I would argue that the deeper problem is that in the many different ways we contribute to the gains from cooperation, effort and luck hard to distinguish. And in fact, the division of labor - the source of the gains from cooperation - guarantees that different social roles contribute in different structures. "Tilers" contribute measurable outcomes in proportion to measurable effort. "Firemen" ideally never contribute at all (it's always better if no fire breaks out). And "Songwriters" may contribute measurable outcomes in unpredictable intervals, possibly never (or posthumously).
Thanks, I found this really interesting. I think we are starting from the same basic point: social life is built on cooperation, and the central question of political economy is how to divide the surplus from cooperation in a way that keeps people willing to cooperate.
The difference is perhaps that your post starts from the problem of contribution: people contribute to social value in very different ways, and institutions such as money, insurance, taxation and property are ways to compensate these different kinds of contribution. The tiler, the fireman and the songwriter illustrate this nicely because their contributions have very different structures: measurable, preventative, or spiky and unpredictable.
My post starts from a slightly different angle: bargaining over the social contract. I am less focused on what each person’s contribution objectively is, and more on how different claims affect the bargaining position of social groups. Contribution matters because, and to the extent that, it changes bargaining power: those whose cooperation is indispensable can demand a larger share of the social surplus. Other claims, based on need, luck, risk, domination, suffering, or status, matter in the same way when they help groups organise, justify their demands, or make their withdrawal from cooperation more credible.
One point where we may differ is on game theory. You write that game theory shows how cooperation can be rational, but then leaves us at the point where ethics begins. I am more bullish here, and I argue, following Binmore, that game theory helps us understand why fairness norms themselves take the shape they do. In that view, ethics is not outside the bargaining problem. It is partly the grammar through which groups negotiate and stabilise the terms of cooperation.
Thanks, Lionel. RE the disagreement: I am persuadable, and I'm looking forward to that persuasion in your next installments! Two questions Binmore's account did not satisfyingly answer for me were:
1. What are the ground rules that make bargaining possible? Does game theory -very compellingly - just push ethics to a different domain (how we communicate with each other)?
2. How do I, as an individual, choose which of the nested set of groups I'm a member of should command my highest loyalty? (I mentioned this in a comment on one of your previous essays).
My intuition says that both, and especially the latter, tie the "what do we owe each other?" problem inextricably to the "what is the good life?" problem.
My view is that game theory provides a framework for explaining ethics itself: why we have ethical principles and intuitions, what role they play in our social interactions, and why they tend to have the features they have. The key point is that ethics is about regulating social cooperation. Ethical principles help us coordinate expectations, stabilise cooperation, allocate duties and rewards, and justify demands on others.
So I would not say that game theory pushes ethics into another domain. Rather, it helps explain why the domain of ethics exists. Once individuals depend on each other for the gains from cooperation, they need shared rules about what each person may claim, what each person must do, what counts as a violation, and what responses are legitimate. That is where ethical intuitions and principles enter the picture.
On your second point, about which group should command our loyalty, I agree that this is crucial. We all belong to many overlapping groups: family, profession, class, nation, religion, political camp, and so on. Which identity becomes salient is not fixed once and for all. It depends on the coalitions available, the gains from belonging to them, one’s expected position within them, and the meaning and status they provide. I talk about these choices in my book Optimally Irrational.
Yes, it is clearly a major dividing issue. Piketty had a paper in 1995 arguing that this was one of the key differences between the US and Europe. I see this as part of the point I make in the post: once people do not have full information about social laws, debates about justice become entangled with debates about facts. People then have incentives to support factual beliefs that are aligned with their interests. For instance, upwardly mobile people might find the effort explanation more congenial and convenient. So the disagreement is not only about whether society should reward effort or compensate for luck, but also about whether inequalities are actually caused by effort or by luck.
The explanation of Binmoores egalitarianism sounds like he gives every disadvantaged subset, no matter how small, the ability to blow up the whole social contract. The workers losing from free trade *can* just reject this outcome... but clearly the criminals losing from going to prison *cant* just reject the outcome. Nor does this depend on how bad the outcome is - if theyre convicted to death, they still couldnt. Some form of power is active here, and not just the sort from threatening exit - its not clear those workers would gain anything from that.
That is a very good point, and I think it clarifies where Binmore’s two-player simplification matters.
In Binmore’s framework, there are only two players, or two cohesive groups. Cooperation requires both groups to participate, so each group is pivotal. If bargaining power is equal, the bargain over the gains from cooperation is pushed towards an egalitarian solution.
Once we introduce more than two groups and allow coalitions to form, things become more complex. A disadvantaged group does not automatically have the power to reject the whole social contract. What matters is how much its participation contributes to the social surplus, and how costly its withdrawal, non-compliance, or opposition would be for others. A group gains bargaining power when it can organise, withdraw cooperation, or join an oppositional coalition that imposes enough costs to make concessions worthwhile.
This is why the worker and criminal cases differ. Individual workers harmed by free trade may have little power on their own, but workers, regions, unions or political movements can sometimes organise and impose political costs. By contrast, an isolated criminal does not gain bargaining power simply by refusing the rules, because the rest of society is better off enforcing punishment than letting him remain outside the social contract.
So the egalitarian conclusion follows cleanly in Binmore’s two-player benchmark. In a more realistic setting with many groups and coalitions, the outcome will depend on coalitional bargaining power: who can organise, who is pivotal, and whose withdrawal from cooperation imposes meaningful costs. In societies where people have equal ability to join coalitions that could shrink the social pie, we can think that it will also lead towards an egalitarian outcomes.
What you are calling the egalitarian option seems to match up with what I would consider the “fair” option. But let me explain myself so you can correct as needed…
The fairest option is the best real world alternative. The most talented, can always take their talents elsewhere. Thus the social contract has to entice them enough to stay. Big contributors get big rewards.
For the least talented, the options are much more limited. They can just opt out of society altogether (risky move likely leading to poverty and death), or emigrate somewhere with better options. Thus the stable option is to give them as much as needed not to opt out or leave, no more, no less.
But this isn’t really what people (even Rawls) consider egalitarian. What am I missing? Or am I just stumbling over what you refer to as coalition forming?
In Binmore’s benchmark framework there are only two groups, and cooperation requires both groups to be part of the social contract. Since no group can simply be forced into cooperation, the split of the gains from cooperation depends on their relative bargaining power. When bargaining power is equal, the outcome is pushed toward an egalitarian solution.
The important point is that “opting out” of the social contract does not necessarily mean emigrating or going off alone into the wilderness. It can also mean refusing to respect the rules of the social contract: property, civility, debt, work discipline, public order, and so on. The plebeians could threaten to leave Rome, but they could also threaten to stop playing by Rome’s rules. In that sense, even a poorer group may have bargaining power if society cannot function without its cooperation.
What you describe, where the talented are rewarded enough to stay while the least talented get only enough not to leave or revolt, is what we would expect when bargaining power is unequal. It is not the egalitarian benchmark. It is closer to a bargaining outcome in which one group has much better outside options or much stronger leverage.
Once we introduce more than two groups and allow coalitions to form, which Binmore largely leaves for later work, the prediction becomes more complex and likely moves away from strict Rawlsian egalitarianism. What each group can regard as “fair” will depend on its bargaining power in the coalitional game. One way to think about this is: how often can this group be part of a winning coalition and impose its preferred terms? A formal measure of this kind of power is the Shapley value. In a relatively uniform society with identical citizens, these values will be roughly equal, which pushes toward equality. But in societies with rigid coalitions, some groups may have more coalitional power than others, and the resulting social contract may be correspondingly unequal.
I wonder how close egalitarianism and Rawls would be if egalitarianism used a loss-averse utility function. Of course, we would need to figure out what the reference point was, but it might be interesting as expected utility would now put greater weight on avoiding a loss which might get closer to Rawls's formulation.
Hi Liam, that is a good question. Loss aversion can indeed make utilitarian bargains harder to sustain. Kahneman, Knetsch and Thaler discuss how loss aversion helps explain the endowment effect and status quo bias: people often demand more to give something up than they would be willing to pay to acquire it in the first place. In a policy context, this can make win-win reforms harder if those who lose from one reform demand very high compensation.
That being said, the framework I discuss is less about isolated policies than about repeated cooperation. If citizens expect a class of policies to be good for them in expectation over time, then utilitarian rules can still work as mutual-advantage arrangements. The problem arises when the losses are concentrated, salient, or unlikely to be offset later. In those cases, loss aversion reinforces the demand for compensation.
On Rawls, I agree with your intuition. If people are strongly loss averse behind the veil of ignorance, they may give much more weight to avoiding the worst outcomes, and this pushes the choice closer to Rawls’s difference principle. But it would not automatically get us all the way there. The difference principle corresponds to a very extreme form of concern for the worst-off. Empirical studies tend to find that people behind the veil do care about the least well-off, but not usually in the uncompromising form proposed by Rawls.
I would like to see this approach applied to a modern capitalist society in which we distinguish between inequalities of wealth, income, and consumption. It should include a definition of "the general welfare" that reflects the diminishing marginal utility of income--or, rather, the diminishing marginal utility of consumption—and all else equal a dollar will always be worth more to a poor man (person) than a rich one. It should also include the assumption that unconsumed income is invested in capital that raises the marginal productivity of labor throughout the economy, and that society is composed of individuals differing in intelligence, education, talents, ambitions, and skills.
I've made a stab at designing the fairest possible system of taxation and income redistribution in such a society, assuming that the goal is to maximize the general welfare of its citizens in this and future generations when defined in utilitarian terms: https://shorturl.at/8d9sj
Thank you! Binmore's resolution of the Rawls/Harsanyi debate is elegant and potentially persuasive. At any event, at the highest level, you have certainly persuaded me, for the little that may be worth.
I am curious about two challenges for what happens when we use game theory to test out what goes on thereafter:
How is justice determined when the degree to which an agreement is binding or non-binding is unknown or itself potentially under contention? (The Algerian Revolution, for instance, started when a few hundred of the ~10m Algerians insisted on a revision to the social compact that up to that point had been impossible. Over the course of the subsequent conflict, the policy preferences of political actors gyrated unpredictably).
What happens when an agreement is binding on one part of the population and non-binding on another due to demographics?
What happens in non-binary societies, featuring individuals with intersectional identities, who are faced with a variety of open questions whose resolution they may prioritize differently depending on their identities, and also depending on the order in which other questions have closed and with what result? It would seem that, if we wanted to use game theory to track societal equilibria in the definition of justice, we'd soon be forced to run linear regressions to hash out the just from the unjust. Or do we just tell everyone: you can pick between exactly two identities—democratic majority, and democratic minority; now go vote!
These are very good questions. I would place the first two under the broader problem of incomplete information and uncertain enforcement. Whether an agreement is binding is not always common knowledge. Actors may be unsure whether others really have incentives to respect the agreement, whether they can coordinate a deviation, or whether punishment would actually follow. In that sense, some social contracts are gambles: their stability depends on beliefs about whether the structure of incentives will make them self-enforcing. The Algerian case can be read in this way: what had looked like an imposed and stable compact became unstable once a coalition was able to make revision of the compact a credible political demand.
The second case, where an agreement is binding for some groups but not for others, is also central. It means that the social contract is not equally self-enforcing for everyone. Some groups may be constrained by the agreement, while others have enough power, numbers, or organisational capacity to renegotiate or ignore it. This is precisely the kind of situation where bargaining power matters, and where the resulting contract will not necessarily have an egalitarian form.
On identities, I think you are right that real societies are not binary. I discussed social identities, and the choices people make between competing identities, in Optimally Irrational. Game theory helps remove some of the mystery surrounding the usual discourse on identity. In short, social identity can be understood as a cognitive and social tool that helps stabilise coalitions. People can, to some extent, identify more strongly with one coalition rather than another, and this depends on incentives: the gains from belonging to a coalition, the position one can expect within it, and the alternatives offered by other coalitions.
This does not mean that we reduce everyone to two identities, majority and minority. Rather, democratic institutions compress a multidimensional coalitional world into temporary majorities and minorities for the purpose of making decisions. The simplicity of the formal model is therefore not meant to describe society in all its complexity. It is a benchmark. Once we add coalitions, overlapping identities and incomplete information, the model becomes much messier, but the basic question remains the same: which coalitions can form, which agreements can be made self-enforcing, and who has enough bargaining power to demand revision of the social contract?
I think that the crux of the issue is more whether people believe that we live in luck village or effort village. https://www.econlib.org/library/Columns/y2016/Klingcapitalism.html
I would argue that the deeper problem is that in the many different ways we contribute to the gains from cooperation, effort and luck hard to distinguish. And in fact, the division of labor - the source of the gains from cooperation - guarantees that different social roles contribute in different structures. "Tilers" contribute measurable outcomes in proportion to measurable effort. "Firemen" ideally never contribute at all (it's always better if no fire breaks out). And "Songwriters" may contribute measurable outcomes in unpredictable intervals, possibly never (or posthumously).
The question is how to apportion the gains from the cooperative surplus generated by the division of labor given these very different contribution structures. https://theruminathans.substack.com/p/a-tiler-a-fireman-and-a-songwriter
Thanks, I found this really interesting. I think we are starting from the same basic point: social life is built on cooperation, and the central question of political economy is how to divide the surplus from cooperation in a way that keeps people willing to cooperate.
The difference is perhaps that your post starts from the problem of contribution: people contribute to social value in very different ways, and institutions such as money, insurance, taxation and property are ways to compensate these different kinds of contribution. The tiler, the fireman and the songwriter illustrate this nicely because their contributions have very different structures: measurable, preventative, or spiky and unpredictable.
My post starts from a slightly different angle: bargaining over the social contract. I am less focused on what each person’s contribution objectively is, and more on how different claims affect the bargaining position of social groups. Contribution matters because, and to the extent that, it changes bargaining power: those whose cooperation is indispensable can demand a larger share of the social surplus. Other claims, based on need, luck, risk, domination, suffering, or status, matter in the same way when they help groups organise, justify their demands, or make their withdrawal from cooperation more credible.
One point where we may differ is on game theory. You write that game theory shows how cooperation can be rational, but then leaves us at the point where ethics begins. I am more bullish here, and I argue, following Binmore, that game theory helps us understand why fairness norms themselves take the shape they do. In that view, ethics is not outside the bargaining problem. It is partly the grammar through which groups negotiate and stabilise the terms of cooperation.
Thanks, Lionel. RE the disagreement: I am persuadable, and I'm looking forward to that persuasion in your next installments! Two questions Binmore's account did not satisfyingly answer for me were:
1. What are the ground rules that make bargaining possible? Does game theory -very compellingly - just push ethics to a different domain (how we communicate with each other)?
2. How do I, as an individual, choose which of the nested set of groups I'm a member of should command my highest loyalty? (I mentioned this in a comment on one of your previous essays).
My intuition says that both, and especially the latter, tie the "what do we owe each other?" problem inextricably to the "what is the good life?" problem.
Thanks, Nathan.
My view is that game theory provides a framework for explaining ethics itself: why we have ethical principles and intuitions, what role they play in our social interactions, and why they tend to have the features they have. The key point is that ethics is about regulating social cooperation. Ethical principles help us coordinate expectations, stabilise cooperation, allocate duties and rewards, and justify demands on others.
So I would not say that game theory pushes ethics into another domain. Rather, it helps explain why the domain of ethics exists. Once individuals depend on each other for the gains from cooperation, they need shared rules about what each person may claim, what each person must do, what counts as a violation, and what responses are legitimate. That is where ethical intuitions and principles enter the picture.
On your second point, about which group should command our loyalty, I agree that this is crucial. We all belong to many overlapping groups: family, profession, class, nation, religion, political camp, and so on. Which identity becomes salient is not fixed once and for all. It depends on the coalitions available, the gains from belonging to them, one’s expected position within them, and the meaning and status they provide. I talk about these choices in my book Optimally Irrational.
Finally, about the question "what is a good life", I wrote a post about it following a naturalistic perspective, here: https://www.optimallyirrational.com/p/happiness-and-the-pursuit-of-a-good
Yes, it is clearly a major dividing issue. Piketty had a paper in 1995 arguing that this was one of the key differences between the US and Europe. I see this as part of the point I make in the post: once people do not have full information about social laws, debates about justice become entangled with debates about facts. People then have incentives to support factual beliefs that are aligned with their interests. For instance, upwardly mobile people might find the effort explanation more congenial and convenient. So the disagreement is not only about whether society should reward effort or compensate for luck, but also about whether inequalities are actually caused by effort or by luck.
The explanation of Binmoores egalitarianism sounds like he gives every disadvantaged subset, no matter how small, the ability to blow up the whole social contract. The workers losing from free trade *can* just reject this outcome... but clearly the criminals losing from going to prison *cant* just reject the outcome. Nor does this depend on how bad the outcome is - if theyre convicted to death, they still couldnt. Some form of power is active here, and not just the sort from threatening exit - its not clear those workers would gain anything from that.
That is a very good point, and I think it clarifies where Binmore’s two-player simplification matters.
In Binmore’s framework, there are only two players, or two cohesive groups. Cooperation requires both groups to participate, so each group is pivotal. If bargaining power is equal, the bargain over the gains from cooperation is pushed towards an egalitarian solution.
Once we introduce more than two groups and allow coalitions to form, things become more complex. A disadvantaged group does not automatically have the power to reject the whole social contract. What matters is how much its participation contributes to the social surplus, and how costly its withdrawal, non-compliance, or opposition would be for others. A group gains bargaining power when it can organise, withdraw cooperation, or join an oppositional coalition that imposes enough costs to make concessions worthwhile.
This is why the worker and criminal cases differ. Individual workers harmed by free trade may have little power on their own, but workers, regions, unions or political movements can sometimes organise and impose political costs. By contrast, an isolated criminal does not gain bargaining power simply by refusing the rules, because the rest of society is better off enforcing punishment than letting him remain outside the social contract.
So the egalitarian conclusion follows cleanly in Binmore’s two-player benchmark. In a more realistic setting with many groups and coalitions, the outcome will depend on coalitional bargaining power: who can organise, who is pivotal, and whose withdrawal from cooperation imposes meaningful costs. In societies where people have equal ability to join coalitions that could shrink the social pie, we can think that it will also lead towards an egalitarian outcomes.
What you are calling the egalitarian option seems to match up with what I would consider the “fair” option. But let me explain myself so you can correct as needed…
The fairest option is the best real world alternative. The most talented, can always take their talents elsewhere. Thus the social contract has to entice them enough to stay. Big contributors get big rewards.
For the least talented, the options are much more limited. They can just opt out of society altogether (risky move likely leading to poverty and death), or emigrate somewhere with better options. Thus the stable option is to give them as much as needed not to opt out or leave, no more, no less.
But this isn’t really what people (even Rawls) consider egalitarian. What am I missing? Or am I just stumbling over what you refer to as coalition forming?
In Binmore’s benchmark framework there are only two groups, and cooperation requires both groups to be part of the social contract. Since no group can simply be forced into cooperation, the split of the gains from cooperation depends on their relative bargaining power. When bargaining power is equal, the outcome is pushed toward an egalitarian solution.
The important point is that “opting out” of the social contract does not necessarily mean emigrating or going off alone into the wilderness. It can also mean refusing to respect the rules of the social contract: property, civility, debt, work discipline, public order, and so on. The plebeians could threaten to leave Rome, but they could also threaten to stop playing by Rome’s rules. In that sense, even a poorer group may have bargaining power if society cannot function without its cooperation.
What you describe, where the talented are rewarded enough to stay while the least talented get only enough not to leave or revolt, is what we would expect when bargaining power is unequal. It is not the egalitarian benchmark. It is closer to a bargaining outcome in which one group has much better outside options or much stronger leverage.
Once we introduce more than two groups and allow coalitions to form, which Binmore largely leaves for later work, the prediction becomes more complex and likely moves away from strict Rawlsian egalitarianism. What each group can regard as “fair” will depend on its bargaining power in the coalitional game. One way to think about this is: how often can this group be part of a winning coalition and impose its preferred terms? A formal measure of this kind of power is the Shapley value. In a relatively uniform society with identical citizens, these values will be roughly equal, which pushes toward equality. But in societies with rigid coalitions, some groups may have more coalitional power than others, and the resulting social contract may be correspondingly unequal.
I wonder how close egalitarianism and Rawls would be if egalitarianism used a loss-averse utility function. Of course, we would need to figure out what the reference point was, but it might be interesting as expected utility would now put greater weight on avoiding a loss which might get closer to Rawls's formulation.
Hi Liam, that is a good question. Loss aversion can indeed make utilitarian bargains harder to sustain. Kahneman, Knetsch and Thaler discuss how loss aversion helps explain the endowment effect and status quo bias: people often demand more to give something up than they would be willing to pay to acquire it in the first place. In a policy context, this can make win-win reforms harder if those who lose from one reform demand very high compensation.
That being said, the framework I discuss is less about isolated policies than about repeated cooperation. If citizens expect a class of policies to be good for them in expectation over time, then utilitarian rules can still work as mutual-advantage arrangements. The problem arises when the losses are concentrated, salient, or unlikely to be offset later. In those cases, loss aversion reinforces the demand for compensation.
On Rawls, I agree with your intuition. If people are strongly loss averse behind the veil of ignorance, they may give much more weight to avoiding the worst outcomes, and this pushes the choice closer to Rawls’s difference principle. But it would not automatically get us all the way there. The difference principle corresponds to a very extreme form of concern for the worst-off. Empirical studies tend to find that people behind the veil do care about the least well-off, but not usually in the uncompromising form proposed by Rawls.
I would like to see this approach applied to a modern capitalist society in which we distinguish between inequalities of wealth, income, and consumption. It should include a definition of "the general welfare" that reflects the diminishing marginal utility of income--or, rather, the diminishing marginal utility of consumption—and all else equal a dollar will always be worth more to a poor man (person) than a rich one. It should also include the assumption that unconsumed income is invested in capital that raises the marginal productivity of labor throughout the economy, and that society is composed of individuals differing in intelligence, education, talents, ambitions, and skills.
I've made a stab at designing the fairest possible system of taxation and income redistribution in such a society, assuming that the goal is to maximize the general welfare of its citizens in this and future generations when defined in utilitarian terms: https://shorturl.at/8d9sj
Thank you! Binmore's resolution of the Rawls/Harsanyi debate is elegant and potentially persuasive. At any event, at the highest level, you have certainly persuaded me, for the little that may be worth.
I am curious about two challenges for what happens when we use game theory to test out what goes on thereafter:
How is justice determined when the degree to which an agreement is binding or non-binding is unknown or itself potentially under contention? (The Algerian Revolution, for instance, started when a few hundred of the ~10m Algerians insisted on a revision to the social compact that up to that point had been impossible. Over the course of the subsequent conflict, the policy preferences of political actors gyrated unpredictably).
What happens when an agreement is binding on one part of the population and non-binding on another due to demographics?
What happens in non-binary societies, featuring individuals with intersectional identities, who are faced with a variety of open questions whose resolution they may prioritize differently depending on their identities, and also depending on the order in which other questions have closed and with what result? It would seem that, if we wanted to use game theory to track societal equilibria in the definition of justice, we'd soon be forced to run linear regressions to hash out the just from the unjust. Or do we just tell everyone: you can pick between exactly two identities—democratic majority, and democratic minority; now go vote!
These are very good questions. I would place the first two under the broader problem of incomplete information and uncertain enforcement. Whether an agreement is binding is not always common knowledge. Actors may be unsure whether others really have incentives to respect the agreement, whether they can coordinate a deviation, or whether punishment would actually follow. In that sense, some social contracts are gambles: their stability depends on beliefs about whether the structure of incentives will make them self-enforcing. The Algerian case can be read in this way: what had looked like an imposed and stable compact became unstable once a coalition was able to make revision of the compact a credible political demand.
The second case, where an agreement is binding for some groups but not for others, is also central. It means that the social contract is not equally self-enforcing for everyone. Some groups may be constrained by the agreement, while others have enough power, numbers, or organisational capacity to renegotiate or ignore it. This is precisely the kind of situation where bargaining power matters, and where the resulting contract will not necessarily have an egalitarian form.
On identities, I think you are right that real societies are not binary. I discussed social identities, and the choices people make between competing identities, in Optimally Irrational. Game theory helps remove some of the mystery surrounding the usual discourse on identity. In short, social identity can be understood as a cognitive and social tool that helps stabilise coalitions. People can, to some extent, identify more strongly with one coalition rather than another, and this depends on incentives: the gains from belonging to a coalition, the position one can expect within it, and the alternatives offered by other coalitions.
This does not mean that we reduce everyone to two identities, majority and minority. Rather, democratic institutions compress a multidimensional coalitional world into temporary majorities and minorities for the purpose of making decisions. The simplicity of the formal model is therefore not meant to describe society in all its complexity. It is a benchmark. Once we add coalitions, overlapping identities and incomplete information, the model becomes much messier, but the basic question remains the same: which coalitions can form, which agreements can be made self-enforcing, and who has enough bargaining power to demand revision of the social contract?
An oxymoron?