John Roemer and plural rationality

Roemer's move, however, is more subtle than it may seem. It is not a moralising call to be more good. He is not contrasting altruism with self-interested rationality, as if there were interests on one side and values on the other. His theoretical gesture is more radical. Roemer questions the idea that there is only one rational way to choose when our actions are intertwined with those of others.

Standard economic theory has accustomed us to the idea that, in a situation of interdependence, each person must ask himself what it is convenient for him to do, given the actions of others. Others thus become part of the environment. They are constraints or opportunities, threats or sources of information. In this framework, I do not decide with them; I decide in spite of them, based on what I imagine they will do. This is the logic of Nash-style optimisation, named after the mathematician John Nash, Nobel laureate in economics and father of the eponymous concept of equilibrium.

In a 'Nash equilibrium', each decision maker takes the choices of the others as given and chooses his own action as the optimal response. Assuming that everyone else will do the same. We reach equilibrium when each person's move is an optimal response to the optimal responses of all the other decision-makers. This is a powerful form of rationality. But it is not the only one possible. Above all, it is not always the one best suited to describe the deep logic of cooperation.

Co-operation does not only mean finding a balance between separate individuals. It means recognising that some actions only make sense if they are thought of as part of a common plan of action. This is where the central idea of Roemer's proposal comes into play. In a co-operative situation, the subject does not just ask himself what is appropriate for me, given what others are doing, he asks, rather - "What is the strategy I would like us all to adopt?" (p. viii). Roemer calls this different form of reasoning 'Kantian optimisation', because he recognises in it an economic translation of the categorical imperative: to act according to a maxim that can be universally desired.

Of course, Roemer is not suggesting that the rational subject should forget himself and sacrifice himself to others. He is telling us something more interesting and that is that in many social situations, my own interest cannot be properly defined if I separate my choice from the collective structure of action. This difference is decisive. In Nash-like reasoning, others remain external to my choice, like given parameters. I have to predict and anticipate their choices in order to be able to adapt my own in the best possible way. In Kantian reasoning, on the other hand, others enter into the very form of deliberation. They are no longer merely what I adapt to. They are part of the action I am trying to think. The question is no longer how is it possible to achieve the best, given what others will do, but is about how would it make sense to behave if everyone in the same situation did?

To understand the extent of this revolution let us try to think of a few examples. Let us start with taxes. According to purely strategic logic, everyone may be tempted to think that if others pay correctly, public services will be financed. If, on the other hand, others do not pay, it makes no sense for me to be the only one to do so. In both cases, evasion may appear the more convenient choice. It is the logic of the free-rider, of the opportunist who benefits from the contribution of others without contributing in turn.

  But this reasoning, if generalised, destroys the very condition of the public good from which everyone benefits. The question, from a Kantian perspective, then takes a different form.

 What is the fiscal behaviour everyone should adopt if we want to live in a society capable of financing schools, hospitals, infrastructure, security, justice, care, quality scientific research? This is not a question for beautiful souls, it is a question about the nature of public goods and the rationality of their production process. A central question in any advanced society. Let us now try to think about elections. From a strictly individual point of view, an individual voter's vote has an infinitesimal probability of being decisive.

 Why then go and vote? Standard economics has often treated this question as a paradox. Roemer dissolves it by changing the frame of reference. I do not vote because I believe that my vote alone will determine the outcome. I vote because I want all those who participate in democratic life to play their part. I vote because democracy is a public good that only exists and can only survive if enough citizens do not reason as if their own actions are separable from those of others. The same logic characterises choices to protect the environment and combat climate change. The action of the individual, taken in isolation, has negligible effects. Hence the temptation to surrender to impotence. Why should I change my consumption, my movements, my habits, if my action does not change the fate of the planet? But even now the problem is misplaced. No collective transformation arises because the individual gesture is sufficient on its own. It arises when the individual gesture is thought within a practice that only makes sense if it is shared. Kantian optimisation does not ask the individual to save the world alone. It asks him not to reason as if the world were merely the mute background of his private existence.

This point allows Roemer to clarify an often overlooked distinction, namely that cooperative behaviour and altruistic behaviour are not the same thing. "Co-operation of an extended kind can be undertaken because it is in everyone's interest, not because everyone cares about everyone else" (p. 5) Roemer writes in this regard.

Co-operation does not necessarily mean sacrificing oneself for the other, nor does it mean giving his or her well-being a direct weight in one's own utility function. Two people can build a house together because they both want to live in it. Two hunters can cooperate because neither could catch the prey alone. A community can finance a public good because everyone knows that, without the contribution of many, that good would not exist and no one could benefit from it.

Co-operation, therefore, does not require a moral basis. It requires the acceptance of a common condition. It requires the recognition that there are goods that cannot be produced by individuals acting only as isolated maximisers. A safe city, a quality school, a liveable environment, efficient public health, a credible tax system, an effective democracy are not goods that can be obtained simply by summing up well-regulated selfishness. They are goods that only exist if a sufficient number of people agree to think of themselves as part of a cooperating community. In this sense, the reasoning on Kantine optimisation adds a decisive element to the more general reflection that Roemer elaborates on the subject of justice.

In the previous three Mind the Economy articles, we have seen that the rhetoric of merit tends to imagine society as a race in which everyone starts, runs, strives and arrives where talent and discipline take them. But the race never really starts from the same point. Circumstances given by family, health conditions, gender, where one is born and the quality of the school one attends shape one's chances long before the individual can be held accountable for his or her choices.

That is why equality of opportunity cannot coincide with uniform, equal treatment for all. If some start further behind, making them run the same distance does not change the situation. Treating them the same often means preserving, and even amplifying, inequalities. Now, however, Roemer proposes that we go a step further because even a fairer race is still a race. Even a competition that is fair in its initial conditions continues to push individuals to think of themselves as adversaries. Of course, a fair competition is better than a rigged competition. A school that compensates for family and territorial disadvantages is better than a school that inherits and certifies them. Better is a labour market that does not turn birth, gender, health or cultural capital into destiny.

But a just society cannot stop there. It cannot only put everyone in a better position to compete. It must also ask why we have decided that so many dimensions of life in common must be thought of and experienced as a permanent competition. This is perhaps the most crucial question in Roemer's argument. Justice is not only about the distribution of resources, but about the very form of social relations. It is not enough to correct outcomes if we continue to produce institutions that teach individuals to distrust one another. It is not enough to compensate for disadvantages if every common good is then treated as an opportunity for individual appropriation. It is not enough to talk about responsibility if institutions make the most responsible behaviour appear unreasonable and unbecoming.

Life in common is not a 'zero-sum' game; not every good is positional; not every value grows if someone falls behind; not every success needs losers. There are goods that only increase when shared: trust, knowledge, public health, security, democratic quality, social peace, environmental sustainability. For these goods, the question "what is good for me?" is reductive. Not because the self does not count, but because it counts within relationships that precede it and make it possible.

To cooperate, then, is not to compete better. It is to remove some decisive dimensions of social life from the logic of permanent competition. It is to understand that a just society is measured not only by the reliability of its rankings, but by the quality of the common goods it manages to produce. Not only by how fairly it distributes individual opportunities, but by how much it makes it possible to recognise oneself as part of a shared destiny. Perhaps this is the most demanding lesson of the latest Roemer. Justice does not only ask where we start from, how hard we work, what obstacles we encounter and what compensation is needed so that the race is not rigged. It also asks what image of others we learn to inhabit: adversaries, competitors, threats, free-riders, or fellow travellers and co-owners of a common home.

A society weakens when inequalities grow, certainly. But it also weakens when collective intelligence is reduced to the art of evasion, when prudence becomes suspicion and every gesture that is not immediately beneficial appears as a naive concession. Then the task of institutions is not only to correct unjust outcomes. It is also to make plausible again the idea that no one really saves himself, that some forms of well-being only exist if they are produced together, and that rationality, when it takes life in common seriously, ceases to be defensive calculation and returns to being the capacity to build a better world