordarson's Website

Methodological individualism is the philosophical position that social phenomena are to be accounted for in terms of the actions of individuals. Social institutions and social development should be understood as no more than the aggregation of individuals and their dispositions. This doctrine traces its roots to the writings of Weber, and is commonly understood as requiring that any sociological account must, in either an explanatory or reductionist sense, be built upon or based on an “atomistic” account, the social atom being the individual and his actions. Ontologically, a strong form of the methodological individualist thesis commits us only to the existence of individuals – all social entities may be described as aggregations of individuals, much as a table may be described as a collection of atoms. Social entities, then, are in a sense the product of individuals, from the microscopic bottom up to the higher level macroscopic analysis.

Methodological individualism is often contrasted with what is called methodological holism. Methodological holism is committed to the claim that there exist social facts which are not merely the aggregation of facts about individuals. Social facts exist in and as of themselves (Durkheim 1994). Furthermore, there are laws which apply to the social system, and which affect individuals and regulate their behaviour. Society should be analysed “from the top down” – starting with irreducible social laws, which in turn may be used to account for the behaviour of individuals.

Which of these philosophical views is a more persuasive foundation on which to base the social sciences? In this paper, I will examine the intuitive and philosophical appeal of methodological individualism. Furthermore, I will evaluate some of the criticisms that have been levelled against it.

Methodological individualism, as I said, has an immediate intuitive appeal: social phenomena are quite obviously composed out of the actions of people. This is trivially true, in the sense that it is possible to have an individual without social phenomena, while the reverse is certainly not possible. In this sense, methodological individualism obviously strikes home -- individuals are necessarily the building blocks of society, and all social phenomena necessarily arise out of the actions of individuals, whether directly or indirectly. The subject of the social sciences is the behaviour of man on a large scale. Reinforcing our inclination to accept this common-sensical view is the alternative to methological individualism, which seems to require that human actions fall under social laws, which is anathema to our notion of autonomy. But do these intuitions mean that any sociological laws are bound to be explicable in terms of the actions of individuals, and that terms over social wholes are to be understood purely as aggregates that add nothing to the members of which they are constituted?

Before we proceed, it is important that we distinguish between two aspects of methodological individualism: First of all, there is the aspect which consists of the claim that all references to social entities can be linguistically reduced to references to individuals, in the same way that the positivists believed that biology was reducible to chemistry, and chemistry to physics. Let us call this the case for reduction. Secondly, there is the aspect consisting of the claim that any explanation of a given social phenomenon is only final and satisfactory once it is provided in language which refers solely to individuals and their actions. Let us call this the case for explanation. Both these claims have been met with severe criticism from different directions.

Let us begin by looking at the the case for reduction. The philosophical literature on reduction is enormous, and it is therefore hardly within the scope of this paper to give an adequate in-depth analysis. However, it is worth considering an example of what a reduction from statements referring to holistic sociological entities to statements referring only to individuals might look like. Take a sentence such as:

The parliament decided to pass the law unanimously

Now, it seems quite clear that this sentence can be reduced to another sentence which does not make use of the holistic term “parliament”. Let us suppose that there were n people in the parliament who passed the law. Then this sentence might be reduced to individualist terms by way of enumeration:

p1 decided to pass the law, p2 decided to pass the law … pn decided to pass the law.

Saying that parliament decided to pass the law unanimously quite simply means that each individual of which it is composed passed the law, or so the individualist would argue. It is the answer we could give if someone asked us “What do you mean by ‘parliament’?”, as opposed to describing it holistically as an ‘institution’ or a ‘political body’ etc. The holistic term is taken to be merely a shorthand of sorts for the set of individuals of which it is composed. Our example, however, is an easy one. There are other holistic terms that may not be as easily reducible. Kincaid points out that “reduction requires equivalences between social and individual terms”. We seem to have a fairly good case of that in our example above, but we run into trouble when we try to deal with terms that can be realized by a multitude of different configurations of individuals. Kincaid mentions terms such as “revolution”, “bureaucracy”, “power elite”, etc. An infinite variety of different individual configurations could become that which any one of these terms refer to. A mapping between these terms and particular configurations may be quite difficult, and therefore “social predicates may be definable, but generally not individualistically”. In order to provide an adequate definition of terms such as these, we are forced to rely on further holistic social terms, thus perpetuating the task of reduction, possibly leading it to an infinite regress. The methodological individualist reductionist therefore faces an “enormous difficulty”, and ultimately, “the burden of proof lies with [him]” (Kincaid 1994).

What are we to make of Kincaid’s critique on the possibilities of reduction? While it is clear that reducing holistic sociological terms to terms about individuals and their actions or dispositions may be far from an easy task, he fails to demonstrate that it is impossible to do so. I believe that this particular debate may very well be irresolvable, much like the logical empiricist debate concerning the reducibility of the language of science to reports on sense-data. The methodological individualist may not wish to undertake the task of reducing all social terms to terms referring only to individuals, but he can always maintain that it is possible in principle, if not now, then at least in the future. Any problematic counter-examples can be dealt with either by maintaining that they are reducible in theory or by rejecting them as meaningless, insignificant or bereft of any worthwhile explanatory power. Unless the methodological individualist is confronted by a holistic term which he can neither maintain is reducible in theory nor convincingly reject, his doctrine of reducibility remains unrefuted.

Let us now move on to the methodological individualist claim that any explanation of a given social phenomenon is only final and satisfactory once it is provided in language which refers solely to individuals and their actions. What could an explanation of a social phenomenon be, other than an account of the actions of the individual actors of which said phenomenon is composed? It could be an explanation which only makes use of holistic social terms. Could such an explanation ever be fully satisfactory? The methodological individualist denies this. But what makes an explanation satisfactory? As far as I can tell, an explanation is satisfactory once we have no wish to probe further. If we ask a question akin to “Why did a dictatorship arise in Germany in 1933?”, we may very well be given an answer consisting solely of holistic social terms, e.g. “The entrenched business elite, combined with fear of Communism amongst the burgeoisie made a considerable part of German people believe that nationalistic dictatorship was a solution to its problems”. But how satisfactory is an answer of this kind? Would not only a person strikingly devoid of curiosity rest satisified with an answer as generic as this one? Don’t we need a further account of these ‘entrenched business interests’, or these ‘Communist-fearing burgeois’? How deep do we have to go? Do we in principle need to refer, individual by individual, to how the members of these groups acted in order to rest satisfied?

Watkins maintains that we will only have a “rock-bottom” explanation for a given social phenomenon once we have an explanation solely in terms of individuals and their dispositions (Watkins 1994). Now, an obvious objection this might be that people often rest satisfied with explanations which refer to social entities, and take it to be a fully satisfactory explanation. Gellner suggests that although “the holistic terms used by the observer may be eliminable”, they are not eliminable “as used by the participant”, and takes this to be a “clue as to their ineliminability” (Gellner 1968). But surely this objection cannot have much weight. The fact that common language includes references to holistic concepts such as society, the nation-state, castes and classes does not in any way vindicate the use of such terms. The validity of the terms is in no way entailed by their popular use, or by the fact that they are widely invoked as explanations of a variety of things. The popular intuition may very well be misguided, unjustified and inappropriate when it comes to rigorous, scientific explanations of social phenonon -- or so a methodological individualist might very well respond.

Watkins’ demand for a “rock-bottom” explanation in terms of individuals and their dispositions can be interpreted in several ways, of which I regard two to be significant for the purposes of our discussion. Either it can be taken as a demand for a rigorous elimination of all holistic terms in any full explanation, or it can be taken as the claim that an explanation in terms of individuals is by itself sufficient and more “basic” than explanations that include holistic terms. The first can only be seen as a radical attack on almost all the social sciences – a great many of the theories include references to social entities for their explanatory power, amongst them quite firmly entrenched disciplines such as microeconomics. Do these fields provide only pseudo-explanations? Kincaid dismisses this position as absurd, but while dismissing current social science theories on this criterion may not be an appealing prospect, I do not think it can be dismissed out of hand. Given e.g. sociology’s generally poor track record of predictions, and the multitude of competing and arguably arbitrary explanations of social phenomena it offers, it strikes me that a demand that nearly all of sociology be taken back to the drawing board is not a completely absurd option, albeit a radically skeptical one.

What can we say about the second interpretation of Watkins, namely that explanation in terms of individuals is more “basic”? The social sciences try to explain social events by way of theories. A call for explanation of social events in terms of individuals is therefore a call for explanation in terms of individualist theory – i.e. in purely individualist terms. Theories referring to holist social entities might thus be mapped to theories about individuals. But this seems to amount to reduction, which we have previously discussed and do not regard as amenable to a solution. However, an individualist might also argue that for any given instance of a social event which calls for explanation, it is possible to provide an an individualist account on a case-by-case basis. Thus, although we can not (or will not) obtain a mapping between holistic theories and theories containing only individualist terms, methodological individualism can still give a more basic explanation for any given instance.

Individualist explanations on a case-by-case basis, if actually adopted, would be highly artificial and run counter to the construction of theories in social science. If the “basic” explanation is provided, not by the theory, but by a seemingly unrelated account made in individualist terms, then we might very well ask what the relevance of the theory is in the first place. And it goes without saying, I think, that without theories to explain social events, our “desert island” individualist explanations will be merely anecdotal, stripped of anything remotely resembling scientific inference. We cannot, under this scheme, categorise social events scientifically into kinds with recognizable features, or derive laws describing regularities. This seems self-defeating to the point where we might just as well abandon altogether the enterprise of engaging social scientific research.

Kincaid suggests that there “might be individualist laws describing human behavior […] realizing a specific social event” without the predicates of those laws being “coextensive with ones in social theory” (Kincaid 1994). Individuals, then, might fall under some sort of description in holistic terms as a group, while each by themselves fall under a description in individualist terms, in such a way that no mapping between the two descriptions obtains, while producing the same result, analogous with the realization of computer program which can (in theory) be instantiated on an infinite variety of hardware, while still being susceptible to laws that explain any given realization. Kincaid proceeds to reject this account, on the basis that, like case-by-case individualist explanations, it cannot refer to kinds of social events, and therefore would also render attempts to derive laws impossible. The inability to classify kinds of social events does seem to greatly reduce potential explanatory power.

We have seen earlier that the doctrine of methodological individualism has some attractive common-sensical aspects. However, there does not seem to be any firm ground on which to defend it from its critics. If we maintain the reductionist position that all statements about social events can be reduced to statements about individuals, we reach an impasse, since both the methodological individualist and his critic will maintain that the burden of proof rests with the other. The more promising path of demanding full explanation in terms of individuals is problematic, and seems to require that the methodological individualist adopt a stance of radical skepticism about the possible predictive and explanatory powers of the social sciences. Perhaps its intuitive general appeal is misleading.

London 2006

Sveinbjorn Thordarson

Sources

Durkheim, E. (1994) “Social Facts” in Readings in the Philosophy of Social Science eds. Michael Martin & Lee C. McIntyre. London, England: MIT Press.

Watkins, J.W.N. (1994) “Historical Explanation in the Social Sciences” in Readings in the Philosophy of Social Science eds. Michael Martin & Lee C. McIntyre. London, England: MIT Press.

Kincaid, H. (1994) “Reduction, Explanation and Individualism” in Readings in the Philosophy of Social Science eds. Michael Martin & Lee C. McIntyre. London, England: MIT Press.

Gellner, E. “Holism versus Individualism” (1968), in Readings in the Philosophy of Social Sciences ed. May Brodbeck.