Note to readers:
In my opinion the new essay by Jon Elster exhibited below (which I have slightly shortened to better fit the email) presents a most important contribution to the general understanding of social science. It offers a synthesis and modification of Elster’s long held view that methodological individualism is necessarily always the essential initial insight in social science. Here Jon Elster shows clearly why the study of social action simply must begin with the individual.
There can be no better way to mark the 2-year anniversary of Social Science Files (Jon Elster joined us on 1st March 2022).
On this day two years ago I exhibited a passage by Leibniz on the ‘Art of Discovery through Synthesis’. I note that Jon’s first publication, in 1975, was a study of Leibniz and emergent capitalist economic rationality.
So now we turn to Elster’s synthesis of the assumption of individual agents…
The Source:
Jon Elster*, ‘What’s the Alternative?’, in The Palgrave Handbook of Methodological Individualism, Volume 1, edited by Nathalie Bulle and Francesco Di Iorio, Palgrave published 27 December 2023
*Subscriber to Social Science Files since 1 March 2022
What’s the Alternative?
1. Introduction
I believe that methodological individualism (MI) is trivially true. Yet eminent scholars, including Kenneth Arrow, have put it in question. I believe most of their doubts and criticisms can be traced back to the following fact. In modern societies, and probably in all societies, individuals meet institutions, belief systems, and value systems that they have to take as given. These are without exception effects or residuals of the actions and interactions of other individuals, but their history is not written on their faces. Denials of MI are largely due, in my opinion, to neglect of this fact.
Although I shall criticise Marx for his violations of MI, I subscribe to his view that in modern societies “the products of the human brain appear as autonomous figures endowed with a life of their own” (Marx 1867 Capital I).
Surprisingly, this view was also stated by Émile Durkheim, who is usually seen as a main opponent of the idea that sociology could, at least in principle, be reduced to psychology. He argued that what he called “social facts” are
“things that have their own existence. The individual encounters them when they are already completely fashioned and he cannot cause them to cease to be or to be different from what they are. Willy-nilly he is therefore obliged to take them into account; it is all the more difficult (although we do not say it is impossible) for him to modify them because in varying degrees they partake of the material and moral supremacy that society exerts over its members. No doubt the individual plays a part in their creation. But in order for a social fact to exist, several individuals at the very least must have interacted together and the resulting combination must have given rise to some new production. As this synthesis occurs outside each one of us (since a plurality of consciousnesses are involved) it has necessarily the effect of crystallizing, of instituting outside ourselves, certain modes of action and certain ways of judging which are independent of the particular individual will considered separately” (Durkheim 1919 Les Règles de la Méthode Sociologique [italics added by Elster])
I now proceed as follows.
In Section 2, I propose a definition of MI and distinguish it from other doctrines to which it bears a more or less superficial resemblance. In Section 3, I cite social scientists whose work violates MI. In Section 4, I discuss some examples that have been proposed to show the inadequacy of MI. In Section 5, I consider the question raised in the title of the chapter. Section 6 has a brief conclusion.
2. Definitions and Misunderstandings
“By [MI] I mean the doctrine that all social phenomena—their structure and their change—are in principle explicable in ways that only involve individuals—their properties, their goals, their beliefs, and their actions” (Elster 1985 Making Sense of Marx).
Today, I would characterize MI as an injunction rather than as a doctrine.
MI is a form of reductionism, which is the engine of progress in science. It enjoins us to seek for microfoundations.
In many cases, a literal attempt to carry out the individualist program will be doomed to failure. The number of agents who contribute to an economic or political decision may be prohibitively large, and the relevant information may be unavailable or unreliable. In such cases, many scholars simply assume that a firm, a political party, or even a nation-state can be considered a single unitary agent.
Often, however, that assumption may be so unrealistic as to make the explanatory enterprise a non-starter. Some “big” problems are too big. Yet whenever it seems at least partly feasible, one should dig in.
[There are e.g.] micro-historical analyses of decisions by large assemblies that were only loosely organised by shifting alliances and factions. At the Federal Convention of 1787, the delegations voted as states, each state delegation aggregating the votes of its individual delegates. The recorded fact that a state voted in favor of a proposal does not allow us to determine whether its vote was close or not. Yet by painstaking triangulations, historians have been able to open these black boxes and make that determination.
Economists tend to respect MI, at least in models of general equilibrium. In their econometric applications, they rarely do so, since the units tend to be households or firms rather than individuals …
[MH: here omitted quote of a typical narrative of economists discussing “intra-household allocations”, “households that have more than one member”, assuming “decisions are made, outcomes are efficient”.]
… Efficiency (Pareto-optimality) [as often described in modern economics] means that when household members bargain over the allocation of household income, the bargaining never breaks down. In the real world, the situation may not always be so idyllic.
To respect MI, efficiency would have to be demonstrated, not stipulated. [The examples quoted show that] MI can be extremely demanding. …
… As I defined it, methodological individualism does not make any substantive psychological assumptions about human behaviour.
It does not assume that individuals are selfish, or that they are rational. Nor does it treat them as “atoms”, if we understand that metaphor to mean that individuals behave like particles in Brownian motion, bumping into each other randomly.
In the belief-desire framework that is standardly used to understand individual behaviour, both beliefs and desires may be oriented towards other people, for instance, through empathy and sympathy.
The preceding paragraph addressed some tedious and trivial issues, mentioned only because they are occasionally raised by critics of MI. Two more interesting misunderstandings arise from the confusion of MI with political individualism and ethical individualism. The former, which we may also refer as “liberalism”, amounts, roughly speaking, to a claim, or a set of claims, that state interference in the economy and in civil society should be minimised.
There is no logical connection between MI and this normative claim, but there may well be a loose sociological connection. In recent French sociology, for instance, Raymond Boudon stood for both methodological and political individualism whereas Pierre Bourdieu was an opponent of both.
The present writer and no doubt many others embrace methodological but not political individualism, whereas it seems harder to find the converse combination. These issues belong to the sociology of science, not to the philosophy of science, which is my subject here.
Ethical individualism is a less well-known idea. It can be defined, or approximated, as the principle that when assessing the welfare of a society, one should count only the properties of individuals, not properties of supraindividual or non-individual entities.
Marx, I believe, endorsed ethical individualism. He always took it for granted that the goal of communism was the self-realisation of human beings, not of humanity (as embodied, for instance, in Gothic cathedrals). Louis XIV, Napoleon, and de Gaulle denied ethical individualism. For them, the greatness of France dominated the welfare of the French.
3. Some Violations of Methodological Individualism
I shall now consider violations of MI. In some cases, these overlap with another violation of the scientific method, namely functional explanations.
Generally speaking, these explain actions by their actual (not merely intended) beneficial consequences for someone (e.g., the capitalist class) or something (e.g., social cohesion). Usually, these explanations are metaphysically impossible, as they explain the present by the future rather than by the past.
There are exceptions, however. B. F. Skinner (1981 [PDF of Skinner article here]) argued for the importance of three ways in which behaviour can be explained by its consequences: by natural selection operating on individuals, by reinforcement, and (although Skinner does not use that term) by group selection. I only state without argument that none of these mechanisms fit the examples I shall discuss.
I shall first discuss some historical examples and then some contemporary ones.
Marx
An example that embodies both violations cited above can be taken from Marx:
“The circumstance that a man without fortune, but possessing energy, solidity and business acumen may become a capitalist [...] is greatly admired by apologists of the capitalist system. Although this circumstance continually brings an unwelcome number of new soldiers of fortune into the field and into competition with the already existing individual capitalists, it also reinforces the supremacy of capital itself, expands its base and enables it to recruit ever new forces for itself out of the substratum of society. In a similar way, the circumstance that the Catholic Church in the Middle Ages formed its hierarchy out of the best brains in the land, regardless of their estate, birth and fortune, was one of the principal means [Mittel] of consolidating ecclesiastical rule and suppressing the laity.” (Marx 1894 Capital III)
The two explananda—upwards mobility into the capitalist class and into the church—have plausible explanations: the absence of a caste system and clerical celibacy. In both cases, the infusion of fresh blood may, arguably, have strengthened the grip of the elites on the people. The agency belongs to the individual upstarts, not to “capital” or the Church. Marx, however, imputed agency to these supraindividual entities.
The use of the word “Mittel’ [means] is especially revealing. A means presupposes an end to which it is a means, and an end presupposes an agent for whom it is an end. Thus capitalists and church leaders may benefit from the upward mobility, but if so, only as an accidental byproduct.
As in many other passages, notably in the Grundrisse, Marx here treats “capital” as an agent. His most explicit denunciation of MI is the following:
“The predominance of capital is the presupposition of free competition, just as the despotism of the Roman Caesars was the presupposition of the free Roman ‘private law’. As long as capital is weak, it still itself relies on the crutches of past modes of production, or of those which will pass with its rise. As soon as it feels strong, it throws away the crutches, and moves in accordance with its own laws. As soon as it begins to sense itself and become conscious of itself as a barrier to development, it seeks refuge in forms which, by restricting free competition, seem to make the rule of capital more perfect, but are at the same time the heralds of its dissolution and of the dissolution of the mode of production resting on it. Competition merely expresses as real, posits as an external necessity, that which lies within the nature of capital; competition is nothing more than the way in which the many capitals force the inherent determinants of capital upon one another and upon themselves.” (Marx 1857 Grundrisse ; his italics)
Marx’s economic theories, notably the labor theory of value, rely on these ideas. Consider a firm that produces shirts. The desire for profit will induce the capitalist to innovate, in order to undersell his rivals. In doing so, he will also reduce the value of labor-power, to the extent that shirts form part of the basket of necessary consumption goods, the labor-value of which determines the labor-value of labor-power in all industries, not only in his own. When the value of labor-power is reduced, profit is increased. This collateral benefit—a kind of gift from the shirt producer to all other producers—is, according to Marx, the real aim of the process of capital accumulation.
He [Marx] writes that
“the shortening of the working day [...], is by no means what is aimed at [bezweckt ] in capitalist production, when labour is economised by increasing its productivity. It is only the shortening of the labour-time necessary for the production of a definite quantity of commodities that is aimed at” (Marx (1867) Capital I).
Note, however, that the shirt producer will not “aim” at shortening the necessary labor time, except in the marginal case that his workers consume a large proportion of the shirts. (Henry Ford is reputed, implausibly, to have said that he made his cars cheap so that his workers could afford to buy them). The benefit to the capitalists of the fall in the value of labor-power occurs only as a collateral byproduct when all producers of consumption goods innovate to cut prices.
Marx’s argument is an instance of the fallacy of cui-bono [MH: who stands to gain] combined with post hoc ergo propter hoc [MH: after this, therefore because of this].
Marx’s theory of history was teleological, explaining the present by the future rather than by the past. As he stated in the Introduction to the Grundrisse, “human anatomy contains the key to the anatomy of the ape”.
He consistently viewed the vicissitudes of nineteenth-century European politics as stepping-stones to the Communist revolution; a defeat of the workers was always seen as reculer pour mieux sauter [MH: to draw back in order to make a better jump].
The only exception is found in the Preface to the 1859 Critique of Political Economy:
“In the social production of their existence, men inevitably enter into definite relations, which are independent of their will, namely relations of production appropriate to a given stage in the development of their material forces of production. The totality of these relations of production constitutes the economic structure of society, the real foundation, on which arises a legal and political superstructure and to which correspond definite forms of social consciousness. The mode of production of material life conditions the general process of social, political and intellectual life. It is not the consciousness of men that determines their existence, but their social existence that determines their consciousness. At a certain stage of development, the material productive forces of society come into conflict with the existing relations of production or—this merely expresses the same thing in legal terms—with the property relations within the framework of which they have operated hitherto. From forms of development of the productive forces these relations turn into their fetters. Then begins an era of social revolution. The changes in the economic foundation lead sooner or later to the transformation of the whole immense superstructure.”
Remarkably, coming from Marx, there is no reference, as there is in the Communist Manifesto, to the role of the class struggle in social revolutions. In fact, there is no mention of any kind of agency.
The textually most plausible reconstruction of what according to Marx motivates revolutions is that in a given set of property relations, there comes a point when the rate of technical progress becomes inferior to what it would be under another set of property relations.
At this point, the class that would gain from a revolution brings about the new property relations. Substantively, however, this torch-relay theory of history is far from plausible.
Although G. A. Cohen’s reconstruction of Marx’s theory brings out its logical structure admirably well, by showing that it depends on a certain kind of functional explanation, it does not provide microfoundations for the theory.
Religion
Methodological individualism is also denied, at least implicitly, when social phenomena are explained by the fact that they satisfy a need.
Consider two statements: “Individuals need religion” and “Societies need religion”. With respect to democratic societies, Tocqueville endorsed both statements, but only the first need as explanatory. On the one hand, he wrote that
“for me, I doubt that man can ever bear complete religious independence and full political liberty at the same time; and I am led to think, that, if he does not have faith, he must serve, and, if he is free, he must believe” (Tocqueville Democracy in America).
This first claim is explanatory: individuals in democracies adopt religion to alleviate the psychological burden of freedom. On the other hand, he stated that
“at the same time that the law allows the American people to do everything, religion prevents them from conceiving of everything and forbids them to dare everything” (ibid .).
This second claim merely states a happy coincidence: because the first claim is true, individuals are not tempted to exploit the opportunities for licentious behaviour that democracies create. Tocqueville never says that religion exists because society needs it, that is, because it has good consequences for society.
In The Elementary Forms of Social Life, Durkheim (1915) wrote that.
“Society could not abandon the categories to the free choice of the individual without abandoning itself. If it is to live, there is not merely need of a sufficient moral conformity, but also there is a minimum of logical conformity that it cannot do without. For this reason it uses all its authority upon its members to forestall such dissidences. Does a mind ostensibly free itself from these forms of thought? It is no longer considered a human mind in the full sense of the word, and is treated accordingly.”
Elsewhere he comments more specifically on religion:
“Howsoever complex the outward manifestations of the religious life may be, at bottom it is one and simple. It responds everywhere to one and the same need, and is everywhere derived from one and the same mental state. In all its forms, its object is to raise man above himself and to make him lead a life superior to that which he would lead, if he followed only his own individual whims: beliefs express this life in representations; rites organise it and regulate its working.” (ibid.)
These and many other needs-based explanations ignore the fact that social needs do not generate their own satisfaction. When a need is satisfied, it may be because its satisfaction eases the mind, as in Tocqueville’s first claim, or, more simply, that the members of the society become aware of their need and act consciously and in concert to satisfy it.
Kenneth Arrow
As noted, and as further discussed below, Arrow expressed doubts about MI. Consistently with these doubts he suggested,
“as one possible interpretation that [norms of social behaviour] are reactions of society to compensate for market failures. [...] Society may proceed by internalisation of these norms to the achievement of the desired agreement on an unconscious level. There is a whole set of customs and norms which might be similarly interpreted as agreements to improve the efficiency of the economic system (in the broad sense of satisfaction of individual values) by providing commodities to which the price system is inapplicable” (Arrow 1971).
As stated, these ideas are blatant violations of MI.
“Society” does not react or act; only people do. People do not agree to establish social norms. In any case, Arrow’s analysis cannot serve as a complete theory of social norms, as many of them are harmful rather than beneficial.
In a restatement of his views, Arrow (2001) seems to refute the view that “social norms arise to meet social needs”. He now associates this view with the work of Talcott Parsons, which he sees as “mostly quite empty of serious content”.
At the same time, he [Arrow] found himself
“compelled by [his] observations to argue that social norms play a major role in understanding the departures of the medical care market from the standards of markets in general”.
This is a weaker and less controversial view than his earlier one. Yet he still maintains the view that “social norms are based on at least perceived mutual gains” (ibid.), contrary to the many cases of norms that are perceived as harmful and yet persist over time.
Political science
This discipline deals partly with the behavior of individuals, notably voters, and partly with the decisions by higher-level actors (institutions), such as the decision by a political party to adopt a certain platform, or the decision by a state to declare war. As everybody knows, in virtually all cases the latter decisions are the result of struggles within the institution. Genuine unanimity is extremely rare, although institutions often present their decisions as unanimous once they are arrived at.
Despite these obvious facts, political scientists routinely present institutions, including states, as unitary actors, who act on the basis of coherent preferences and rational beliefs—as rational individuals writ large.
Scholars rarely, or virtually never, take the trouble of defending the unitary-action assumption, but one can imagine two procedures. First, one could identify the preferences and beliefs of the individuals, and, using an appropriate aggregation mechanism, construct the preferences and beliefs of the collective actor.
In assembly politics, aggregation by majority voting seems the most natural procedure. Second, one could impute preferences to the collective actor that coincide with its objective interests and beliefs that correspond to the facts. …
… Some [political scientists] may also know that belief-aggregation is vulnerable to what is variously called the discursive dilemma or the doctrinal paradox, but which should really be called the Poisson paradox after its first expositor.
[MH: I believe the article by Elster re. Poisson can be downloaded here at SSRN].
This paradox, too, can leave belief formation by majority voting in a multi-member body essentially indeterminate. Put crudely, a group may not know what it wants or what it believes. The second procedure, which completely disregards subjective elements and the possibility or relevance of disagreement among group members, characterizes the Realist school of international relations.
In this approach, which is properly called “methodological statism” [Tamar Megiddo 2019] states “are assumed to be unitary actors, their internal dynamics of little importance for studying their behaviour internationally”.
[MH: there seems to be a little muddle in the Handbook editing. I believe this article by Tamar Megiddo can be downloaded free at SSRN here]
I am not a scholar of international politics, but this assumption strikes me as highly implausible. It is violated, for instance, when political leaders go to war to distract from domestic unrest. It is also violated when different branches of government have different interests and different beliefs concerning the conduct of a war, as illustrated by the conflict between the Pentagon and the State Department during the Vietnam War. A special case that undermines the assumption of the unitary actor is the struggle for resources among the branches of the military.
4. Some Alleged Obstacles to MI
As noted at the outset of this chapter, the practical obstacles to MI may be unsurmountable. Because I view MI as an injunction to dig deeper rather than as a research program, these obstacles merely reflect the fact that “ought implies can”.
A more forceful objection to MI would arise if one could show that there exist social phenomena that in principle cannot be addressed within the framework of MI, even when all relevant individual-level information is available.
I do not think there are any such cases. As it is hard to prove a negative, I can only address the most prominent obstacles and objections.
Social norms
A very common objection is that MI cannot account for social norms.
As noted earlier, it is certainly true that individuals mostly grow up in an environment where they face strict social norms that are enforced by the avoidance or punishment of norm-violators. It is also true (i) that social norms are shaped by the interaction of individuals and (ii) that even when violations are punished, individuals often violate norms when unseen by others. Unlike moral norms, social norms are not internalised. (These are part of my stipulative definitions.)
It is also true, finally, that no one has proposed a clear and convincing account of how the interaction of individuals causes new norms to appear. If, as Arrow (1971) suggested, social norms arise because of a need to remedy a market failure, this need must be embodied in the perceptions and actions of individuals, including their propensity to avoid or punish norm-violators. But why would anyone punish norm-violators before the norm exists?
A possible answer could be found in the simultaneous provision of new information to many individuals about previously unknown harm caused by activities that now become the target of negative norms. Spitting on the floor before the understanding of bacterial dissemination of illness could be an example. The invention of social norms has also been imputed to “norm entrepreneurs”, although it is not always clear exactly what the diffusion mechanism(s) would be. For the time being, feature (i) of norms is stipulated rather than demonstrated. What is certain, however, is that Arrow’s 1971 proposal doesn’t work. “Society” does not create norms.
Feature (ii) of norms also counts against Durkheim’s claim that immaterial norms have a compelling and irresistible character, comparable to the circle in the sand that prevents the chicken from going outside it. However, norms can disappear almost overnight if the victims of harmful norms organise themselves to discontinue the compliance. Even if the norm persists in the population, some individuals may secretly persist in the proscribed behaviour. …
In what sense are social norms social? Not in the sense that they are independent of individuals. They are shared individual expectations about how other people would react to certain actions. The expectations are not necessarily correct. As Tocqueville wrote, the belief in the existence of a norm maybe just “the empty phantom of a public opinion [which] is enough to chill innovators” until someone shows its non-existence by violating it without suffering avoidance or punishment. …
Institutions
Another common objection is that MI cannot account for institutions.
These are of course created by individuals and staffed by individuals. They operate by rules such as the principle that an official cannot give orders to other officials more than one step below him or her in the hierarchy or the principle that institutional decisions are subject to appeal. These rules exist on paper and in the minds of the members and the clients of the institution. If the latter do not know (and cannot learn) the rules … we are dealing with a practice rather than with an institution. Institutions exist as networks of mutual expectations among officials and between officials and clients. To individuals subject to their decisions, they can appear as supraindividual, monolithic entities, but to those who operate them, they appear as what they are.
To explain their decisions and their changes over time, the interaction of officials with each other and with their clients provides the necessity and sufficient data. A term such as “institutional sclerosis” is merely an umbrella shorthand for a variety of individual-level processes.
Arrow (once more)
Kenneth Arrow (1994) states a number of objections to MI. The paper is brief, programmatic, enigmatic, and frustrating since it leaves the impression that he could have said much more.
[MH: This paper by Arrow can be downloaded at the uberty.org site — this link goes straight to the PDF]
He [Arrow] writes notably that “limitations on individualistic methodology appear very strongly when considering the role of information”, adding that “the individualistic viewpoint is not to be completely neglected”. He returns to this issue towards end of the paper, where he states that it may be easier to think of information breeding information and to suppress the role of individuals. This is very much like the role of genetic information. As the late nineteenth century writer, Samuel Butler, said: “A chicken is an egg’s way of making an egg”. Unfortunately, Arrow does not spell out the argument.
[MH Bonus: You can read my Hayekian 2021 essay on a seventeenth century systems theory of egg/chicken formation at my other website, which is safe and family-friendly]
Arrow also writes that “the failure to give an individualistic explanation of price formation has proved to be surprisingly hard to cure, though there are some ingenious stories”. A powerful example concerns the prices under duopoly:
“What [..] are the strategies of the two players? Cournot assumed that firms choose quantities. Is it not equally reasonable that firms choose prices? If we assume rational consumers, they will all buy from the sellers with the lowest price. The outcome of this game is entirely different from Cournot’s; under simple assumptions it is in fact the competitive equilibrium even if there are only two firms. What this example shows is that the rules of the game are social. The theory of games gets its name and much of its force from an analogy with social games. But these have definite rules which are constructed, indeed, by a partly social process. Who sets rules for real-life games ? More generally, individual behavior is always mediated by social relations. These are as much a part of the description of reality as is individual behavior.” [quotation from Arrow ibid.]
I have little competence to assess this issue. However, a perusal of the literature confirmed my intuition that the form of competition—by prices or quantities?—may be endogenous. … [and] perfectly consistent with MI.
Sally Haslanger [2020] discusses a series of alleged failures of MI. A representative example is the following:
“Why did the Native Americans in what is now the northeastern United States rely on a diet of root vegetables in the winter months, rather than leafy vegetables and fruits? Answer: Because the climate and storage techniques available wouldn’t support the preservation of leafy vegetables and fruits through the winter. Note that the explanation of the Native American diet is in terms of the climate, technology, and properties of leafy vegetables and fruits, not in terms of individuals and their interactions. But surely the diet of Native Americans is a social phenomenon to be explained. So explanatory individualism is unwarranted.”
[MH: This paper by Haslanger, a critique of which may be relevant to theorisations of type 2 hunter-gatherers, can be downloaded direct from the author’s website]
This passage shows the need for a clarification of explananda in science. Haslanger asks for the answer to a “Why-question” rather than for an explanation. Since all explanation is causal and since an absence, such as the absence of a diet, cannot be an explanandum, the objection is misplaced.
5. What’s the Alternative?
Critics of MI feel or should feel an obligation to present an alternative.
Durkheim’s proposal amounts merely to a stipulation, namely that there exist irreducibly social facts. Marx’s statement that “Capital” is prior to the many individual capitals is not a causal claim, but a confused conceptual one. Arrow’s ideas are not sufficiently spelled out.
Here I shall discuss four attempts to provide an alternative to MI: Pierre Bourdieu’s notion of a “field” (champ), Robert Putnam’s concept of “social capital”, various attempts to formulate an idea of “group agency”, and Christian List’s and Kai Spiekermanns proposal to “reconcile” individualism and its apparent opposite, holism. In the vast literature on the subject, there are many other proposals, often couched in impermeable terminology that makes assessment difficult. Although this characterisation is also to varying degrees appropriate for the three first alternatives mentioned above, their prominence in the literature justifies the discussion. …
[MH: We will omit Bourdieu and Putnam here and skip to ‘group agency’.]
Group agency
A number of scholars have asserted, more or less explicitly denying MI, that groups can act in ways that are not reducible to individual actions.
Perhaps more appropriate terms would be that groups can decide to act or have joint intentions. While shared intentions are consistent with MI, joint intentions are not.
Scholars who argue along these lines include Michael Bacharach, Michael Bratman, Robert Sugden, Margaret Gilbert, John Searle, Philip Pettit, and Christian List. They do not argue for exactly the same conclusion, but their arguments have a family resemblance.
A recurring phrase is “I intend that we...”. The idea has gained some prominence, as shown by the publication of a Handbook of Collective Intentionality (Jankovic and Ludwig, eds. 2017). Since I claim that MI is trivially true, I may seem to be committed to the claim that these contributions are trivially false.
Matters are more complicated, however, since advocates of this view engage only in conceptual analysis. I do not know of a single example of a group-agency scholar trying to (1) explain an empirical phenomenon and (2) argue that individual-level explanations are in principle unable to do the job. I confess to finding group agency and joint intentions unintelligible ideas, at least in their more extreme forms which include the stipulation of irreducibly collective emotions (references given). I shall not pursue these issues further.
[MH: I’m so glad Jon Elster argues this so clearly. In celebration I have reopened my August 2023 critique of the “collective action” paradigm in archeology.]
Instead, I shall ask whether something like a group agency might explain cooperation in a one-shot Prisoner’s Dilemma. I have in mind “magical thinking” and the closely related “Kantian optimization”. These occur when the agents—not the scholars who study them —ignore or violate MI.
Magical thinking is a well-documented tendency, notably in problems of cooperation, to think that “If I cooperate, others, being like me, will cooperate: therefore I shall cooperate”. The greater the similarity with other members of the relevant group, the stronger the tendency. Kantian thinking is the imperative: “Do what would be best if everybody did the same”. These hypothetical forms of “group agency” can induce and explain the action.
List and Spiekermann
[MH: Their article can be downloaded at Spiekermann’s university webpage]
In the abstract of their article, Christian List and Kai Spiekermann (2013) write that
“Political science is divided between methodological individualists, who seek to explain political phenomena by reference to individuals and their interactions, and holists (or nonreductionists), who consider some higher-level social entities or properties such as states, institutions, or cultures ontologically or causally significant. We propose a reconciliation between these two perspectives”.
… I submit that the individual-level account is a complete causal explanation.
What List and Spiekermann, propose is an argument about necessitation (Elster 2015). To illustrate, consider a house of cards exposed to occasional wind gusts. That the house will collapse is a certainty. Close observation will tell us which gust made it collapse and the order in which the cards fell. Only the latter provides an explanation of the collapse of the house of cards as it happened. …
… Assuming —a strong assumption—that one can identify “structural reasons” that necessitate the outcome, would they provide better insight into the explanandum than a well-supported narrative would do? My inclination is to say No, but do I have an argument to support it?
My preference is partly esthetic: it is more satisfying to follow the actions and interactions of specific and idiosyncratic human beings than to subsume them under abstractions such as “the lack of a common interest”.
I have also intellectual grounds, however, for my preference. Only the study of individuals close up can suggest psychological mechanisms for the scholar’s toolbox.
Perhaps the victory of the American colonies in the War of Independence can be explained by “structural reasons”, such as the tendency of all large empires to break down because of their very size. In France, Turgot (Oeuvres, volume 1) observed in 1750 that “colonies are like fruits that remain attached to the tree only until they are ripe: once they are self-sufficient they do what Carthage did [when gaining its independence from the Phoenicians in 650 B.C.] and what America will do one day”. He was right, but I find it more interesting to open that particular black box and find that the Americans were moved by the enthusiasm and not simply by anger (Elster 2023). …
6. Conclusion
I still believe that MI is trivially true, but I appreciate the fact that others disagree. One person’s trivial conclusion may be another’s deep mistake. Even the interpretive norm of respecting “the plain meaning of the text” does not always give a clear-cut answer. In twenty-one Supreme Court cases decided with an opinion during the spring of the 1993 term, a conflict between majorities and dissents derived at least in part from disagreements over the plain meaning of the statute at issue.
In rethinking the issue four decades after my first engagement with it, a major source of reflection has been the four articles by Kenneth Arrow in which he briefly touches on MI. …
From the pieces I have cited [above], one can conclude that he was both attracted to and worried about MI. He was attracted to it, because as an economist he built on the principle that individual agents are the building blocks of analysis. He was worried about it because as a citizen he deplored that in modern societies the actions of these agents often lead to market failures. Since they did not always do so, he tried to identify the non-market mechanisms that compensate for or prevent such failures. The article on social capital is perhaps the one in which he comes closest to doing so without violating MI.
In my opinion, history and psychology are, or ought to be, the main pillars of the social sciences. If one accepts that view, MI follows almost automatically, at least as an injunction. In all of this, I follow Marc Bloch (1964 The Historian’s Craft), when he wrote that
“the good historian resembles the giant of the fairy tale. He knows that wherever he catches the scent of human flesh, there lies his quarry (son gibier)”.
[END]
Elster references various of his earlier works in support of the arguments made:
Elster, J. (1985). Making Sense of Marx. Cambridge: Cambridge University Press.
Elster, J. (1993). Ethical individualism and presentism. The Monist 76, pp. 333–48.
Elster, J. (2013). Securities Against Misrule. Cambridge: Cambridge University Press.
Elster, J. (2015). Explaining Social Behavior. Cambridge: Cambridge University Press.
Elster, J. (2017). On seeing and being seen. Social Choice and Welfare 49, pp. 721–34.
Elster. J. (2021). Enthusiasm and anger in history. Inquiry 64, pp. 249–307.
Elster, J. (2023). America Before 1787: The Unraveling of a Colonial Regime. Princeton: Princeton University Press.
Les Trois Personnages, by Fernand Leger [Date: 1924]
Please send comments or corrections to me directly at mgs.heller@gmail.com
Tell your friends and colleagues about Social Science Files
Social Science Files displays multidisciplinary writings on a great variety of topics relating to evolutions of social order from the earliest humans to the present day and future machine age