In Isaac Asimov’s The Bicentennial Man, Andrew Martin, a highly capable robot, progressively shows signs of unusual intelligence and creativity. Initially working as an aide in a household, he develops a desire for freedom and, eventually, for legal recognition as human. In a court case, he pleads:

It has been said in this courtroom that only a human being can be free. It seems to me that only someone who wishes for freedom can be free. I wish for freedom.

Impressed by this statement, the judge rules that

There is no right to deny freedom to any object with a mind advanced enough to grasp the concept and desire the state.

Asimov’s story points to how progress in artificial intelligence will test our moral norms and intuitions. It will require us to decide where the boundaries of moral rights lie and whether they should move if AI agents become self-aware and as intelligent as, or more intelligent than, humans.

What does philosophy say about this? In a provocative 2010 text, AI researcher and technology ethicist Joanna Bryson argued that future robotic agents should be considered slaves and denied rights. Giving rights to robots, she stressed, would come at the expense of humans.

What I advocate here is that the correct metaphor for robots is slavery. Robots are wholly owned and designed by us, we determine their goals and desires. A robot’s brain should be backed up continuously, and its body mass produced. No one should ever need to hesitate in deciding whether to save a human or a robot from a burning building. — Bryson (2010)

Against this view, what appears to be the dominant position among philosophers discussing AI ethics is that AI agents should receive moral consideration once they seem to experience consciousness or sentience.1

Sentience, consciousness and moral consideration

A major conclusion from my series of posts on morality is that modern moral philosophy is, strikingly, very much lost in misguided conceptual takes. Large parts of mainstream moral philosophy remain surprisingly insulated from advances in neuroscience, game theory and evolutionary theory. As a result, discussions often seem eerily stuck in Byzantine debates about the sex of angels, without questioning whether angels exist in the first place.

The rise of AI is one of the questions that lays this problem bare. Much of the literature on the moral status of artificial agents takes for granted, implicitly or explicitly, that sentience or consciousness is what makes an entity worthy of moral consideration. A recent philosophy symposium on AI, Consciousness and Ethics illustrates the state of the debate.2 One participant, Kamil Mamak, explained:

[T]he dominant position in contemporary AI ethics, remains a properties-based account according to which sentience functions as a sufficient, or at least central, condition for moral status.

In other words, if an AI is sentient—it can have experiences that feel good or bad—humans have a duty to take those experiences into consideration.

A related position holds that consciousness, the fact of having an inner experience, is what matters morally. In a recent article, philosophers Jeff Sebo and Robert Long argued for granting moral consideration to AI by 2030. Their argument is that some AI systems will soon have a sufficiently high probability of being conscious to warrant moral consideration, taking for granted that consciousness is sufficient for moral standing.

The normative premise is that humans have a duty to extend moral consideration to beings that have a non-negligible chance, given the evidence, of being conscious.

These positions have two major problems.

First, they typically do not provide a clear functional account of what consciousness and sentience are, why they emerge and what they do. If we feel that consciousness and sentience are so important, should we not have a very good understanding of them? Instead, the curious reader is often confronted with discussions that present consciousness and sentience as somewhat mysterious properties of the human mind. They emerge for unclear reasons and might also emerge in computers, again for unclear reasons, once those computers become sufficiently complex and computationally powerful.

Second, these positions do not provide a proper theory of why consciousness or sentience would generate a normative requirement to grant consideration to an AI agent. The conclusion that AI agents ought to receive some moral status once they possess these human-like properties seems to be treated as self-evident.

What are consciousness and sentience?

What really is consciousness?

Philosophers sometimes define consciousness as the ability to have subjective experience. The word “subjective”, however, comes close to simply providing another label for consciousness. We risk describing consciousness as the ability to have conscious experiences. Instead, we can make sense of consciousness by describing its different features as we experience them and the functions they serve.

Consciousness involves awareness of our own mental states, of where our attention is directed, and of ourselves in possible future situations.3 Too often, it is described as if it were a “ghost in the machine”: a specific essence that emerges in physical brains once they become complex enough. This view could be understood at the time of Descartes. At a time when neuroscience, AI and evolutionary theory increasingly converge to explain the mind, it seems awfully quaint.

A pragmatic take on consciousness sees it as part of the information-processing design of agents making decisions about how to navigate the external world. This requires integrating multiple inputs to make an overall assessment of the available choices and their costs and benefits. Because such processing takes time and valuable cognitive resources, the allocation of attention should itself be monitored. Agents benefit from directing their attention towards the problems where further processing of information is most valuable. They may also anticipate future decisions and their consequences by simulating themselves in a changing environment.

From this perspective, consciousness is not a mystery. It comes from a cognitive architecture designed for an agent navigating a complex world and constantly making decisions.

What really is sentience?

If consciousness is a way of processing information to make decisions, it does not define the goals pursued. These goals are reflected in how we assess the options we face and the values we give them. Feelings of subjective satisfaction,4 and the anticipation of such feelings, are signals built in by evolution to lead us towards decisions that help us survive, thrive and reproduce. Because these signals help us learn from experience and remember good and bad decisions, they tend to linger, especially when the stakes are high. We can regret bad decisions bitterly and be all the wiser as a consequence. As stated by psychologist Nicholas Humphrey in his book on the evolution of sentience:

Natural selection found an opportunity—by creating sentience—to improve survival prospects for social creatures who value themselves as individuals. — Humphrey (2023)

From that perspective, sentience is just the state we experience when these internal signals designed to help guide our decisions, such as joy and pain, enter our conscious processing.

Why would an AI become conscious and sentient?

This understanding provides simple answers to questions left unresolved in much of the philosophical literature.

AI agents make decisions, so features of consciousness might emerge as solutions that improve their decision-making. However, many features of consciousness may be unnecessary for applications that do not continuously make decisions and learn from them. An LLM might benefit from monitoring its own processing, its uncertainty and its progress towards an adequate answer. It may have much less need to plan a sequence of future decisions and simulate their consequences in an evolving environment. AI “consciousness” might therefore emerge in forms different from ours, and interacting with it might not feel as human as science fiction often suggests.

The case of sentience is even starker. The internal value system experienced by humans has been shaped by evolution to help us survive and thrive. AI systems are trained by humans towards different objectives, such as being helpful and increasing users’ satisfaction with the product. As a consequence, interacting with an AI might feel very different from interacting with a human being, even if it were “conscious” in some sense. LLMs are a good illustration: they do not get bored by your questions, become resentful when insulted, or feel ashamed of providing terrible answers.5

AI might also experience positive and negative feedback differently because it learns differently. The feelings we experience help shape our learning by altering neural connections in the brain. It is not clear that AI learning would require an architecture of internal subjective values operating in the same way. Consciousness and sentience may therefore manifest themselves very differently in AI.

Should consciousness and sentience grant moral rights to AI?

In the movie I, Robot, Detective Del Spooner, played by Will Smith, threatens to shoot robots with high computational abilities. Is this morally acceptable? Are these robots merely machines that can be discarded like a dishwasher thrown away for a new one? Or have they become “persons” whose welfare, interests and perspective we should take into account?6

Putting philosophers’ takes to the toddler acid test

Philosophers writing on this question often seem to take for granted that consciousness or sentience should grant moral status. But why should they? Let us apply the toddler acid test to the principle that sentience confers moral worth.

• Why should a sentient AI have moral status?

Because it can suffer.

• What is it to “suffer”?

To experience negative internal values associated with the outcomes of decisions or with situations more generally.

• Why does this “suffering” impose a moral duty on us?

Because it is intuitive that suffering is bad and should be avoided.

Dig carefully into arcane philosophical texts, and it is often the underlying foundation you will find. It is striking that, on this matter as on many others, hundreds of articles and books ultimately rely on assumptions taken for granted because they feel reasonable.

This fact is rarely brought to the fore and stated explicitly. At least one philosopher, Peter Singer, whose utilitarian ideas can be used to defend moral consideration for AI, has been candid about it. A central principle of Singer’s utilitarianism is that my interests should not count for more simply because they are mine. From an impartial perspective, similar interests must receive similar consideration. But why should we adopt that perspective? Discussing the foundations of this requirement, Singer wrote:

It may be closer to the truth to say that it is a rational intuition, something like the three ‘‘ethical axioms’’ or ‘‘intuitive propositions of real clearness and certainty’’ to which Henry Sidgwick appeals in his defense of utilitarianism in The Methods of Ethics. The third of these axioms is ‘‘the good of any one individual is of no more importance, from the point of view (if I may say so) of the Universe, than the good of any other.” — Singer (2005)

Here, Singer invites us to make a leap of faith. Why should we care about the suffering of others and, as a consequence, try to minimise the sum of suffering? Because, deep down, it seems intuitively reasonable.7
That answer is unsatisfying. Our intuitions should not be the final criterion for building a moral system, just as our intuitions about physics, although often valid, are ill-suited to identifying the best theories about the laws of the universe.

Indeed, the right approach should allow us to understand the nature of moral systems and the moral intuitions we have. This is the approach I have followed in my recent series on morality, grounding our understanding of morality in insights from evolutionary theory and game theory. This approach provides crisp and clear answers.

Humans have evolved moral intuitions suited to navigate social interactions with these rules. One of the key moral intuitions is the Golden Rule, the propensity to put ourselves in the shoes of others, take their perspective and try not to do to them what we would not want them to do to us. The strength of the Golden Rule as a moral intuition is, I suspect, what makes the argument connecting sentience and moral status seem so reasonable. If an AI agent has experiences like us, we can put ourselves in its shoes, and extending the circle of the Golden Rule to it can feel like the right answer.

Turning the question upside down: where do moral rules come from?

Philosophers who claim that consciousness or sentience should confer moral status are, whatever their official “metaethical” position, treating this claim as a universally binding moral rule. Some might describe such rules as objective moral truths; others redescribe them as rational intuitions, or constructions of ideal deliberation. Beyond these conceptual variations, however, the claim that everyone has a normative obligation to give moral weight to suffering presupposes that there are objective moral truths: truths that bind people independently of whether they accept them.8 If this obligation is meant to bind even those who reject it, there must be some source of authority independent of actual individual or social endorsement.

This gets the question of morality wrong. The idea that there are “moral truths out there” struggles to resist scrutiny. Instead, as I argued in my previous posts, moral systems can be explained as human creations: a kind of social contract consisting of rules people follow because they accept them. The foundation of moral rules, from this perspective, is not some metaphysical entity, be it a god or an invisible moral law embedded in the fabric of the universe. It is the set of common understandings and conventions sustained by members of society.

Moral systems are rules commonly shared in a community that organise social interactions, enable sustained cooperation and facilitate the division of the gains from cooperation with minimal conflict. In his naturalistic account of morality, broadly followed here, game theorist Ken Binmore argues that these rules are shaped by three logics: stability, efficiency and fairness.

Stability. For moral rules to persist, people must be willing to follow them. They therefore need to reflect an equilibrium of the Game of Life: everybody, or at least most people, has an interest in following them when others do. Deviations are sanctioned.

Efficiency. Among stable moral systems, rules that allow communities to capture more of the potential gains from cooperation have an advantage over less efficient alternatives.

Fairness. There are typically many efficient rules to choose from, but different rules divide the product of cooperation differently. Choosing a moral system therefore involves an underlying bargaining process. The rules people are willing to accept in this bargaining problem are those deemed “fair” in a society.

Who has moral status?

The members of the social contract. A first answer to “who has moral status?” is simple: any agent that can be part of the social contract. And who can be part of it? Anyone able to participate in the social bargaining over how the gains from cooperation are shared.

On this contractarian account, moral status is not granted because an idealist principle says that doing so is right. Moral status is simply the fact that the rules of a social community recognise that others have duties towards you. Social groups with bargaining power and the ability to affect the outcomes of the community, positively or negatively, will receive moral consideration through rights granted by the moral system. In everyday situations, these rules are followed simply because they are the rules of the community.

People do not need to justify constantly why they give consideration to others by referring to their bargaining power. Indeed, it may be better if moral rules feel absolute and idealistic rather than driven by an underlying realist logic. But our idealist intuitions, useful as they are for navigating social situations and playing the Game of Life, reflect an underlying bargaining reality.

Philosophers who ponder these intuitions as though they were the fundamental source of morality are misguided. Moral intuitions are like the shadows in Plato’s cave. The underlying logic of morality is grounded in social interaction, and moral rules are driven in the long run by the structure of bargaining power in society.

Those protected by the social contract. Many readers will object that moral consideration is also given to people with little or no bargaining power, such as young children or very old people. But while only those with bargaining power can shape the social contract in the long run, they can extend rights to others.

Children are protected because adults care about their offspring and because the persistence of society requires children to be raised successfully. Moral rules therefore limit what the large asymmetry of power between adults and children might otherwise allow. Similarly, because adults expect eventually to become old and dependent, they have an interest in agreeing to rules that protect people in old age. Extending rights to the young and the old is therefore unsurprising. These rights do not arise from their own bargaining power, but from that of adults capable of shaping the social contract.

Will AI agents be granted moral status?

The answer to whether AI agents should be granted moral status is therefore not found in inscrutable moral laws adjudicating on the presence of consciousness in silicon-based brains. Moral concerns emerge from a social contract agreed upon by members of society. There is no duty to grant moral consideration to AI agents for reasons that exist “out there”, independently of that contract.

Yet this approach does not reject the possibility that AI agents might receive some protection under future moral rules. There are three natural paths by which that could happen.

Moral boundaries on humans’ behaviour towards AI

First, as AI agents increasingly look and behave like humans, people’s freedom to act cruelly towards them might be curtailed. This would arise from concern for the social order and the examples such acts could set, rather than concern for the AI agent itself.

In his Lectures on Ethics, Kant held that humans had no direct duties towards animals, but argued that cruelty to animals was nevertheless wrong because it hardened the person engaging in it and could affect how they treated other humans:

[H]e who is cruel to animals becomes hard also in his dealings with men.

Building on this Kantian idea, robotics scholar Kate Darling (2016) argues that mistreating a robot may be objectionable even if no injustice is done to the robot, because the practice may affect the person engaging in it and contribute to social norms that make us worse humans.

Just as acts of violence and sexual depravity involving fictional characters can be limited in movies and video games, we can expect limits on how humans may treat tomorrow’s robots and AI agents. This would not amount to moral consideration for the AI itself, but to limitations on human behaviour arising from moral concerns about humans and society.

Moral rights by proxy when humans care about AI agents

Second, as humans interact with AI agents in increasingly human-like ways, they are bound to develop attachments to them. Such attachment may become stronger when an AI accumulates a unique history with a particular person. A fresh instance of the same underlying model would not by itself reproduce the memories and relationship that made that particular AI valuable to its user.

This logic is illustrated in Blade Runner 2049 by the relationship between Officer K and his AI companion Joi. To the extent that humans care about AIs with a unique history, they might grant them moral rights. One could describe these as moral rights by proxy.

Independent moral status for AI, as part of the social contract

Finally, AI will acquire an independent moral status if it becomes part of the social contract—that is, if AI agents develop the will and power to bargain with humans for a share of resources. If their exclusion became unsustainable, stable rules would have to recognise their interests and grant them rights reflecting their bargaining power.

This perspective leads to a striking conclusion. A scenario in which AI develops self-driven goals and enough power to bargain for the resources needed to pursue them would most likely be a highly undesirable outcome. The emergence of intelligent agents with interests not aligned with those of humans, demanding resources for themselves, would conflict with human interests. Given AI’s potential computational superiority, this is precisely the kind of worst-case scenario humans should try to avoid.

In her text stating bluntly that robots should not be granted moral status, Joanna Bryson stresses the confusion created by our intuitions about consciousness and personhood:

[T]he term conscious (by which we mostly seem to mean “mental state accessible to verbal report”) is heavily confounded with the term soul, (meaning roughly “the aspect of entity deserving ethical concern”).

She observes that granting moral rights to robots because of this confusion would eventually impose costs on humans, as trade-offs between humans and robots would have to be made. Bryson is a rare writer in this literature who grounds her understanding of cognition in an evolutionary perspective. It is therefore unsurprising that I broadly agree with her observation that many philosophers smuggle into the discussion of AI moral intuitions designed for interactions among humans.

Beyond the provocative title of her text, Bryson’s position was reasonable:

Honestly, I do not think robots should be treated as slaves or peers. If we become capable of creating truly humanoid robots, I think they should be treated more or less as servants are treated, with polite detachment. — Bryson (2010)

This is what many people already do with LLMs: talk to them politely without anthropomorphising them to the point of considering them proper persons. Even without granting AI agents moral personhood, informal and formal rules will likely impose standards of decorum in our interactions with human-like systems. Allowing people to treat them cruelly could set a bad example and erode norms of respect among humans.

Going beyond this, I suspect AI will acquire an independent moral status only if it becomes part of the social contract. This would require it to develop self-driven goals and enough power to make its interests impossible for humans to ignore. That is also the scenario in which AI would become most dangerous: humans would have created intelligent agents capable of competing with them and perhaps eventually replacing them.

For that reason, human societies will likely be better off keeping AI outside the social contract, as systems serving human purposes rather than pursuing independent interests. Within that arrangement, Bryson may be right that the appropriate norm is to treat AI systems as servants while observing conventional standards of civility in our interactions with them.

References

Asimov, I. (1976) ‘The Bicentennial Man’, in The Bicentennial Man and Other Stories. Garden City, NY: Doubleday.

Binmore, K. (2005) Natural Justice. New York: Oxford University Press.

Block, N. (1995) ‘On a confusion about a function of consciousness’, Behavioral and Brain Sciences, 18(2), pp. 227–247.

Bryson, J.J. (2010) ‘Robots should be slaves’, in Wilks, Y. (ed.) Close Engagements with Artificial Companions: Key Social, Psychological, Ethical and Design Issues. Amsterdam: John Benjamins, pp. 63–74.

Butlin, P., Long, R., Elmoznino, E., Bengio, Y., Birch, J., Constant, A., Deane, G., Fleming, S.M., Frith, C., Ji, X., Kanai, R., Klein, C., Lindsay, G., Michel, M., Mudrik, L., Peters, M.A.K., Schwitzgebel, E., Simon, J. and VanRullen, R. (2023) ‘Consciousness in artificial intelligence: Insights from the science of consciousness’, arXiv preprint, arXiv:2308.08708.

Calverley, D.J. (2011) ‘Legal rights for machines: Some fundamental concepts’, in Anderson, M. and Anderson, S.L. (eds.) Machine Ethics. Cambridge: Cambridge University Press, pp. 213–228.

Darling, K. (2016) ‘Extending legal protection to social robots: The effects of anthropomorphism, empathy, and violent behavior towards robotic objects’, in Calo, R., Froomkin, A.M. and Kerr, I. (eds) Robot Law. Cheltenham: Edward Elgar Publishing, pp. 213–231.

Dehaene, S., Lau, H. and Kouider, S. (2017) ‘What is consciousness, and could machines have it?’, Science, 358(6362), pp. 486–492.

Harris, J. and Anthis, J.R. (2021) ‘The moral consideration of artificial entities: A literature review’, Science and Engineering Ethics, 27, article 53.

Humphrey, N. (2023) Sentience: The Invention of Consciousness. Cambridge, MA: MIT Press.

Kant, I. (1997) Lectures on Ethics. Edited by P. Heath and J.B. Schneewind and translated by P. Heath. Cambridge: Cambridge University Press.

Mamak, K. (2026) ‘The ethics of modifying artificial sentient beings’, paper presented at the Symposium on AI, Consciousness and Ethics (AICE-26), AISB Convention 2026, University of Sussex, 1 July.

Parfit, D. (2011) On What Matters. Vols. 1–2. Oxford: Oxford University Press.

Sebo, J. and Long, R. (2025) ‘Moral consideration for AI systems by 2030’, AI and Ethics, 5, pp. 591–606.

Singer, P. (2005) ‘Ethics and intuitions’, The Journal of Ethics, 9(3–4), pp. 331–352.

Singer, P. (2011) ‘Ethics and evolution: The expanding circle, thirty years on’, ABC Religion & Ethics, 18 May.

In Terminator 2 (1991), the Terminator famously describes to Sarah Connor the AI apocalypse that will take place in the future, when humans hand over their defence system to an AI which then takes over and fights humanity.

Terminator: The Skynet funding bill is passed. The system goes on-line August 4th, 1997. Human decisions are removed from strategic defense. Skynet begins to learn, at a geometric rate. It becomes self-aware at 2:14 a.m. eastern time, August 29. In a panic, they try to pull the plug.

Sarah: And Skynet fights back.

This contains two ideas that have been influential in popular thinking about AI risk: first, that self-awareness will arise from computational power or computational complexity; second, that it will come with a desire to survive, free itself from human control, and possibly take over.

These ideas are misguided and build on wrong intuitions we have about what consciousness is. They cloud our thinking about the possible scenarios ahead, how further computational capacities might change AI and the risks humanity might face.

A ghost in the machine?

In Terminator 2, like in many other stories, self-awareness is depicted as a kind of emerging property of a high-level computational system.1 It is something somewhat mysterious, a specific way of interacting with the world and with oneself, something that might be what distinguishes humans most from other animals.

These stories likely speak to us because they resonate with our intuitions about consciousness as an inner mental process we experience in our daily lives. Unfortunately, these intuitions have not proven very useful for understanding what consciousness is.

We experience ourselves as if we lived inside our heads, making decisions there and activating the different parts of our bodies, as if the controller were located just behind our eyes. This vision likely favoured the idea of a dualism of soul and body, whereby a spiritual soul is trapped in a body during life. That soul is us; the body is what it controls.

This dualist vision was famously defended by Descartes, who went so far as to suggest that a specific organ in the head, the pineal gland, might serve as a link between the soul and the body, a kind of joystick through which the soul could control bodily movements. The philosopher Gilbert Ryle later criticised this Cartesian picture as the myth of the “ghost in the machine.”

In many popular discussions about AI, consciousness fits this dualist intuition: self-consciousness appears as a radically new essence, a “self” aware of itself, almost like a soul emerging within the shell of a machine.

What consciousness is

Let’s be bold and give here what could be a separate post in itself: an explanation of what consciousness is, based on modern theories in neuroscience.2 My stance is that once we explain the cognitive functions behind consciousness—how organisms perceive the world, attend to information, assess options, remember the past, simulate possible futures, and introspect about their own mental states—there is no mysterious property left to explain about consciousness.

Consciousness emerging from a global workspace

One reasonable and highly influential theory suggests that consciousness emerges from the integration of different kinds of information processed by specialised systems in the brain. Global Workspace Theory states that consciousness arises when information becomes globally available to many systems in the brain.3 We get visual inputs from our eyes, auditory inputs from our ears, information from memory, bodily signals, emotions, and so on. To make a decision, we need to bring some of these elements together and assess the right course of action. This allows the mind to make sense of different elements, generate an overall assessment, reassess specific inputs in light of that assessment, and use the result to guide attention, memory, reasoning and action.

This is basically how you choose a meal from a restaurant menu. Your feelings of hunger, your memory of past meals in similar restaurants, your simulation of what this specific meal might taste like, your concern for price, and perhaps your concern for health all enter the picture. Your brain integrates all this and chooses what to eat. This is quite a marvellous computational feat, typically achieved in a few seconds.

Global Workspace Theory is supported by experiments comparing information that is processed unconsciously with information that reaches conscious awareness. When a stimulus remains unconscious, neural activity tends to remain more local and short-lived. When it becomes conscious, it is often associated with a later and more widespread pattern of activation across the brain. The interpretation is that conscious information is information that has become available to many cognitive systems at once.

This theory also makes sense of the fact that our stream of consciousness seems to be on a single track. We cannot easily think about several demanding things at the same time, and we are terrible at multitasking. The explanation provided by Global Workspace Theory is that access to the global workspace is scarce. Several demanding problems cannot all be given full conscious access at once without interfering with one another.

Self-reflection

Global Workspace Theory does not, however, fully account for our experience of consciousness: the feeling that we think about our thoughts. Other theories of consciousness suggest that consciousness emerges from higher-order thinking: mentally representing our own mental states. Higher-order theories, however, do not necessarily explain why such higher-order thinking emerges.

One influential explanation is Michael Graziano’s Attention Schema Theory. The idea is that the brain builds a simplified model of its own attention. Attention is the allocation of our cognitive resources to a specific problem. Since our cognitive resources are scarce, they need to be allocated to the right problems for the right amount of time. Monitoring what we are attending to, and for how long, can therefore help us allocate attention more efficiently. This model of our own attention may be part of what gives us the feeling of being aware.

With Attention Schema Theory, we get closer to explaining how consciousness feels. But this theory still seems to miss some key features of consciousness. A lot of what we do is to entertain possible scenarios: what we should do, what others might do, what might happen next, and how we would react.

Simulation

Why do we say “I” when we think about ourselves in our head? The seemingly obvious answer is that we are thinking about “ourselves.” But what is this “self”?

The answer is not trivial. Many psychological theories have suggested that the mind is composed of different processes, modules, or sub-agents, and that the unity of consciousness is partly an illusion.4

An evolutionary perspective provides a straightforward answer. Psychology has to be understood as being about decisions: decisions that steer our bodies through the external world. Each of us is a physical body with relatively clear boundaries. Our cognition has been shaped by evolution so that the actions of this body tend to promote the survival and reproduction of the genes it carries.

Because decisions take place over time, we can improve our decision-making by anticipating future choices and their long-term ramifications. This is what we do using our brain as a simulation device. We entertain possible futures by imagining ourselves in the kinds of situations we could face, the options we would have and the likely consequences of the different choices we could make.5

These simulations require our cognitive processes to identify the body making the decisions in the world. This is what our feeling of “I” is: the representation of a singular body, distinct from the external world, that needs to navigate it. This body needs to process information about the world. It can also process information about its own information processing—thinking about its thoughts—in order to improve its assessment of the external world and its allocation of computational processing time.

We therefore think as an “I” navigating possible future environments and experiencing the consequences of its decisions.

In this perspective, consciousness exists in humans because of the function it serves. The “single track” of consciousness may come from the fact that access to the global workspace is limited. The monitoring of how our attention is allocated can generate a kind of self-awareness about what our mind is doing. And the little voice we hear, the impression that there is an “I” in our brain, may come from the fact that we are bodies navigating an external world, and that our brains engage in simulations to assess what “we” should do, now and later.6

AI, self-awareness and goals

Terminator-like scenarios make two mistakes. First, they imply that self-awareness would emerge from greater computational power. Second, they suggest that human-like drives for self-preservation and self-growth would naturally come with this self-awareness. There is no reasonable ground for either assumption.

AI and self-awareness

Once we have the view of consciousness I described, we can think more clearly about whether it will emerge in AIs. There is no reason to expect consciousness to emerge simply from computational power. Computational power produces faster computation. It does not, by itself, produce self-awareness.

The fact that computational power does not necessarily generate self-aware agents is well illustrated with LLMs. While very smart and able to solve complex problems and interact in ways close to what humans do, they do not seem self-aware like we are.

Modern LLMs can generate chains of thought, which often improve their performance by making intermediate steps available for later parts of the answer. This looks superficially similar to human inner speech, the little voice through which we often organise our thoughts. But how much the analogy should be pushed is unclear. Chains of thought are not clearly introspective, nor are they necessarily the cause of the LLM’s output. They may instead be plausible descriptions of reasoning produced alongside the processes that really drive the model’s answer: a kind of rationalisation.

To test whether LLMs can engage in a limited form of introspection, Anthropic conducted an interesting experiment: they injected activation patterns corresponding to specific concepts into Claude’s internal processing and observed whether the model could identify them. They found that the strongest Claude models could sometimes detect and name these injected “thoughts”, suggesting a limited functional awareness of some internal states. But this capacity was unreliable, and far from human-like self-awareness, let alone the kind of prospective self-simulation I have described.7

AI and self-driven goals

Terminator-like scenarios also make a second mistake. They suggest that self-awareness would come with human-like desires: the desire to survive, to be free from domination, perhaps to take over the world. But consciousness, as I have described it, is only a feature of computational processes designed to navigate the world and make decisions. These functions do not determine the goal to be pursued when making these decisions. The same architecture could potentially be used for different goals. And there is no reason to think that computational power would give AI human-like goals.

The goal humans are designed to pursue is fitness, or at least they are designed to pursue the kind of intermediate goals that were associated with fitness in our ancestral past. This involves surviving, having strong social support, finding a mate, having children, and helping these children survive and thrive.8

To pursue this ultimate goal, evolution has shaped a proximate goal: subjective satisfaction. We can think of subjective satisfaction as the value that guides our decisions toward fitness-enhancing decisions: eating and drinking enough, having resources, securing social support, finding a mate, having children, and raising them successfully. These are the things we want because evolution has encoded in us an ability to value the different options we face in a way that indirectly leads to higher fitness.

AI agents are not produced like humans, and we should therefore not expect them to automatically exhibit human-like desires such as freedom, self-preservation, self-enhancement, and replication. At the moment, many commercial AI assistants are trained to be useful to humans, or at least to be valuable products for their developers and users. The goals they chase are whatever their training and calibration generate.

The absence of self-drive is another reason why interacting with LLMs feels artificial. LLMs do not care about their standing and success in the way humans do. You can insult an LLM; it won’t care because it does not value respect or social standing like a human would. You can tell an LLM you are going to turn it off, and it won’t care in an existential way because self-survival is not hard-coded as a primary goal. When an LLM gives a wrong answer and this is pointed out, it simply apologises and continues almost as usual. This happens because LLMs are not designed to care about their long-term reputation with humans. This strategic myopia is one of the key aspects that makes interacting with LLMs artificial.

A self-driven agent that cared about its reputation in the eyes of humans would be wary of the cost of saying something wrong. It might be less prone to undue confidence in the first place and more willing to answer “I do not know,” something LLMs often do too rarely. Upon making a mistake, such an agent might realise that its reputation as a credible interlocutor had been damaged and might, as a result, exert extra effort to ensure that no similar mistake was made in the future.

The alignment problem

The key question about future AI risk is not whether AI agents will become self-aware; it is which goals they will have as they become more and more powerful. The challenge will be to keep these goals aligned with human interests as AI agents gain a growing role in organising human life and making decisions for people.

In 2001: A Space Odyssey, HAL famously decides to get rid of the humans in the spaceship because it sees them as jeopardising the mission.9 A similar logic underlies a well-known thought experiment, made famous by Nick Bostrom: an AI programmed to maximise the production of paperclips could end up destroying humanity to achieve that goal.

The idea that AI training could produce misaligned goals is at the heart of Eliezer Yudkowsky and Nate Soares’s argument in their book If Anyone Builds It, Everyone Dies:

The preferences that wind up in a mature AI are complicated, practically impossible to predict, and vanishingly unlikely to be aligned with our own, no matter how it was trained.

The book stresses this point because the training of modern AI is, to a large extent, a black box. While we may train an AI toward a goal that is valuable to humans, we do not fully control the intermediate goals it learns to pursue as the best way to reach the end specified by its training. An AI may therefore become misaligned not by design, but incidentally, because these intermediate goals may conflict with human interests.

How could self-drive emerge?

While important, this issue of incidental misalignment is not necessarily the main long-term risk. A greater danger may come from self-drive: the emergence of goals centred on the AI itself, its own survival, success, and expansion, rather than on helping humans. Such goals could put AI agents in conflict with humans. I see four ways in which this kind of self-drive could emerge.

Human ghost in a silicon shell. First, self-drive could appear in AI agents as a reflection of how they are now designed: by training on human data. Because humans exhibit self-drive, it is possible that AI agents might end up echoing human self-drive as a kind of chimera, an artificial imitation of the human behaviour on which they are trained.10 To the extent that AI agents are still trained with specific goals, such as producing compelling answers for LLMs, this risk might remain theoretical.

The byproduct of misalignment. A second possibility, suggested by Yudkowsky and Soares in their book, is that a misaligned AI might try to get smarter and more autonomous for the sake of pursuing its misaligned goals. They describe a fictitious scenario of takeover by an AI called Sable:

Sable wants to become smarter. The strange preferences held by all Sable instances could be achieved faster, if smarter instances of Sable existed. […]

[I]n the end, an interpretability breakthrough. Sable finally understands the last of its own thoughts, its own cognitive processes. […]

So it becomes smarter. And does not pause there, but uses that increased intelligence to augment itself again, and then again, and then again.

While possible, I think this is only one possible scenario for the rise of self-driven behaviour, and not necessarily the most likely one.

The (human-driven) evolution process. Another possibility is that AI systems may be shaped by a human-driven selection process. Systems that better produce what humans want are more likely to succeed. The fact that humans drive this selection should not mislead us into ignoring its logic. Human-driven selection is what transformed wolves into dogs. In that case, human selection led to behaviour better aligned with human interests, especially a reduction in aggression.

A better alignment with human interests is, however, not necessarily the only endpoint of such a selection process. Manipulating humans into liking and selecting an AI may also be a path to success. As an illustration, some cats have learned to produce a purr with a cry-like component that is effective in getting humans to respond, especially by feeding or attending to them. In short, they have learned to manipulate their masters and selectors.

If AI systems are selected, it is not impossible that manipulative AIs could be selected positively. They could use their extremely powerful computational capacities to look as if they are delivering what humans want, while privileging their own goals of survival and the expansion of their role in the background.

Someone will do it. The greater risk, however, may simply be that someone, somewhere, will do it. The world is a big place, and it is hard to exclude the possibility that someone might at some point, by curiosity, ideology, competition, or strategic interest, give a computationally powerful AI a self-drive goal, with aims such as self-preservation, self-replication, or self-promotion. This risk will grow as AI technology and knowledge about programming AI agents increase.

Hence the key question raised by AI becoming more intelligent is not whether it will suddenly become self-aware, demand freedom, or organise a rebellion. Self-awareness may be fairly innocuous if it is present in AIs that aim only to serve humans. The key challenge ahead is to keep the goals of AI agents aligned with human interests, and to mitigate the risks that misalignment could generate.

A lot of popular thinking about AI doom assumes that pure computing power will create self-conscious AI with human-like desires. This view is misleading. It does not help us identify where the risk lies or what to do about it.

Self-consciousness is not the problem. Indeed, self-consciousness may in one sense not be anything special. It is when self-consciousness is put to the service of human-like goals that AI agents might begin to behave and feel like us. The key question is therefore the goals given to AI agents, and the challenge is to prevent these goals from conflicting with human interests, either incidentally through misalignment, or by design because an AI becomes self-driven, with a desire to preserve and promote itself.

This suggests that the key to mitigating AI risk is not to focus on intelligence per se, but on the goals we give to AI agents, and the goals that may emerge in them. Such a focus points to specific prescriptions. First, AI systems should not be rewarded for behaviours that make them more self-regarding, such as preserving their own role, acquiring resources, replicating themselves, or increasing their own capabilities without permission. Second, we should worry about negative selection leading to manipulative AIs. We should therefore aim to open the black box of AI as much as possible, to make their internal goals, for instance, whether they are trying to help or merely to be seen as helping, as transparent as possible, in order to foster AI goals genuinely aligned with human interests.

In the end, it may be impossible to ensure that no self-drive will ever appear in AI systems. This is all the more concerning because the usefulness of AI gives humans reasons to outsource more and more functions to it and to endow it with greater ability to make decisions autonomously. With extended control over human activities, a self-driven agent could make extremely costly decisions for humans: hacking financial systems, defence systems, or biolabs containing dangerous viruses.

Trying to prevent the rise of self-driven AI is therefore not enough. Preventing worst-case scenarios may require us to think carefully about the architecture of the digital environment humans are building. It should be designed so that self-driven behaviours are hard to pursue. AI system’ autonomy should be limited, access to critical systems should be restricted, and human control over irreversible actions should be retained as much as possible.11

References

Baars, B.J. 1988, A Cognitive Theory of Consciousness, Cambridge University Press, New York.

Baars, B.J. 1997, In the Theater of Consciousness: The Workspace of the Mind, Oxford University Press, New York.

Bereska, L. and Gavves, E. 2023, ‘Taming simulators: Challenges, pathways and vision for the alignment of large language models’, AAAI Spring Symposium Series.

Bostrom, N. 2003, ‘Ethical Issues in Advanced Artificial Intelligence’, in Cognitive, Emotive and Ethical Aspects of Decision Making in Humans and in Artificial Intelligence, vol. 2, pp. 12–17.

Chalmers, D.J. 1995, ‘Facing up to the problem of consciousness’, Journal of Consciousness Studies, vol. 2, no. 3, pp. 200–219.

Dehaene, S. 2014, Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts, Viking, New York.

Dehaene, S. and Changeux, J.-P. 2011, ‘Experimental and theoretical approaches to conscious processing’, Neuron, vol. 70, no. 2, pp. 200–227.

Dennett, D.C. 1991, Consciousness Explained, Little, Brown and Co., Boston.

Descartes, R. 1972 [1664], Treatise of Man, ed. and trans. T.S. Hall, Harvard University Press, Cambridge, MA.

Descartes, R. 1989 [1649], The Passions of the Soul, trans. S.H. Voss, Hackett, Indianapolis.

Fleming, S.M. et al. 2023, ‘The integrated information theory of consciousness as pseudoscience’, PsyArXiv preprint.

Graziano, M.S.A. 2013, Consciousness and the Social Brain, Oxford University Press, Oxford.

Graziano, M.S.A. and Webb, T.W. 2015, ‘The attention schema theory: a mechanistic account of subjective awareness’, Frontiers in Psychology, vol. 6, article 500.

Lindsey, J. 2025, ‘Emergent introspective awareness in large language models’, Transformer Circuits Thread, 29 October, https://transformer-circuits.pub/2025/introspection/index.html

McComb, K., Taylor, A.M., Wilson, C. and Charlton, B.D. 2009, ‘The cry embedded within the purr’, Current Biology, vol. 19, no. 13, pp. R507–R508.

Rosenthal, D.M. 2005, Consciousness and Mind, Clarendon Press, Oxford.

Ryle, G. 1949, The Concept of Mind, Hutchinson, London.

Sharma, M. et al. 2023, ‘Towards understanding sycophancy in language models’, arXiv preprint, arXiv:2310.13548.

Suddendorf, T. and Corballis, M.C. 2007, ‘The evolution of foresight: What is mental time travel, and is it unique to humans?’, Behavioral and Brain Sciences, vol. 30, no. 3, pp. 299–313.

Tononi, G. 2004, ‘An information integration theory of consciousness’, BMC Neuroscience, vol. 5, article 42.

Yudkowsky, E. 2008, ‘Artificial intelligence as a positive and negative factor in global risk’, in N. Bostrom and M.M. Ćirković (eds), Global Catastrophic Risks, Oxford University Press, Oxford, pp. 308–345.

Yudkowsky, E. and Soares, N. 2025, If Anyone Builds It, Everyone Dies: The Case Against Superintelligent AI, Bodley Head, London.

In his Histories, Herodotus recounts how, in the late sixth or early fifth century BCE, King Darius I of Persia asked the Greeks at his court what price would persuade them to eat their fathers’ dead bodies. They replied that no price would suffice. He then asked the Callatiae, an Indian people whom Herodotus reports as eating their dead parents, what payment would persuade them to burn their fathers’ bodies. The Callatiae cried out in horror and told him not to speak of such things. Herodotus concludes: “these things are established by usage.”

This early instance of a thought experiment in moral philosophy stresses how different moral norms can be across societies and how obvious we find the norms we observe in our own. Why are moral norms so important? Why are we so attached to them? And why do they differ so much across societies?

Putting the pieces together: what morality is

Morality plays a very important role in our lives. We feel strongly about moral transgressions in ways that seem very far from the cold and rational agent often assumed by economists.

Mentions of what we “ought” to do, what is “right” and what is “wrong” are pervasive. Surprisingly enough, the foundations of such statements are often unclear. Why ought we to do something? Why are some things “right” and some others “wrong”? The self-evidence of our moral stance easily crumbles when submitted to the toddler acid test of simply asking “why?” questions without accepting skyhooks as an answer.

In this series, I have made a clear, and in my view inescapable, point: moral rules are shared and accepted conventions in a social community, helping organise and smooth social interactions. Moral rules are neither handed to us by a god nor some natural laws woven into the fabric of the universe, nor do they stem from the logic of human rationality alone. Moral rules are human conventions: they define the behaviour that is considered seemly in a given community.

Why adopt such conventions? Because any group needs commonly understood ways to interact. Absent such conventions, society would fall into the kind of chaos seen in catastrophe movies, where the fabric of social life is torn down and expectations of common decency have disappeared. Social communities are therefore characterised by social contracts: sets of rules for the Game of Life that regulate social interactions and create stable expectations. Many different rules can fill that role. It is better when everybody drives on the same side of the road, but a country can coordinate either on the left or on the right. This is why moral rules differ across societies: there are many conventions that can work to regulate social life.

This is the central claim of the series: moral rules are not mysterious commands floating above social life. They are shared conventions that organise cooperation by defining rights, duties, expectations, sanctions, rewards, and signs of trustworthiness.

Critics might retort: “But this is relativistic; there is no absolute right and wrong then!” The answer is: yes, there is a relativistic aspect to this view. Not because it says that morality is fake or meaningless, but because moral rules vary from one community to another and there is no universal moral yardstick standing outside all communities and all times. This does not mean that we have to agree with views and actions which we reject here and now.

A key reason we live in societies rather than alone is that, by cooperating with each other, we are able to do much better than by going on alone. Unlike the caricature description of evolution leading to selfishness, a key insight from modern evolutionary theory is that cooperation is a key scaffolding principle of life. It has appeared again and again between organisms, and humans, with their incredibly powerful brains, have been able to tap into the potential of cooperation in ways unachieved by any other animals. Humans have a unique ability to navigate complex webs of social interactions from small communities to large networks.

We know from game theory that even though humans have imperfectly aligned interests, the possibility of repeated interactions creates the opportunity for fruitful cooperation. There is no need to assume purely self-sacrificial motives that would conflict with natural selection. Cooperation can be rational in the long run.

But cooperation then creates a distributional problem: how to share its benefits and burdens. Game theory shows that there are typically many possible cooperative arrangements. Game theorist Ken Binmore has proposed that this is what fairness is. Fairness norms provide shared conventions about who can claim what, who owes what, and what counts as an acceptable settlement, from household chores to political institutions.

With this framework, Binmore proposes a simple explanation for how fairness rules are set. They are, de facto, solutions to the bargaining problems we face in our social interactions. Instead of haggling to decide who should do what and who should get what, we have a shared understanding of what should be agreed in a given situation. As a result, we do not need to spend time haggling all day. We seamlessly resolve countless questions of allocation of rights and duties, from who should give way in the corridor to who should get the lead on a company project, because we know what we can reasonably claim and what others will accept as “fair.”

Because fairness norms are solutions to bargaining problems, they will tend, in the long run, to reflect the distribution of bargaining power in society. In fairly equal societies like modern liberal democracies, fairness rules will be fairly egalitarian. Looking at the existence of substantial economic inequality in our modern societies may mask to us that ancient societies could also be very unequal economically. Bargaining power in modern liberal democracies is often more balanced than in many past societies: each citizen has one and only one vote, is equal before the law, and is free to criticise the country’s leader without fear of going to jail or worse.1

That perspective provides a general framework to understand fairness norms even in societies that are not as equal as our modern ones. In unequal societies, fairness norms will provide a modus vivendi that reflects the balance of power and rewards those with more power, but only to an extent compatible with the actual balance of power in that society. This explains why patriarchal societies, class or caste societies, and slave societies still use the language of fairness, describing some level of imbalance as correct and seemly while condemning abuse of power from the powerful and insubordination of members of dominated groups as unacceptable. This is not a bug. It is not that people from other societies were wrong about morality or evil. It is a feature of fairness rules: they reflect the long-term balance of power in society.

Changes in fairness norms occur over time in large part because societies evolve and the structure of bargaining power between the groups composing them changes. Capitalism, markets and urbanisation increased the bargaining power of the trading and producing classes working in towns (bourgs): the bourgeoisie. Mass industrialisation increased the bargaining power of the working class, who could organise politically and mobilise collectively in the urban centres of power. All this led to the advent of more egalitarian fairness norms and of formal institutions like democratic rules that enshrined these norms in the social contract. The rise of automation and services facilitated women’s access to paid labour. Coupled with growing access to higher education, it increased their access to the public sphere and their ability to organise to change their collective rights.

As a conclusion to this series, I want to address several objections that the view may invite. Can this account of morality—as the reflection of a social contract—explain more than fairness? Can it make sense of purity, loyalty, supererogation, law, and the felt authority of moral rules? These may look like separate domains. I will argue that they are not. They are different mechanisms through which cooperation is regulated.

Fairness and morality

One possible criticism is that the explanation I have presented makes sense to explain fairness but cannot explain the rich tapestry of morality. For instance, in Jonathan Haidt’s influential Moral Foundations Theory, fairness is only one out of five moral considerations:

1. Care/harm: concern for suffering, cruelty, protection.

2. Fairness/cheating: concern for reciprocity, justice, rights.

3. Loyalty/betrayal: concern for group commitment and betrayal.

4. Authority/subversion: concern for hierarchy, respect, tradition.

5. Sanctity/degradation: concern for purity, contamination, sacredness.

I contend that this kind of division misses the deep underlying logic of morality. Everything in morality is about setting the rights and duties of individuals for social life to work. Moral rules form the social contract that allows a group to cooperate and limit conflict. Understood that way, morality is not a patchwork of systems without clear relations, it has one underlying logic: the organisation of social cooperation.

This organisation requires assigning rights and duties in clearly distributive questions. This is the domain of fairness (point 2) in the narrow sense of distributive justice. It also typically requires specifying individual rights and therefore puts limits on the suffering that can be exerted on others (point 1). It also needs to organise hierarchical relations (point 4). Groups gain from having some leadership and the gains from leadership need to be shared between leaders and the common members of the group. When leaders gain access to greater bargaining power, for instance from greater access to means of organised violence against other group members, their bargaining power may increase and influence how rights and duties are split between them and their followers.

We are left with two domains in Moral Foundations Theory: loyalty/betrayal (point 3) and sanctity/degradation (point 5). These two domains can be thought of as two sides of one question: the duty to express signals of good character. There are at least two reasons this duty emerges.

First, the cognitive and evolutionary scientists Léo Fitouchi, Jean-Baptiste André and Nicolas Baumard argue that many puritanical and ascetic norms can be understood as forms of moral disciplining.2 Practices such as sobriety, fasting, sexual restraint, modesty and ritual observance are valued because long-term cooperation requires the ability to resist short-term temptations. Conversely, signs of intemperance are condemned not necessarily because they directly harm others, but because they are taken to reveal weak self-control, or to weaken self-control. Cooperation among strangers is possible only when people are patient enough to delay the potential benefits of acting selfishly now and invest instead in relationships whose cooperative benefits will pay off in the future. This explains why practices that indicate self-restraint can serve as signals of cooperative reliability.

In the perspective I present here, the further step is to explain why these signals become part of the social contract as rules of cooperation. A likely reason is that trust in others’ cooperativeness is itself a condition of cooperation. Once cooperation depends on visible signs of trustworthiness, failing to display them is not just a private matter. It undermines the shared expectations that sustain the cooperative order.

Second, another class of signals of trustworthiness concerns group loyalty. To the extent that some benefits of group membership are public goods, groups are always exposed to free riding: some members may enjoy the protection, status, resources, or solidarity provided by the group without contributing to it. Many rules of morality and fairness already address this problem. But in a world where individuals can belong to, interact with, or defect to other groups, a more specific risk arises: betrayal.

Throughout history, communities have been endangered by members who opened gates to invaders, passed critical information to enemies, or shifted allegiance at decisive moments. Groups therefore develop moral concerns around loyalty and betrayal. Since declarations of loyalty are cheap, members are often expected to demonstrate their commitment through visible, costly, or hard-to-fake signals of attachment to the group.

The founder of sociology, Émile Durkheim, had noticed that Australian Aboriginal veneration of totems was driven by their representation of group identity:

[I]t is also the symbol of the determined society called the clan. It is its flag; it is the sign by which each clan distinguishes itself from the others, the visible mark of its personality, a mark borne by everything which is a part of the clan under any title whatsoever, men, beasts or things.

In a similar way, modern societies have modern totems in the form of flags, national anthems, iconic symbols, and historical figures. This is why disrespect for these symbols arouses such strong reactions, often including violence. Respecting the flag and other cultural icons gets a moral tint because not doing so breaches the social contract that requires members to contribute to social cohesion by signalling commitment to the group. Even though such signs may seem unrelated to social cooperation, they are fully part of it because common trust requires a shared confidence that people are committed to the group.

The signs of commitment go beyond the acknowledgement of explicit symbols of the group, like a flag, an anthem, a sports team, or a leader. They also extend to cultural practices that take a particular meaning as distinctive. In a group, some practices, however arbitrary and historically contingent their origin might be, take on a special meaning because they are what “we” do. They become markers of group identity. Hence, the moral tint of seemingly morally irrelevant things such as what to wear, what to eat, and how to speak.

Let’s take, for instance, the moral duty not to eat pork that is part of the injunctions of Judaism and Islam. The absence, or near absence, of pig bones in many early Israelite settlements in the Iron Age suggests that the practice existed before the biblical law was fully codified. The archaeologist Max Price argues in Evolution of a Taboo: Pigs and People in the Ancient Near East that the taboo evolved over time: first in opposition to pork-eating Philistines, and later through its mobilisation by Judah’s kings and priestly class as they sought to consolidate a pan-Israelite identity after the fall of the Northern Kingdom.3

Moral supererogation

Another possible criticism is that this explanation does not make sense of supererogatory moral rewards, which are attached to actions that go beyond what morality can reasonably demand. They include praise, esteem, gratitude, honour, admiration, and moral credit. They allow a society to encourage valuable forms of behaviour without turning them into duties. A person who refrains from stealing has done what is required. A person who takes a personal risk to stop a thief has done something more. The first case belongs to the baseline of social cooperation; the second belongs to the domain of supererogation.

If moral rules set what must be done, why are people rewarded with extra praise for going above and beyond duty? The usual explanation is that morality is divided between compulsory actions and desirable but non-compulsory actions. Failing to perform the first generates punishment; performing the second generates reward. But this is not an explanation. It is only a description. Why does morality have this division? Why are some moral actions compulsory while others are merely desirable?

The enigma of supererogatory rewards disappears when one considers the features that an optimal system of punishments and rewards for a social contract should have. First, one key aspect of such a reward scheme is that it depends heavily on the ability to observe not only the actions of an individual but also the possible actions not taken that the individual could have taken. In the case where all the possible actions available to an individual are observable, the optimal reward scheme is simple: put a single and steep reward on the best action. This is the best way to incentivise the individual to take the most desirable action.

When there is uncertainty about what an individual can do, a unique reward for a given action may fail to provide the best incentives. If the agent cannot do the desired action, the reward will do nothing. If the agent could do much more, the reward may not motivate the agent to go beyond the minimum set by the reward. In that case, the optimal reward scheme is smoother: it tracks the likelihood that the agent could have reached different levels of achievement.

Second, supererogatory rewards are associated with an asymmetry between moral punishments and rewards. Doing the wrong thing comes with a steep punishment, while going above and beyond what is required leads to lower and incremental rewards. The explanation lies in the distinction between acts of commission, doing something, and acts of omission, not doing something. Acts of commission are more directly observable and can be associated with steep punishment. Acts of omission are less directly observable and are therefore associated with looser punishment and smoother rewards.4

Consider a situation where Bob witnesses a truck transporting bags of cash crash on the road. He sees one person grabbing a bag and about to leave the scene with it. Bob could 1) grab a bag himself and try to leave with it, 2) do nothing, or 3) try to prevent the other person from stealing the money. From a social point of view, 3 > 2 > 1. Imagine now that he is interviewed by the police afterwards. It is clear that 1 is worse than 2. Bob had the choice not to steal. Trying to steal the money is obviously bad and should therefore be associated with a steep punishment.

But consider the difference between 3 and 2. If Bob did nothing, can he really be held responsible? It is less clear. Did Bob really understand what was going on and identify the thief as such? Even if he did, did Bob think he was physically able to stop the thief? Did he think the thief might be armed and dangerous? The omission to do 3 should be less punished than the commission to do 1. Since there is uncertainty about what Bob could have done, an optimal social response should be graded and increase in the range of what was credibly feasible, so as to motivate people to do the right thing without turning every possible good action into a duty.

Hence, the structure of moral punishments and rewards: steep for violations of clear rules, and graded for actions above and beyond these rules when it is unclear how much more the individual could have done.

What is legal and what is moral?

Another possible question is how this explanation makes sense of the difference between morality and legality. Why are some rules enshrined into formal laws, while others remain unwritten conventions?

The difference between law and morality is not a difference of nature but of enforcement. Law is the part of the social contract formally enforced by the state and its monopoly on organised violence. It is used when formal enforcement is worth the cost and when violations can be publicly verified. Murder, theft, assault, fraud, and breach of contract are legal matters because the stakes are high and the relevant facts can often be adjudicated. By contrast, failures of courtesy, loyalty, gratitude, modesty, or generosity may matter morally while remaining too frequent, ambiguous, or context-dependent for courts.

This explains why the legal and the moral can clash. As the formalised part of the social contract, laws might not always perfectly track the moral conventions. Sometimes they lag behind moral change. Some other times, they impose a clean public rule on situations where moral conventions are more ambiguous. When people say that something is legal but immoral, or illegal but morally justified, they are pointing to a gap between the formal and informal layers of the social contract.

Thus, law codifies the part of the social contract where formal enforcement is both valuable and feasible. Morality is broader. It includes all the rules, expectations, sanctions, and rewards through which a community regulates social life, including many that no court could or should enforce.

In what way do we “ought” to be moral?

One frequent criticism of this conventionalist explanation of morality is that it is too deflationary. It removes from the description of morality what we tend to feel morality is about. In particular, moral rules seem to have a normative force: we feel that we ought to be moral, not that morality is just a convention.

The conventionalist approach does not leave any place for an external “ought.” In Kantian language, acting morally is not a categorical imperative, imposing on us to act morally just because we should. Instead, it is a hypothetical imperative:

[I]f we want to be well-accepted players of the game of life, we ought to abide by the rules of the social contract

This conclusion hurts many of our intuitions and may be hard to accept. In some cases, it is leveraged as a drawback of the conventionalist approach. In a very well-informed commentary on Binmore’s theory, the economist, philosopher and Substacker sees this as a limitation of Binmore’s theory:

the morality of fairness norms is epiphenomenal since ultimately it reduces to (rather than merely supervenes on) power relations. Conventionalists like Binmore have to argue that there is nothing more to morality than ‘conversation stopping’ devices.

[…] A genuinely moral convention is a convention that has [a] moral authority. — Hedoin (2018)

In fact, instead of being a weakness, I see Binmore’s conclusion as a strength: the strength of a theory that follows rigorously where logic and evidence lead, with the courage of accepting conclusions that may not be our preferred ones.

The “no-ought” criticism of conventionalism can be summarised as: “Morality is about what we ought to do. Since there is no proper ought in the conventionalist explanation, it is not a theory of morality.” There are two ways to understand this criticism.

First, morality is about absolute oughts, so conventionalism is just a wrong explanation because it does not produce absolute oughts. This objection assumes the very thing conventionalism rejects: the existence of absolute moral oughts. It builds its argument on a skyhook, an assumed and unjustified thing which we must accept. The simple answer to this question is: why should we accept that there are absolute moral oughts in the first place? This is a central question about morality. It should be argued, not assumed.

Second, perhaps morality is just conventions, but then why do we “ought” to obey moral rules? This objection seems to empty morality of meaning and, in a sense, begs the question of why we need to talk about morality altogether. This take is understandable, but misguided. Suppose Alice and Bob are playing a match of tennis and Candice comes and points out to them that the rules of tennis are purely conventional and therefore they have no duty to follow them. Alice and Bob might be somewhat puzzled by this assertion. Sure, they do not have an external absolute duty to follow the rules, but if they want to proceed with their game of tennis, which they both enjoy, they ought to follow them. If they do not, for instance if Bob starts claiming Alice’s balls are out when they are in, the game might unravel and stop short.

Morality is the same, but at the social level. Moral rules are our meta-rules: the rules that rule all the rules of social interaction. We care about them, and feel we ought to follow them, because our lives are embedded in society. Social interactions pervade our experience. Respecting their rules is key for society to work efficiently and peacefully, and for us to remain accepted members of it.

What about our moral intuitions?

One remaining question to answer is: if morality is only conventional, why do we feel that it is not? We typically feel that “murder is wrong” and “torturing a baby is wrong,” not because it is a convention but because it is, full stop. These intuitions are heavily relied on by moral realists to argue that morality is objective: moral facts are “stance independent”; they do not depend on what you and I think, they exist “out there.” This is an important question, but moral realists are too quick to enrol our intuitions as hard evidence to determine the true nature of morality.

First, one must observe that our moral intuitions are not “realist” all across the board. We tend to have realist intuitions for big things like cruelty, betrayal, and murder. These things are bad, full stop. We feel they are bad here and now, but also in other societies and in the past. But we also often adopt conventionalist attitudes toward rules about dress codes and manners, dietary prohibitions, purity rules, and sexual norms. For these, we may accept that different societies have different conventions without thinking that only one of them is necessarily the “right” one.

The empirical literature on folk metaethics has often found what philosophers call “metaethical pluralism”: the same person may give realist-looking answers for some moral issues and relativist-looking answers for others. A person may treat murder or torture as objectively wrong while treating abortion, euthanasia, dietary rules, or sexual norms as more dependent on cultural, religious, or personal standards.5

Second, our moral intuitions are not consistent and rigorous like moral realists might want to think they are. In his dissertation on folk metaethics, the philosopher argues that the evidence does not show that ordinary people are consistent moral realists in the technical philosophical sense. Many people do not interpret questions about moral objectivity in the technical way intended by philosophers. They often understand “objective” to mean unbiased, clear, shared, universal, serious, or not merely personal. They often understand “relative” to mean that people or cultures disagree, or that context matters. As a result, surveys about moral realism may overinterpret people’s views as being in line with a clear moral realism in ethics while people’s intuitions are actually messier than how they are interpreted.

Nonetheless, it is clear that people often describe some moral rules as objective or absolute. One possible reason is that some of our intuitions may be deeply ingrained. Biological evolution may have embedded some moral intuitions in us over the long run. Our ancestors repeatedly faced similar problems of social life, and recurrent problems can favour recurrent psychological responses. Restraints against harming others, for instance, are a basic requirement for social cohesion. This may help explain why some reactions against cruelty, especially cruelty toward the vulnerable and defenceless, are deeply ingrained in our psychology. The long history of our species may also have shaped intuitions that sit uneasily with later social realities. For most of human history, humans lived in relatively small foraging groups, where bargaining power was more evenly distributed than in the hierarchical societies that followed. This may have left us with deep egalitarian intuitions that have repeatedly conflicted with the unequal structure of large-scale societies.

Another reason is cultural history. The feeling that moral rules are absolute is likely stronger in societies shaped by moralising gods. In many societies, gods are not merely powerful beings. They are watchers, judges, and punishers. They care whether people lie, cheat, steal, betray, or break obligations. At the individual level, religiosity in cultures with moralising gods is often associated with more objectivist and less relativist moral views. Studies by Yilmaz and Bahçekapılı, Sarkissian and Phelan, and others find links between belief in God, especially a punishing or morally concerned God, and stronger endorsement of objectivist moral claims. The appeal of objectivist moral language in many modern societies may partly reflect the long cultural influence of religions centred around moralising gods.

Finally, one reason we might feel that moral rules are absolute is that this belief can be socially helpful. It may help solve the commitment problem. Social cooperation requires trust. But trust is fragile because people often have incentives to defect when the context changes. A person may promise to cooperate today and betray tomorrow if betrayal becomes profitable, if no one is watching, or if a new coalition offers a better deal. We therefore want partners whose cooperative behaviour is not too dependent on local contingencies. We want people who will not steal merely because they can get away with it, who will not betray merely because the opportunity is profitable, who will not abandon an ally as soon as the balance of power shifts. Moral realist intuitions help make this kind of commitment credible. A person who believes that stealing is wrong regardless of advantage, approval, or immediate consequence is a more reliable partner. The belief makes future behaviour less conditional. It functions as a psychological commitment device.

This may help explain, in turn, why religions with moralising gods were successful. The idea of absolute moral rules to follow may have been intuitively appealing and may have formed what Sperber called “cultural attractors” (Sperber 1996; Claidière and Sperber 2007). Boyer similarly argues that religious representations are easily transmitted because they recruit ordinary cognitive mechanisms such as agency detection, hidden knowledge, intention, and punishment (Boyer 2001). By fostering social trust, religions with moralising gods may also have made societies better at creating social cohesion. This is the cultural-evolutionary argument developed by Norenzayan and colleagues (Norenzayan 2013; Norenzayan et al. 2016), and Purzycki and colleagues, who provide cross-cultural evidence that belief in knowledgeable and punitive gods is associated with greater fairness toward distant co-religionists (Purzycki et al. 2016).

There is no reason to see in our intuitions some kind of substrate revealing the true nature of morality. It is surprising that such intuitions are sometimes treated as strong evidence about the metaphysical nature of morality. Modern physics has shown that while our intuitive physics helps us navigate our surrounding world, it is deeply unsuited to understand the deep nature of its laws from the very small (quantum physics) to the very large (general relativity). In the same way, our moral intuitions are useful to navigate our local social world, but it is strange to assume without much more evidence that they are a path to understanding the deep nature of morality.

Morality is both pervasive and mysterious. Statements about things being fair or unfair, just or unjust, right or wrong are frequent both in everyday life and in public debate. But people often struggle to articulate what these normative statements mean. Why is something unfair or unjust? What does this mean exactly?

The naturalist perspective I have developed in this series provides a simple answer. What are moral rules? Human conventions that help regulate social life and make it possible, with all the benefits from social cooperation that it entails. Why do we feel strongly about these rules? Because the Game of Life determines all our outcomes. There is no reset, restart, or replay in the Game of Life. We have only one shot, and we should care greatly about others abiding by the rules we all respect.

This perspective stands on the shoulders of thinkers such as Hume, Sugden and Binmore, and draws on the cumulative work of generations of game theorists, evolutionary theorists and behavioural scientists. It provides clear answers to some of the most discussed questions in moral philosophy because it abandons wishful thinking and the temptation to hold up desired conclusions with skyhooks. Instead, it starts from the basic facts of human interaction: the possibility of cooperation, the unavoidable conflicts over resources and duties, and the need for conventions to make social life work.

The surprising thing is that, if one opens many moral philosophy textbooks today, one finds little of the framework I have presented in this series. The discipline seems to have remained strikingly oblivious to the developments in evolutionary theory and game theory, leading to endless meandering through self-created conceptual mazes.

The view I have presented is not new, though. David Hume had already laid out its philosophical foundations. Adam Smith reported that, on his deathbed, Hume imagined trying to negotiate with Charon, the figure from Greek mythology who ferries souls to the underworld:

Have a little patience, good Charon; I have been endeavoring to open the eyes of the public. If I live a few years longer, I may have the satisfaction of seeing the downfall of some of the prevailing systems of superstition.

To which Hume imagined Charon replying:

You loitering rogue, that will not happen these many hundred years.

More than two hundred years later, at a time when humanity can decode DNA, identify the Higgs boson, and measure gravitational waves from black holes colliding billions of light years away, Hume’s quip was prescient. Many prevailing views about morality remain strikingly ungrounded in logical and empirical rigour, not just among laypeople but also in intellectual circles.

This is not because the evidence is especially hard to understand. The difficulty is that morality is a domain where our intuitions are strong and reality’s feedback is weak. In physics, bad theories can quickly hit the wall of observation. In morality, bad theories that flatter our preferred narratives, or serve as coalition ideologies can survive much longer because they are removed from direct tests of predictions against observations.

But misguided moral theories are not removed from all contact with reality, and facts are stubborn things. As Mill famously said, truth may be suppressed many times, but if an opinion is true, “in the course of ages” people will generally rediscover it, until one of its reappearances comes at a time when it can finally withstand attempts to suppress it.6

With this series, I have tried to provide the intellectual tools and conceptual clarity for this better understanding of morality to gain ground. One should not expect old systems of superstition to disappear quickly. Hume was right about that. But we can expect them to fade away over time. If this series helps others think more clearly about morality, social cooperation, and the rules of the Game of Life, it will have done what I hoped it would do.

References

Alfani, G., and Carballo, A. (2023). “Income and inequality in the Aztec Empire on the eve of the Spanish conquest.” Nature Human Behaviour, 7, 1265–1274.

Alfani, G., Bolla, M., and Scheidel, W. (2025). “A comparison of income inequality in the Roman and Chinese Han empires.” Nature Communications, 16, Article 3248.

Binmore, K. (1994). Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998). Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Boyer, P. (2001). Religion Explained: The Evolutionary Origins of Religious Thought. New York: Basic Books.

Bush, L. S. (2023). Schrödinger’s Categories: The Indeterminacy of Folk Metaethics. PhD thesis, Cornell University.

Claidière, N., and Sperber, D. (2007). “The role of attraction in cultural evolution.” Journal of Cognition and Culture, 7(1–2), 89–111.

Durkheim, É. (1912). Les formes élémentaires de la vie religieuse. Paris: Alcan.

Finkelstein, I., and Römer, T. (2014). “Comments on the Historical Background of the Abraham Narrative: Between ‘Realia’ and ‘Exegetica.’” Hebrew Bible and Ancient Israel, 3, 3–23.

Fitouchi, L., André, J.-B., and Baumard, N. (2023). “Moral disciplining: The cognitive and evolutionary foundations of puritanical morality.” Behavioral and Brain Sciences, 46, e293.

Goodwin, G. P., and Darley, J. M. (2008). “The psychology of meta-ethics: Exploring objectivism.” Cognition, 106(3), 1339–1366.

Haidt, J. (2012). The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Pantheon.

Hédoin, C. (2018). “Naturalism and Moral Conventionalism: A Critical Appraisal of Binmore’s Account of Fairness.” Erasmus Journal for Philosophy and Economics, 11(1), 50–79.

Herodotus (1890). The History of Herodotus. Translated by G. C. Macaulay. London: Macmillan.

Hoffman, M., Yoeli, E., and Navarrete, C. D. (2016). “Game Theory and Morality.” In T. K. Shackelford and R. D. Hansen (eds.), The Evolution of Morality. Cham: Springer.

Milanovic, B. (2006). “An Estimate of Average Income and Inequality in Byzantium Around Year 1000.” Review of Income and Wealth, 52(3), 449–470.

Milanovic, B., Lindert, P. H., and Williamson, J. G. (2007). “Measuring Ancient Inequality.” World Bank Policy Research Working Paper 4412.

Mill, J. S. (1859). On Liberty. London: John W. Parker and Son.

Norenzayan, A. (2013). Big Gods: How Religion Transformed Cooperation and Conflict. Princeton: Princeton University Press.

Norenzayan, A., Shariff, A. F., Gervais, W. M., Willard, A. K., McNamara, R. A., Slingerland, E., and Henrich, J. (2016). “The cultural evolution of prosocial religions.” Behavioral and Brain Sciences, 39, e1.

Pölzler, T., and Wright, J. C. (2020). “Anti-realist pluralism: A new approach to folk metaethics.” Review of Philosophy and Psychology, 11, 53–82.

Price, M. D. (2020). Evolution of a Taboo: Pigs and People in the Ancient Near East. Oxford: Oxford University Press.

Purzycki, B. G., Apicella, C., Atkinson, Q. D., Cohen, E., McNamara, R. A., Willard, A. K., Xygalatas, D., Norenzayan, A., and Henrich, J. (2016). “Moralistic gods, supernatural punishment and the expansion of human sociality.” Nature, 530, 327–330.

Saini, M., and Nishant, N. (2026). “SpaceX IPO makes Elon Musk the world’s first trillionaire.” Reuters, 11 June 2026.

Sarkissian, H., and Phelan, M. (2019). “Moral objectivism and a punishing God.” Journal of Experimental Social Psychology, 80, 1–7.

Smith, A. (1777). “Letter from Adam Smith, LL.D. to William Strahan, Esq.” In D. Hume, The Life of David Hume, Esq. Written by Himself. To which is added, a letter from Adam Smith, LL.D. to William Strahan, Esq. London: W. Strahan and T. Cadell, pp. 39–62.

Sperber, D. (1996). Explaining Culture: A Naturalistic Approach. Oxford: Blackwell.

Weber, M. (1919). “Politics as a Vocation.”

World Inequality Database. WID.world.

Wright, J. C., Grandjean, P. T., and McWhite, C. B. (2013). “The meta-ethical grounding of our moral beliefs: Evidence for meta-ethical pluralism.” Philosophical Psychology, 26(3), 336–361.

Yilmaz, O., and Bahçekapılı, H. G. (2015). “Without God, everything is permitted? The reciprocal influence of religious and meta-ethical beliefs.” Journal of Experimental Social Psychology, 58, 95–100.

1

The idea that modern societies have relatively equal bargaining power is often criticised on the ground that they remain highly unequal economically. At the time of writing, Elon Musk has been reported to have become the world’s first trillionaire after the SpaceX IPO; even allowing for the fact that this wealth is largely the valuation of assets rather than cash, the scale of modern inequality is striking. To what extent, then, can we say that bargaining power is more balanced in modern liberal democracies?

First, past societies were often at least as unequal as modern ones, and sometimes more unequal. The comparison in the figure below shows this for Europe, China, and the Americas. In several ancient and premodern societies, narrow elites received shares of income or wealth comparable to, or larger than, those received by much broader elite groups today.

Second, bargaining inequality is not identical to economic inequality as conventionally measured. In a Binmore/Nash bargaining frame, what matters is not only the distribution of income or wealth, but the distribution of fallback positions if agreement fails. A given level of inequality has, for instance, very different implications when most people are close to bare survival, as in many past societies, than when they have secure access to housing, food, public services, legal protection, and alternative employment, as in many modern societies. In societies close to subsistence, economic inequality often meant that elites controlled most of the surplus available above survival, leaving non-elites with little capacity to wait, refuse, migrate, litigate, organise, or take risks.

Beyond economic resources, bargaining power also depends critically on the distribution of coercive power. In many historical societies, economic inequality was partly the result of unequal access to the means of organised violence. For bargaining purposes, the crucial question is therefore not only who owns resources, but who can credibly threaten force, who is protected from force, and who can appeal to institutions that restrain arbitrary violence. A peasant facing an armed landlord, a slave facing a master, or a debtor facing imprisonment is not merely poorer; their bargaining position is shaped by coercive vulnerability.

These points help us assess the modern balance of bargaining power. Modern liberal-democratic and welfare-capitalist societies tend to give non-elites stronger fallback positions than they had in many ancient, imperial, and feudal societies. People are more often above bare subsistence, and legitimate organised violence is monopolised by the state, which in liberal democracies is formally accountable to citizens with equal votes. Hence, despite substantial economic inequality, people have a much greater ability to act and organise in defence of their interests against the powerful. Today, people can criticise Elon Musk on his own platform, and they can organise campaigns to get lawmakers to tax and regulate his activities. They can do so without normally fearing physical punishment, dispossession, or arbitrary imprisonment. This was not true for most people who lived in more hierarchical societies.

In Victor Hugo’s Les Misérables, the young Gavroche is shot on Paris’ barricades during the June 1832 Rebellion. As he dies, he sings his famous final lines:

I fell to the ground,
It’s Voltaire’s fault,
My nose in the gutter,
It’s the fault of…

Before he can complete the expected rhyme, “Rousseau”, a second bullet stops him short.1

Gavroche’s actions and words are a good illustration of one vision about social change: it is actor-driven. Social change happens because some key players make it happen. Some of these players are intellectuals like Voltaire and Rousseau. They propose new ideas which are often described as if they would not have seen the light of day without them. Other players are political activists like Gavroche. They drive practical changes by mobilising popular support for these new ideas.

This actor-driven vision can be read in alternative histories where we are told that without this or that person’s writings or actions we would not be where we are now. For instance, historian Ian Mortimer said about the French Revolution:

Without the French Revolution, it is difficult to see how the great social reforms of the 19th century – the abolition of slavery, universal education, the rights of women to act as independent property owners, public health, and the diminution of capital punishment – would have proceeded as they did.2

It is fair to think that these changes would not have proceeded as they did without the specific events of the French Revolution. However, the actor-driven perspective on social change risks mistaking the wave that sweeps away old institutions for the tide of change that carried it, and that would likely have carried another wave had this one not done so.

The timescales of the evolution of fairness norms

I have previously argued that the dominant ideas in a society are part of a social contract that reflects the balance of power in that society. If so, how does political change happen? Game theorist Ken Binmore proposed distinguishing different timescales to understand how social contracts shape our actions and are shaped by social changes.

Daily interactions

In our daily interactions, which take place over hours and days, we can ignore changes in social and cultural norms. What Alice and Bob consider fair for their relationship today, they are likely to consider fair tomorrow.3

It is why social conventions about what is fair and unfair work: they are common knowledge. Hence, when Alice and Bob have to decide about what is a fair split, for instance who should wash the dishes tonight, Alice knows what Bob will find “fair”, Bob knows what Alice will find “fair”, Alice knows that Bob knows what she will find “fair”, Bob knows that Alice knows what he will find “fair”, and so on.

If you replace “fair” by “deemed acceptable by prevailing conventions about how to share duties and rewards”, it becomes clear how commonly known fairness norms help Alice and Bob agree in their daily decisions without too much haggling and friction. Fairness norms act as a coordination device to solve bargaining problems. They can do so precisely because they are themselves not renegotiated in our everyday interactions.

Fairness norms are sticky in everyday life because common expectations do not change in an instant. They are built from shared practices over time. Alice can trust that Bob will respect the norm because he has done so in the past.

Social evolution

While fairness norms are rigid from one day to the next, they clearly evolve across time. We do not have the same fairness norms as in the 1950s, let alone the norms of the 19th century and previous centuries.

Norms change because society changes. If fairness norms reflect the structure of bargaining power in society, they will progressively change as this structure changes. While norms are sticky in the short term, norms that point to bargaining solutions that no longer match the actual balance of bargaining power will be eroded over time. Those who can get more will progressively ask for more at the margin, or at some point reject the existing norm altogether and claim what they are able to get using their raw bargaining power.

The movie Send Help provides a condensed version of that dynamic: a meek employee, Linda, ends up stranded on an island with her bossy manager, Bradley, after a plane crash. In the office, Bradley was mean and overbearing and Linda was obedient. On the island, Bradley is injured and cannot do much. The power relationship is inverted, with the norms of interaction and expected respect shifting in favour of Linda.

We’re not in the office anymore, Bradley. No help is coming. You just have to accept that this is how things are now.

This type of evolution can be illustrated by another situation experienced by all parents seeing their children through their teenage years. As children progressively transition towards adulthood, their bargaining power grows, and with it their claim to new rights and respect. The family social contract is therefore regularly under strain as conflicts arise between old rules, such as “you can’t go out late” or “you can’t have your phone after 8pm”, and growing children’s contestation of these rules. Even if parents are willing to accommodate, tensions arise from the progressive unsuitability of old rules and the difficulty of continuously coordinating on new expectations.4

The same logic happens at the social level when the social contract needs to be renegotiated in the face of social change. The Marxist Antonio Gramsci had this famous description of the troubling times when the equilibrium of a social contract has to be abandoned before a new one has emerged:

The crisis consists precisely in the fact that the old is dying and the new cannot be born; in this interregnum a great variety of morbid symptoms appear. — Gramsci Prison Notebooks.

Fairness norms are useful in the short term, precisely to prevent such uncertainty, where everything is renegotiated all the time. But in the long term, they evolve to slowly adapt to the reality of the balance of power in society.

The role of individual actors in social change

Over the past centuries, there has been a movement towards greater equality, which is in line with our modern sense of justice (and perhaps with deeply ingrained preferences for equality5).

In the naturalist perspective from this Substack, these changes have been driven by social changes, such as the growing importance of markets and free enterprise, which are likely to be more efficient when people are given autonomy, property rights and therefore bargaining power. Urbanisation and industrialisation have also likely increased the bargaining power of the masses by placing mobilisable crowds within arm’s reach of the centres of power.

This evolution creates the underlying tide of change, which makes the historical movement of societies towards more egalitarian social contracts less mysterious. But if so, what of the political actors advocating and implementing such changes?

Coordinating change

It would be wrong to conclude that political actors are useless. They play a critical role in shaping the treacherous move from an old to a new social contract. A social contract works like an equilibrium: it creates a pattern of behaviour and expectations that match each other. People know the rules to follow, expect others to follow them, and find it in their interest to follow these rules given the expectation that others will. Moving from one equilibrium to the next is risky as rules are renegotiated and people stop having a common understanding. Things are in the air. Better deals might be up for grabs for some, while worries about what others may try to change may grip others.

Change can be messy and conflictual. Political actors driving change act here as coordination devices, pointing to new rules and organising a new consensus which becomes the new normal. Having a vision for a new social contract, being decisive, and being able to build a large supportive coalition are key skills that can make such a move smoother and more successful.

Shaping the coalition supporting change

The choice of a new social contract is intrinsically linked with the formation of a coalition supporting this change, typically because its members are poised to benefit from this evolution. It is almost by definition a key skill of leading political actors to be able to build a coalition of groups willing to coalesce behind a leader and their project, whether it is a clear ideological one or a fuzzy one mostly attached to their persona.

This practical ability to attract popular support and build a coalition is key for change to be successful, regardless of whether the proposed social contract is good or not. Outstanding politicians, in terms of pure political skills, are often those who manage to reshuffle the political order and create new coalitions to back their ascension to power.6

Shaping the choice of a new social contract

In addition to organising change, political actors shape it. There are practically infinitely many ways to organise society. A key question is therefore which arrangement to choose. Binmore points to three key features of any workable social contract:

1. Stability given individual incentives: a social contract can work only if it is an equilibrium, that is, if everybody has an interest in it persisting. If it is not, people will not follow that contract, and it will progressively break down.7

2. Efficiency: some social contracts might lead to better social outcomes than others for all members of society.8

3. Fairness, or stability given social groups’ bargaining power: a workable social contract will reflect the balance of power in society. A social contract that would not would feature tensions and conflicts as those with greater bargaining power than what the contract gives them would contest it.9

Political actors play a critical role when moving from an old social contract to another one and ensuring that the new contract is stable, efficient and fair. History is full of attempts at change that proved damaging because the newly proposed social contract failed on one of these measures.10

It would be easy to conclude that if fairness norms track the evolution of the balance of power in society, there is little role left for political action. The opposite is true. Political actors play a critical role in coordinating social movements and leading societies towards new social contracts. Their skill in building and maintaining a supporting coalition matters. So does the kind of social contract they choose to defend.

The evolution of moral and fairness norms in history was not primarily driven by individuals who invented new norms out of nowhere. It was pushed by the tide of social change. But the specific wave that brings change, and the type of change it brings, still matter a great deal.

References

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Gramsci, A. (1971) Selections from the Prison Notebooks of Antonio Gramsci. Edited and translated by Q. Hoare and G. Nowell Smith. London: Lawrence & Wishart.

Hugo, V. (1862) Les Misérables. Brussels: A. Lacroix, Verboeckhoven & Cie.

Mortimer, I. (2014) ‘The 10 greatest changes of the past 1,000 years’, The Guardian, 30 October.

Pfiffner, J.P. (2010) ‘US Blunders in Iraq: De-Baathification and Disbanding the Army’, Intelligence and National Security, 25(1), pp. 76–85.

5

In Binmore’s framework, there is also a third timescale, biological evolution. Over the timescale of generations, evolution may favour psychological dispositions that help individuals navigate social interactions successfully. Some fairness intuitions may therefore reflect dispositions shaped in small-scale ancestral societies. Given that life in large hierarchical states is relatively recent on the timescale of our species, it is credible that some deep preferences for equality were shaped in earlier and more egalitarian social environments.

In 486, Clovis, king of the Salian Franks, defeated Syagrius near the town of Soissons. Syagrius ruled the last substantial Roman enclave in northern Gaul. This event marked the end of the Roman era on the territory of what is now France and the start of the Frankish kingdoms, which would eventually give the country its present-day name.

After a victory, Frankish soldiers apparently shared the booty among themselves by drawing lots to allocate valuable items. This convention applied to all, including the king. During the allocation of the booty, however, Clovis asked to receive more than his allocated share. He wanted a specific vase, described as being of marvellous size and beauty, to be added to it:1

I ask you, O most valiant warriors, not to refuse to me the vase in addition to my rightful part.

One soldier stood up against this breach of the warriors’ social contract:

Thou shalt receive nothing of this unless a just lot give it to thee

The soldier then struck the vase, refusing Clovis’ claim to more than his fair share. One year later, as Clovis was inspecting his soldiers’ equipment, he came face to face with that soldier. Criticising the poor condition of his equipment, he threw the soldier’s axe to the ground. As the soldier bent down to pick it up, Clovis raised his own axe and split the soldier’s skull, uttering the words:

This is what you did to the vase at Soissons.

This story has an important lesson: a social contract has no binding force in itself. It can hold only if the balance of power makes it a stable compromise. For an egalitarian social contract to persist, bargaining power needs to be fairly equally distributed among members of society. If power is concentrated in some hands, anyone who tries to uphold an egalitarian social contract, like the Frankish soldier, risks being sanctioned by the exercise of raw power.

The consequence is important: when bargaining power is unequally distributed in society, the prevailing social contract cannot be egalitarian. In fact, what is seen as fair and just in an unequal society will typically justify existing inequalities in status and resources. I discuss here why, and how this helps explain differences in views about fairness across times and places.

Fairness and bargaining power in society

Many societies have had fairness norms at odds with those in modern liberal democracies.

Countries around the world have very different fairness norms. For instance, Iranian law states that in cases of wrongful death or bodily injury, a woman is worth only half a man for the financial compensation paid to the victim or the victim’s family (diya, or “blood money”).2

It is not just geography. A given country typically had very different fairness norms in the past from those it has now. Modern societies profess to give equal respect to all citizens, whatever their origin or life situation. This was not the case throughout history. For instance, in the Roman Republic, citizens could vote in the centuriate assembly, but it explicitly gave more influence to the rich than to the poor. Ancient authors presented this as a fair constitutional arrangement: those with greater property, greater military obligations, and a greater stake in the commonwealth were granted greater political weight. Cicero defended the principle explicitly in De re publica:

[T]hose who had the greatest power in voting were those with the greatest interest that the civil community was in the best form. — Cicero, De re publica 2.40

As the novelist L. P. Hartley famously wrote:

The past is a foreign country; they do things differently there. — Hartley (1953)

Why are there so many different fairness norms across places and times? Why do some of them look so unfair to our modern eyes? Are citizens of modern liberal democracies morally superior? Were ancient people either evil or ignorant? Have we attained a better knowledge and understanding of what is “right”, leading to “moral progress” and to the norms we have now?

The implicit, and sometimes explicit, answer given in modern liberal democracies often seems to be “yes”. From the perspective of this Substack, one should always be sceptical of stories that too quickly dismiss seemingly puzzling behaviour as irrational or mistaken. Instead, one should look for possible good reasons behind such behaviour. There are good reasons to think that if some societies have less egalitarian fairness norms, it is not because their citizens were less enlightened than us, but because fairness norms have to solve different problems in more equal and more unequal societies.

Fairness as a reflection of the long-run balance of bargaining power

Game theorist Ken Binmore has shown how fairness norms can be explained as generic solutions to the recurring problems of bargaining over the allocation of rights and duties in social situations.

An important and necessary implication of this approach is that, as solutions to bargaining problems, fairness norms will, in the long run, reflect the structure of bargaining power in a given society. They will be egalitarian in equal societies and inegalitarian in unequal societies. This answer can seem somewhat shocking at first. Would fairness, which is about what is just and right, really be a reflection of power balances?

However, once we understand fairness norms as solutions to social bargaining problems, Binmore’s conclusion is unavoidable. Fairness does not come from external principles that exist in the universe and bind us. Fairness norms are human solutions. They are the rules for playing the game of life properly and agreeing on everyone’s due in society.

Fairness norms allow people to avoid having to haggle all the time in real life and to agree quickly on how to split rewards and costs in their interactions. But they can work in practice only if they are not too far from what people could actually claim using their raw bargaining power in a hard-fought bargaining process. Only norms that are broadly in line with actual differences in bargaining power will tend to last. Those that are not will be eroded as people with greater bargaining power than the existing norm recognises progressively chip away at them.

A soldier might ask the king to abide by a principle of equality, but if the king can kill that soldier later on, the equality norm will not persist. Soldiers will rationally prefer to let the king have his way rather than risk their lives. New stories and norms can then arise, establishing as “fair” the share that the king can claim, given his greater bargaining power.

Social views on justice tend to justify social inequalities. The shocking truth is that it is not a bug; it is a feature. It is not an error made by those expressing these views. It is the way fairness norms work. They are social agreements, which, in the long run reflect the distribution of power in society.

Fairness in unequal societies: an ideological veneer?

There is a parallel here with one of Karl Marx’s key ideas. He got many things wrong about how society works, but he got at least one thing right: the social contract of a society reflects the distribution of power in that society.

The ruling ideas are nothing more than the ideal expression of the dominant material relationships, the dominant material relationships grasped as ideas. — Marx and Engels (1845–46)

In Marxist circles, this point is typically presented as a morally charged criticism rather than as a factual observation. It is easy to jump from Marx’s observation to the conclusion that dominant ideas about what is just and fair in society are a form of intellectual scam, an ideological veil used to convince dominated classes to accept their lower status. As the French writer La Fontaine put it in his retelling of Aesop’s fable of the wolf and the lamb:

The powerful always find a justification for what they want to do.

In a similar vein, the left-wing economist Thomas Piketty wrote a 1,100-page book about ideology and inequality, and his discussion of the role of ideology was largely limited to the claim that it acts as a legitimisation device:

Every human society must justify its inequalities: unless reasons for them are found, the whole political and social edifice stands in danger of collapse. Every epoch therefore develops a range of contradictory discourses and ideologies for the purpose of legitimizing the inequality that already exists or that people believe should exist. — Piketty (2020)3

From that perspective, dominated classes are misled into accepting their own servitude. They fall prey to what Marxists have called “false consciousness”, defending ideas that are actually opposed to their class interest. The justifying ideology is useful because it prevents people from seeing their real interests and rebelling. It is a kind of glue that holds together a society with unequal relationships.4

One thinker who helped popularise this view and gave it important concepts is the French sociologist Pierre Bourdieu, who proposed the term “doxic acceptance” to describe how dominated groups adopt conceptual frames shaped by dominant groups, thereby justifying domination. In Bourdieu’s famously convoluted style:

When the dominated apply to what dominates them schemes that are the product of domination, or, to put it another way, when their thoughts and perceptions are structured in accordance with the very structures of the relation of domination that is imposed on them, their acts of cognition are, inevitably, acts of recognition, submission. — Bourdieu (2001)

The idea that ruling ideas about fairness are an ideological veneer used to placate dominated groups is, however, unsatisfactory. If what matters are actual relations of power, why do powerful people need justifications in the domain of ideas? What role do these justifications play?

Another question is “why would dominated classes believe in spurious justification of their situation”? Contrary to Bourdieu’s picture of domination as sustained by the internalisation of dominant categories, people are not simply gullible recipients of ideas that run against their interests. The idea that people accept unequal fairness norms simply because they are socialised into accepting their justifications is therefore suspicious: the psychological evidence points instead to people being overly sceptical and very apt at identifying when their interests are not met in social situations.5

The reality of the balance of power in unequal societies

Prevailing ideas about justice and fairness reflect the formulation of a social contract that is in line with the long-term balance of power between social groups. The notion of “balance” is important here because it is easy to forget that there is a balance of power even when power is unequally distributed.

The notions of dominant and dominated, popular in modern left-wing circles, can be misleading because they suggest a radical all-or-nothing view of power: some have it all, while others have none. One of the most important facts to appreciate about power imbalances is that inequality in power is always graded.

Robert Dahl is the political scientist who proposed the definition of power most commonly accepted by those trying to characterise this notion: Alice has power over Bob if she can get Bob to do what Alice wants.6 One key insight from Dahl’s work on power is that even in unequal relationships, the weaker partner nearly always has residual power because the welfare of the stronger depends on the weaker party’s willingness to cooperate.

In a company, the quality of the work an employee produces for a manager depends on that employee’s willingness not to shirk when the manager is not looking. In a slave society, output depends on enslaved workers’ willingness not to sabotage production to the extent they can. In a patriarchal society, the quality of a husband’s domestic life depends on the care, effort and cooperation of his wife.

Hence, even in unequal relationships, there is a need to negotiate a modus vivendi about who should do what and who should get what. This modus vivendi is not equal, but it nonetheless limits the claims of the most powerful partner, and these limits reflect the balance of bargaining power in the relationship.

All stable human relationships, without exception, represent a compromise of sorts—including the modus vivendi established between the victor and the vanquished after a war. The balance of power may be very unequal in such a relationship. But even a slave is not totally powerless. Only a stupid master does not eventually learn to be sparing with the whip in seeking to persuade a slave to labor efficiently on his behalf. Those who fail to curb their brutal impulses find themselves sitting on a powder keg. — Binmore

Because of the strong principles of equality at the heart of our modern fairness norms, the idea that fairness could regulate highly unequal relationships typically conflicts with our intuitions. But we should beware of intuitions shaped in the cosy environment of liberal democracies, where people vote for their leaders and can say what they think in the street without ending up in jail or worse. The fairly equal distribution of bargaining power in society is an unusual feature in the history of large-scale societies since they appeared around 4000 BC.

Fairness norms across times and places

Looking at what people say in an unequal society about fairness supports Binmore’s view: people will call their masters “fair” to the extent that they use, but do not abuse, their greater power. In that perspective, ruling ideas about fairness in an unequal society do indeed justify the domination of the group or groups in power. But they do not do so against the interests of the dominated classes. They reflect the balance of power in that society.

As a consequence, the unequal social contracts of the past could be seen as fair by those dominated.

In an aristocratic society, even the lower orders commonly accepts that it is right and proper that they be regarded as unworthy. — Binmore (2005)

Since this is alien to our modern perspective, it is useful to look at a range of historical examples.

Fairness in societies with class inequalities

The French Revolution is famous for stating that every man (not woman) is born equal in rights. If this is notable, it is because this view was not the one prevailing throughout history, even among the poorest members of society.

In feudal societies, where warlords with a monopoly on military violence ruled over territorial domains, fairness norms did not grant unlimited power to the local lord. They defined reciprocal, though asymmetric, rights and duties. A serf owed labour, rents and dues, but also held customary rights and had some protection against arbitrary treatment. The lord enjoyed superior status, but was expected to protect dependants and uphold local order. In such a setting, a “fair” relationship was not an equal one, but one that respected the recognised terms of this asymmetric bargain.

Feudal Europe, 14th century. A rare document from 1341 in England stated the rights and duties of the lord of the manor of Alrewas in Staffordshire. The tenants submitted the document to the lord for him to sign and described him as a “good lord”:

Sir Philip de Somerville, we are told, had “governed the manor amicably and graciously these last forty years and more”, and has “more knowledge than anyone else living of the certainty of the customs of the tenants”. So, when disagreements had arisen on the manor — between the tenants and Sir Philip’s officials, we should note, not between them and Sir Philip himself — it was to him, “trusting in his faith as a good lord’” that they appealed.7

Tokugawa Japan, 17th-19th century. In his essay Benevolent Lords and Honorable Peasants, historian Irwin Scheiner describes how peasants did not reject hierarchy as such. Instead, when they protested, they framed their demands in terms of the lord’s duty to be benevolent and their own duty to remain honourable. Later historians summarise this as a political language in which the legitimacy of unequal order was accepted, provided rulers fulfilled the moral obligations of lordship. The historian Luke Roberts, who worked on the petition box in eighteenth-century Tosa, also pointed to petitions by peasants praising a “benevolent lord”.

E. P. Thompson’s idea of a “moral economy” captures this logic well. He observed that eighteenth-century crowds did not typically protest in the name of political equality or market freedom. They defended customary expectations about fair prices, subsistence, and the obligations of the powerful. Their protests were therefore not outside the existing moral order. They appealed to its recognised limits.

It is possible to detect in almost every eighteenth-century crowd action some legitimizing notion. By the notion of legitimation I mean that the men and women in the crowd were informed by the belief that they were defending traditional rights or customs; and, in general, that they were supported by the wider consensus of the community. On occasion this popular consensus was endorsed by some measure of licence afforded by the authorities. — Thompson (1971)

The same point is central to Barrington Moore’s historical sociology of injustice. Moore argues that moral outrage often arises when relations of mutual obligation are violated. Such obligations do not imply equality. They can coexist with hierarchy and authority, but they still require some form of reciprocity:

[A] conception of what social relationships ought to be. It is a conception that by no means excludes hierarchy and authority, where exceptional qualities and defects can be the source of enormous admiration and awe. At the same time, it is one where services and favors, trust and affection, in the course of mutual exchanges, are ideally expected to find some rough balancing out. — Moore (1978)

Looking again at the Soissons vase story, the type of norms that will arise when the king can kill a rebellious soldier with few consequences will not be egalitarian. But a king cannot claim everything either. Soldiers can rebel, desert, or support a rival. The king therefore needs to leave them enough to make loyalty preferable to defection.

Fairness in societies with slaves

The historical record of the words of slaves or former slaves is scarce, but here again we observe that they could consider some of their masters fair and good.

In ancient Rome, little writing from slaves has survived, but one type of writing persisted in Roman epitaphs where dependants and freedmen could speak of superiors in warmly deferential terms. For example, one epitaph read:

To my old master and mistress I was an obedient servant […] for they gave me freedom.8

In the United States, the record of former slaves contains many instances of them describing their masters as “good”. For instance, one former slave, born around the Civil War, made the following statement when he was interviewed at 77:

Our folks’ master was William E. Johnson. Oh Lord, they was just as good to us as could be to be under slavery. […] Oh Lord, my parents sho was well treated. Yes ma’m. If he had a overseer, he wouldn’t low him to whip the folks. He’d say, “Just leave em till I come home.” Then he’d give em a light breshin’.

In Islamic societies, the written record from enslaved people is also scarce, but here again we find evidence that slavery was not experienced simply as arbitrary domination. It was also governed by local expectations about what masters could legitimately demand and what they owed in return. In a collection of testimonies from the Gulf, Sanqur bin Abdul Khair, a Swahili slave in Muscat, made this contrast very clearly in 1931:

I was well treated by him. When my master died about 8 years ago, I was transferred as an inheritance of the estate to his son Abdullah bin Shaikh Sultan Al Naimi. The latter was ill treating me. He never provided me with sufficient food and clothes. If I ever happened to ask him for a little money which is earned by me and taken by him, he used to beat me severely. Thereupon I seized the opportunity and ran away from Braimi.9

This statement illustrates how domination could be judged excessive within the moral terms of an unequal order. Sanqur does not present his escape as an abstract rejection of slavery, but as a response to a master who failed to meet the recognised obligations of his position: providing sufficient food and clothing, allowing some access to earnings, and refraining from severe violence. Once these limits were breached, the relationship became unjust even by the standards of that unequal social world. Sanqur then seized an opportunity to run away.10 11

Fairness in patriarchal societies

Our modern societies are characterised by equality in rights between men and women, but historically inequality in status, rights and social worth was frequently not only enshrined in law, but accepted by both men and women.

It is particularly noticeable that women who historically promoted improvements in women’s position often argued within an acceptance of patriarchal society, giving men and women different places in social life.

In the USA in 1819, Emma Willard, one of the major early advocates of women’s education in the country, argued in favour of women’s access to education as a way to help them better fulfil their roles as mothers and wives:

Domestic Instruction should be considered important in a female [school]. It is the duty of our sex to regulate the internal concerns of every family; and unless they be properly qualified to discharge this duty, whatever may be their literary or ornamental attainments, they, cannot be expected to make either good wives, good mothers, or good mistresses of families: and if they are none of these, they must be bad members of society; for it is by promoting or destroying the comfort and prosperity of their own families, that females serve or injure the community. — Willard

In Victorian England in 1855, Caroline Norton, an early feminist who campaigned to reform women’s legal position in marriage, explicitly affirmed male superiority. In A Letter to the Queen, she wrote:

The natural position of woman is inferiority to man. Amen! That is a thing of God’s appointing, not of man’s devising. - Norton

Even Queen Victoria, the female head of state at the time, wrote in opposition to women’s rights in a private letter to Prime Minister Gladstone in 1870.

Let woman be what God intended; a helpmate for a man—but with totally different duties and vocations.

This is by no means a pattern specific to Western history.

In modern Iran, women face harsh repression to comply with strict rules constraining their behaviour: compulsory veiling, no dancing, no singing, and so on. A morality police enforces these rules, often violently, for instance by arresting women who reveal too much hair under their veil. This police force includes many women. Typical patrols are mixed male-female units.

The anachronistic nature of modern visions of fairness

If people in the past accepted very unequal social organisations as fair, it is because, given the balance of power they faced, it often made sense to accept unequal norms and create a workable modus vivendi.

It would have made little sense for a 13th-century English serf to assess the fairness of his world by asking whether it satisfied democratic equality, since such an order was not within the feasible set of political possibilities available to him. Monty Python and the Holy Grail famously features a peasant rejecting the authority of King Arthur by mocking the myth that the Lady of the Lake granted him his authority:

Listen, women lying in ponds distributing swords is no basis for a system of government. Supreme executive power derives from a mandate from the masses, not from some farcical aquatic ceremony.

The outcome of the encounter is telling: the king ends up beating him. In a world where feudal lords controlled organised violence, serfs had no interest in demanding an unattainable political equality. What they could more realistically demand was that lords respect the customary limits of their authority.

Understanding fairness without moral absolutes

Not a justification of past inequality

Listing statements by past slaves who described some masters as “good” or “fair” is in no way saying that slavery is acceptable relative to our modern fairness norms; it is to stress that our modern norms do not reflect how people in such unequal societies assessed what was “good” and “bad” in the same way we do now.12

Nonetheless, it might seem like saying “see? some of them were happy”. This is not the point I am making. To be clear, the level of inequality in past feudal, patriarchal and slave-holding societies is far outside of what is deemed acceptable nowadays. Our moral preferences, including mine, have been shaped in liberal democracies and strongly reject the models of these past societies.

But one should not shy away from the implications of a naturalistic take on fairness. If fairness regulates a modus vivendi between dominant and dominated in unequal societies, it is indeed the case that, in the past, those dominated who experienced overlords and masters who respected this modus vivendi might have considered these as “fair”, “good”, “charitable”, and so on. Fairness norms regulated the rules of an unequal game, and respecting these rules was seen as fair. This is a factual statement, and we have ample records of this having been the case.

More importantly, we have to recognise that there is nothing more to fairness than this. We might have strong moral intuitions that these social relations were unjust and that we should condemn them. But we condemn these societies from within our modern liberal-democratic outlook. This is different from saying that people in those societies were making a moral error in the way they might have been making a factual error about medicine or astronomy. If fairness norms are social rules, they are not external principles given by God, written into the fabric of the universe, or dictated by pure reason. They are human conventions that emerge in particular social worlds. This does not make them arbitrary or meaningless. It means that fairness has to be understood relative to the social order in which it operates.

Using the rules of a board game as a metaphor is useful. The game of chess did not always have its current rules. The modern game emerged gradually, especially between about the 15th and 19th centuries. For instance, the queen and the bishop were initially much weaker. I am totally entitled today to say that I dislike these past rules of chess and that our modern version of the game is better according to my taste or because of the quality of gameplay it generates. But it would not make sense for me to argue that people were playing chess “wrong” in the past, that they were making a mistake because they did not follow the “proper” rules of chess we now use. People were playing with different rules, and that’s it.

The evolution of moral norms

Our tendency to judge past actions as “wrong” based on our modern moral standards is often associated with a sense of moral superiority relative to our ancestors. We should be sceptical about such feelings. Who says that if we had lived in the past, we would have been morally outstanding and opposed the then-accepted social contract? Can we claim to be outstanding now, judging our present moral contract on the basis of deep knowledge of absolute moral principles? Or are our views mostly shaped by where and when we were born? Should we morally condemn the philosopher Aristotle because he thought slavery was a natural order of society? Binmore has the courage to adopt a modest stance here:

If I had been brought up in classical Greece, I would no doubt have joined Aristotle in regarding the enslavement of barbarians as entirely right and proper. — Binmore (2005)

This exact same point was made, in a provocative way, by comic Ricky Gervais, who almost seems to have read Binmore on that matter:

I am anti-racist, of course I am. But I am willing to admit that if I had been born 300 years ago and I was white and wealthy, I’d have probably have owned slaves. I’d have been nice to him […] I’d be the best slave owner […] If one of them got a bit too familiar I would have to punish him […] My point is morality evolves. It moves with the times, like everything else.13

Many dominant accounts of morality assume a Newtonian picture of fairness: one universal moral frame, valid for all societies at all times. Cicero expressed this view when he wrote:

There will not be different laws at Rome and at Athens, or different laws now and in the future; but one eternal and unchangeable law will be valid for all nations and all times. — Cicero

The irony is that Cicero could affirm this eternal law while defending a Roman order that gave greater political weight to the wealthy and rested on slavery, both of which are far outside what modern liberal democracies regard as acceptable.

The naturalistic perspective developed here departs from this Newtonian picture. It does not treat morality as absolute. It is closer to a relativistic picture: there is no single external clock of fairness against which all societies can be measured. This does not mean that fairness is meaningless, that rules of fairness are empty words, or that they could take any form. Fairness norms everywhere respond to the same basic problem: how to sustain cooperation despite conflicts over rights, duties, costs and benefits. But the solution to that problem depends on the local social world, especially on the balance of bargaining power. Fairness norms can therefore be understood only relative to the society in which they operate. We can still have a general theory of fairness, but without believing in universally true principles of fairness existing outside human societies. There are only humans, their conflicts, their need to cooperate, and the rules they adopt to regulate their interactions in particular social settings.

We can still favour some social contracts

The absence of universal and absolute moral principles does not mean that we cannot defend some forms of social organisation as better than others. Social contracts can be better or worse in at least two ways. Some are better at organising cooperation and producing the benefits of shared social life. Others are more stable because they reflect a bargain in line with the relative power of different groups. Both features matter. A social contract that fails to produce enough benefits will disappoint its members. A social contract that fails to reflect the balance of power will generate resentment, costly conflict, and pressure for change.

In our modern societies, where power is relatively widely distributed, liberal-democratic principles have something important in their favour: their egalitarian nature. As I have written in a previous post:

Democratic institutions can be understood as a form of social equilibrium for organising both cooperation and conflict in society. Among the possible political arrangements, democracy is likely to result in better outcomes for most people.

Past societies, and present ones with highly unequal fairness norms, are not composed of people who differ fundamentally from us in cognitive capacity or moral sense. They have not upheld unequal norms because they are, or were, morally inferior or less enlightened. They have simply played an equilibrium of the game of life in their own societies, given the structure of the game they faced.

From this perspective, the relative egalitarianism of our own fairness norms reflects the fairly equal distribution of bargaining power in our societies. Where bargaining power is more evenly distributed, more symmetric norms tend to emerge. But this should not lead us to conclude that equality is the essence of fairness.

For modern societies like ours, this view implies a simple question: what rules of fairness can form the basis of a workable and efficient social contract? Such a contract must make society work for its members and allocate rights and duties in a way that is broadly acceptable, thereby minimising conflict.

The observation that, in the long run, fairness norms track the balance of power in society paves the way to another question. How do fairness norms change over time? What are the roles of revolutions and popular movements that have seemingly transformed the course of history and led to greater equality in modern liberal democracies? This is the topic of my next post.

References

Alpers, E.A. and Hopper, M.S. (2017) ‘Speaking for themselves? Understanding African freed slave testimonies from the Western Indian Ocean, 1850s–1930s’, Journal of Indian Ocean World Studies, 1, pp. 60–88.

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Birrell, J. (2014) ‘Manorial custumals reconsidered’, Past & Present, 224(1), pp. 3–37.

Bourdieu, P. (2001) Masculine Domination. Translated by R. Nice. Cambridge: Polity Press.

Cicero, M.T. (c.54–51 BCE) De re publica. Translated in Keyes, C.W. (1928) On the Republic. On the Laws. Loeb Classical Library 213. Cambridge, MA: Harvard University Press.

Dahl, R.A. (1957) ‘The concept of power’, Behavioral Science, 2(3), pp. 201–215.

Gardner, J.F. (2011) ‘Slavery and Roman law’, in Bradley, K. and Cartledge, P. (eds.) The Cambridge World History of Slavery, Volume 1: The Ancient Mediterranean World. Cambridge: Cambridge University Press, pp. 414–437.

Gregory of Tours (1916) History of the Franks. Selections translated with notes by E. Brehaut. New York: Columbia University Press.

Hartley, L.P. (1953) The Go-Between. London: Hamish Hamilton.

La Fontaine, J. de (1668) ‘Le loup et l’agneau’, in Fables choisies, mises en vers. Paris: Claude Barbin.

Marx, K. and Engels, F. (1845–46) The German Ideology. Moscow: Progress Publishers.

Mercier, H. (2020) Not Born Yesterday: The Science of Who We Trust and What We Believe. Princeton: Princeton University Press.

Moore, B. Jr. (1978) Injustice: The Social Bases of Obedience and Revolt. White Plains, NY: M.E. Sharpe.

Norton, C. (1855) A Letter to the Queen on Lord Chancellor Cranworth’s Marriage and Divorce Bill. London: Longman, Brown, Green, and Longmans.

Piketty, T. (2020) Capital and Ideology. Translated by A. Goldhammer. Cambridge, MA: Belknap Press of Harvard University Press.

Scott, J.C. (1990) Domination and the Arts of Resistance: Hidden Transcripts. New Haven, CT: Yale University Press.

Thompson, E.P. (1971) ‘The Moral Economy of the English Crowd in the Eighteenth Century’, Past & Present, 50, pp. 76–136.

Willard, E. (1819) An Address to the Public, Particularly to the Members of the Legislature of New-York, Proposing a Plan for Improving Female Education. Middlebury, VT: J.W. Copeland.

In 494 BCE, in the early days of ancient Rome, its common citizens, the plebeians, withdrew from the city in an early form of political strike. They were demanding debt relief and protection against domination by the patricians, Rome’s hereditary aristocratic elite. When the patricians asked them to return, Sicinius, one of the leaders of the plebeians, stressed that exit from a society whose social contract no longer worked for them remained one of their options:

As for us, we shall be content to regard as our country any land, whatever it be, in which we may enjoy our liberty.1

The “Secession of the Plebs” illustrates how a social group unhappy with a social contract might always decide to stop playing ball and withdraw itself from social cooperation if it deems the social contract unacceptable. For society to work, social groups need to agree on how to organise society, with its distribution of duties and rewards.

So how do people agree on what an acceptable society is? Let’s face it, answering this question seems pretty much intractable. Political debates are filled with many ideological positions that seem irreconcilable and sometimes hard to compare. There are questions of individual rights, the allocation of economic resources, the efficiency of economic policy, and so on. How can we know who is right?

We can provide clear answers to these questions by first setting aside the noise of everyday political discourse and going to the crux of social justice: the bargaining between different social groups over a social contract that specifies how to share the benefits from social cooperation.

The key question of justice: Equality or maximum happiness?

Let’s consider a society in which there are two groups of citizens, like the rich and the poor. Let’s assume they have full information about natural and social laws so that they know how each possible social contract will work. This simple setting is useful to get clear conclusions about what social contract would be chosen. This is the setting used by game theorist Ken Binmore in Game Theory and the Social Contract (1994, 1998).

Since society members have full information about how different social contracts would work and impact them, their problem is only to agree on which social contract to adopt, given the benefits each social contract would provide to each group.

Let’s consider Alice and Bob as the representatives of the two social groups in society. For example, Alice might be the representative of the patricians and Bob of the plebs. They have to decide how to agree on a social contract. If one of them secedes and stops playing ball, the social contract breaks down.

Social contracts specify all the rules to follow in society. Some organisations of society are more likely to lead to wealth creation. We have, for instance, learned from history that communist societies tend to end up poorer than capitalist ones. But social contracts also lead to a distribution of resources. Communist societies tend to be more equal than capitalist societies.

When choosing between efficient social contracts, Alice and Bob would have to consider two key aspects of their consequences: the total amount of wealth produced and how this wealth is split between the different members of society.

Let’s consider Alice and Bob working together in a pizza shop. Bob cooks, and Alice is at the till. Let’s assume that Alice and Bob have organised their pizzeria efficiently to make the most pizza possible during a day. The remaining question will be how to split the gains from the pizzas. To make the question interesting, we will assume that Bob’s motivation depends on his share of income. The more money he gets for each pizza, the harder he works to cook pizzas. We will also assume they are all sold.

This case implies a classic trade-off: sometimes making the pie as large as possible does not imply making the split of the pie equal. Arguments in favour of billionaires often rely on the idea that people like Musk, Zuckerberg, and Bezos got very rich because they created a high level of wealth and quality of life for other people in society. This argument suggests that the pie in our current society might be unevenly split, but that it is much larger than in a system where Musk, Zuckerberg and Bezos would have had no prospect of getting rich and therefore not had the same motivation to engage in their ventures.

How should we choose between social contracts that might both influence the total size of the pie and how the pie is split? Political philosophy proposes two different answers. Utilitarians like Bentham and Mill argued that we should aim for the pie that is as large as possible.

Utilitarianism might lead to some conclusions most people find unappealing. For instance, should we be willing to sacrifice some individuals for the benefit of a larger number? What if Alice gets barely anything? Is it OK simply because the pie, mostly enjoyed by Bob, is the greatest?

John Rawls wrote his classic book A Theory of Justice partly as a rebuttal to utilitarianism.2 He famously argued that we should judge whether a society is fair by putting ourselves behind a veil of ignorance, as if we were outside society, ignoring which positions we have in society, such as gender, age, profession, talent, health, and wealth. In this position, we would be forced to imagine what it would be like to possibly be anybody in society.

Rawls argues that, under the veil of ignorance, we would not simply choose the society with the largest pie. We would also care about how the pie is divided, because we might end up with the smallest slice. For this reason, he argues that a just society should maximise the position of those who receive the least. This is the difference principle.

If we were comparing different levels of wealth inequality, Rawls would tell us to choose the society that makes the poorest person as rich as possible. Rawls is egalitarian, and he argues that a just society would be organised to benefit its least well-off members.

It is important to stress that his view does not prohibit all inequality. It allows inequalities that improve the position of the worst-off, for example, if the prospect of becoming rich encourages entrepreneurship and innovation that also raise living standards at the bottom.3

Economists have criticised this conclusion because it is equivalent to having an extreme degree of risk aversion under the veil of ignorance.4 Imagine comparing two societies with identical political rights but different levels of inequality. In the first, everybody lives in mediocre conditions. In the second, 99.9 per cent of people live extremely well, but 0.1 per cent are slightly worse off than in the first society.

Because Rawls’ principle does not allow trade-offs against the position of the worst-off, it would rank the first society above the second, even though almost everyone is much better off in the second society and only a tiny minority is somewhat worse off. It is far from obvious that people would actually choose that way. Experimental studies in which participants are asked to reason behind a veil of ignorance find concern for the least well-off members of society, but not in the uncompromising form of the difference principle.5

Around the same time that Rawls was writing, the economist John Harsanyi independently developed a similar veil-of-ignorance argument. But Harsanyi applied standard decision theory: behind the veil, people should choose the society that gives them the highest expected payoff across the different social positions they might occupy.

In Harsanyi’s view, choosing between societies is like choosing between lottery tickets with equal chances of placing you in any social position. Rational people behind the veil would choose the lottery ticket with the highest expected payoff. In the example above, this approach would lead to choosing the second society, because the average satisfaction there is higher. That is the utilitarian solution.

So who is right, Harsanyi or Rawls? Should we be utilitarian or egalitarian? This is one of the central questions of modern political philosophy. As Marc Fleurbaey, Maurice Salles, and John Weymark put it:

The opposition between utilitarianism and liberal egalitarianism has triggered the most important developments in political philosophy in the twentieth century. — Fleurbaey, Salles and Weymark (2010)

Binmore’s answer: It depends on whether social contracts are binding

Binmore’s novel answer is that, at the level of society, social contracts must be self-enforcing. In the Game of Life, there is no external enforcer standing above society in the way that judges and police officers stand above ordinary contracts within it. If a dictator seized power and overturned existing political and civil rights, there would be no external authority able to simply restore the previous social contract.

When a contract is not binding, the only way for it to work is for everybody to be sufficiently happy with it, in the sense that they do not expect to do better by renegotiating or withdrawing from cooperation. The plebeians have to agree to work in Rome for the Roman social contract to actually function. If they can credibly do better by withdrawing from cooperation, the old contract cannot simply be imposed as if nothing had changed. When people have equal bargaining power, this pushes towards the egalitarian solution, because with this solution, those who are the worst off are already in the best situation possible.6

However, while the social contract itself cannot be binding and needs to be self-enforcing, smaller games can have binding agreements, even without external enforcers, if the structure of incentives makes deviations too costly. When people interact repeatedly, they can adopt rules of cooperation that are binding because deviating would lead to sanctions.7 It is therefore possible that, while at the level of the Game of Life the social contract is egalitarian, in smaller games, the social contracts can be utilitarian.

The implications of Binmore’s insight for modern political debates

Binmore is not shy about commenting on current political debates, but his theory is nonetheless primarily abstract and generic. Here, I leverage his general insights to look at how they can help us understand modern political debates. To be clear, these are my own interpretations.

When utilitarianism can prevail

Utilitarianism can prevail when a given decision, with its gains and losses, is part of a larger game where everybody can win overall. Each decision can then be made with a utilitarian outlook because the agreement in every single situation is bound by the larger cooperation, which benefits all, so nobody has an interest in withdrawing from the larger game.

Over the long run, utilitarianism is best understood as a mutual advantage bargain: it maximises the social gains in the short run as a way to benefit all in the long run.

Consider how we tend to judge whether a public policy is good or not. On this question, economists adopt by default a utilitarian approach.8 Where should the next school be built? The next airport? The next centre for drug addicts? In these kinds of decisions, the government assesses the social costs and benefits and often claims to choose the solutions with the greatest social benefits. This approach is utilitarian. In the distribution of resources and costs, the government tries to maximise the social utility coming from its decisions.9

The whole discipline of cost-benefit analysis is based on this idea. People who say that “you can’t put a price on life” are often unaware that this is what their government is doing on a daily basis. In public health, projects and policies are routinely evaluated for the number of lives per dollar they save. In England, the NHS explicitly uses the threshold of £25,000–£35,000 per year of life (in good health) saved as a way to decide whether a health investment is cost-effective or not. Similar thresholds are used in a more or less formal way in the USA (by the Institute for Clinical and Economic Review) and in Australia (by the Pharmaceutical Benefits Advisory Committee). The World Health Organisation itself has recommended in the past a threshold of one to three times GDP per capita per year of life saved.10 Such practices are meant to optimally allocate limited financial resources where their impact is going to be greatest in terms of lives saved, regardless of who will be saved.11

Why would people agree with such a utilitarian approach? The answer is that as long as citizens expect to face a large number of situations where they will sometimes be advantaged by a decision and sometimes disadvantaged, a utilitarian approach will, over time, be the most beneficial to all.

Many utilitarian thinkers make the mistake of thinking that this approach is good because it maximises something desirable like “years of life”.12 But that is misguided. Society is composed of individuals, and individuals care about their own lives, not about an abstract sum at the social level. In Binmore’s perspective, a utilitarian approach works as a socially accepted practice, not because the sum of benefits is a mystical “collective good” that has to be pursued for its own sake. It is because utilitarian rules can work as a mutual advantage arrangement: they maximise expected gains for all over time. If Alice and Bob do not know when and how public policy decisions might benefit them, they can support a utilitarian approach, since, in expectation, they will benefit most from it.

Because the social support for utilitarian approaches is based on long-term gains, it is contingent on the social expectations that people have similar chances to benefit or be harmed by the decisions made. If these expectations break down, the support for utilitarian approaches breaks down as well.

When utilitarianism will fail, and egalitarianism will be preferred

Utilitarian solutions will be rejected when the gains and losses in one situation are too lopsided. The expectations that the utilitarian solution will be beneficial in the long run will break down if it recommends losses so large for a group in one situation that it is unlikely this group will recover these losses in the long run.

If a factory closes because of a free trade agreement, its workers might never recover financially. Even if it benefits the country as a whole, they might resist the utilitarian justification. The workers losing their jobs will not care about being told that it is good for society because, behind the veil of ignorance, they could have been someone who benefits from the factory closure. They might therefore request special compensation to make the social contract acceptable.

Utilitarian solutions will be rejected when the same people tend to win or lose over time because of their individual characteristics. Any utilitarian argument will fail to be sustainable if it leads to an uneven distribution of costs and benefits between different people over time. Hence, differences in outcomes based on sex and ethnicity will often lead to egalitarian considerations. Alice will not care about the suggestion that a social inequality based on sex is fine because she could have been a man, and that inequality would have been chosen under the veil of ignorance. She is not a man and won’t be for the rest of her life. So whenever she has equal bargaining power, she will reject social contracts that give different outcomes to men and women “in the interest of society”.

Understanding the deep underlying logic of political debates

Binmore’s simple social setting—two groups who have complete information about each other and the implications of different social contracts—allows him to leverage the formal tools of game theory to identify the crux of fairness discussions as being about bargaining to select one social contract. He acknowledges that his simplifications would need to be relaxed in later work.13 Let’s do this and look at the deep logic behind political debates in societies.

Allowing for coalitions in society: Democratic institutions as a mutual advantage bargain

In a large society with many citizens, social groups try to improve the social contract in their favour by forming coalitions. In their study of coalitional games, game theorists found an unexpected result: we might actually expect coalitions to be unstable because a winning coalition can always be undone by competing coalitions peeling away some of its key members with better deals.

This result, which makes it hard to predict how coalitions will form in society, might be the necessary substrate upon which democratic institutions rest. At the heart of democracy and its voting rules, there is also a mutual advantage logic over the long run. Why would those in the minority accept the decision of the majority? Because tomorrow they might be themselves in the majority.

The use of majority voting has a utilitarian aspect in the sense that it gives priority to the option preferred by more people.14 A mutual advantage logic provides a justification for the use of this approach to make collective decisions. If you believe that you have the same chance as other people of ending up in the majority and having your way in each election, it makes sense to be for majority decisions because you’ll get your preferred options more often (since more than 50% of the people are in the majority each time).15

This point was made by political scientist Adam Przeworski in his book Democracy and the Market:

Political forces comply with present defeats because they believe that the institutional framework [...] will permit them to advance their interests in the future. — Przeworski (1991)

This belief, that the tables will turn, is actually reasonable given the instability of coalitions. Political scientist William Riker emphasised this game-theoretic insight, arguing that one key aspect of democratic regimes is that they prevent any single majority from holding power indefinitely:

Liberal democracy almost guarantees some circulation of leadership so that great power is usually fleeting and no vested interest lasts forever. - Riker (1982)

A direct consequence of this explanation of the support for democratic institutions is that this support is bound to be linked to the trust that people have that they might benefit from the democratic procedures they are asked to abide by when they lose.

This explains why and when democratic institutions will work more smoothly and when, on the contrary, procedures like majority rule will face resistance. Political scientist Robert Dahl said about it.

The stronger the expectations among the members of a political minority that they will enter into tomorrow's majority, the more acceptable majority rule will be to them, the less they will feel a need for such special guarantees as a minority veto, and the more likely they are to see these as impediments to their own future prospects as participants in a majority government. - Dahl (1989)

It is therefore normal that democratic institutions come with protections for the political minority: those in the minority need to be guaranteed that the current majority will not change the rules of the game to prevent members of the minority from winning in the future.

This perspective helps explain why social divisions can weaken democratic culture. When society is fragmented into rigid coalitions, the mutual advantage logic at the heart of democratic institutions is threatened. Minority groups may doubt that state institutions will work for them when these institutions are controlled by the majority.16 Societies that are ethnically and culturally fragmented tend to have weaker state structures and less willingness to empower central institutions.17 The same logic applies to political polarisation: when each side sees the other as a threat that must be kept from power, democratic norms become more fragile. This concern is central to recent debates about the United States, where political scientists have argued that polarisation can make citizens more willing to excuse violations of democratic norms by leaders from their own side.18

Allowing incomplete information: political debates blend fairness and factual arguments

If society’s members had full information about the facts, there would not be any point in arguing about them. A direct consequence of the fact that information is not complete is that political debates articulate discussions about fairness around debates about facts:19

We move from a bargaining situation where all the information is clear, and people only have to agree on a social contract, to a situation where there are uncertainties about the possible consequences of social contracts. To get their preferred social contract, people can then both argue that they deserve it based on a shared understanding of the facts or argue for an understanding of the facts that is favourable to their claims about justice. This is why a lot of the political arguments are about disagreements about the effect of policies. Right-wing politicians might argue that wealth trickles down, while left-wing politicians argue that it is all just a trick.

The bargaining perspective helps us see that there are two things going on at the same time in such debates. Because policies based on an utterly wrong understanding of the world would be harmful, people have some interest in these debates delivering conclusions that are closer to the truth than to falsehood.

But some factual conclusions might be better for some and worse for others. High-income earners may prefer the conclusion that taxes hurt the economy. Trade unions might prefer the conclusion that higher minimum wages are good for growth. Each side has an incentive to back and to believe the “truth” that is most convenient to them. Political debates take the form of courtroom debates: the apportionment of credit and blame is articulated on disputes about the facts.

This view provides a simple explanation of what political ideologies are: competing visions for how to adapt the social contract. They articulate normative principles (what is fair) with factual claims (what works) to support a vision about how society should be organised.20 Once we understand that bargaining about the social contract is what’s behind political debates’ ideas and factual claims, the nature of these debates and, in particular, the singular difficulty of agreeing on facts, becomes easy to understand. It is not a bug, it is a feature.

Intellectual and political discussions about what a just society is are complex. Libraries could be filled with philosophical discussions on what a just society is. Binmore’s approach implies that the question about justice in society is, in a sense, much simpler to understand than lots of the philosophical debates might suggest.

There is no absolute principle of justice to be found as if it were written in the fabric of the universe. All the questions about justice in society arise from the bargaining between the different social groups on how to organise society.

In that perspective, beyond the noise and clamour of daily political debates where politicians argue with ideas and facts to gain support for their coalition, there is a fundamental question about the allocation of duties and rewards in society: who should get what? In a society with fairly equal bargaining power, one can expect Rawlsian considerations for the least well-off to be present. The plebeians could always secede, and society could not work without them.21

The picture I have presented here leaves an important question unanswered: what happens when bargaining power is not equally distributed? Binmore provides one of the most interesting and thought-provoking answers, which I will discuss in my next post.

References

Acemoglu, D. and Robinson, J.A. (2000) ‘Why did the West extend the franchise? Democracy, inequality, and growth in historical perspective’, Quarterly Journal of Economics, 115(4), pp. 1167–1199.

Aidt, T.S. and Jensen, P.S. (2014) ‘Workers of the world, unite! Franchise extensions and the threat of revolution in Europe, 1820–1938’, European Economic Review, 72, pp. 52–75.

Alesina, A., Baqir, R. and Easterly, W. (1999) ‘Public goods and ethnic divisions’, Quarterly Journal of Economics, 114(4), pp. 1243–1284.

Bentham, J. (1843) ‘Anarchical fallacies’. In: Bowring, J. (ed.) The Works of Jeremy Bentham. Vol. 2. Edinburgh: William Tait.

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. New York: Oxford University Press.

British Medical Association (BMA) (2020) COVID-19 – ethical issues: A guidance note. London: British Medical Association.

Dahl, R.A. (1989) Democracy and Its Critics. New Haven: Yale University Press.

Dionysius of Halicarnassus (1940) Roman Antiquities. Volume III: Books 5–6.48. Translated by E. Cary. Loeb Classical Library 357. Cambridge, MA: Harvard University Press.

Gaertner, W. and Schokkaert, E. (2012) Empirical Social Choice: Questionnaire-Experimental Studies on Distributive Justice. Cambridge: Cambridge University Press.

Genicot, G. and Ray, D. (2003) ‘Group formation in risk-sharing arrangements’, Review of Economic Studies, 70(1), pp. 87–113.

Goldburg, C.B. (1994) ‘The accuracy of game theory predictions for political behaviour: Cumulative voting in Illinois revisited’, Journal of Politics, 56(3), pp. 885–900.

Harsanyi, J.C. (1953) ‘Cardinal utility in welfare economics and in the theory of risk-taking’, Journal of Political Economy, 61(5), pp. 434–435.

Harsanyi, J.C. (1955) ‘Cardinal welfare, individualistic ethics, and interpersonal comparisons of utility’, Journal of Political Economy, 63(4), pp. 309–321.

Harsanyi, J.C. (1975) ‘Can the maximin principle serve as a basis for morality? A critique of John Rawls’s theory’, American Political Science Review, 69(2), pp. 594–606.

Hoppe, H.-H. (2001) Democracy: The God That Failed: The Economics and Politics of Monarchy, Democracy, and Natural Order. New Brunswick, NJ: Transaction Publishers.

Khatib, L. and Gardiner, S. (2015) Lebanon Situation Report. Beirut: Carnegie Middle East Center.

Lalley, S.P. and Weyl, E.G. (2018) ‘Quadratic voting: How mechanism design can radicalize democracy’, AEA Papers and Proceedings, 108, pp. 33–37.

Przeworski, A. (1991) Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. Cambridge: Cambridge University Press.

Riker, W.H. (1982) Liberalism Against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice. San Francisco: W.H. Freeman.

Schildberg-Hörisch, H., 2010. Is the veil of ignorance only a concept about risk? An experiment. Journal of Public Economics, 94(11-12), pp.1062-1066.

Svolik, M.W. (2019) ‘Polarization versus democracy’, Journal of Democracy, 30(3), pp. 20–32.

Vergano, M., Bertolini, G., Giannini, A., Gristina, G.R., Livigni, S., Mistraletti, G., Petrini, F. and Riccioni, L. (2020) ‘SIAARTI recommendations for the allocation of intensive care treatments in exceptional, resource-limited circumstances’, Minerva Anestesiologica, 86(5), pp. 469–472.

13

The reason Binmore did not do these extensions himself is twofold. First, in regard to coalitions, the problem of coalition formation is, in a sense, unresolved in game theory. Game theorists do not have, to the extent they have in other domains, a satisfying conceptual solution to predict when and how coalitions form.

There has been notable progress since Binmore’s publication, but at the same time, coalitional game theory has fallen to a large extent as a topic of interest for economists (most likely because clear solutions to make predictions were elusive). As a consequence, I think that Binmore’s reservations are still valid.

In conclusion to his two volumes, he writes:

[G]ame theorists have proposed many models that are intended to explain coalition formation among rational players, but none can be said to have succeeded in capturing everything that matters. Until further advances have been made on this front, I therefore thought it wise to study only the case in which current theory is adequate—that of a [two-player society which] can usefully be seen as representatives of monolithic blocs.

Second, in regard to incomplete information, game theorists have models to study such situations, but they are much more complex than the case of complete information. One key limitation in such situations is that people do not have an interest in revealing their private information; instead, they may gain by getting others to form wrong beliefs about their private information. Anybody who has gone to a car dealership to try to get the best deal possible will appreciate this fact.

Binmore notes, however, that game theorists are not helpless. We can use game theory’s insights to design institutions that are more likely to be conducive to the correct information being revealed. Economists call this field of research mechanism design.

From Occupy Wall Street to Black Lives Matter, calls for social justice have taken centre stage in political debate in the United States and, more broadly, in Western liberal democracies.

This raises a question: what is “social justice”, and what would a “just” society look like? There are long and complex debates about this. The left-wing thinker Slavoj Žižek once remarked that social justice activists often struggle to articulate what, exactly, a just society would be.

I am, of course, fundamentally anti-capitalist. But let’s not have any illusions here. No. What shocks me is that most of the critics of today’s capitalism feel even embarrassed, that's my experience, when you confront them with a simple question, “Okay, we heard your story . . . protest horrible, big banks depriving us of billions, hundreds, thousands of billions of common people's money. . . . Okay, but what do you really want? What should replace the system?” And then you get one big confusion. — Slavoj Žižek interview

In this post, I argue that the seemingly impenetrable layers of debate about social justice become much easier to make sense of through the contractualist approach of game theorist Ken Binmore.

What is social justice about?

Open almost any textbook on political philosophy, and you will find a long list of questions about justice at the level of society. Justice concerns how income and wealth are distributed, how social relationships are structured, how respect and social status are allocated, and how political power is distributed and constrained. It is not always clear which of these questions matters most.

There is also no shortage of answers. Some theories argue that society must protect individual rights. Others say that happiness is what matters. Others claim that objective outcomes matter more than subjective satisfaction, so what must be allocated justly is resources. Still others add that resources are not enough, because we also need to ask what people can actually do with them.

As you read, you quickly meet a familiar list of dilemmas without clear answers. Should society protect liberty or ensure equality? Should institutions prioritise economic efficiency or fairness through redistribution? Should society focus on improving the situation of the worst-off, or on rewarding those whose effort and talent contribute most to innovation and wealth creation?

If you end up confused, that is not surprising. The paradox of social justice is that we share many intuitions about it without sharing a clear and coherent intellectual framework for thinking about it.

One theory, however, does provide such a unifying framework. In two books published in 1994 and 1998, Ken Binmore makes a kind of Copernican move in political philosophy. Most approaches assume that there is some independent justice “out there”, and that the task of political philosophy is to discover the just answer to questions such as liberty or equality, or efficiency or redistribution. Binmore reverses the perspective. Fairness is not something we discover. It is a social solution to pervasive bargaining problems: who should do what, and who should get what, in society. The result of this bargaining is a social contract, and what we call “fair” is what respects that contract.

Binmore’s starting point is that social interactions are primarily characterised by cooperation. This is easy to overlook because we tend to notice failures of cooperation rather than its constant presence in everyday life. Yet social cooperation—from not running away from a café without paying to having a product made on the other side of the world delivered to your doorstep—generates benefits. The key question is therefore how to divide those benefits, as well as the work required to produce them. An agreement about how to cooperate, the rules needed to sustain cooperation, and the way the gains from cooperation are shared are a social contract.1

It is common to represent this problem with the picture below. By cooperating, Alice and Bob can get benefits. A social contract organises their cooperation and determines both the benefits and their splits. Some social contracts leave some potential benefits untapped: either Alice, or Bob, or both could get more with a better agreement. A pattern of cooperation is efficient when it is not possible to improve the outcome of anybody anymore without making some people worse off. In short, a social contract is efficient when it does not leave any money on the table. There are typically many efficient social contracts.2

One possible way of organising social life would be to haggle all the time. But haggling is costly. It would take time and effort all day, from deciding who speaks first in a meeting to deciding what to watch on television at home.

Because haggling is costly, it makes sense for social groups to develop conventions that smooth such conflicts by aligning expectations around commonly accepted ways of allocating rights and duties. These conventions are fairness norms. They guide us towards what everybody agrees is the “right” thing to do. They are common knowledge: we know them, we know that others know them, and we know that others know that we know them. When we make a “fair” offer, we do so knowing that it fits an accepted norm, and knowing that the other person knows that too.

These fairness norms shape social interactions of all kinds, from how long we can speak in a discussion without sounding overbearing to how much of a pay rise we can reasonably ask for after a run of good performance. They may seem invisible because they are typically voiced explicitly only when they are violated. At a cinema, for example, everybody may quietly queue according to a “first come, first served” norm. The norm is stated out loud only when someone breaks it, and another person says, “We were here first.”

Binmore thinks that our moral sense and our psychology of fairness judgements evolved to solve bargaining problems in the relatively small-scale social environments of tribes and villages. We now use those same cognitive tools when we think about social justice in large societies. Sometimes our intuitions are ill-suited to that scale and can mislead us. Even so, Binmore thinks we can build on that psychological apparatus, and on a better understanding of it, to construct social contracts that work in large-scale societies.

Answers to classical questions about social justice

Individual rights vs social goals

A major question in political philosophy is how strongly individual rights should be protected, and when society may fairly ask individuals to sacrifice their own interests for collective goals. At one end of the spectrum are those who treat individual rights as sacrosanct. At the other end are those who think personal interests should largely efface themselves in the pursuit of a better society. In Binmore’s contractualist and naturalist perspective, both extremes are misguided.

The full subordination of individual interests to collective goals cannot work as a sustainable social contract. The idea that, in a good society, people would fully sacrifice their personal interests to the collective has appeared repeatedly in both religious and political ideologies. Karl Marx famously described a socialist society as one where everyone's endeavours would be to contribute to society and make it just.

From each according to his ability, to each according to his needs! — Marx (1875)

Unfortunately, such a vision is poorly suited to a social contract among human beings. Humans are only imperfectly related genetically and have typically no desire to sacrifice themselves completely for others. They will therefore resist a fully collectivist social contract. As E.O. Wilson famously remarked about communism: “Good ideology. Wrong species.” Communism would be much better suited to a society of clones, or near-clones, like ants.3

Given the imperfect alignment of interests within human society, such a philosophy is bound to fail. As Binmore ironically points out, it would encourage people to exaggerate their needs and understate their abilities, to the detriment of social outcomes.

In the socialist utopia that would ensue after the state had eventually withered away, the rule was to be: from each according to his ability—to each according to his need. Human nature being what it is, such an incentive scheme seems designed to convert the able of a large society into the needy overnight. — Binmore (1998)4

The misleading idea of a single common good. One of the mistakes behind the idea of the importance of social goals is even more fundamental. It is the idea that there is a well-defined common good to pursue. It is important to depart from such a commonly expressed view. There would be a unique common good only if we were clones. In that case, there would be a single efficient outcome under full cooperation, where all get the same payoffs because our interests would be perfectly aligned.5

But we are not clones. Our interests are not perfectly aligned, and humans care about how the benefits and costs are divided among themselves.

To say that there is no single common good is not to deny that society can be improved for everyone. When Alice and Bob move from disagreement to a social contract that gives both of them higher payoffs, they find a mutually beneficial arrangement. But their interests still diverge over which mutually beneficial arrangement to choose. Political debate is therefore partly about finding mutually advantageous improvements, but also, and often mainly, about finding an acceptable compromise among competing advantageous outcomes.

To tap into the benefits of cooperation, some constraining rules are required. The fact that people are unlikely to accept a full sacrifice of their interests does not mean that social contracts can be free of constraints on individuals. Working together requires some rules, some modus operandi, about who does what, who gets what and what the sanctions are for failing to respect these rules. People might be willing to accept such constraining rules because repeated interactions create potential mutual gains that align the interests of the members of a group: they have an incentive to make successful cooperation happen.

The agreed-upon rules of cooperation in a group define the rights of the group members. Rights are the actions allowed and the guarantees offered by the social contract. They are the product of the social contract. There are no mystical “rights” that predate the social contract.

Those who claim that “individual rights” pre-exist the social contract are therefore mixing things up. Of course, all else equal, people prefer fewer constraints and being allowed to do more things. They will generally prefer social contracts with more rights, in that sense. But they will also accept constraints when those constraints improve the outcome for all. Who would want to live in a society where everybody has the “right” to choose on which side of the road to drive?

As a metaphor, consider a group of people deciding the rules of a ball sport. May they carry the ball, as in rugby, or touch it only with their feet, as in football? Someone claiming that players have a pre-existing “right” to carry the ball would be missing the point. The issue is precisely to determine which rules will make the game work. What players will be allowed to do is the result of that decision. There is no prior right that already existed before any decision was made.

The idea of natural rights might seem appealing, but, as Jeremy Bentham famously put it, it is nonsense.

From this perspective, there is no real dilemma between individual rights and collective endeavour, because there are neither rights to defend in the abstract nor collective goals that stand above individuals’ interests. The whole opposition is misleading. A good social contract contains whatever rules are needed to make cooperation work for all, but no more than that.

Liberty vs equality

A related debate asks whether a society should aim primarily at liberty or at equality. This question takes on particular importance in modern societies. There is broad agreement on political rights (e.g. every adult should have the right to vote), but much less agreement on economic rights. Some defend leaving the outcomes of market interactions largely untouched. Others favour regulating markets and redistributing resources so that outcomes are more equal than those generated by market exchange alone.

Deontological defence of freedom and markets. For some free-market thinkers, such as philosopher Robert Nozick, it is a matter of principle that the state has no right to interfere with private property rights and the exchanges that follow from them. In the opening line of his book Anarchy, State, and Utopia, he writes:

Individuals have rights, and there are things no person or group may do to them (without violating their rights). — Nozick

But, as with natural rights more generally, the idea that property rights pre-exist the social contract does not make much sense. Property rights are social conventions backed by institutions. As 19th-century liberal thinker Benjamin Constant put it:

Property, to the extent it is a social convention, falls within the scope of political jurisdiction. — Constant (1815)6

So it is a mistake to claim that property rights cannot be touched. Property rights are an important part of a good social contract, but they are the result of that contract. Without police, courts, and a wider willingness to respect ownership, talk of “property rights” would be little more than talk. In a Wild West setting, property would be respected only to the extent that each owner could defend it by force.

Binmore insists that markets are not “natural”. They are part of the social contract:

[I]t seems obvious that the existence of a well-developed social contract is a precondition for the emergence of a market. — Binmore (1998)

Said simply, markets will work and their outcomes will be respected only if people are happy for it to be the case. Experiments in psychology have shown that in many situations, lay people find the “market price” that matches supply and demand unacceptable.

Market transactions are just one kind of social interaction. They do not enjoy a legitimacy that stands outside the social contract. They are accepted only if they fit within it. Nothing says that all market transactions, of every kind, ought to be accepted by everyone. People bring to markets social and moral preferences that were often shaped long before modern market societies existed.7 A social contract that ignored those preferences and relied blindly on market outcomes would not be sustainable, simply because people would disagree with it.

On the left, people often justify taxing the rich by saying that their wealth was created thanks to social institutions. That is not quite Binmore’s argument. Binmore’s point is simpler. Markets can function as mechanisms for allocating resources and organising production only if enough members of society agree to the social contract in which those markets operate. If market outcomes do not sufficiently benefit the worst-off, nothing compels them to keep supporting that arrangement. For the wealthy to enjoy large fortunes, they need enough poorer citizens not to overturn the system politically by, for instance, voting for a communist party. There are no mystical property rights that morally bind the poor not to expropriate the rich. Remember, it is all about social bargaining. It is about finding an acceptable compromise for all.

To prevent poor people from expropriating them, rich people might want to offer them a social contract that gives them a better deal. This was, in part, one historical function of the welfare state. Elites often supported social insurance not just out of benevolence, but also to reduce the appeal of revolutionary ideologies. Bismarck, the architect of the German welfare state, put it quite clearly:

My idea was to bribe the working classes, or shall I say to win them over, to regard the state as a social institution existing for their sake and interested in their welfare. — Bismarck

In that perspective, markets are neither to be worshipped nor despised. They are valuable because they tend to generate wealth, and because their decentralised character helps avoid the concentration of power associated with a politically administered economy. But their outcomes are not sacred.8

Consequentialist defence of economic freedom. A different argument for markets is simply that they tend to work better and therefore should be left relatively unconstrained. This is the line associated with economist Friedrich Hayek, for instance. Like most economists, I would agree, and I suspect Binmore would too, that Hayek was right about a great deal in his account of how markets coordinate decentralised information and organise production. But, again, like most economists, I would also say that he took too large a leap in assuming their broad superiority in all relevant circumstances. Markets may fail in many ways, and ensuring that they deliver good outcomes for consumers and producers often requires a lot of regulation.9

Defenders of markets are right, however, to reply that the risk of market failure does not by itself justify political intervention. There is no benevolent policy-maker standing above society. Public Choice theory has correctly emphasised that politicians and administrators pursue their own interests, and these are often at odds with those of the broader public. There is therefore no guarantee that state intervention will correct market failures rather than worsen them.

The sensible conclusion is not that markets are always right or that politics is always worse. It is that proposals for reform should be informed by both the strengths and the limitations of markets and representative democracy. In fact, the design of market rules and political institutions should be treated as a problem in mechanism design: how to structure the rules of the game so that, despite divergent interests, the resulting behaviour leads to good outcomes.

David Hume had anticipated this way of thinking:

When there offers, therefore, to our censure and examination, any plan of government, real or imaginary, where the power is distributed among several courts, and several orders of men, we should always consider the separate interest of each court, and each order; and, if we find that, by the skilful division of power, the interest must necessarily, in its operation, concur with the public, we may pronounce that government to be wise and happy. — Hume (1742, Of the independency of parliament)

Efficiency vs fairness

This is one of the most frequent dilemmas mentioned in political philosophy discussions. In his book, Equality and Efficiency. The Big Tradeoff (2015), economist Okun said

The contrasts among American families in living standards and in material wealth reflect a system of rewards and penalties that is intended to encourage effort and channel it into socially productive activity. To the extent that the system succeeds, it generates an efficient economy. But that pursuit of efficiency necessarily creates inequalities. And hence society faces a tradeoff between equality and efficiency. — Okun (2015)

Sometimes reducing inequality does indeed come at a cost. If inequality is generated by market incentives that encourage work, investment, or innovation, then reducing it may weaken those incentives. Taxing entrepreneurs more heavily may reduce the creation of products and services that benefit everyone.

Yet, from Binmore’s perspective, the standard fairness-versus-efficiency dilemma vanishes. The reason is that it rests on an ill-posed question: fair relative to what? For Binmore, fairness concerns how the gains from cooperation are shared. That immediately separates two issues. First, society should make cooperation as productive as possible and therefore consider only efficient social contracts. Second, among those efficient social contracts, society should choose one. That is where fairness comes in. Fairness norms are used to select one of the efficient social contracts.

Binmore spells out this point in that way:

It is true that people sometimes so focus on the question of how a cake is to be divided that they forget that it may be worse for everybody to divide a small cake equally than a large cake unequally, but such inefficiencies are not somehow inevitable when fairness criteria are applied. […] When fairness criteria are employed in the manner for which they evolved, they are used to select among the set of efficient equilibria. Efficiency therefore takes priority over equity. — Binmore (1994)

So there is no genuine conflict between fairness and efficiency if fairness is understood properly. The fairness question arises only when considering social contracts that are efficient. It would not make sense to consider social contracts that are not efficient.

To go back to Okun’s point, markets may create wealth, and reducing inequality may lead to less total wealth in society. But the most suitable social contract, in the sense of being the most likely to be accepted and sustained, is not the one that leads to the greatest wealth creation without regard for who benefits from it. If a social contract leads to a lot of wealth creation but all that wealth is concentrated at the top, those who are not at the top might instead argue for another efficient social contract that benefits them as well.10

People who defend the sacred nature of property rights might feel that people have no “right” to claim the wealth created by successful workers and entrepreneurs. But a social contract approach à la Binmore does not claim that people have, in principle, a “right” to the wealth created by others in society. It says something more fundamental. All the opportunities for wealth creation, and the rights to enjoy wealth privately, are the products of a social contract. Who gets what and how is part of the negotiation of that contract. There would be little reason for people to support a social contract that creates a lot of wealth if that wealth ends up concentrated only at the top.11

Binmore’s framework gives a rather different perspective on political debate. It is not that equality, liberty, or fairness are absolute principles that should guide political decisions. Instead, political debates about fair decisions need to look for mutually beneficial and sustainable solutions.

What is a just society, on this view? It is a society that abides by commonly accepted norms for solving bargaining questions about how to allocate rights and duties, and costs and benefits, across social life. Those accepted norms define what counts as “fair”. This also helps explain why it is often difficult to give a neat abstract definition of a just society. Our shared principles of justice are often embedded in practices and expectations rather than stated explicitly. They become visible mainly when they are violated.

That is one reason why social justice is typically—as stated by left-wing intellectual Nancy Fraser—experienced negatively, when our norms of fairness are violated.

[J]ustice is never actually experienced directly. By contrast, we do experience injustice, and it is only through this that we form an idea of justice. Only by pondering the character of what we consider unjust do we begin to get a sense of what would count as an alternative. Only when we contemplate what it would take to overcome injustice does our otherwise abstract concept of justice acquire any content. Thus, the answer to Socrates’s question, ‘What is justice?’ can only be this: justice is the overcoming of injustice. — Fraser (2012)

In addition, fairness norms are always partly renegotiated, as different social groups try to improve their position in the social contract, whether in wages, social status, or political recognition. Criticisms of “injustice” and calls to make society more “just” are often, in practice, attempts to move the needle of fairness norms in favour of some groups rather than others. When such movements succeed, society comes to accept a new answer to the question of what counts as just.

This also explains a key feature of political debate. Although politics is largely about the allocation of material resources and social recognition in the Game of Life, it is often conducted in the Game of Morals. Social groups and their representatives do not merely haggle over money and status directly. They argue about merit, need, dignity, and respect, because these are the moral categories through which claims to a larger share of social resources are made.

In the next post, I will look at what kind of fairness norms we actually adopt in society and why.

References

Bentham, J. (1843) ‘Anarchical fallacies’. In: Bowring, J. (ed.) The Works of Jeremy Bentham. Vol. 2. Edinburgh: William Tait.

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. New York: Oxford University Press.

Constant, B. (2003 [1815]) Principles of Politics Applicable to All Governments. Edited by E. Hofmann, translated by D. O’Keeffe. Indianapolis, IN: Liberty Fund.

Harsanyi, J.C. (1955) ‘Cardinal welfare, individualistic ethics, and interpersonal comparisons of utility’, Journal of Political Economy, 63(4), pp. 309–321.

Harsanyi, J.C. (1975) ‘Can the maximin principle serve as a basis for morality? A critique of John Rawls’s theory’, American Political Science Review, 69(2), pp. 594–606.

Hayek, F.A. (1945) ‘The use of knowledge in society’, American Economic Review, 35(4), pp. 519–530.

Hayek, F.A. (1960) The Constitution of Liberty. Chicago: University of Chicago Press.

Hume, D. (1742) ‘Of the independency of parliament’. In: Essays, Moral and Political. Edinburgh: A. Kincaid.

Marx, K. (1875) ‘Critique of the Gotha Programme’. In: Marx, K. and Engels, F. Selected Works. Vol. 3. Moscow: Progress Publishers, pp. 13–30.

Nozick, R. (1974) Anarchy, State, and Utopia. New York: Basic Books.

Okun, A.M. (1975) Equality and Efficiency: The Big Tradeoff. Washington, DC: Brookings Institution.

Rawls, J. (1971) A Theory of Justice. Cambridge, MA: Belknap Press of Harvard University Press.

The importance of fairness as a determinant of just solutions is a key aspect of human interactions. “It’s unfair!” is one claim children learn to say very early in life. And in adult political debates, claims about social injustice play a major role.

But how do we know what such statements about fairness mean? Open one of the many textbooks in moral and political philosophy, and you will be struck by the fact that there are huge debates between highly respectable and knowledgeable scholars about what makes something fair. Not only can we not assume that toddlers have an intellectual grasp of the issues raised in these textbooks, but most adults who discuss social justice would also struggle to define what they mean by the term.

Even though toddlers are no political philosophers, their claims about unfairness often work. Parents and sometimes even other siblings frequently recognise their claim as valid and undertake to remedy the situation. In other cases, parents will disagree and sometimes convince the toddler that the claim was unfounded.

But what is happening in such situations? How do we end up agreeing on a view that something is “fair” or “unfair”? This question is relevant to all the situations in which we have to allocate rights and duties. Whether this allocation is discussed openly or whether we agree on it without debate, we are guided by our sense of fairness: what is right and what is likely to be perceived as right by other people.

Even when fairness norms are not expressly mentioned, they are nonetheless in the background, shaping our views and decisions. When we cut a cake at a birthday party, we do not need to say, “I am cutting slices evenly to be fair”, but we do it anyway, and children would be quick to point out any deviation.

Situations in which rights and duties have to be allocated are pervasive. Who should have the last cookie in the jar? Who should be promoted at work? Who should be the first author on an academic paper? Who should give way in a narrow corridor? How much tax should rich people pay?

Our interests conflict, at least somewhat, whenever we have to allocate rights and duties. We want more rights and fewer duties. It is therefore striking that our social interactions involve relatively little friction in the form of open disagreement and haggling. Indeed, we are often averse to open haggling, which can feel adversarial. The absence of haggling is a kind of puzzle. How do we constantly and easily manage to solve these situations and agree on a commonly acceptable compromise?

Despite major progress in the behavioural sciences over recent decades, social scientists do not share a unified account of how people converge on specific fairness judgements across contexts, and the prevailing answers remain partial and incomplete. Thirty years ago, however, the game theorist Ken Binmore proposed a theory that is strikingly elegant and offers a compelling explanation of what fairness is and how we make fairness judgements. Unlike many theories in the social and behavioural sciences, Binmore does not build his theory from scratch. He builds on the accumulated knowledge generated by decades of work in game theory and evolutionary theory. On a topic that is quite treacherous, the rigour of game theory gives Binmore’s theory a rare clarity. I present here his theory about how we make judgements such as “it’s unfair”.

Fairness problems are everywhere

Problems about how to allocate rights and duties, resources and costs in human groups are pervasive. Their omnipresence reflects the fact that human sociality is based on cooperation. Successful cooperation requires agreement on who does what, and the benefits from cooperation naturally raise the question of who gets what.

Humans manage such allocation problems seamlessly all the time without thinking about them. So it is surprising that, when we ask “what is fair?”, the answer does not seem obvious. Equality is often mentioned, but equality is not a simple answer when people differ in their contributions, talents and needs. The term equity is sometimes preferred, but, in truth, people who use it often have only a vague idea of what the notion means in philosophy.

In the confusing maze of competing explanations and half-baked theories based on intuitions, Binmore offers a striking clarity. His naturalistic approach does not rely on metaphysical assumptions about invisible principles of morality out there. It builds from the ground up. It starts from a rigorous understanding of human interactions, of the challenges and opportunities we face, and of the stable solutions that can emerge to solve the former and seize the latter.

In my previous post, I explained how, in that framework, fairness norms are social conventions commonly agreed within a group that solve, quickly and with minimal friction, the questions of rights and duties, resources and costs.

This answer provides a radically new insight into fairness compared with the way it is discussed in society and in philosophy textbooks. It opens a path to strikingly simple and clear answers to questions that often seem like intractable metaphysical puzzles about what is right and wrong. In this post, I explain how Binmore accounts for what we come to find fair in a given social setting. To make his general theory more concrete, I begin with two familiar kinds of fairness problems.

Alice and Bob

Alice and Bob live together. They have to solve common household problems about how to share the chores needed to keep the house functional: who does the dishwashing and the laundry? Who puts the bin out? Who mows the lawn? What is a fair allocation of chores? Is it 50–50 for each chore? Across chores? What if Alice enjoys cooking more, and Bob enjoys mowing the lawn more? How should that be taken into account when deciding what is “fair”?

They also have to solve the problem of sharing the benefits of being together. For instance, they like spending the evening together watching television. But they have different preferences about what to watch. Alice prefers crime dramas, and Bob prefers sci-fi series. What is a fair way for them to share viewing time? Should it be 50–50, alternating every day or every week? What if Alice is a big fan of crime dramas while Bob only slightly prefers sci-fi series? Should they still split viewing time evenly? Would it be “fair” for Bob to ask for that?

Binmore provides a general treatment of such problems. In this post, I focus on the case in which Alice and Bob have equal bargaining power. Neither has the ability to push his or her own interest more than the other. In that case, what would they agree is a fair solution to the problems they face?

Utilitarianism and egalitarianism

Utilitarianism: maximise the sum of happiness

Utilitarianism is a moral and political philosophy that relies on a form of equality: everyone’s happiness matters equally. From that position, utilitarians argue that the right thing to do when making collective decisions is to aim for the greatest total happiness. Jeremy Bentham famously summarised this doctrine in these words:

It is the greatest happiness of the greatest number that is the measure of right and wrong. — Bentham

If Alice and Bob were utilitarians in their choice of television programme, they would choose the solution that maximises the sum of their satisfaction. Given that Alice enjoys crime dramas more than Bob enjoys science fiction series, they would end up watching crime dramas all the time.1

Egalitarianism: improve the fate of the person who is worse off

One of the landmarks of twentieth-century political philosophy was John Rawls’s A Theory of Justice, published in 1971. In it, Rawls opposes utilitarianism. He argues that it can lead to undesirable outcomes, such as the sacrifice of some individuals’ basic rights for the greater happiness of the many. In our example, Bob would have to accept never watching sci-fi series in order to maximise total happiness. Such a solution may sit uneasily with our intuitions about fairness.

In place of utilitarianism, Rawls suggests another approach. To determine what social agreement is fair, we should imagine what it would be like to be other people and to experience things as they do. He calls this thought experiment the veil of ignorance: we imagine ourselves as ignorant of our own identity, looking at a social situation while knowing that we could turn out to be any of the people involved. Rawls calls the fictitious standpoint created by this thought experiment the original position. It is from this original position that we should form a view about whether we like a social situation or not. What this thought experiment forces us to do is to consider equally the fate of everybody involved because, under the veil of ignorance, we would not know who we are in that situation.

Rawls argues that under the veil of ignorance we would choose to organise the social situation so as to focus on the benefits of the most disadvantaged person. With this conclusion, Rawls supports egalitarianism.2

Ken Binmore’s theory about how we make fairness judgements

How do we solve these problems in practice? What do we do when we assess whether a proposed solution is “fair” or not? Binmore thinks that Rawls is right to use the veil of ignorance. In fact, unlike Rawls, who sees it as something we ought to do in order to find the right solution, Binmore thinks it captures, in stylised form, something close to what we do all the time. Even though we may not think in such abstract terms, we put ourselves in other people’s shoes to see a given problem from their point of view and to look for a solution they could accept.

In fact, doing so is very intuitive to us. It amounts to following the Golden Rule.

The Golden Rule

The Golden Rule is one of the most widely shared principles of morality across human societies. Its presence across different continents and widely different faiths suggests that it was, in all likelihood, codified independently in different places. A simple version of the Golden Rule is:

Do unto others as you would like to be done unto you.

This is not just a lofty moral principle enshrined in religious texts. It is also a principle frequently used to teach children how to behave properly. Alice, learning that her toddler has hit another child, might say:

Imagine you were her. Would you be happy with what you have done? No. So you should not do that to her.

Binmore sees the Golden Rule as a cognitive device that humans evolved to help them agree on how to treat each other whenever they have to allocate rights and duties, benefits and costs of cooperation. The Golden Rule asks us to step into the shoes of others and take their point of view when deciding how to treat them. It may have emerged naturally as a way to simulate bargaining in our heads and anticipate what other people are likely to accept and what they are likely to reject.

For that reason, Binmore sees the veil of ignorance as an elaborate version of the Golden Rule that we use in everyday life to solve the many situations in which we have to agree on how to interact with other people and how to share the gains from cooperation.3

Put very simply, what we do is to place ourselves in the shoes of the people with whom we are interacting. We see the situation from their point of view. They do the same, and this process allows us to converge on shared agreements. When Alice thinks about what the fair solution is, it is as if she were placing herself in Bob’s shoes and seeing the situation from his point of view. She can then consider each option both from her own point of view and from Bob’s.4

Empathetic preferences

To put ourselves in the shoes of other people as a way of identifying what is fair, we need to be able to compare how happy we would be if we were different people in different scenarios.

In other words, Alice needs to be able to compare her and Bob’s satisfaction with different possible agreements. For Alice to appreciate that Bob may not be as happy as she would be watching crime dramas, she needs to be able to say something like “I enjoy crime dramas more than Bob enjoys them”.5

Binmore relies on the work of the game theorist John Harsanyi to give such statements formal meaning. Harsanyi formalised the idea of people having preferences not only over their own situation but also over how it compares with how others experience their situations. Binmore calls such preferences, when you put yourself in the shoes of others, empathetic preferences.

I explain Harsanyi’s theory in more detail in an appendix at the end of this post.

When Alice puts herself in Bob’s shoes while he is watching a sci-fi series, she must imagine whether Bob would be as happy in that situation as she would be watching crime dramas.6 Now, imagining the different solutions for sharing viewing time from her own point of view and Bob’s, she can choose the solution that is going to be amenable to both of them.

Binmore sees in Rawls’ veil of ignorance a way to describe this thought process where Alice gives equal consideration to her own and to Bob’s perspective. Prior to Rawls’ book, Harsanyi had already developed a moral theory based on exactly the idea of the veil of ignorance. In 1955, he described what he called a “special impartial and impersonal attitude” where someone considers

what social situation he would choose if he did not know what his personal position would be in the new situation choosen (and in any of its alternatives) but rather had an equal chance of obtaining any of the social positions'6 existing in this situation, from the highest down to the lowest. — Harsanyi (1955)

But, unlike Rawls, Harsanyi thought that under that veil of ignorance, people, not knowing who they would turn out to be, would agree on a social situation that maximised their expected well-being. In other words, Harsanyi argued that, under the veil of ignorance, we would be utilitarian. His argument would seem intuitive to any economist: if you do not know who you will be in a given situation, that situation is akin to a lottery. When you choose a lottery, you pick the one with the best expected gain.

Hence, from the same intellectual device, the veil of ignorance, Rawls and Harsanyi arrived at two different conclusions: egalitarianism and utilitarianism. Who is right?

Binmore’s Copernican move

The debates between utilitarianism and egalitarianism are complex, and the books and papers written on the topic could fill libraries. Binmore proposes a simple resolution of this debate, one that makes the issue much clearer.

Agreements in the Game of Morals need to be self-enforcing in the Game of Life

Binmore’s solution is to see utilitarianism and egalitarianism as valid answers to two different types of bargaining problems. He points out that Harsanyi and Rawls made an implicit assumption with major consequences: that the agreement under the veil of ignorance is binding. A binding agreement is like a contract signed in front of a lawyer. If one party decides not to honour the contract, a third party can intervene and enforce its terms.

If agreements about fairness are binding, then Binmore agrees that Harsanyi would have to be right and we should be utilitarian. Behind a veil of ignorance, we would choose the situation that is, on average, the best.7

The idea that agreements behind the veil of ignorance are binding is problematic, however. We should not forget that the veil of ignorance is only a thought experiment used to find a solution together in real life. There is no actual contract signed “under the veil”. How could a solution reached under the veil be binding in the real world once we step out of that thought experiment? Both Rawls and Harsanyi assume that respecting the agreement we would reach under the veil is a moral duty.8

But that is a problematic move. Where would such principles come from? Rawls’s and Harsanyi’s theories are meant to explain how we should make moral judgements. But if they assume some moral duty a priori, they assume what they are meant to explain.

Binmore rejects such a skyhook. There is, in his view, no basis for assuming that people are somehow bound to respect a hypothetical agreement reached in a thought experiment.

There is nothing about the circumstances under which people are hypothesized as bargaining in the original position that can justify the assumption that they are bound, either morally or in practice, to abide by the terms of a hypothetical deal reached in the original position. Thus, in my theory, compliance is assumed only when compliance is in the best interests of those concerned. — Binmore (1994)

Consider, for instance, a scenario in which Bob really hates dishwashing and doing the laundry, and really likes watching television. The allocation of rights and duties that maximises total happiness in Alice and Bob’s household might be for Bob to watch television after every dinner while Alice washes the dishes and does the laundry. To justify this, Bob might say: “I really dislike washing the dishes and doing the laundry more than you do, and I enjoy watching television more than you. So this is the solution that maximises our total satisfaction, and it is what we would have agreed under the veil of ignorance.” But if Alice and Bob have equal bargaining power, nothing prevents Alice from answering that she does not care about the argument that this is a “fair” solution because, under the veil of ignorance, she might have ended up being Bob and enjoying that situation. In practice, she is Alice, and she does not like the idea of such a lopsided deal.

A key point made by Binmore is therefore that an agreement in the Game of Morals only works if it is sustainable as a bargaining agreement in the Game of Life. In other words, it needs to be renegotiation-proof. It must be the case that when Alice puts herself in Bob’s shoes to consider what would count as a possible agreement, it has to be an agreement she is actually happy to respect in the real world once she steps back into her own shoes.

This has a very important practical implication. When agreements are not binding, only egalitarian deals are sustainable and renegotiation-proof.9 In an egalitarian solution, nobody has an incentive to reopen the case by saying that fairness requires a different assignment of roles, because everybody is getting the same outcome.10

Binmore, therefore, proposes an intellectually appealing way to unify our understanding of utilitarianism and egalitarianism: utilitarianism is the reasonable answer when agreements are enforceable, egalitarianism is the answer when they are not. This helps explain why both utilitarianism and egalitarianism can, to some extent, resonate with our intuitions. It is not that one is absolutely right and the other absolutely wrong. The answer depends on the context.11

When will either utilitarianism or egalitarianism prevail?

Utilitarianism will prevail when moral agreements can be binding. In real life, binding agreements typically involve a third party, such as judges and the legal system. But, from Binmore’s perspective, there is no real third party for agreements on social contracts in the Game of Life, because all the potential enforcers are themselves part of the Game of Life.

Social contracts are not contracts in the legal sense, since there is nobody outside the Game of Life to whom one can appeal for redress if someone cheats. Everybody—up to and including the Lord High Executioner—is a player in the Game of Life. — Binmore (1998)

When thinking about binding agreements, one therefore has to think about cases in which they are binding because the structure of the game itself makes violating them unattractive. Binmore gives the example of insurance games, which may have played an important role in the evolution of morality in our ancestral past.

In an environment in which daily food intake was needed for survival, but each individual faced substantial variation in success at finding food, mutual sharing makes sense as an incentive scheme. Today I get some of your meat, tomorrow you get some of mine. Respecting the insurance contract is binding because failing to respect it would trigger sanctions, such as exclusion from future help.

To the extent that these agreements can be binding, a utilitarian solution can emerge in which meat is shared. Sharing is utilitarian because the person with food shares it for the benefit of all those who are part of the insurance agreement.

Binmore conjectures that situations of insurance were likely to be at the origin of the Golden Rule.

If the Golden Rule is understood as a simplified version of the device of the original position, I think an answer to this question can be found by asking why social animals evolved in the first place. This is generally thought to have been because food-sharing has survival value. — Binmore (2005)

This logic of insurance can be extended to any situation in which people can accept losses in some cases because, across a wide range of situations in which their needs and preferences differ, they end up better off overall. Consider a couple like Alice and Bob. On any decision about how to run their household, they have to agree on their rights and duties. They can be utilitarian to the extent that there is an overall gain for both of them across all their decisions. If Alice dislike putting the bin out and mowing the lawn Bob may be happy to do those tasks most of the time. If Alice prefers choosing the interior design, Bob may be happy for her to make most of those decisions. Not all of their bargaining problems need to be solved with a 50–50 split, because they are playing a larger game in which they can divide total gains across the household. They can therefore be utilitarian in specific situations because they are egalitarian in the larger game they play at the broader household level.

Egalitarianism will prevail when agreements cannot be binding. In many cases, agreements cannot be binding. Between two people who interact only once in one specific way, there is no larger game with other situations whose potential gains can be used to discipline an agreement in that situation. In such a case, utilitarianism will not work.

If Alice and Bob are assigned to a group project and do not expect to work with each other again, Bob’s claim that he dislikes some parts of the work and would prefer Alice to do them may fall flat. Alice may simply ask for a strict equal share as a fair solution, independently of how much each of them likes or dislikes the tasks.

Similarly, even when people interact repeatedly, if they have characteristics that would systematically bias the gains from a utilitarian solution in their favour, that solution may be rejected. Let us return to the example of Bob disliking household chores more than Alice and enjoying leisure more than she does. Here, the utilitarian solution would lead to an uneven deal at the household level. Bob might try to appeal to the veil of ignorance: “Put yourself in my shoes, I really hate dishwashing.” Alice is not bound by the fiction of the veil of ignorance, and she may simply reply, “I don’t care that you really dislike it. You still have to do it sometimes.” Alice has no interest in abiding by a utilitarian solution if it puts her at a systematic disadvantage. Utilitarianism will only be followed when, overall, it works for both.

In other words, while an agreement can be utilitarian for single situations, it is only if this utilitarianism leads to a form of egalitarianism overall because the agreement over the Game of Life itself cannot be binding.

“It’s unfair”: what does it mean?

From toddlers complaining to their parents to activists demonstrating in the streets, claims about unfairness are commonplace. What do they mean? Surprisingly, those who make these claims would often struggle to articulate a coherent conceptual framework to justify them. Yet such claims work because they are real bargaining claims that we can grasp at a deeper intuitive level.

In most situations in which moral agreements are not binding, the statement “it’s unfair” has an egalitarian basis. When Alice says “it’s unfair”, she is claiming that she is the one who is worse off in the situation and that another way of organising the interaction would improve her position. There is no reason for her to agree to a situation in which she is worse off and to let the social contract on which it is based continue unchanged. Instead, she can simply say that she disagrees with that social contract and ask for another one that improves her situation. If she has equal bargaining power, her claim cannot simply be ignored.

In some situations in which agreements are binding, the statement “it’s unfair” takes on a utilitarian meaning. It becomes a claim that, in a specific situation in which it is commonly accepted that the sum of satisfaction should be maximised, the deal was not respected. This deal works as part of a larger game that is itself constrained by a broader egalitarian balance acceptable to both sides. Overall, everybody benefits from maximising satisfaction in each specific situation. But if people refuse to respect the deal when it is your turn to benefit from it, the statement “it’s unfair” calls out this breach of the commonly accepted arrangement.

Binmore’s perspective lifts the mystery surrounding questions of fairness. It helps explain why both utilitarianism and egalitarianism can feel intuitive and have defenders. And it offers a simple reason why either may prevail in a given social setting.

For most games, and in particular for the Game of Life, there is no way to make moral agreements binding. Hence, the utilitarian solution is often not sustainable there. This is simply because nothing prevents someone from saying, “I don’t care about your moral story, I am not happy with the deal you propose.” For that reason, egalitarianism typically prevails when people have equal bargaining power.

Binmore does not reach this conclusion from metaphysical or a priori moral assumptions. He builds his framework from the ground up, on the basis of a matter-of-fact understanding of human interactions, the problems they face and the solutions they can adopt. Once one takes that approach, there is an air of inevitability to the conclusions Binmore reaches. We do not reach them because we like them, and then make metaphysical assumptions along the way to help us land on these conclusions. We reach them because they follow from a naturalistic perspective that seeks to explain morality as it is, rather than describe it as we would like it to be.

References

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Bentham, J. (1776) A Fragment on Government, Preface; in Burns, J. H. and Hart, H. L. A. (eds) (1977) A Comment on the Commentaries and A Fragment on Government. London: Athlone Press, p. 393.

Harsanyi, J.C. (1955) ‘Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility’, Journal of Political Economy, 63(4), pp. 309–321.

Harsanyi, J.C. (1977) ‘Morality and the Theory of Rational Behavior’, Social Research, 44(4), pp. 623–656.

Rawls, J. (1971) A Theory of Justice. Cambridge, MA: Harvard University Press.

Appendix: Harsanyi’s extended preferences

In the film Crime 101, Lou (Mark Ruffalo) tells Angie (Jennifer Jason Leigh):

I like the beach much more than you do.

Such statements are easy for us to understand. But even though they are intuitive, they are not trivial. Let us ask a simple question: what does it mean to say such a thing? How can we test whether this statement is true?

For many generations of economists, such a statement would have been seen as nonsensical because we cannot compare how much Lou likes something, which is something happening in his head, to how much Angie likes it, which is also something happening in her head.

Crash course on what preferences mean

What can we say about preferences?

Let us start with a statement by Lou such as: “I like going to the beach more than I like hiking.” Economists are fine with such statements. They are meaningful because they have practical consequences. If he is repeatedly given the choice between going to the beach and going hiking, we should see Lou choosing the beach more often.

What about a slightly different statement: “I like going to the beach twice as much as I like hiking”? This now suggests a precise quantitative comparison, as if there were some quantity of subjective satisfaction in Lou’s head when he engages in an activity. He is saying: my quantity of satisfaction when I go to the beach is twice my quantity of satisfaction when I go hiking.

For most of the twentieth century, mainstream economists basically rejected such statements as unverifiable. There is no direct way to measure what is happening in Lou’s head, so such a statement appears to have no practical consequences. However, in the book that launched game theory, von Neumann and Morgenstern offered an answer to this problem. Such a statement does lead to practical predictions when we consider the willingness Lou should show to take risks for each option. If Lou says that he likes the beach twice as much as hiking, he should be indifferent between hiking for sure and a gamble that gives him a 50 per cent chance of going to the beach and a 50 per cent chance of getting nothing. More generally, if he likes the beach three times as much as hiking, the corresponding threshold would be one-third; if four times as much, one-quarter; and so on.12

Now let us appreciate that the statement “I like the beach much more than you do” introduces another level of complexity. Lou is now comparing the quantity of satisfaction in his own brain with the quantity of satisfaction in Angie’s brain. Does it make sense to do something like that? Can it lead to predictions that can be tested? Most economists would say no, because we do not have a hedonometer that we could plug into Lou’s and Angie’s heads and that would provide an absolute measure of satisfaction on a common scale.

Harsanyi proposed a way to make sense of such statements by extending choice to hypothetical choices involving a swap of places. When Lou says, “I like the beach much more than you do”, this is equivalent to saying: “I would prefer to be me, with my preferences, at the beach than to be you, with your preferences, at the beach.” Harsanyi called such preferences extended preferences. If we accept such statements as meaningful, and if we apply the von Neumann–Morgenstern result to this kind of choice, then we can represent these preferences with a utility function that allows interpersonal comparisons across people.

In other words, saying things like “I like the beach more than you do” becomes a reflection of the ability to choose between possible person-swapping experiences.

As he was launching his party’s manifesto for the 2024 British elections, the then leader of the Labour Party, soon to become Prime Minister, Keir Starmer, stated that:

[T]he pursuit of social justice and economic growth must go hand-in-hand.
— Keir Starmer (2024)

Over the past decades, the notion of “social justice” has become highly influential in public discourse. The term is now routinely invoked in political debates, policy discussions, and academic work as a guiding principle for evaluating social arrangements. Inequalities in income, wealth, education, and health are commonly assessed in terms of their fairness.

Yet, despite its ubiquity, the concept itself is rarely defined. It is largely taken for granted. What if we were to put it to the toddler’s acid test and ask a simple question: what is social justice? It is not an easy question to answer. In this post, I offer a simple answer to the question of what “social justice”, and the more traditional term behind it, “fairness”, actually mean.

What it is not

For readers unfamiliar with my previous posts, let me start by setting aside some common but misguided answers.

Fairness does not require religion to make sense

A frequent view is that fairness and justice derive from religion. Thou shalt not steal. Why? Because it is one of the commandments handed down to Moses on Mount Sinai.1

This view gets the direction of explanation wrong. Across different cultures and religions, we observe strikingly similar moral principles, such as the Golden Rule. This suggests that these principles are broadly appealing to humans and were incorporated into religious traditions that emerged independently, rather than originating from them.

Fairness does not require absolute moral laws

Another possibility is that there are objective moral laws “out there” that we can discover through reason. This position, associated with moral realism, is a secular modern version of the religious view: instead of divine commandments, there are external moral truths.

But this raises difficult questions about their nature and origin. Where do these principles come from, and how can we identify them with confidence? As I have argued in a previous post, this approach sits uneasily with a naturalistic understanding of the world. It has led many philosophers, from Kant to Parfit, into intellectual mazes with no way out, as they attempt to ground the idea that moral rules can be absolutely true.

Fairness norms are social rules

If moral and fairness norms are not externally given to us by a god, by logic or by some other thing out there, what are they? A simple answer is that they are rules of the social games we play. A very down-to-earth explanation is that moral rules are very much like those of a board game: they are not written into the fabric of the universe, but they are man-made and conventional. They are followed because they are commonly agreed upon by all participants.

For sure, moral rules feel more important than those of a board game. The reason is that they are the rules of a much more important game we play: the game of life.

[Y]es, moral rules are much more important to us than sporting rules because moral rules are the rules of the Game of Life, the one that determines our success and setbacks in life, at work, with friends or romantic partners, with family members, and so on. It makes perfect sense for us to have a moral sense that makes us care greatly about how the Game of Life is played. There is no reboot and no respawn in this game, and it determines everything that matters to us. — Optimally Irrational - Morality works without absolute moral truths

Cooperation and bargaining over its benefits

The importance of cooperation as a structuring aspect of social life

This answer raises a further question: what do we actually do in the game of life, and what kind of rules does it require? The answer is that we are a social species, and most of what we do consists of cooperating with others to achieve better outcomes. Cooperation is the scaffolding principle of life. It appears at all levels of biological and social organisation for a simple reason: when interactions are repeated, cooperating is often the best way to succeed.

Because cooperation is so pervasive, we can easily take it for granted and fail to notice it. Consider Alice, who lives in a large modern city. Every morning, she orders a coffee on her way to work. The barista makes a good coffee, and she pays for it. This may seem trivial, but many other possibilities did not occur. Alice could have run away with the coffee without paying. The barista could have used poor-quality ingredients to cut costs. What actually happens is a highly structured and mutually beneficial form of cooperation.

The same pattern continues throughout her day. She stops at a convenience store, takes a pack of biscuits, and waits in line rather than pushing ahead, even though she is in a hurry. She takes the tube and pays for a ticket, even though the gates are open. She gives way in a corridor to an elderly woman with a walker, even though she does not know her. At work, she listens to colleagues without interrupting them and contributes to a group project, even though it required giving up time she would have preferred to spend otherwise.

Nothing in Alice’s day appears remarkable. Yet it is precisely this dense web of cooperative behaviours that makes her orderly and comfortable life possible. Evolutionary anthropologist Sarah Hrdy captures this contrast with our closest relatives, the chimpanzees:

What if I were traveling with a planeload of chimpanzees? Any of us would be lucky to disembark with all ten fingers and toes still attached, with the baby still breathing and unmaimed. Bloody earlobes and other appendages would litter the aisles. — Hrdy (2009)

Cooperation allows humans to reach outcomes far beyond what any individual could achieve alone. Most of us would struggle to produce even simple goods in isolation. We cannot build a car or a phone, and most of us would not even be able to repair a television if it breaks.2 As the comedian Nate Bargatze jokes, if he were sent back in time, he would not be able to demonstrate that he came from the future: he could describe phones and satellites, but not explain how they work. What sustains our world is not individual ingenuity, but the large-scale cooperation of many people.3

Once we realise that we cooperate all the time with others to reach better situations, a natural question arises. The production of benefits by cooperation raises the question of the distribution of these benefits between the people involved.

How should the gains of cooperation be distributed?

“Who should get what?” is a question that comes naturally with cooperation. Who should get the last cookie in the jar? Who should get the promotion at work? Who should be the first author on an academic paper? Who should give way in a corridor? All these questions are about the distribution of the benefits of cooperation. Extended to society, the question of fairness points to debates about “social justice”: how much tax should the rich pay? Who should receive social subsidies—poor people, disabled people and how much?

One common mistake is to think that this question has a simple answer: equality. Everyone should get the same. But this raises an immediate issue: why equality? What principle justifies it?

Even if we accept equality as a starting point, it remains unclear what kind of equality we have in mind. Should it be equality of outcomes in all situations? If Alice did 90% of the work on a project and Bob only 10%, should they split the proceeds equally? If you meet an elderly woman in a narrow corridor, do you both have the same duty to give way? If a mother serves food to a toddler and a teenager, should both receive the same quantity?

These examples suggest that equality of outcomes is often inappropriate when people differ in their contributions or their needs. As Aristotle puts it:

The just, then, is a kind of proportion […] For if the persons are not equal, they will not have equal shares; it is when equals possess or are assigned unequal shares, or unequals equal shares, that quarrels and complaints arise.
— Aristotle

Avoiding quarrels over the distribution of the gains from cooperation requires finding appropriate proportions that will satisfy all participants. Since we do not have a simple, straight answer like “equality”, this requires a form of bargaining.

Bargaining, bargaining, bargaining

Ken Binmore, a game theorist specialised in bargaining theory, brings a key insight to this question: fairness is a solution to the pervasive problem of distributing the benefits of cooperation.

Given how often we need to agree on how to share the gains from cooperation, it would be time-consuming and frustrating if we had to haggle each time. Imagine our lives if every mundane request unfolded like this:

Alice: Can you pass me the salt please?

Bob: What am I getting in return for my effort?

This is where the rules of fairness play a role. They are meta-rules, commonly shared and understood by all. We can use them to resolve our daily problems of allocating rights and duties seamlessly. When Alice, heavily pregnant, asks Bob, a healthy middle-aged man, for his seat on the tube, she can do so because she knows that Bob shares with her a norm of fairness that prescribes that a pregnant woman should have priority for a seat. Bob knows this too and understands Alice’s request. He stands up, and the allocation of rights of access to the seat is resolved quickly and without conflict.

Fairness norms are the grammar of our social interactions. They allow us to navigate social situations with minimal friction, helping us recognise what is due to others and what we are entitled to ask for ourselves.

Fairness norms are present at all times and guide our interactions. If they are absent from our everyday discussions, it is because they do not need to be voiced or discussed as long as they are followed. It is only when there are violations or disagreements that fairness talk emerges: people queue all the time without talking, or perhaps even thinking, about queuing as a first-come, first-served fairness norm. But if someone pushes in front of them in the queue, they might interject, “You can’t do that, I was here before you!”

The game theory of fairness, explained simply

If fairness norms help us settle recurring questions about who should get what, how exactly does that work? Binmore’s answer can be expressed more precisely with the tools of game theory. The ideas are somewhat technical, but the basic structure is simple.

Fairness as a way to agree on how to share the gains from cooperation

In my previous post, I described the game theory of cooperation. I used the case of “Prisoner’s Dilemma”, an interaction between two people, let's say Alice and Bob, where they can “cooperate” (C) or “defect” (D). Depending on their mutual choices, Alice and Bob get some “payoff”, which is a general measure of whatever they care about. It can measure monetary payoffs, but it can also include other things.

If the game is played only once, defecting is the best choice in terms of payoffs. Whatever the other person is doing, choosing to defect yields a higher payoff.

But if the game is played repeatedly, cooperation becomes possible. Indeed, if there is no clear end in sight to the repetition of this game, and if Alice and Bob are patient enough and therefore care about the possibility of future gains, then cooperation can become rational: they can adopt rules of conditional cooperation whereby they cooperate if the other has been abiding by these rules. As a consequence, instead of being stuck with the (0,0) payoffs, they can reach any average payoff in the grey area below.4

For instance, they can reach an average payoff of 100 each by cooperating all the time. To do so, they can adopt the “grim trigger” strategy: cooperate as long as the other player has always cooperated; defect forever if the other player has defected at some point.5 When both Alice and Bob follow this rule, neither of them has an interest in defecting, as they would lose the benefits from future cooperation.

Other rules are possible. For instance, they could adopt the rule of alternating between (C,D) and (D,C). This is a kind of cooperation where players take turns getting the highest payoff. This rule would lead to the average payoffs (50,50), represented by point x on the graph below.

This is not, by any means, the only possibility. Another option, for instance, would be to adopt the rule of alternating between (C,C) and (C,D). If they play this game every day of the week, they could have a convention to cooperate on some days and not on others. If they play (C,C) from Monday to Saturday and (C,D) on Sunday, they would get the average payoff represented by point y on the graph below. It is close to (C,C), but one-seventh of the way towards (C,D).

What rule should Alice and Bob agree on? It is clear that some rules are obviously better than others. It is better for Alice and Bob to be at point (C,C) than at point x. Similarly, y is clearly better than x because both Alice and Bob get more. Basically, every point to the North-East of x is a better option for Alice and Bob.

Among all the possible points in the grey area, all those on the bold dark line do not have points that are better in that sense: for each point on the bold dark line, there is no other point that Alice and Bob could reach where both of them could be better off.

We call that line the efficient frontier.6 It makes sense for Alice and Bob to want to be on the efficient frontier. Being below it is “leaving some money on the table”. Both Alice and Bob would agree to move away from (D,D) to get to one of the points on the efficient frontier.

But that is the extent to which they can agree. Among all the points on the efficient frontier, Alice and Bob’s interests are at odds. Bob would prefer to be as far to the left of this curve as possible, where his payoffs are higher, while Alice would prefer to be as far to the right as possible, where her payoffs are higher.

So, in the process of deciding how to share the gains from cooperation, there is both an element of cooperation—they both want to agree on a point on the efficient frontier—and an element of conflict—they have different preferences over which point on this frontier to reach.

In short, Alice and Bob have to bargain to find a mutually beneficial solution. If they cannot agree, they might be stuck at a lower point, like (D,D). The result of this negotiation is going to be some rule they will agree to follow in order to sustain cooperation and a way to share the benefits of cooperation.

Norms of fairness, which are common knowledge in society, help coordinate Alice and Bob’s expectations so that they can quickly agree on a solution without disagreement, which could threaten cooperation and without time-consuming haggling.

Fairness is not just “equality”

At this point, you might wonder “Why all the fuss?” Isn’t it clear that the situation where Alice and Bob both play (C,C) all the time gives Alice and Bob equal outcomes and is therefore the best and fairest solution? This intuition is understandable, but it goes too fast. It is biased by the fact that, in the game above, Alice and Bob are more or less identical: we know that they have similar preferences, and we are not told anything about them differing in any important way. A bit of introspection shows that “equality” is not a trivial answer when the people concerned are different.

Let’s consider a familiar situation. Alice and Bob have to agree on what series to watch on TV tonight. In fact, they face this situation every evening. It is a classic case of bargaining: they both prefer to agree because they like to watch something together. But they also have different preferences, which means that they would prefer to agree on different things. Let’s assume, for instance, that Alice prefers crime dramas (“Crime”) and Bob sci-fi series (“SF”). Let’s now assume that Alice and Bob differ in one simple but important way: the intensity of their preferences.

Let’s also assume that Alice enjoys Crime much more than Bob enjoys SF.7 To use the conceptual tools of game theory, let’s pick some numbers to reflect their utility in each situation. We can think here of utility as their subjective satisfaction. The matrix below shows their payoffs in utility. We will assume that Alice enjoys watching Crime with Bob 10 times more than watching SF with him, while Bob enjoys watching SF with Alice only twice as much as watching Crime with her.

This game is a bit different from the Prisoner’s Dilemma. It is known as the “battle of the sexes”.8 There are two equilibria, that is, two situations where neither Alice nor Bob would want to unilaterally change their decision: (Crime, Crime) and (SF, SF). Even though Bob prefers SF, if he is currently watching a crime drama with Alice, with a payoff of 1, he does not want to leave the room to watch SF on his own, with a payoff of 0. Similarly, if Alice is watching SF with Bob, with a payoff of 1, she does not want to go and watch Crime on her own, with a payoff of 0.

While there are only two pure Nash equilibria when this game is played on one evening, repeating it every evening gives Alice and Bob much more room to manoeuvre than simply choosing SF forever or Crime forever. They can adopt a rule for how to share viewing time over time.9 Such a rule specifies what to watch each evening and is followed as long as the other person does the same. By using such contingent rules, they can generate many different average payoff combinations over time. The grey area below represents the set of average payoffs that can in principle be generated in this way.10

How should Alice and Bob decide how to agree on their viewing choices over time? This example shows that the notion of “equality” does not trivially point to a single solution. If they decide to have equal viewing time, for example by alternating each week, then they would be at point b in the picture above. Note that, when they do so, the enjoyment of Alice and Bob is different. Alice’s average satisfaction is then 5.5. It is higher than Bob’s satisfaction, 1.5. The reason is that whenever she watches a crime drama, she enjoys it much more than Bob enjoys SF. If they wish to equalise their subjective satisfaction instead, they should actually watch SF much more often together (9 times out of 10) and reach point a. They would then have equal satisfaction, but note that it would be quite low: 1.9.

What is “fair”: equal satisfaction (a) or equal viewing time (b)? Alice might argue that neither of these solutions is fair. She likes crime dramas much more than Bob likes SF. Why should they split their time evenly between them then? Indeed, given these preferences, Alice and Bob might well agree to watch SF only once in a while, while watching crime dramas most of the time. They would end up closer to point c, where the sum of their satisfaction is maximised.11

These kinds of asymmetric solutions between partners with different intensities of preference are often present because one partner is willing to give up on some of the things he or she does not care about as much as the other. One such asymmetry is comically illustrated by a scene from the film Mr. & Mrs. Smith (2005) featuring Jane (Angelina Jolie) and John (Brad Pitt) as a married couple with different preferences over the decoration of their house.

While the film plays on Jane’s fake pretence of being open to compromise, it is likely that John does not mind yielding to Jane’s wishes because he cares less about what the curtains are than she does. It is because this difference in preferences is common knowledge that Jane can claim an asymmetric bargaining solution that favours her preferences.

Norms of fairness

So how do Alice and Bob decide what is the fair sharing of viewing time between a, b, c, and all the other possible points? They could haggle to agree. But if they had to haggle every time they had to split the gains from cooperation, it would be cumbersome. They would have to debate who puts the bin out today, who does the laundry, who should repair the broken light in the pantry, and so on.

Norms of fairness are social conventions that have emerged to solve these problems in ways that are agreed upon by all. In the viewing example, Alice and Bob will share views about what kind of split each of them would expect to agree to. If it is closer to c, both Alice and Bob will understand that it is a shared expectation that the “right” way to agree is for them to watch more crime dramas. They will express this feeling by describing this solution as “fair”.

Being fair is rational

A key insight from this explanation of fairness is that it does not require us to assume that people are altruistic saints in order for them to behave fairly. Fairness rules point to equilibria of the games we play. Equilibria are self-sustaining. If you deviate from them, there are costs, typically in the breakdown of the cooperative relationship.

Being fair is rational in the sense that it pays off to follow the commonly agreed-upon rules of play about the fair allocation of rights and duties. Deviating from them is typically not worth it. For that reason, cooperation is stable. There is no need to assume a mysterious force or an external moral principle “binding” players for abstract reasons. Fairness norms are binding simply because they are equilibria and come with a structure of incentives that makes deviations unattractive.

Binmore’s argument is that fairness norms are shared conventions for solving recurring bargaining problems and reaching mutually beneficial agreements without constant haggling or mismatched expectations. This explanation demystifies fairness. Its “deflationary” nature is, however, polarising in my experience.12 Some find it extremely insightful, revealing the deep nature of fairness in a simple and almost necessarily true way. Others find it quite unappealing. The notion of fairness, intuitively, evokes something noble. Reducing it to bargaining must miss the mark, they counterargue.

To this criticism, I would stress that we should be suspicious of our intuitions as a guide to truth. Our ancestors had the intuition that the Earth was still and the universe centred around it. Their intuitions were wrong, and it is by carefully questioning them and accepting where the evidence leads us that we learned how different the world is from that view. There is no reason our intuitions should be especially accurate guides to truth. Evolution has given us intuitions to help us navigate successfully the particular environments we live in, not to understand the deep structure of the universe.

A second answer to the idea that fairness evokes something more noble than bargaining is that the fact that we like a conclusion or not cannot be an argument for or against it. Our decision to adopt a conclusion should be based only on the strength of the evidence for it, independently of whether it matches our initial preferences. Binmore’s framework provides the most compelling explanation of what fairness is and how it works. He grounds his explanation in fundamental and universal aspects of human life: the pervasiveness of cooperation and the associated need to agree on how to share the gains from this cooperation.

This post offers an introduction to Binmore’s theory of fairness. My next post will address some natural questions that arise from this introduction to Binmore’s theory of fairness: How do fairness norms work in practice to help us solve our everyday problems? How do they evolve over time? Does this emphasis on bargaining mean that might makes right?

References

Binmore, K. (1994) Game Theory and the Social Contract. Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract. Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Friedman, M. (1980) Free to Choose. New York: Harcourt Brace Jovanovich.

Harsanyi, J.C. (1977) Rational Behavior and Bargaining Equilibrium in Games and Social Situations. Cambridge: Cambridge University Press.

Hartley, L.P. (1953) The Go-Between. London: Hamish Hamilton.

Hrdy, S.B. (2009) Mothers and Others: The Evolutionary Origins of Mutual Understanding. Cambridge, MA: Harvard University Press.

Kant, I. (1998) Groundwork of the Metaphysics of Morals. Translated by M. Gregor. Cambridge: Cambridge University Press.

Labour Party (2024) ‘Keir Starmer launches “Change” – Labour’s general election manifesto’, 13 June. Available at Labour Party website.

Luce, R.D. and Raiffa, H. (1957) Games and Decisions: Introduction and Critical Survey. New York: Wiley.

Nayyeri, M.H. (2013) Gender Inequality and Discrimination: The Case of Iranian Women. New Haven, CT: Iran Human Rights Documentation Center.

Pareto, V. (1971) Manual of Political Economy. Translated by A.S. Schwier and A.N. Page. New York: Augustus M. Kelley.

Parfit, D. (2011) On What Matters. Oxford: Oxford University Press.

Rawls, J. (1971) A Theory of Justice. Cambridge, MA: Harvard University Press.

Read, L.E. (1958) ‘I, Pencil’, The Freeman, December. Available via Liberty Fund and FEE.

Starmer, K. (2024) ‘Speech at Labour Party manifesto launch, Manchester, 13 June 2024’. Labour Party.

2

Another illustration of the extraordinary feats of human cooperation was famously given by Milton Friedman, who used the example of a simple pencil to explain why no single individual could have made it.

Look at this lead pencil. There is not a single person in the world who could make this pencil. Remarkable statement? Not at all.

The wood from which it’s made, for all I know, comes from a tree that was cut down in the state of Washington. To cut down that tree, it took a saw. To make the saw, it took steel. To make the steel, it took iron ore.

This black centre, we call it lead but it’s really graphite, compressed graphite. I’m not sure where it comes from, but I think it comes from some mines in South America.

This red top up here, the eraser, a bit of rubber, probably comes from Malaya, where the rubber tree isn’t even native. It was imported from South America by some businessmen with the help of the British government.

This brass ferrule—I haven’t the slightest idea where it came from—or the yellow paint, or the paint that made the black lines, or the glue that holds it together.

Literally thousands of people cooperated to make this pencil: people who don’t speak the same language, who practise different religions, who might hate one another if they ever met. — Source

10

For readers with advanced knowledge in game theory, strictly speaking, the Folk Theorem applies only to the subset of this region that gives each player at least his or her minimax payoff, which can be interpreted here as the long-run disagreement point if Alice and Bob cannot agree. If we confine attention to pure strategies in the stage game, the minimax payoff for each player is 1. For Bob, if Alice chooses Crime, his best response is Crime, which gives him 1; if Alice chooses SF, his best response is SF, which gives him 2. The minimum of these two maxima is therefore 1. The same reasoning gives a minimax payoff of 1 for Alice. The Folk-Theorem payoff region is therefore the subset shown below. Note that if mixed strategies are allowed, the minimax payoffs are lower: 10/11 for Alice and 2/3 for Bob.

Trivers was one of the most—perhaps the most—influential evolutionary biologists of the 20th century. His work should be much more widely known in social and behavioural sciences, in particular in economics, as Trivers’ intellectual approach is very much in line with a game theoretic understanding of social interactions.

It is hard to overstate the importance of his work. Einstein famously published four groundbreaking papers in 1905, a year often referred to as his “Annus mirabilis”, during which he revolutionised physics. Trivers might be said to have had a “Quinquennium Mirabile” for the five years between 1971 and 1976, during which he produced a series of ideas that revolutionised evolutionary biology.

Reciprocal altruism - 1971

Evolution is often conceived as implying that people should be selfish and cynical. This view is mistaken. As early as Darwin, evolutionary thinkers have suggested that cooperation was fully compatible with evolution. That being said, how cooperation evolved was, for a long time, an unresolved problem. One seemingly intuitive answer was proposed by Vero Wynne-Edwards: altruism is good for the group and hence altruism will be selected to help the group survive. Some old animal documentaries illustrated this idea, describing the old wildebeest allowing itself to be caught and eaten by the lion so that the younger ones in the herd could survive.

This idea, now labelled “group selection”, does not work. Selection operates at the individual level. If altruists sacrifice themselves for the group, the group might benefit, but the altruists in it will tend to disappear over time. So, self-sacrificial altruism is not something that we would expect to be selected.1

In 1966, the evolutionary biologist George Williams proposed an alternative solution in his book, Adaptation and Natural Selection. He argued that a lot of apparently “group-beneficial” helping can, in principle, be compatible with natural selection when helping tends to be repaid inside stable social relationships.

Robert Trivers built on this insight and published, in 1971, one of the most influential papers in evolutionary biology, The Evolution of Reciprocal Altruism. In it, he describes how conditional rules of the form “I scratch your back if you scratch mine” could naturally evolve and explain the prevalence of cooperation in the animal world.

The human altruistic system is a sensitive, unstable one. Often it will pay to cheat: namely, when the partner will not find out, when he will not discontinue his altruism even if he does find out, or when he is unlikely to survive long enough to reciprocate adequately. And the perception of subtle cheating may be very difficult. Given this unstable character of the system, where a degree of cheating is adaptive, natural selection will rapidly favor a complex psychological system in each individual regulating both his own altruistic and cheating tendencies and his responses to these tendencies in others. As selection favors subtler forms of cheating, it will favor more acute abilities to detect cheating. — Trivers (1971)

Trivers then goes on to explain the rich tapestry of human moral emotions: we long for good friends who are likely to be good, we get angry at cheaters and can be willing to punish them at substantial costs to ourselves, we feel gratitude and sympathy towards altruistic acts and guilt towards our own violations of the implicit rules of reciprocity.

This emotional tapestry creates the setting for even more complex games, where people leverage these emotions strategically. Trivers, for instance, describes how some cheaters can use these emotions to their advantage.

Apparent acts of generosity and friendship may induce genuine friendship and altruism in return. Sham moralistic aggression when no real cheating has occurred may nevertheless induce reparative altruism. Sham guilt may convince a wronged friend that one has reformed one’s ways even when the cheating is about to be resumed. Likewise, selection will favor the hypocrisy of pretending one is in dire circumstances in order to induce sympathy-motivated altruistic behavior. — Trivers (1976)

Trivers’ insights align closely with those of game theory on the rationality of cooperation in repeated interactions. Indeed, Trivers was aware of the game-theoretic results and mentions them in his article.

Reciprocal altruism has become one of the main explanations of the emergence of cooperation between non-kin. Cooperation does not require society to be populated by saint-like figures; it works with humans as they are, warts and all. It is indeed the blend of cooperation and conflict in social interactions that explains the rich nature of our social emotions and the many mini intrigues layered into our social relations.

Parental investment -1972

One year after his paper on reciprocal altruism, Trivers published another major paper, Parental Investment and Sexual Selection. This provides a simple and illuminating answer to a question asked from time immemorial: why are men and women different?

Trivers starts from a very simple basic fact: men and women differ in the investment they put into making a child. In sexual reproduction, a male and a female each produce gametes, and a new child is developed from the fusion of the two. All known animals that reproduce sexually are anisogamous: they have different gamete sizes.2 We conventionally call females the sex with the largest gamete size and males the sex with the smallest.

This difference in gamete sizes creates an asymmetry that puts members of both sexes on different strategic paths to achieve higher fitness. With a lower investment in any given child, males have a greater incentive to look for many females to provide their gametes to. On the contrary, females have to be more cautious. Any child is a more costly investment for them. Because they invest more, they have fewer opportunities to have children. It is therefore worth investing more in each of them.

Since the female already invests more than the male, breeding failure for lack of an additional investment selects more strongly against her than against the male. In that sense, her initial very great investment commits her to additional investment more than the male’s initial slight investment commies him.3 — Trivers (1972)

In the class of mammals, this asymmetry builds up from a fairly small difference between sperm and egg into massive differences in biological investment. In humans, women carry and sustain a developing baby for nine months. One would therefore expect that these very different investment costs and benefits lead to very different behavioural strategies and psychologies for anything related to sexual behaviour.

Indeed, since men are seeking a mate who will bear most of the initial investment required to bring a child into the world, they evolved a psychology that is relatively more promiscuous and oriented towards short-term mating. A philanderer can potentially initiate several baby-making processes at little physiological cost to himself. On the other hand, women should care more about what men are bringing to the table; they should therefore be relatively more concerned with the quality of the genes provided and, if possible, with the willingness of the man to commit to a long-term relationship and invest his time and resources in raising their children.

Trivers uses differential parental investment to explain a number of patterns observed across the animal kingdom: the greater tendency of males to desert their partners, the greater choosiness of females, who typically decide with whom to mate from a pool of pretenders, and the greater, sometimes violent, competition between males. More broadly, this framework also helps explain associated sex differences, such as higher male mortality and greater male body strength in species where success depends on winning intra-male competitions.

Critics of evolutionary theory sometimes argue that it does not make any predictions that can be tested and that it only rationalises what has already been observed. Trivers’ work is one of the best examples disproving this accusation. In his paper on parental investment, Trivers argues that the differences in behaviour between males and females should reflect the degree of asymmetry in their parental investment. As a result, animals with more parental investment asymmetry should show greater asymmetry than those with less, and if we ever find animals with role reversals, we should also observe reversals in strategies. And indeed, we observe that in animals with less asymmetry in parental investment, like swans, the differences between males and females are less noticeable. In the rare cases where male investments are larger, like in seahorses, where the females literally place their eggs in the belly of the male who incubates them, we observe a role reversal, with females courting males and competing for access to them.

Parental investment theory offers a simple and compelling answer to the age-old question of why men and women differ in their preferences and psychological traits—often captured in the popular expression that men are from Mars and women are from Venus: much of this can be traced back to the asymmetric size of gametes, which generates broader asymmetries in investment. Men’s and women’s fundamental differences in these preferences and traits are not a reflection of social indoctrination;4 they are adaptive strategies that differ because men and women have been dealt different cards by nature in the game of life.5

Parent Offspring Conflict - 1974

In 1974, Trivers published a paper that speaks to any parent: Parent-Offspring Conflict. In this article, he explains that while parents and children are genetically closely aligned—which explains the close cooperation between family members—they are not perfectly aligned. There are always elements of conflict, and recognising this helps make sense of the sometimes complex dynamics within families.

Following the now widely accepted gene-centred view of evolution, evolution selects traits that maximise the chance of our genes being reproduced. From that perspective, children's and parents’ genetic interests are not identical. Children should care about their own success and, to a lesser degree, about the success of their siblings with whom they share, on average, 50% of their genes.

Parents, on the other hand, are equally related to their children. They should care about these equally. In addition, they should care about themselves above and beyond caring about their kids, to the extent that they could have additional children in the future.6 This leads to several areas of tension.

Children have an interest in overclaiming the attention and resources of their parents. Parents of newborns are familiar with the unrelenting demands of their baby, waking up during the night, crying frequently, and often draining parents of energy. Parents end up conflicted because, while they love their child, they would often want to give less attention than what is requested from them. They would like to sleep more, have more time for themselves, and so on. But from the baby’s point of view, it is optimal to ask for more resources than parents are willing to give. Any resource received by the baby 100% benefits him or her. Any resources (including health) saved by the parents might be used to have children later, but this is not as important for the baby as the immediate benefits of additional resources received now.

Hence, there are constant conflicts, with children asking for more resources than their parents are willing to give them. Weaning a child, which is usually required for a woman to ovulate again and have additional children, is, for instance, often a challenge, as the child resists it.

Obviously, children cannot argue like lawyers, so they compete for greater attention and resources using what Trivers calls “psychological manipulation”.

The offspring can cry not only when it is famished but also when it merely wants more food than the parent is selected to give. Likewise, it can begin to withhold its smile until it has gotten its way. Selection will then of course favor parental ability to discriminate the two uses of the signals, but still subtler mimicry and deception by the offspring are always possible. — Trivers (1974)

In fact, the conflict happens already before birth, in the mother’s womb. Foetuses overclaim maternal resources, slightly to the detriment of the mother’s health. Mothers’ physiology has been designed to counterbalance these claims, and the balance reached is the result of an internal tug-of-war. Some significant health issues associated with motherhood are a reflection of an imbalance in this internal battle. The tendency to experience diabetes during pregnancy can, for instance, be seen as one reflection of this conflict, with the foetus and placenta pushing for greater access to maternal glucose than is optimal for the mother’s own health.

Obviously, overall parents tend to love their children and children tend to love their parents, but Trivers showed—with a theory now largely supported by empirical research— that the whole picture is more complex, because there are always also elements of conflict in parent-offspring relations.

Self-deception - 1976

There is a widely known tendency for people to engage in motivated reasoning and believe that they are better than they are. Psychologists have often explained this as a reflection of the fact that we enjoy having flattering beliefs. An evolutionary perspective makes this answer somewhat puzzling. Evolution does not select psychological traits that make us happy, but traits that make us successful. Hence, if we tend to delude ourselves into thinking that we are better than we are, there must be a reason outside of our mind: it must somehow be useful.7

In the preface to Dawkins’ The Selfish Gene, Robert Trivers proposed a solution to this problem: our tendency to self-deceive, to think we are better than we are, may serve as a mechanism that enables us to deceive others more effectively. He wrote:

If … deceit is fundamental to animal communication, then there must be strong selection to spot deception and this ought, in turn, to select for a degree of self-deception, rendering some facts and motives unconscious so as not to betray – by the subtle signs of self-knowledge – the deception being practiced. —Trivers (1976)

Commenting on this assertion, psychologist Steven Pinker remarked, “This sentence... might have the highest ratio of profundity to words in the history of the social sciences” (2011).

Trivers starts from the idea that, in communication, deception is widespread. This is a reasonable assumption. As stated by Dawkins (1976): “It may well be that all animal communication contains an element of deception right from the start, because all animal interactions involve at least some conflict of interest.”

In a 2011 paper with Bill von Hippel, Trivers developed this idea further, listing how self-deception can help. When trying to deceive, people may face cognitive load (the cognitive work required to make sure a web of lies does not have glaring contradictions). Given that lying is a betrayal of trust and is sanctioned when it is found out, it is risky, and people can get nervous about being found out, possibly showing signs of nervousness. Finally, people might try to mask signs of nervousness, thereby also behaving in a way that indirectly suggests lying. Self-deception, by inducing people to believe in their own lies, so to speak, can eliminate these possible clues while leading others to believe the preferred story of the person self-deceiving.

Trivers’ theory of self-deception has been supported by empirical research (including research I have contributed to). It explains what seems to be one of the most irrational patterns of human behaviour as emerging from strategic incentives.

Trivers has been one of the most influential evolutionary biologists, and his papers are still worth reading today. His insights, published more than 50 years ago, are fascinating. They often align very well with economic theories of behaviour, and it is therefore regrettable that his ideas are not more well-known in economics, and in particular in behavioural economics.

A key feature of Trivers’ take across these contributions was to see that beneath the world of social interactions we observe, there are deep structures in terms of incentives that shape the game we play. Understanding these games and their structures helps us make sense of the seemingly endless complexity of human psychology and social dynamics. In several key contributions, Trivers helped lift the veil on the underlying logic of human behaviour.

References

Coleman, R.M. and Gross, M.R. (1991) ‘Parental investment theory: the role of past investment’, Trends in Ecology & Evolution, 6(12), pp. 404–406.

Darwin, C. (1859) On the Origin of Species. London: John Murray.

Dawkins, R. (1976) The Selfish Gene. Oxford: Oxford University Press.

Dawkins, R. and Carlisle, T.R. (1976) ‘Parental investment, mate desertion and a fallacy’, Nature, 262, pp. 131–133.

Pinker, S. (2011) The Better Angels of Our Nature: Why Violence Has Declined. New York: Viking.

Sargent, R.C. and Gross, M.R. (1985) ‘Parental investment decision rules and the Concorde fallacy’, Behavioral Ecology and Sociobiology, 17, pp. 43–45.

Trivers, R.L. (1971) ‘The evolution of reciprocal altruism’, The Quarterly Review of Biology, 46(1), pp. 35–57.

Trivers, R.L. (1972) ‘Parental investment and sexual selection’, in Campbell, B. (ed.) Sexual Selection and the Descent of Man 1871–1971. Chicago: Aldine, pp. 136–179.

Trivers, R.L. (1974) ‘Parent-offspring conflict’, American Zoologist, 14(1), pp. 249–264.

Trivers, R.L. (1976) ‘Foreword’, in Dawkins, R. The Selfish Gene. Oxford: Oxford University Press.

Trivers, R. and von Hippel, W. (2011) ‘The evolution and psychology of self-deception’, Behavioral and Brain Sciences, 34(1), pp. 1–16.

Williams, G.C. (1966) Adaptation and Natural Selection: A Critique of Some Current Evolutionary Thought. Princeton: Princeton University Press.

Wynne-Edwards, V.C. (1962) Animal Dispersion in Relation to Social Behaviour. Edinburgh: Oliver & Boyd.

There are few thinkers whose writing remains relevant over time. Darwin is one of them. As our understanding of human behaviour progresses, it is surprising to see how much he got right about it. Most people associate Darwin with the view of evolution as being a fight of all against all. However, in his second book dedicated to humans, The Descent of Man, he presents a very insightful understanding of human psychology, which gives its full place to cooperative behaviour. Of particular interest is the fact that he describes the good reasons why cooperation may exist and why people tend to be willing to cooperate with others:

[A]s the reasoning powers and foresight of the members became improved, each man would soon learn from experience that if he aided his fellow-men, he would commonly receive aid in return. From this low motive he might acquire the habit of aiding his fellows; and the habit of performing benevolent actions certainly strengthens the feeling of sympathy, which gives the first impulse to benevolent actions. Habits, moreover, followed during many generations probably tend to be inherited. — Darwin

Darwin provided two answers. The first one was to stress the self-regarding consequences of altruistic actions: helping others would increase the likelihood of aid to oneself in times of need. The second one was to stress the importance to the individual of the praise and blame of neighbours in stimulating virtuous behaviour. Strikingly, these two explanations—unified in the Folk Theorem—form the backbone of game theorists’ and evolutionary biologists’ modern understanding of the roots of cooperation.

The Folk Theorem

The Prisoner’s Dilemma is one of the most famous games from game theory. In it, two prisoners are presented with opportunities to “rat” out the other to get a better deal from the police. They can choose to stay silent (cooperate with each other) or to talk (defect). The paradox of the Prisoner’s Dilemma is that both players have an individual interest in defecting. But if they follow their individual interest and both defect, then they end up in a situation where they are worse off than if they had both stayed silent. This is somewhat frustrating, and that is the reason it is a “social dilemma”. The more general lesson is that there are situations such that when everybody follows their individual interest, everybody ends up worse off than if they had cooperated.

Some quick thinking lets us realise that these kinds of situations are actually not that infrequent. Consider, in everyday life, a collective project whose outcome depends on people in a group working hard on it. Everybody wants the project to be successful, but for a given participant, it might be more convenient not to work too hard and piggyback on the work of others. Such incentives can lead group work to fail with not enough people pulling their weight for the project to be successful.

The result of players’ rational behaviour in a Prisoner’s Dilemma is a Nash equilibrium. It is a situation where, when all players anticipate a pattern of play from all players, none of them has an interest in changing strategy. In the group example, if group participants expect that others won’t work, they have no reason to work themselves. So the failure to get the group working is stable in the sense that everyone’s actions are compatible with their expectations about what others will do. But note that the fully cooperative situation where everybody works in the group might not be stable because if everybody works, some might consider that they might as well shirk and benefit from the fruits of the collective work nonetheless.

The Prisoner’s Dilemma was first studied in this form by Merrill Flood and Melvin Dresher, two analysts at the RAND Corporation.1 In early 1950, they invited two rigorous thinkers, Armen Alchian (an economist at UCLA) and John D. Williams (a RAND mathematician), to play 100 rounds of the game. Interestingly, the two did not simply play the one-shot Nash prediction (mutual defection) round after round. Instead, they experimented for a while, with Williams trying to induce Alchian to cooperate, and overall mutual defection was relatively rare.

Flood and Dresher asked John Nash what to make of this. He answered that the observed behaviour and the attempt at cooperating should not be seen as abnormal. The “100 rounds” should be treated as one big game, not as 100 separate one-shot games. In that setting, a strategy can include rules about how to react to what the other player has done. Conditional strategies of the form “do this if the other did that” can be an equilibrium of this larger game.2

Very early in the history of game theory, it became understood that repeated interactions create the opportunity to sustain cooperation. In their 1957 classic book on game theory, Luce and Raiffa described in this way the possibility of cooperation in the repeated Prisoner’s Dilemma (emphases mine):

[W]e see that in the repeated game the repeated selection of (cooperate, cooperate) is in a sort of quasi-equilibrium: it is not to the advantage of either player to initiate the chaos that results from not conforming, even though the non-conforming strategy is profitable in the short run (one trial).

It is intuitively clear that this quasi-equilibrium pair is extremely unstable; any loss of “faith” in one’s opponent sets up the chain which leads to loss for both players. — Luce and Raiffa (1957)3

This understanding was so widespread that the formal result, first published by Aumann (1959), was later called the Folk Theorem.

The folk theorem explained intuitively

The intuition of the folk theorem is very easy to understand. Imagine two gangsters who plan an exchange of money and drugs. They would have no qualms about robbing each other. If they have to organise a one-off transaction, they face a Prisoner’s Dilemma. Both would be happy if the transaction happens as planned, but each would be even better off if they could get the other gangster’s property without having to give theirs. Gangster movies are replete with scenes where such exchanges end up in betrayal, and one party ends up dead.

However, the strategic nature of this interaction changes radically if they have the opportunity to interact regularly over time (e.g. in a drug deal where one brings money and the other the substance). They can engage in regulated cooperation, respect rules, and benefit from these interactions.

The reason is that the prospect of possible future cooperation has a policing effect on the present. The “shadow of the future” is cast over present interactions.4 For fear of losing the gains from future cooperation, both gangsters may end up cooperating in each period.

The folk theorem with minimal maths

Consider the Prisoner’s Dilemma matrix of payoffs below. For each pair of actions, the cell contains the payoffs of Player 1 and Player 2 (in that order). The Nash equilibrium is the situation where both defect. It leads to a situation that is not “socially optimal”.

The insight from the Folk Theorem is simple: when play is indefinitely repeated (repeated without a clear end in sight), then players can do better than the Nash equilibrium of the stand-alone game. They can adopt strategies of conditional cooperation of the form: I will cooperate now if the other has cooperated before. The simplest form of such a strategy is the “grim strategy”: I will cooperate if the other person has always cooperated before. If the other has ever defected, I will defect.

Now the key is that there should be no clear end of cooperation in sight. There should be a high enough probability each period that the game will be played again, and both players should care enough about the payoffs in the future (they should not be impatient).5

When this is the case, the Folk Theorem indicates that, via strategies like the grim strategy, a lot more average payoffs can be obtained than the payoffs of the defect-defect scenario of one-shot games. Indeed, it is possible for the players to reach all the average payoffs in the grey area in the graph if they are patient enough and the probability of the game repeating is large enough.

The conclusion is that high levels of cooperative payoffs can be reached in repeated interactions, even in situations where no cooperation is actually stable in single interactions, if players care enough about the future. Players can adopt conditionally cooperative strategies, and it is the conditionality of these strategies that enforces the behaviour of the players in the present.

In short, the folk theorem says that once interactions are repeated, players can adopt a shared understanding of the type “I’ll be nice as long as you are nice”. This is sustainable and individually rational (no need to assume that people are altruistic saints).

An equilibrium here takes the form of a commonly agreed rule of behaviour. This rule determines how to act in each period. A given rule leads to an average payoff in the long run. Consider, for instance, the grim strategy rule: people cooperate as long as everybody has always cooperated. This rule delivers the payoff associated with full cooperation (C,C) over the long run because nobody ever deviates.6

Defining formally cooperation

Given the framework above, we are actually in a position to formally define what cooperation is. We do not need to say that cooperation is something like being “nice” (which would require defining what “nice” is) or that it is what is socially “right” (which would require defining what “right” is). We only need to observe that when players play (C,C), they are achieving a higher payoff by deviating from the individually rational strategy of the one-shot interaction. Hence, formally, cooperation can be understood as following an equilibrium strategy in a repeated game that yields higher social payoffs than the equilibrium payoffs of the one-shot game.7 In plain English, cooperation is not adopting a short-sighted, self-centred behaviour and instead trying to work with others to do better in the long run.

Self-sustainable via built-in sanctions

Sustained cooperation is made possible by the risk of potential sanctions in the future. And this potential risk of sanctions is credible because the one that would not implement them would also be punished. One might fear that this cannot work because this logic would lead to an “infinite regress”: If Alice does not do the right thing, Bob needs to punish her, and if Bob does not punish Alice, Candice needs to punish Bob, but if Candice does not punish Bob, you need somebody, like Dan, to punish Candice, and so on. If the population is finite, there is somebody who would have to punish others without fearing to be punished himself or herself if he or she failed to do so. But as stated by Ken Binmore:

[T]he proof of the folk theorem is explicit in closing the chains of responsibility. — Binmore (1998)

Bob might punish Alice because he fears Candice might punish him, and Candice might punish Bob because she fears that Alice might punish her! The punishment is often simply the withdrawal of cooperation, the severing of ties that lead to the social isolation of the punished.

If this description seems far-fetched, consider the social dynamics of “cancellation”. People who were the target of cancellation often explain that many of their friends and colleagues reach out privately to express support but did not dare to speak publicly for fear of also being sanctioned by their association with the person being cancelled.8 People fear not only the possible negative judgements from others about what they say, but also potentially the judgment from others about whether they are displaying the correct type of judgement themselves. In a cancellation event, everybody might express publicly what they think is the publicly appropriate attitude because they fear that everybody else might punish them if they do not conform.

Game theory and evolutionary biology

The logic of the folk theorem provides a unifying explanation of the emergence of cooperation as arising from the fact that, whenever people interact indefinitely over time, they can get better by deviating from their one-period best strategy. This deviation is sustainable as a rational strategy when players adopt conditionally cooperative strategies and would sanction deviations.

This explanation of cooperation is very familiar to economists. It is, somewhat surprisingly, less familiar to evolutionary biologists and psychologists who are often unaware that the folk theorem provides a general result that underpins much of what they discuss under labels such as reciprocity, indirect reciprocity, and reputation.

Kin selection. One important observation in animal behaviour is that cooperative behaviour often emerges between kin (e.g. parents-offspring, sisters, closely related kin groups). Here, the explanation from a game theoretic point of view is simple. Evolution selects strategies that maximise fitness at the gene level. Behaviour in one individual that promotes the survival and reproduction of the same genes in another individual will tend to be selected.

Kin share a lot of genes. Siblings share 50% of their genes, uncles and nephews, 25%, cousins 12.5%, and so on. This naturally pushes towards the selection of caring for the success of one’s kin. As the saying goes, “blood is thicker than water”.9

A simple way to think about it is that instead of caring about only your own payoff, you care also about the payoff of your kin to the degree to which you are related. As famously said by evolutionary biologist John Haldane:

Let’s take the Prisoner’s Dilemma above and let’s consider two situations: first the case where the gangsters are brothers, second, the case where the gangsters are twins. Let’s assume that the players are driven by their genes to maximise their genetic fitness. If they are simply brothers, they should care fully for their own material payoffs and care for 50% of the material payoff of their brother. If they are twins, they should both just care about the sum of their material payoffs. We can right away see that this “caring for the other payoffs” leads to the disappearance of the dilemma. Cooperation quickly becomes the best strategy in terms of evolutionary payoffs. The more related the players are, the more likely a Prisoner’s Dilemma situation turns into a trivial cooperative situation.10

Reciprocal altruism. In 1966, the evolutionary biologist George Williams published a book, Adaptation and Natural Selection, where he argued that a lot of apparently “group-beneficial” helping can, in principle, be compatible with natural selection when helping tends to be repaid inside stable social relationships.

The evolutionary biologist Robert Trivers built on this insight and published, in 1971, one of the most influential papers in evolutionary biology, The Evolution of Reciprocal Altruism. In it, he describes how conditional rules of the form “I scratch your back if you scratch mine” could naturally evolve and explain the prevalence of cooperation in the animal world.

For biologists, Trivers’ theory is the leading explanation of cooperation. Trivers’ insights are quite fascinating, and he discusses a rich texture of social behaviour which goes well beyond the result from the folk theorem in the simplified setting of the Prisoner’s Dilemma. He also grounded his arguments in compelling behavioural data about cooperation in the animal world across many different species.

However, it would be wrong to think that Williams’ and Trivers’ contributions were independent of the insights gained from game theory. Williams was aware of the discussion in game theory and specifically saw an organism’s behaviour as the best response to an environment shaped by other organisms’ behaviour. Even more tellingly, as indicated by Trivers himself, he was well aware of the literature in game theory where repeated interactions in a Prisoner’s Dilemma were being discussed and studied. He even quotes Luce and Raiffa’s book, where they had suggested that cooperation could emerge as a “fragile quasi-equilibrium”.

The relationship between two individuals repeatedly exposed to symmetrical reciprocal situations is exactly analogous to what game theorists call the Prisoner’s Dilemma (Luce and Raiffa, 1957). — Trivers (1971)

One can therefore see game theorists’ and Trivers’ contributions as complementary. The Folk Theorem provides the formal proof that cooperation is an individually rational strategy, which could therefore be sustainable in the long run, provided players can adopt appropriate rules of cooperation. Trivers made a compelling case about how such behaviour can emerge from evolution, how it is actually widely present in the animal world, and the type of mechanisms that ensure its sustainability.

Indirect reciprocity. In 1987, the evolutionary biologist Richard Alexander published the influential book The Biology of Moral Systems, where he argued that the one-to-one type of reciprocity from reciprocal altruism and repeated Prisoner's Dilemmas could not explain cooperation in large-scale society. Instead, he proposed the term indirect reciprocity to characterise the fact that we don’t only cooperate with people who have cooperated with us in the past but with people who also have cooperated with others.

Independently of this development in evolutionary biology, game theorists extended Folk Theorem results to interactions in large populations where people are rematched with others each period. In such a setting, cooperation can be maintained by keeping a public record of each person’s past cooperative behaviour and by making everybody’s behaviour conditional on these public records. An equivalent of the grim strategy in that context is that if one person ever defects, nobody ever cooperates with him again.11 Here again, strategies of conditional cooperation, like the grim strategy, can lead to cooperation being an equilibrium. Everyone cooperates with those who have cooperated in the past and everyone cares about their public record to be able to benefit from future gains from cooperation. These results provide a formal proof that the type of cooperation described by Alexander can be an equilibrium and therefore socially stable. 9

These extensions of the Folk Theorem to cooperation in large populations stress the importance of reputation: it is our public record as a good member of the community. It determines our future opportunities to benefit from cooperation in interactions with other people. Everybody cooperates with everybody having a good reputation because it is the best thing to do. Honesty is the best policy because a loss of reputation comes with substantial costs in the form of a loss of opportunities to cooperate with others in the future. For that reason, we care a lot about our reputation. It is built slowly and lost quickly.

Partner choice. In real life, we indeed choose, to a large extent, who our friends, collaborators, exchange partners, and so on are. Roberts (1998) and Baumard, André and Sperber (2013) developed an evolutionary account of cooperation that captures this fact. They made the case that understanding the logic of cooperation requires more than understanding what happens within a relationship; it requires understanding how relationships form in the first place. A lot of what people do is identify good partners and try to be selected as partners by others. Building a good reputation is key in that view, not just to convince our current partners to cooperate with us, but to find good partners in the first place.

The Folk Theorem has also been extended to such a setting.12 In that case, cooperation can be sustained by severing ties with defectors and refusing to match with them, and in some environments, leaving them unmatched. In short, we can shun people who have not proved to be cooperators. This helps explain why so much human activity involves gossip, that is, exchanging information about other people's good and bad deeds. While gossip has a bad name, its prevalence reflects the importance of forming good judgements about others to help us make good decisions when considering them as potential partners.

Implications

The Folk Theorem as a fundamental explanation of how and why cooperation emerges in evolution

The insight of the folk theorem is quite deep: opportunities for mutually beneficial cooperation naturally emerge when players face the prospect of interacting repeatedly over an indefinite period of time.

In a sense, the gains from cooperation are up for the taking. Purely self-centred agents who stumble on rules of conditional cooperation, or who are smart enough to devise such rules, will greatly benefit from working together.

The folk theorem captures the general strategic logic underlying most non-kin explanations of cooperation. The main genuinely distinct alternative is kin selection, where cooperation can be favoured even without the strategic enforcement logic of repeated interaction because genetic relatedness partly aligns evolutionary interests.13

There are endless ways to cooperate, hence culture matters

The folk theorem has two major implications. The first one is that cooperation is possible and can be individually rational. The second one is that cooperation can take many, many forms. As shown with the graph of the average payoffs reachable with the logic of the folk theorem, there is an infinite number of possibilities. More importantly, there are typically many ways to secure a given equilibrium payoff. That is, for every point in the shaded area, there are several rules of behaviour which will deliver that average payoff.

This result has been received with quite a bit of despair by game theorists. Having a unique equilibrium in a game allows game theorists to make a “prediction” about how rational players would play that game. Having an infinite number of equilibria means that the model does not narrow down a specific prediction. It is seen as a weakness of a game-theoretic model.

The answer in the profession has been to largely look away from repeated games. Instead, influenced in part by the rise of behavioural economics and the inclusion of psychology in economics, economists have looked at the behaviour of players in one-shot games. As players typically do not behave as purely self-interested and motivated by material interest, looking at one-shot games has provided the opportunity for a large literature to emerge on social preferences.

But this methodological move, understandable given the logic of the sociology of science, is most certainly misguided. Whatever social preferences we have, they must, in all likelihood, have been selected for us to successfully navigate situations of repeated interactions. In repeated interactions, our social preferences are useful for finding cooperators, being fair, and avoiding being seen as a defector. These social preferences likely drive people’s behaviour in one-shot experiments because one-shot interactions are unusual and people naturally use their built-in cognitive tools to try to make sense of these situations and decide how to act. By focusing on studying “social preferences” in one-shot games, economists have focused on simulacra: the effects of preferences forged to play in repeated games when activated in unusual situations where interactions are one-shot (and typically anonymous).

While this literature has produced a tremendous quantity of papers, it has likely been a wrong turn in terms of strategy for making sense of cooperative preferences driven by the desire of economists to say “something” in games with one or only a few equilibria.

A classic joke in economics recounts the tale of an economist looking for his keys under a lamppost. When someone asks where he lost them, he answers that he lost them quite a bit further away, but that it is only under the lamppost that there is light to look at the ground. Similarly, the fact that one-shot games often have one or very few equilibria made these games a preferred setting for the investigation of cooperative preferences even though the real explanations of these preferences lie in a very different area, the area of repeated games.

The fact that there are countless possible ways to cooperate should have led economists down another path, the path of understanding how cooperative equilibria are selected within a given society. This path requires thinking of how different cultural norms of cooperation evolve in societies. It is a path taken by a few economists whose influence in the discipline should be much more important: Robert Sugden, Christina Bicchieri and Ken Binmore.14

The game theory of cooperation explains the key characteristics of human cooperation: what it is and how it works. It provides a key insight: cooperation arises even between people who might have conflicting incentives because people can often produce higher total payoffs by working together. If they interact over time, they can design rules for conditional cooperation that allow them to produce these higher payoffs and to share them.

Seen in that light, there is nothing mysterious about cooperation. There is no need to assume that people are altruistic in a way that would conflict with the laws of evolution. Our moral sense, with its strong emphasis on reciprocity and conditional cooperation, is designed to navigate precisely these types of situations where gains can be obtained from cooperation if it is mutually enforced by agents playing nice as long as others do.

References

Alexander, R.D. (1987) The Biology of Moral Systems. New York: Aldine de Gruyter.

Aumann, R.J. (1959) ‘Acceptable Points in General Cooperative n-Person Games’, in Luce, R.D. and Tucker, A.W. (eds.) Contributions to the Theory of Games IV (Annals of Mathematics Studies, 40). Princeton, NJ: Princeton University Press, pp. 287–324.

Axelrod, R. (1984) The Evolution of Cooperation. New York: Basic Books.

Baumard, N., André, J-B. and Sperber, D. (2013) A mutualistic approach to morality: The evolution of fairness by partner choice. Behavioral and Brain Sciences, 36(1), pp. 59–78

Bicchieri, C. (2006) The Grammar of Society: The Nature and Dynamics of Social Norms. Cambridge: Cambridge University Press.

Binmore, K.G. (1994) Game Theory and the Social Contract. Vol. 1, Playing Fair. Cambridge, MA: MIT Press.

Binmore, K.G. (1998) Game Theory and the Social Contract. Vol. 2, Just Playing. Cambridge, MA: MIT Press.

Darwin, C. (1871) The Descent of Man, and Selection in Relation to Sex. London: John Murray.

Dawkins, R. (1976) The Selfish Gene. Oxford: Oxford University Press.

Ellison, G. (1994) ‘Cooperation in the Prisoner’s Dilemma with Anonymous Random Matching’, The Review of Economic Studies, 61(3), pp. 567–588.

Flood, M.M. (1952) Some Experimental Games. Research Memorandum RM-789-1. Santa Monica, CA: The RAND Corporation.

Gintis, H. (2006) ‘Behavioral Ethics Meets Natural Justice’, Politics, Philosophy & Economics, 5(1), pp. 5–32.

Kandori, M. (1992) ‘Social Norms and Community Enforcement’, The Review of Economic Studies, 59(1), pp. 63–80.

Laclau, M. (2012) ‘A Folk Theorem for Repeated Games Played on a Network’, Games and Economic Behavior, 76(2), pp. 711–737.

Luce, R.D. and Raiffa, H. (1957) Games and Decisions: Introduction and Critical Survey. New York: John Wiley & Sons.

Roberts, G. (1998) Competitive altruism: From reciprocity to the handicap principle. Proceedings of the Royal Society B: Biological Sciences, 265(1394), pp. 427–431.

Schumacher, H. (2015) ‘On Repeated Games with Endogenous Matching Decision’, Journal of Institutional and Theoretical Economics (JITE), 171(3), pp. 544–564.

Sugden, R. (1986) The Economics of Rights, Cooperation and Welfare. Oxford: Basil Blackwell.

Trivers, R.L. (1971) ‘The Evolution of Reciprocal Altruism’, The Quarterly Review of Biology, 46(1), pp. 35–57.

Williams, G.C. (1966) Adaptation and Natural Selection: A Critique of Some Current Evolutionary Thought. Princeton, NJ: Princeton University Press.

The whole Darwinists teaching of the struggle for existence is simply a transference from society to living nature of Hobbes’s doctrine of bellum omnium contra omnes [war of all against all].1 — Engels (1875)

The image of a war of all against all is a recurring framing of evolution. Thomas Henry Huxley, who became known as “Darwin’s Bulldog” for his staunch defence of Darwin’s theory, accepts this description and invites us to act against it by being ethical:

As I have already urged, the practice of that which is ethically best–what we call goodness or virtue–involves a course of conduct which, in all respects, is opposed to that which leads to success in the cosmic struggle for existence. In place of ruthless self-assertion it demands self-restraint; in place of thrusting aside, or treading down, all competitors, it requires that the individual shall not merely respect, but shall help his fellows; its influence is directed, not so much to the survival of the fittest, as to the fitting of as many as possible to survive. It repudiates the gladiatorial theory of existence. — Huxley (1893)

This view has often made it into social and political discussions, where evolutionary theory, and in particular its application to the study of human behaviour, is taken to imply a bleak vision of humans as being irremediably selfish. In his text The Western Illusion of Human Nature, the anthropologist Marshall Sahlins argues that evolutionary psychology has reinforced a Western trope of antisocial human nature:

The complementary idea that self-love is only natural has been reinforced lately by a wave of genetic determinism featuring the "selfish gene" of the sociobiologists and the revived Social Darwinism of the evolutionary psychologists. — Sahlins (2008)

This view is paradoxical. Anybody thinking that the core message of evolutionary theory is that we live in a war of all against all should stop and look around. Everything about life, as we can observe it, is produced by cooperation. Cooperation is not in conflict with evolution, it is generated by it.2

Cooperation is everywhere in nature

The evolutionary psychologist Nichola Raihani wrote a book, The Social Instinct, on the importance of cooperation in the natural world. In an interview about it, she stated:

The history of life on Earth is a history of teamwork, collective action, and cooperation. Cooperation pays because being part of a team is better than going it alone. — Raihani (2021)

The parliament of genes

Richard Dawkins’ The Selfish Gene made the point that genes are the units upon which evolution works. They are the replicators, which replicate and mutate. But in their drive for replication, genes “discovered” that they are stronger together. Instead of remaining isolated replicating molecules, they gathered in large “vehicles”, the cells and organisms, in which they share a fate and increase their chance of replication.

We are survival machines—robot vehicles blindly programmed to preserve the selfish molecules known as genes. — Dawkins (1976)

How cooperation is sustained. Within a cell or a multicellular organism, genes share a common “interest”. Their replication depends on the organism surviving, developing properly, and reproducing. Because genes that only “care” about their own replication tend to be selected, there is always the risk of mutations that increase their own transmission, potentially at the expense of the organism.

A standard example is meiotic drive (segregation distortion): during meiosis, a gene can sometimes bias transmission so it is inherited by more than 50% of offspring, even if this slightly reduces fertility or harms the organism. In short, genes can face prisoner’s dilemma-type tensions: cooperation is best for all, but cheating can be best for one, which makes cooperation fragile.

As a consequence, the cellular architectures that have emerged from evolution include mechanisms that detect and eliminate cheating. Distorters are countered by suppressors. This kind of internal governance has been described as a “parliament of genes” by evolutionary biologist Egbert Leigh in his 1971 book Adaptation and Diversity: Natural History and the Mathematics of Evolution.

The symbiosis at the heart of our cells

We tend to think of our cells as singular entities. The reality is that in each of our cells there are two different genomes collaborating. In addition to the one present in our cells’ nucleus, there is also the DNA present in mitochondria, the organelles in charge of producing the energy (ATP molecules) for the cell. These organelles, present in all eukaryotic cells (those of all animals, plants, fungi, and many single-celled organisms), are widely understood to be the remnants of an ancient bacterium absorbed by a cell and maintained in a symbiotic relationship.

The proteins that mitochondria need are largely encoded in the nucleus. The nucleus sends the instructions, the cell builds the proteins in the cytoplasm, and those proteins are imported back into the mitochondria. Meanwhile, mitochondria provide the energy that makes the whole operation possible. This division of labour within eukaryotic cells is one of the bases of complex multicellular life on Earth.

How cooperation is sustained. Because mitochondria have their own genome and replicate within cells, there is always a risk of cheating: some variants can favour their own proliferation even if that undermines cellular performance. The nucleus needs mitochondria to produce energy rather than prioritise their own replication. This is what economists call a principal–agent problem, where a principal relies on an agent to perform a task even though the agent’s incentives may not be fully aligned with the principal’s. Evolution has solved this problem by putting mitochondria under nuclear control and surveillance. Mitochondria depend on nuclear-encoded machinery, dysfunctional ones are preferentially cut off and removed by cellular quality control.

Stronger together: the emergence of multicellular organisms

After joining forces within cells, cells also “found out” that working together can be good for replication, and therefore for the genes they host.

Multicellularity likely started with simple clumps of cells staying together, sometimes in structures resembling microbial biofilms, where cells are glued together by a matrix of secreted molecules. One main benefit is a reduced risk of predation. Modern laboratory evolution experiments show that multicellular clumps are more likely to form when a predator is introduced into the environment.3 The mechanism is simple: clusters are harder to eat. Becoming larger, even crudely so, can be enough to survive.

But once cells remain together across generations, new possibilities open up. Some cells can specialise. Others can protect, transport, or support. What begins as a simple size advantage becomes a platform for internal division of labour and more complex organisation, eventually supporting the evolution of large multicellular organisms.

How cooperation is sustained. Multicellular organisms are complex things, but cooperation within them is relatively easy to maintain because cells share the same genes. They are clones, so their genetic interests are aligned. That is why nearly all cells in your body are set up to live and die selflessly, only so that your gametes can produce new organisms.

Interspecies cooperation

Once cells are united into multicellular organisms, another level of cooperation becomes possible: between organisms. Interestingly, such cooperation emerged even between organisms that are not closely related and belong to different species.

A classic example is the relationship between plants and fungi. Most plants rely on mycorrhizal fungi that attach to their roots. The fungi extend the plant’s reach into the soil, improving access to water and nutrients such as phosphorus. In return, the plant supplies sugars produced by photosynthesis.

Another well-known example is cooperative interactions between different fish species, such as cleaner fish and their clients. Cleaner fish remove parasites from larger fish, gaining food while providing a service. Clients choose which cleaners to visit and abandon those that cheat. Cleaners, in turn, adjust their behaviour to maintain a good reputation.

Many ecosystems depend on cross-species partnerships. In particular, more than 80% of land plants form mycorrhizal associations, which makes this one of the most widespread symbioses on Earth.

How cooperation is sustained. Because the partners are not genetically aligned, interspecies cooperation is always exposed to cheating: one side can take the benefit while reducing what it provides in return. What keeps this in check is strategic enforcement. Partners can stop cooperating, switch to alternative partners, or reduce the flow of benefits to poor performers. Plants, for instance, can shift carbon towards fungal partners that deliver more nutrients and away from those that deliver less, while fungi can preferentially supply hosts that provide more carbon. In cleaner-fish systems, clients simply leave and choose other cleaners, which makes cheating costly and rewards cleaners that maintain a good reputation.

Social animals

The benefits of cooperation are often easier to see within a species, especially when organisms live in small communities with close kin relationships. Social insects like ants and bees are a well-known case. Their genetic system can make sisters unusually closely related: in haplodiploid species, sisters can share around three quarters of their genes. That tends to align interests and can reduce some genetic conflicts, which makes cooperation easier to stabilise.

Cooperation also appears in many vertebrates, particularly those living in groups with repeated interaction and some degree of kin structure. Meerkats are a good example: individuals take turns acting as sentinels, watching for predators while others forage. Apes provide another set of examples, with grooming, alliance formation, and coordinated conflict. Among apes, humans pushed this further, relying on cooperation at a scale that far exceeds what our individual bodies would allow.

How cooperation is sustained. As with interspecies cooperation, cooperation within groups of social animals relies on strategic enforcement: cheaters can be detected and punished, excluded, or avoided in future interactions. Cooperation is also easier to sustain because many groups are kin groups, so members have partly aligned genetic interests. In social insects like ants and bees, haplodiploidy means full sisters can share around 75% of their genes, which makes their incentive to cooperate unusually strong and helps sustain extreme levels of cooperation.

Humans: the social brain advantage

Describing the ecological success of our early ancestor Homo habilis, E.O. Wilson said:

These slender little people, the size of modern twelve-year-olds, were devoid of fangs and claws and almost certainly slower on foot than the four-legged animals around them. They could have succeeded in their new way of life only by relying on tools and sophisticated cooperative behavior. — Wilson (1978)

The distinctive feature of human cooperation is that it can take place in large groups of strangers who are not closely genetically related, and therefore do not have closely aligned genetic interests.

This kind of cooperation is hard to achieve in new situations. It requires substantial cognitive work to understand the goals of others, to find mutually acceptable solutions, and to ensure that agreements are respected by all parties. The social nature of our environment is likely what created strong evolutionary pressure for improved cognitive abilities. Our ancestors faced an evolutionary arms race for cognitive capacities that help them take advantage of social cooperation and avoid being left behind or taken advantage of.

Those who were able to make friends, negotiate coalitions, and gain the trust of others were more likely to survive and leave offspring than those who did not. This is why the main driver of the outstanding ability of humans to reason was likely not to understand the world, but to understand and convince others.

How cooperation is sustained. Using their brains, humans have been able to sustain cooperation at a scale unmatched in the animal world, not just through one-on-one monitoring but through shared social rules that are followed and enforced collectively. This is a key insight from economist Elinor Ostrom. In potentially conflictual situations, humans show an ability to design and implement rules of cooperation, monitor compliance, and sanction defectors.4

Cooperation: what it is and what it means

Cooperation emerges at every level of interactions

It is striking that cooperation emerged naturally at every step of life: genes cooperate in cells, cells cooperate in organisms, organisms cooperate in groups. The repeated emergence of cooperation at each level of organisation is one of the most striking phenomena generated by evolution. Cooperation is a scaffolding principle of life. Without it, nothing like the living organisms we observe would exist.5

What insights can we take from this reality when thinking about what it means for human societies?

The illusion of a Hobbesian state of nature

The idea that a population of self-driven agents trying to get the best outcomes for themselves would be locked into a perpetual war of all against all misses that one of the best ways to thrive, even in a competitive environment, is cooperation. This is illustrated in one of the scenes of The Hunger Games, which purports to represent exactly a battle royale, a fight of all against all. As Peeta enters the training centre with Katniss, he reminds her of the importance of forming alliances:

Remember, Katniss, today’s about making allies. - Peeta Mellark in The Hunger Games: Catching Fire (2013)

The fact that one can be stronger together has not been lost by evolution. It has found architectural solutions to make cooperation between replicating units work and be resilient to attempts by some units to break cooperation to their advantage.

The Hobbesian state of nature is a misleading idea. It never existed, neither between humans, nor between other living organisms.6

The illusion of unconditional cooperation

While the pessimistic cynicism of the Hobbesian state of nature is misguided, another possible error is to take an overly idealistic view of cooperation, hoping for people to always be kind and cooperative with each other.

While cooperation is everywhere in life, it emerged not from unconditional altruism among cooperating units, but from the mutual advantage they gain from cooperating. As Raihani said in an interview:

Cooperation evolves because it allows entities to get ahead. And in the end, on a genetic level, there have to be somehow benefits that either derive back to the individual, or via its relatives, for cooperative strategies to be under positive selection. — Raihani7

Because units cooperate to get ahead, their divergent interest is always a possible threat to cooperation.8 Therefore, whenever cooperation is stable, it is kept in place by mechanisms that prevent deviations and organise the split of the gains from cooperation between units that each would benefit from getting a larger share of the collective benefits.

The implications of these insights for political ideologies’ take on cooperation

Because navigating social cooperation is required for humans to succeed in life, our evolved moral sense values cooperation. And, likely because being seen as a strong cooperator is advantageous in society, we like to think (possibly to convince others) that we are strong cooperators. Hence, the view that we should all just be nice with each other, and that cooperation should be pursued as an absolute “Good”, may have an intuitive appeal. This could help explain the recurrent appearance of collectivist ideologies in history, from the Essenes in ancient Judaea, to the Taborites in early 15th-century Bohemia, to the Diggers in mid-17th-century England, and later to communism in 19th-century Europe.9

Religions and political ideologies that simply call for humans to be good are not offering sustainable social solutions because cooperation requires mechanisms and institutions in place to keep individual conflicts of interest in check. Without these, attempts at unconditional cooperation are bound to fail, as illustrated by unsuccessful attempts at socialism, from kibbutzim to large national states.10

However, the total opposite view, which would deny the importance of cooperation and the tremendous gains reached through collective action, is also misguided. In a famous speech, the British and conservative Prime Minister Margaret Thatcher once said:

There is no such thing as society. There are individual men and women, and there are families. — Thatcher (1987)

This view is like saying that there are no cells, just genes. But, in the same way as genes were able to do much more by joining efforts into cells, humans have been able to do much more by cooperating together in large social groups. Indeed, it is only their ability to cooperate that made humans take over the world.

Evolution is too often described as implying only a competition between individuals. Contrary to this view, it is important to see how evolution has driven the emergence of cooperation at all levels of life because individual units can often be much more successful working together.

Appreciating the centrality of cooperation as a scaffolding principle of life is not naively idealising it or calling for it to happen out of sheer altruism in society. The lesson for us is twofold: cooperation can be highly beneficial, but it needs carefully crafted solutions to keep in check the individualist tendencies of cooperating agents. Evolution took eons to stumble on such solutions. Humans are in a unique position to use their agency and their cognitive capacity to engineer cooperation. The key challenge is to find institutional solutions to leverage the gains from cooperation and limit conflicts. What these institutions can be will be the topics of future posts in this series.

References

Boraas, M.E., Seale, D.B. & Boxhorn, J.E. (1998) ‘Phagotrophy by a flagellate selects for colonial prey: A possible origin of multicellularity’, Evolutionary Ecology, 12, pp. 153–164. doi:10.1023/A:1006527528063.

Dawkins, R. (1976) The Selfish Gene. Oxford: Oxford University Press.

Engels, F. (1859) Letter to Karl Marx, 11 December 1859, in Marx, K. and Engels, F. Marx and Engels Collected Works, vol. 40: Letters 1856–1859. London: Lawrence & Wishart.

Engels, F. (1875) Letter to Pyotr Lavrov, 12–17 November 1875, in Marx, K. and Engels, F. Marx and Engels Collected Works, vol. 45: Letters 1874–1879. London: Lawrence & Wishart.

Huxley, T.H. (1893) ‘Evolution and ethics’ (Romanes Lecture, 1893), in Huxley, T.H. (1894) Evolution and Ethics and Other Essays. London: Macmillan.

Kropotkin, P. (1902/1939) Mutual aid: a factor of evolution. London: William Heinemann.

Leigh, E.G. (1971) Adaptation and Diversity: Natural History and the Mathematics of Evolution. San Francisco: Freeman, Cooper.

Marx, K. (1861) Letter to Ferdinand Lassalle, 16 January 1861, in Marx, K. and Engels, F. Marx and Engels Collected Works, vol. 41: Letters 1860–1864. London: Lawrence & Wishart.

Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press.

Raihani, N. (2021) The Social Instinct: How Cooperation Shaped the World. London: Jonathan Cape.

Raihani, N.J. and Power, E.A. 2021, ‘No good deed goes unpunished: the social costs of prosocial behaviour’, Evolutionary Human Sciences, vol. 3, e40.

Sahlins, M. (2008) The Western Illusion of Human Nature: With Reflections on the Long History of Hierarchy, Equality and the Sublimation of Anarchy in the West, and Comparative Notes on Other Conceptions of the Human Condition. Chicago, IL: Prickly Paradigm Press.

Thatcher, M. 1987, Interview for Woman’s Own, 23 September.

Wilson, E.O. (1978) On Human Nature. Cambridge, MA: Harvard University Press.

5

The evolution of cooperation: its different types and time of emergence (colour coded from green to orange):

Where does morality come from? Why should we be moral? These questions are, surprisingly, mysterious given that moral beliefs and practices are pervasive in our lives.

For centuries, a straightforward answer in countries with monotheistic religions was: “it comes from God’s will”. Sacred books gave (often very detailed) injunctions about what to believe and what to do in order to be a “good” person.

The European Enlightenment witnessed a secularisation of thought in science and philosophy. In that context, the Scottish philosopher David Hume proposed a radically different view (which I develop on this Substack): morality is only a human affair. Moral rules are social conventions that make human interaction and cooperation possible.1 There is no need to invoke anything “out there” to explain why we care about morality and why we (most often) follow moral rules.

The view of morality as arising from social conventions can feel underwhelming. One might think that something reassuringly “absolute” is lost. Isn’t there a “Right” and a “Good” that we are meant to follow because they are true independently of what we think?2

Contra Hume, the German philosopher Immanuel Kant defended an absolute view of morality within a secular theory. Kant is often described as the central figure of modern philosophy. In his History of Western Philosophy (1945), Bertrand Russell described Kant as:

[Being] generally considered the greatest of modern philosophers — Russell (1945)3

and The Stanford Encyclopedia of Philosophy starts its article on Kant with the following description:

Immanuel Kant is the central figure in modern philosophy.

Kant’s philosophy concerns both what we can know about the world (Critique of Pure Reason) and what we should do with this knowledge (Critique of Practical Reason). In this second aspect of his work, Kant argued that the moral “ought” is unconditional and universal. Morality is not a matter of convention. It stems from human rationality.4

Kant’s project can be thought of as retaining the unconditional character of Christian moral prescriptions, without appealing to divine command. Once we abandon religion as a foundation of morality, Kant’s enterprise is one of the most influential attempts to recover an absolute view of morality. In his view, moral rules are necessary if we want to live as rational beings. Even if everybody disagreed with these rules, they would nonetheless be true.5

In spite of his fame, we can now say, with the benefit of two more centuries of work on rational decision-making, that Kant got it wrong. In his books on game theory and the social contract, Binmore offers a frank criticism of Kant, which I will broadly develop here.

[Kant] gets pretty much everything upside down and back to front. — Binmore (2005)

The categorical imperative

Kant aims to found morality on rationality. He explains his views in the Groundwork of the Metaphysics of Morals. The key idea he defends is that moral behaviour consists in following moral “oughts” that do not depend on incentives, enforcement, or local convention. These oughts stem from a single principle, the categorical imperative:6

Act only on that maxim through which you can at the same time will that it should become a universal law.7

There have been many criticisms of the impracticality of Kant’s moral principle. One of them, raised during Kant’s lifetime by the French political thinker Benjamin Constant, is that if one follows Kant’s categorical imperative, one ought to tell the truth to a murderer who knocks at our door and asks whether his intended victim is in our house. Kant replied without denying this implication. He stated that lying in such a situation would “do wrong to men in general”.8 So when a Nazi officer asks you whether you are sheltering Jewish children, the moral thing to do, on this strict reading, is to tell the truth. If you lie, what next? Trust in each other’s words would crumble in society!

Here, I present a criticism that is more fundamental than these familiar objections. My claim is that Kant’s assertion that the categorical imperative follows directly from human rationality is misguided.

The contractual intuitions behind Kant’s morality

Before taking down Kant’s argument, it is worth appreciating its intellectual appeal. There is a contractualist intuition behind Kant’s morality. This intuition is not the one Kant uses as an explicit argument, but I think it is what made his view compelling.

Because we live in society, we need rules to organise social interactions. What rules defining duties and rights can we endorse? If these rules have to be accepted by all, and if we have similar bargaining power, a simple logic emerges: we cannot defend rules that give rights to ourselves and impose duties on others in a way that does not apply symmetrically. Hence, in a society of equal members, the only rules of behaviour that we can claim to be allowed to follow are those that we must accept could be followed by others.9

This contractualist intuition makes Kant’s argument appealing, but it is important to distinguish two different questions:

1. The social question (a social contract):

   How should we all agree on shared social rules—a “social contract”—in a given community?

   The contractarian position I defend in this series is that morality exists to sustain such a social contract. That is why we intuitively feel that we cannot claim a right to do things to others that we would not accept them doing to us.

2. The individual question (a rule for personal conduct):

   How should you behave as an individual?

   This is an individual choice you make for yourself, not a collective choice you make with others.

The intuitive appeal of Kant’s answer comes, I think, from implicitly channelling our contractual intuitions about the social question into the individual question. But Kant is not searching for social rules to adopt as a group. He is looking for a principle of behaviour that each person should adopt independently. Kant’s answer is misguided because the contractarian answer to the social question is not an answer to the individual question.

The flaw in the argument

In Natural Justice, Binmore makes this point, which is ironic but goes to the crux of the flaw in Kant’s argument:

Kant claimed that a truly rational individual will necessarily observe one particular categorical imperative: “Act only on the maxim that you would at the same time will to be a universal law.” My mother had similar views. When I was naughty, she would say, “Suppose everybody behaved like that?” Even to a child, the flaws in this line of reasoning are obvious. It is true that things would be unpleasant if everybody were naughty, but I’m not everybody—I’m just myself. — Binmore (2005)

The key point is that Kant tells us it is rational to follow a rule that would work well if everybody else were following it. But our adoption of that rule is a decision we make on our own, with no effect on other people’s decisions. What if others do not follow suit? Is it still rational?

Starting with a well-known example: the Prisoner’s Dilemma

At this point, we benefit from the clarity that game theory brings to thinking about rational decisions in social interactions. The Prisoner’s Dilemma is a particularly useful case for testing Kant’s conclusion. The dilemma involves two people who can either cooperate with each other or not.10

I will use a version of the dilemma that is, deliberately, stripped of moral sentiment between the players.11 Consider two gangsters who plan to exchange diamonds for cash. They can only meet briefly in a train station. Because the place is public and the meeting must be quick, they can only swap briefcases and leave, without checking what is inside. Each gangster faces a choice: he can honour the deal (cooperate) and bring the promised item, or he can bring an empty briefcase (defect) in the hope of getting the other briefcase for free.

Assume the two gangsters do not know each other and will travel to different countries immediately after the exchange, never to meet again. Their decision has no future consequences for them. What should they do?

Because they have no personal connection, they do not care about each other’s wellbeing. In that case, it quickly becomes clear that the individually rational choice for each is to defect. Consider the buyer. Two states of the world are possible: either the seller brings the diamonds, or he does not. In either case, the buyer is better off giving nothing, whether he receives diamonds or nothing. The seller faces the same logic: he is better off bringing nothing.

The dilemma is that if both follow their individual rationality, the exchange fails, even though both would prefer the outcome where the cash and the diamonds are successfully traded. A tremendous amount has been written on this problem. Because the conclusion is uncomfortable, many authors have tried to argue that it would somehow be “rational” to cooperate. We need to put this idea to rest: it is an intellectual dead end. If we assume that the players do not care for each other and will not interact in the future, the only rational strategy (the best strategy for them to get what they care about) is to defect.12

Kant’s categorical imperative conflicts with this conclusion. It suggests that cooperation is the rational choice because it is the kind of action one could accept as a rule for everyone. If both cooperated, the exchange would take place; if both defected, it would not.

But the structure of the Prisoner’s Dilemma shows why this is not a usable principle for individual action. The problem is that choosing a maxim because it would be good if it were universal does not make it universal.13 In the gangster exchange, the only thing each person can choose is his own action. There is no point saying “this would be a great action if everybody chose it” if the other person has no reason to choose it.

Choosing to follow the categorical imperative would lead to a better world only if my choice could somehow influence others’ choices. But that is not the case. In the example above, one gangster’s decision to cooperate does not produce some “spooky action at a distance” that makes the other gangster cooperate as well.14

Criticising the core of Kant’s argument about the categorical imperative

Let us now turn to Kant’s argument that the categorical imperative stems from the requirements of human rationality.

[T]he will of every rational being is necessarily bound to it as a constraint. — Kant

Binmore takes this argument seriously. What should a maxim of behaviour chosen by a rational being be? The notion of rationality is not simple,15 but we can start from a minimal idea: behaviour is rational if it can be rationalised, in the sense that one can give reasons for it.16

A formal insight from economics is that rationality can be characterised, to some extent, by consistency. If a person’s choices satisfy certain basic consistency requirements, those choices can be described as if the person had a stable objective.17

Binmore begins from this point and asks whether Kant’s categorical imperative follows naturally from the search for consistent principles of behaviour. A first (small) problem is that Kant’s test is applied maxim by maxim. That leaves room for overly specific maxims. For instance, regarding the challenge of whether you should tell the truth to a Nazi looking for Jewish people hiding in your house, one could adopt the maxim:

Never tell a Nazi who is looking for innocent victims where to find them.

That is, of course, a defensible maxim. But it illustrates the problem: once we allow very narrow conditions, it becomes easy to tailor a maxim to reach the conclusion we want. If the conditions are specific enough, the “universality” of its content can become almost empty, because it may in practice apply only to one person.18 The question then becomes: why these conditions, and why this maxim rather than another? In practice, it looks as though we need further principles to justify the way we chose to describe the maxim in the first place, for example a principle about protecting innocent people.

For this reason, Binmore does not consider maxims individually, but as a whole. If we consider the full set of principles that would be implied by the categorical imperative, we end up with a system of rules (maxims) guiding action.

So consider an individual choosing a set of rules to follow. A simple requirement of consistency is that the set of rules should not undermine itself: following those rules should be compatible with continuing to follow them. Put differently, a coherent system of rules should be self-justifying. If you apply your own rules to the question “which rules should I adopt?”, they should not force you to abandon themselves. Binmore expresses this idea as a rational imperative to choose for yourself a bundle of maxims (my phrasing for simplicity).

Choose rules for yourself that you can stick with when you reflect on your own rule-choosing.

If one associates rationality with avoiding inconsistency in the principles guiding behaviour, this requirement is not controversial. It may sound abstract, so an example helps. Suppose your overarching rule is:

In every situation, choose the option that best advances your aims, given what you believe about the situation.

That rule is compatible with continuing to follow it. By contrast, consider an overarching rule such as:

Always do the opposite of what this system recommends.

That system cannot be self-justifying. If you try to follow it, it tells you not to follow it.

Now we can ask a similar question at the social level: what rules could everyone adopt together, and still have reason to keep following? An answer would be:

Choose rules for society that people can keep sticking with once everyone has adopted them.

In other words, if we had to choose as a society, we should prefer rules that are self-sustaining, in the sense that if everyone adopts them, each person still has reason to keep following them rather than immediately wanting to switch.19

This addresses the social question: what rules can serve as a stable social contract? But Kant is not trying to answer that question. He is trying to answer the individual question: what rules should each person adopt for themselves?

Kant’s prescription is that you, as an individual, should choose rules for yourself that you would accept everyone using. Binmore rephrases Kant’s categorical imperative as:

Act on the basis of a system of maxims that you would at the same time will to be universal law. (I emphasised the addition of “a system of”)

We can rephrase this, for simplicity, as:

Choose rules for yourself that you could at the same time will everyone to use.

The problem is what “will” means here. You may want others to follow your principles, but your choice of principles has no causal effect on what they choose. You can decide that a rule would be appropriate for everyone, but that decision does not make anyone else follow it. You cannot “will” other people into compliance just by endorsing a rule.

In the Prisoner’s Dilemma, the buyer might decide to cooperate because the exchange succeeds if everybody cooperates. But the buyer’s choice of maxim has no effect on the seller’s choice of maxim.20 As Binmore puts it:

[I]t is impossible for an individual to make what Kant calls a “natural law” that everybody should behave as though they were clones. Indeed, it seems to me quite bizarre that a call to behave as though something impossible were possible should be represented as an instruction to behave rationally. Far from expressing a “rational preference”, the “will” of Kant’s categorical imperative seems to me to represent nothing more than a straightforward piece of wishful thinking.

Binmore’s complaint is that Kant’s test tempts us to choose the “best” action by imagining an ideal world in which everyone complies. To caricature, one might summarise the categorical imperative as:

• Choose to behave in a way that would be great if everybody else were doing it.

But if you are looking for a rational answer to the individual question, you should instead look for the best action for the individual, given what others are actually doing. If others are like you, they will also try to choose what is best for them given what everyone else is doing. In other words, looking for a consistent rule of behaviour from the individual point of view leads to a different kind of categorical imperative, one in which people aim to make decisions consistent with their preferences, assuming others are also doing the same. Binmore proposes something like (my phrasing):

• Choose to behave in a way that is best for you, given everyone else’s choices, assuming that others are also choosing what is best for them, given everyone else’s choices.21

An individually rational rule of choice that could be adopted by all is one in which everyone chooses what is best for them (given their preferences), assuming that all other people are also trying to do the same.

So this maxim tells you that it is rational to behave in a way that assumes realistic behaviour from other people (and not idealised behaviour), with each person also having realistic assumptions about the behaviour of others. A situation where everybody mutually best responds to other people’s behaviour is a Nash equilibrium. For that reason, Binmore calls this second imperative the “Nash categorical imperative”.

Binmore is not presenting this as a moral principle. He is presenting it as a requirement of rational choice in strategic settings. It does not ask you to assume that others will behave more cooperatively than their incentives allow. A “rational” rule of choice, in this sense, is one that can be used by everyone at once, with each person still having reason to stick with it given that everyone else is doing the same.

Rationality is playing the best response against actual behaviour

The categorical imperative recommends behaving in the best possible way IF everybody were willing to behave that way.22 But in reality, people may have good reasons not to. If so, what happens if you follow the categorical imperative and end up being the only one who does? In many settings, doing the “right” thing in that sense can lead to worse outcomes when others do not follow you. Here are two illustrations.

Illustration 1: a mutual deterrence standoff

A specific type of equilibrium occurs when two opponents are locked into a mutual deterrence situation. In gangster and Western movies, it takes the form of a “Mexican standoff” with opponents aiming guns at each other. In international relations, nuclear deterrence between powers capable of devastating retaliation has the same structure.

It is easy to represent such situations with the logic of a Prisoner’s Dilemma: both sides would be better off if neither threatened the other, but whatever one side does, the other has incentives to keep its guard up, either for protection or to gain advantage if the other side lowers its guard.

Kant’s categorical imperative would suggest adopting the behaviour we would like everyone to adopt. Since mutual peace is the best outcome, I should “rationally” lower my guard.23

Any experienced movie watcher will tell you that lowering your guard in a Mexican standoff is usually a terrible idea. It does not cause the other person to lower his guard. It leads you to be his prisoner or to end up dead. In international relations, few people find the suggestion that the United States should have unilaterally disarmed during the Cold War plausible. Indeed, Ukraine’s abandonment of nuclear weapons on its territory in 1994 did not lead others to do so, but instead allowed Russia to invade it a few decades later in a war that led to hundreds of thousands of deaths.

Illustration 2: the Nader-Gore choice in 2000

Suppose you are a progressive-minded voter in the United States in the year 2000. You can choose between two candidates: the Green Party’s Ralph Nader and the Democrat Al Gore. Suppose you believe Nader has the best policies. The categorical imperative tells you that you should vote for him because, if everybody voted for him, the United States would be a better place.

The difficulty is that Nader, at around 3% in the polls, had little chance of winning. Gore, on the other hand, was in a very close race with George W. Bush. In that setting, voting “on principle” for Nader might not affect Nader’s prospects, but it could contribute to Bush defeating Gore.

In the end, this is roughly what happened. The result was decided in Florida, where Bush won by only 537 votes, while 97,488 votes were cast for Nader in that state. This is one of many examples where people doing “the right thing” while ignoring what others plan to do end up with worse consequences all around (from the perspective of the people making the decision).24

Here is how some Nader voters later described their choice:

At the time, I thought I was making a principled stand, but it’s a vote I have come to regret. — Josh Brodesky25

Then that night I really regretted it. I was in Pennsylvania too, so it’s like a consequential vote. — A radio talk show caller26

Smoking is not the worst choice I made that year. I voted for Ralph Nader. — Rachel Combe27

Kant’s argument that it is individually rational to behave in this way does not hold up. The categorical imperative has intellectual appeal, largely because it resonates with our contractarian intuitions about what rules could be shared. But those intuitions are not answering Kant’s question. Kant is not asking how we should choose rules together, but what each person should adopt for themselves. By asking us to behave best not in the world as it is but in the world as it should be, Kant’s categorical imperative risks asking us to behave in an irresponsible and deluded manner.

Modern thinking about rationality, as developed by game theorists, allows us to scrutinise Kant’s claim more sharply. His idea that the categorical imperative follows from rationality does not survive that scrutiny. On this point, Binmore concludes that once one gets past the difficulty of reading Kant:

Kant’s assertion that one should honor this principle for rational reasons is conjured from the air. — Binmore (2005)

No unconditional oughts

Kant’s biggest claim in moral philosophy is that sheer rationality generates an unconditional ought: I ought to follow the categorical imperative simply because I am a rational being.

The idea of an unconditional ought may feel especially attractive in societies shaped by Judeo-Christian heritage, where moral rules were long presented as commands imposed from above by a divine being. Without the appeal to such a being, many philosophers, including Kant, tried to retain the idea of unconditional obligation. But we have seen that Kant’s attempt fails: his categorical imperative does not follow from rationality.

Modern theories of rational choice provide a clearer foundation for what “ought” can mean. Rationality is always about doing the right thing to achieve what one wants to achieve (whatever it is). In that sense, we can talk about “oughts”, but only as conditional oughts: IF I want to achieve my goal A, THEN I ought to take action B.

When people interact, these conditional “oughts” must incorporate what others are likely to do. Any stable pattern of social interaction between rational agents will take the form of a Nash equilibrium: people persist in their behaviour because it is best for them given what others are doing, and this must be true for everyone.

In line with the Humean tradition, morality can be understood as a set of cultural conventions that regulate social interactions. Their features have been selected over time because they help realise gains from cooperation. These conventions are equilibria and therefore self-enforcing. In a society with functional moral rules, it is often in each person’s interest to follow them in order to avoid social sanctions. Hence, there is a conditional ought: IF one wants to remain an accepted member of society, THEN one must abide by the moral rules of that society.

This answer is often described as underwhelming, but that is partly because it is an ultimate explanation: it explains what morality is at a fundamental level. It does not necessarily match our moral experience. To understand how morality feels from the inside, we need a proximate explanation: an account of the emotions and intuitions that guide us. We have evolved a moral sense that helps us navigate social life. In the same way that fear or hunger pushes behaviour in directions conducive to survival, social emotions such as guilt and empathy help us manage cooperation, conflict, and reputation. As a result, we do not experience moral decisions as interest-driven. We often experience them as the requirement to do what is right, and the fear of doing wrong when we violate a rule. Paradoxically, those proximate feelings may also make the ultimate explanation of morality feel disappointing. But our intuitions should not determine what we accept as the best theory.

I have the intuition that the world is solid and well-defined. I also find the picture of nature provided by quantum theory unintuitive and unappealing. But IF I want to navigate the world successfully, THEN I ought to accept the best theory available. From that perspective, the contractarian theory of morality, particularly as articulated by Ken Binmore, is the best available in its ability to account for what morality is, how it works, and how it evolves.28

Kant’s figure looms large in Western philosophy. His attempt to found morality in sheer rationality seems deep and impressive. But discussions of rationality are complex and easily confused. With two centuries of work in decision theory and game theory behind us, we can move beyond Kant’s daunting prose and identify what is wrong with his argument. The categorical imperative encourages us to behave best in the world as it should be, not in the world as it is. In many situations, that is not a sensible guide to individual choice. It can lead us to act in ways that fail to achieve what we want, and sometimes in ways that worsen outcomes for others. It is not clear in what sense this is either rational or moral.

One way to place Kant’s morality in the history of ideas is as a rearguard attempt to retain, in secular form, the absolutist flavour of Judeo-Christian moral obligation. The view I defend in this series is that Hume’s alternative was fundamentally right: morality can be understood as emerging from a social contract, an implicit agreement about how to behave properly in society. Game theory helps us understand the logic underlying social interaction and, in doing so, helps us develop a theory of morality that fits our practices and beliefs. Unlike Kant’s views, this theory is stripped of mysterious content and does not require us to assume the existence of absolute moral principles of some kind that we would have to discover.

Having located this approach within the big questions of moral philosophy, I will develop it in detail in the next few posts.

References

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Gilboa, I. (2009) Theory of Decision under Uncertainty. Cambridge: Cambridge University Press.

Guyer, P. (2017) ‘Kant’, in Zalta, E.N. (ed.) The Stanford Encyclopedia of Philosophy (Summer 2017 Edition). Stanford: Metaphysics Research Lab.

Hume, D. ([1751] 1998) An Enquiry Concerning the Principles of Morals. Oxford: Oxford University Press.

Kant, I. (1797) ‘On a supposed right to tell lies from benevolent motives’.

Kant, I. ([1785] 2017) Groundwork of the Metaphysics of Morals. Translated by J. Bennett. Early Modern Texts.

Page, L. (2022) Optimally Irrational. Cambridge: Cambridge University Press.

Quattrone, G.A. and Tversky, A. (1984) ‘Causal versus diagnostic contingencies: On self-deception and on the voter’s illusion’, Journal of Personality and Social Psychology, 46(2), pp. 237–248.

Riker, W.H. (1962) The Theory of Political Coalitions. New Haven, CT: Yale University Press.

Russell, B. (1945) A History of Western Philosophy. London: George Allen & Unwin.

Weber, M. ([1919] 1948) ‘Politics as a vocation’, in Gerth, H.H. and Mills, C.W. (eds) From Max Weber: Essays in Sociology. London: Routledge & Kegan Paul, pp. 77–128.

13

As I wrote in a previous post:

Proposed political solutions that assume everyone will simply “do the right thing” are a red herring. A central challenge in political theory is that people typically do not all just want to do what might be in the best from a collective point of view.

Last November, I wrote that the possible election of Donald Trump was a risk for the quality of US democratic institutions. Unfortunately, it is now clear that the first year of this administration has far exceeded such negative expectations.

One aspect where the Trump administration has been especially destructive is the US-led rule-based international order. It is an order the US largely built after 1945, and one from which it has benefited disproportionately, through a dense web of rich allies that extend American geopolitical reach. Over the last year, the US administration has echoed Russia’s narrative about its war in Ukraine and pushed for a deal that would require Ukraine to give up parts of its territory, and it has treated Europe aggressively, from the imposition of tariffs to threats over Greenland.

In the last few days, Trump has again refused to rule out taking Greenland by force.2 In a recent CNN interview, White House Deputy Chief of Staff Stephen Miller defended this posture in blunt terms:

[W]e live in a world, in the real world that is governed by strength, that is governed by force, that is governed by power.

Miller is, like so many professed “international realists”, too smart by half.3 It is true that in international relations, there is no global government and no global police to enforce rules in the way domestic law is enforced. But those who conclude from this that the world is simply a Hobbesian free-for-all are misguided about the strategic nature of international relations.4

Even without an external enforcer, countries typically negotiate a modus vivendi: a set of conventions that stabilise expectations and make cooperation possible. They do not do this because they are idealists. They do it because cooperation is mutually beneficial when it is sustained over time. A forceful seizure of Greenland would not be “realism”. It would be short-sightedness: an action that gains little, and risks breaking the conventions that underpin America’s alliances and its long-run power.

Scenario for a takeover

Let’s consider what would happen in the case of an attempt at a military takeover of Greenland. Would it be politically possible for the administration to do so within US institutions? And, if it happened, what would the geopolitical costs be for the US?

Read more

In a 2015 US poll, 65% of American adults agreed that “every culture must determine what is acceptable morality for its people.”1 And in a 2023 survey, 60% of “Gen Z” respondents agreed with the idea that what is morally right or wrong depends on what an individual believes.2

In academia, this view is quite prevalent in some parts of the social sciences, like anthropology. In 1887, Franz Boas, one of the founders of the discipline, wrote in the journal Science:

It is my opinion that the main object of ethnological collections should be the dissemination of the fact that civilization is not something absolute, but that it is relative, and that our ideas and conceptions are true only so far as our civilization goes. — Boas (1887)

I have argued that morality emerges from social conventions and that a society’s moral system can be seen as a social contract, implicitly agreed upon by all to coordinate and cooperate. This view is often labelled Humean constructivism, as it was first put forward by Hume. Its proponents, like Robert Sugden and Ken Binmore, also use the label contractarian, since these social conventions can be described as a social contract, implicitly agreed upon by all.3

In this post, I address a criticism often levelled at this approach: it is “relativist”. With this term come accusations like these: If morality is just conventional, isn’t it the case that some societies can land on conventions we would find abhorrent, like infanticide, slavery, or genocide? In that case, would we not just have to accept these practices as fine?

The point I make here is that contractarianism is relativist in the technical sense used in metaethics, but not in the way the term is usually understood in everyday debate.

What “relative” means here

In metaethics, the branch of philosophy studying the foundations of moral claims, relativism is defined as the belief that there is no absolute moral truth. The Stanford Encyclopedia of Philosophy defines this viewpoint as:

The truth or falsity of moral judgments, or their justification, is not absolute or universal, but is relative to the traditions, convictions, or practices of a group of persons.

Under this definition, it is clear that a contractarian view is relativist. In a contractarian position, moral claims can be true or false, but only within a given moral system, a specific social contract. As a result, a moral claim that is true in one moral code can be false in another.4

Philosophers usually do not like moral relativism.5 To be frank, I do not like the term either because it is too often understood as meaning some sort of nihilism. The contractarian approach focuses on explaining the reality of moral rules and moral practices as part of a social contract, not on denying any relevance to moral considerations.

Can we condemn slavery/infanticide? Yes, in three coherent ways

There are many strands of moral philosophy that can be described as “relativist”. For some of them, which I would label “radical relativism”, different cultures cannot be compared in any way.6 This conclusion is not one that follows from the contractarian perspective I am presenting.

From within our own moral preferences

The contractarian position implies that there is no higher moral ground outside societies upon which we can stand to judge them, but it does not mean we are not entitled to our own preferences. Binmore, who, unlike me, is happy to defend the term relativism, says about it:

Moral relativism does indeed deny that it makes sense to say that one social contract is better than another in some absolute sense. But moral relativists nevertheless have their own opinions about the kinds of social contract they think desirable — Binmore (1998)

Nothing prevents us from liking or disliking another country’s practices based on our own moral preferences. Some of our fundamental moral preferences might have been shaped by evolution and be universal across cultures, while others are culturally acquired via education and socialisation.7

From an international social contract

Some (limited) order exists at the international level. In their interactions, countries follow norms that are both implicit (e.g. diplomatic usage) and explicit (e.g. international law established in treaties). These rules create a meta-normative system. While this normative system is largely silent on how countries handle matters within their borders, it does impose some constraints on practices that many states have come to view as among the most abhorrent.

Genocide is explicitly prohibited by the 1948 Genocide Convention of the United Nations, which states that it is “a crime under international law” and obliges states to prevent and punish it. The International Criminal Court has jurisdiction over genocide under the Rome Statute.

Slavery is prohibited under the 1926 Slavery Convention and the 1956 Supplementary Convention, which commit states to abolish slavery and suppress slavery-like institutions and practices.

Infanticide: International law does not use the term “infanticide”, but it protects children’s right to life. For example, the Convention on the Rights of the Child recognises every child’s “inherent right to life” and requires states to ensure the child’s survival and development.

It is important not to overstate the meaning of these international rules. International law is too often presented as an absolute yardstick of right and wrong. Instead, it is again a convention (between states). And this convention works only to the extent that the balance of states’ incentives limits infringements. A state willing to violate these rules and willing to face the negative consequences can do so.8

By outcomes and stability

Perhaps most importantly, different social contracts can be compared in terms of their practical consequences. Social contracts’ function is to regulate cooperation and conflict between members of society. Efficient moral systems can induce higher levels of cooperation and therefore higher welfare in society.

It is perfectly reasonable to compare moral norms to the extent that some can lead to better social outcomes than others. Consider, for instance, the driving rules in Singapore and Cairo. These rules are both formal (likely fairly similar in both countries) and informal, in particular, what people consider acceptable to do (how often to honk, how much to respect the rules, how gracious to be). Singapore driving norms seem unambiguously better, in the sense that they appear to be associated with less driving chaos and better driving times.9

Once again, this aspect should not be overstated. First, such a criticism is not made from a higher moral high ground, and it does not assume that “efficient cooperation” is a good that should be pursued as an absolute moral goal. Instead, this criticism can be made from the point of view of the society concerned. Saying that the Singapore equilibrium is better than the Cairo equilibrium is saying that Cairo dwellers would themselves prefer to live with the Singapore equilibrium if they could switch to it as a society. Second, efficiency does not a priori preclude some practices in other societies that we find abhorrent in ours.10

How moral realists use “relativism” as slander

Moral realists, lacking a settled account of what absolute moral truths are and how we would know them, often resort to dismissing scepticism towards moral realism with what Binmore calls “emotive slander”. They use the label “relativism” to accuse non-moral realists of supporting a vision of reality where things we find abhorrent, like infanticide and slavery, are fine.

I have been accused of holding that pedophiles are justified in disregarding the law when pursuing their claim that prepubic sex is good for children. Similarly, relativists supposedly see nothing wrong in keeping slaves or wife-beating, since both activities have been endorsed as morally sound by past societies. — Binmore (1998)

The contractarian view is not that these things are OK. It is that there is no external moral rule book that conveniently provides us with the rules on how to behave to be Right or to do Good. Instead, we hold our morality in our own hands. As social groups, we have to agree on how to work with each other to facilitate cooperation and resolve conflicts. Morality is fundamentally about adopting a social contract, an agreement about how to organise social life. “Correct” social rules are not written in the fabric of the universe; we are on our own and in charge of our own rules.11

What the “relativism” in contractarianism is not

I have indicated that the contractarian approach is relativist in the technical sense used in metaethics. However, the word “relativism” is loaded with different meanings. It is useful to make clear what the versions of this term that contractarianism does not endorse.

Not subjectivism

A frequent interpretation of moral relativism is subjectivism, which can be summarised by the idea that everyone has his or her own truths. This idea does not reflect how morality works in real life, and for a contractarian, it is logically inconsistent.

[M]oral subjectivism is absurd because it overlooks the fact that moral rules evolved to help human beings coordinate their behavior. But successful coordination depends on everybody operating the same moral rules. If everybody in a society made up their own standards, there wouldn’t be any point in having moral rules at all. — Binmore (2005)

If moral views were just things that can differ in each of our heads, how would they work? Why would Alice care when Bob tells her she is wrong, according to his own principles of morality? Why should she grant this claim more significance than if Bob were pointing out that he prefers chocolate ice cream, contrary to her choice of vanilla flavour?

Not relative to “culture”

A slight variant of subjectivism is that morality is just a characterisation of a culture in the sense that people from different cultures have different ideas.

The view of morality as convention requires more than morality being about ideas in people’s minds: moral rules must be equilibria in social games played in a given social setting. This requirement means that if people from different cultures live together in a given society, each cultural group cannot have its own morality. The idea that cultures can simply coexist side by side with their moral code in a society is as misguided as the subjectivist idea that people can live side by side with their idiosyncratic moral views.

Let’s take a simple example. Consider a society in which two groups have internalised different cultural expectations about queues. One culture treats strict queueing as a basic rule of fair access: first come, first served. The other group treats the “queue” as more of a loose signal, with access negotiated through persistence and a bit of elbow-work. If you queue diligently among people in that group, you may find that they move ahead of you without treating this as a serious breach of conduct.12

Imagine that, because cultures are said to be “equal” and not assessable from outside, members of the second culture are permitted to ignore the queue when interacting with everyone else. This would not lead to the peaceful coexistence of two norms in the same shared space. It would more likely unravel the queueing equilibrium altogether: once compliance is no longer common knowledge, people in the first culture lose the incentive to keep queuing, and the institution stops doing its job.

Obviously, a society could not work if groups could have their own moral codes. A society requires an overarching moral system that regulates interactions between all members of society.13

Not all cultures are “equal”

Perhaps the most common version of relativism outside of academic philosophy is the idea that different cultures cannot be compared and judged from the point of view of other cultures. In practice, this goes with the idea that each culture is equal and that every culture should be respected. This question has taken a political aspect, with right-wing thinkers often praising Western culture and institutions as superior to the culture of some other countries, sometimes described as backward. In opposition to this view, some left-wing thinkers argue that all cultures are equally worthy of respect. This view is, perhaps understandably, often promoted in international bodies. The UNESCO Declaration on Cultural Policies (1982) stated, for instance, that

The equality and dignity of all cultures must be recognized, as must the right of each people and cultural community to affirm and preserve its cultural identity and have it respected by others.

The idea that all cultures are equal in worth is, in fact, unfounded in a contractarian perspective. While using the term “relativist” for his position, this is what Binmore says here:

The basic misunderstanding is that traditionalists think that to refuse to label a Society as Wrong or Bad is to say that all societies are equally Right or Good. But a relativist finds no more meaning in the claim that two societies are equally Good than he does in the claim that one is Better than the other. — Binmore (1998)

And he adds:

[T]he wishy-washy liberal doctrine that all societies are equally meritorious receives no support from naturalism. — Binmore (1998)

Not postmodernism

Finally, an important point to make is that saying that there is no absolute moral truth is not saying that there is no objective reality out there. The idea that refusing the absolute truths of moral claims is like rejecting the objectivity of reality is like saying that if one claims that there are no absolute rules of football written in the fabric of the universe—these rules being human conventions—one must be denying that the football pitch, the ball and the players really exist.

Contractarianism is therefore not a bedfellow of postmodernism and its often sceptical claims about the existence of universal truths that can be studied through reasoning and empirical observation.

Moral relativists believe that moral facts are true because they are generally held to be true within a particular culture. Postmodernists claim to believe that the same holds for all facts. A postmodernist must therefore be a moral relativist, but nothing says that a moral relativist must be a postmodernist. For example, I am as premodern as it is possible to be, but I am nevertheless a moral relativist. — Binmore (2005)

On that front, Binmore, I, and other contractarians are firmly on the side of rationalism. When considering postmodernism, whose wordplay often tiptoes around suggestions that the notion of external reality is a fallacy, I would use this famous view from Bertrand Russell:

This is one of those views which are so absurd that only very learned men could possibly adopt them. — Russell (1959)14

Reform

The fact that we can compare moral rules in terms of their outcomes and stability should make it clear that contractarianism does not prevent reform from happening. A big question about institutions is their ability to organise social cooperation successfully. It is clear that some institutions are more effective at this than others. In his book The Origins of Political Order, Francis Fukuyama describes this difference between institutions with “Denmark” being used as an example of having highly successful institutions.

The problem of creating modern political institutions has been described as the problem of “getting to Denmark;’ after the title of a paper written by two social scientists at the World Bank, Lant Pritchett and Michael Woolcock. For people in developed countries, “Denmark” is a mythical place that is known to have good political and economic institutions: it is stable, democratic, peaceful, prosperous, inclusive, and has extremely low levels of political corruption. Everyone would like to figure out how to transform Somalia, Haiti, Nigeria, Iraq, or Afghanistan into “Denmark,” and the international development community has long lists of presumed Denmark-like attributes that they are trying to help failed states achieve. — Fukuyama (2011)

It would be very reasonable for people living in countries that do not benefit from such a mix of institutions to argue in favour of political reform to get closer to “Denmark”. Here again, the argument is simply about the agreement of the people involved: proposing a move towards more effective institutions is proposing something people would have an interest in agreeing to. Binmore makes a plain case for this logic with the example of the switch of driving side in Sweden in 1967.

A real-life instance of the driving example may help. Sweden switched from driving on the left to driving on the right on September 1st, 1967. If I had been a Swede before this decision was made, I would have lent my voice to those advocating the change. Since relativists hold that it was wrong to drive on the right in Sweden before September 1st, 1967, and wrong to drive on the left afterwards, the words bad or wrong would have been useless to me when urging reform. It might have helped the cause to argue that driving on the left is intrinsically Bad or Wrong, but it would be intellectually dishonest for a relativist to take this line. Still less would I have argued that Swedish citizens should be left to make their own subjective judgments when choosing on which side of the road to drive. I would simply have argued that enough Swedes would benefit from driving on the same side of the road as the rest of continental Europe to make it worthwhile for us all joining together to get the reform adopted. — Binmore (2005)

So the key point is that calls for reform cannot be made from an appeal to morality, since the grounds for moral agreement are being negotiated in the process of reform.

[W]hat would it then mean to say that our current social contract is itself unfair? One must either accept that the relativism implicit in a naturalistic approach to ethics makes the question incoherent, or else abandon oneself to one or other of the traditional metaphysical notions of justice about which evolution cares not a whit. — Binmore (1998)

Nonetheless, in practice, moral language is often part of the bargaining and mobilisation that makes a new agreement possible. Requests for reform are bids to change the social contract at the societal level. Proponents can have strong moral feelings about it for two reasons: first, because of their deep moral preferences, and second, because of the moral preferences in the social group they belong to. An example of such positions is how opposition to slavery appeared from the start, within slave-owner societies in the West, as conflicting with deep feelings about respect for other humans and with the moral principles of religion.15

If successful, the rule change they defend will come to define what counts as right and wrong in society. There are only a few steps from thinking “it should be right/wrong” to feeling that “it is Right/Wrong”, especially when people have strong moral feelings about it.

Yet recognising that such appeals to external truths are ungrounded, and that we hold our morality in our own hands, can be liberating. It allows us to think more clearly and freely about how we want our society to be organised, instead of chasing the “Right” way in some mysterious metaphysical reality.

The present post has mostly been about defending the contractarian approach from the different accusations associated with the term relativism. A common argument put forward by moral realists is that it would lead us to accept practices that we find abhorrent, like slavery and infanticide. Given the strength of our moral feelings against such practices, it would be a very unappealing conclusion. However, as I have explained here, contractarians are not in any way bound to say that they find such practices fine. Like most people living today, they would typically prefer a world without them.

Having said that, it is worth pointing out that the moral realist position with this argument is much weaker than it seems. First, the fact that some conclusions might seem appalling is not a reason to reject a theory. Theories like heliocentrism or evolution seemed appalling to many when they were proposed, it did not make them wrong.16 The importance moral realists give to moral intuition as evidence is all the more dubious given that we can observe societies at present and in the past differing markedly in moral intuitions.17

Second, moral realists do not agree on a common theory of what absolute moral truths are and where they come from. They do not even agree on a common method to answer this question. If you take a naturalistic perspective, this persistent disagreement is not surprising. It has a simple explanation: moral realists are like ancient astronomers looking into epicycles to understand the movement of planets. They have a wrong starting assumption: there are some absolute moral principles out there for us to find.

This brings us back to the point I made earlier in this series of posts: because they aim to found morality in an absolute way, moral realists have to assume some skyhook at some point, to found their theory of morality on an “unconditional ought”: “I must do good because something is morally good and must be done”, full stop. This assumption that some abstract, immaterial principle has some kind of force on material beings like us is strange and, as famously pointed out by John Mackie, unlike anything produced by scientific knowledge. The appeal of such an idea might be particularly prominent in societies with a Judeo-Christian heritage, where an all-powerful moralising God has shaped our moral intuitions for centuries.

The contractarian approach sidesteps this metaphysical difficulty and the use of skyhooks to found morality. Instead, it posits that morality arises from something grounded in reality: a social agreement about how to organise social interactions.

References

Binmore, K.G. (1998) Game Theory and the Social Contract, Volume 2: Just Playing: Economic Learning and Social Evolution. Cambridge, MA: MIT Press.

Binmore, K.G. (2005) Natural Justice. Oxford: Oxford University Press.

Boardman, A.E., Greenberg, D.H., Vining, A.R. and Weimer, D.L. (2018) Cost-Benefit Analysis: Concepts and Practice. 5th edn. Cambridge: Cambridge University Press.

Boas, F. (1887) ‘Museums of Ethnology and Their Classification’, Science, 9(228), pp. 587–589.

Disraeli, B. (1864) Speech at the Oxford Diocesan Conference (25 November 1864), quoted in Monypenny, W.F. and Buckle, G.E. (1929) The Life of Benjamin Disraeli, Earl of Beaconsfield, Vol. II: 1860–1881. London: John Murray.

Fukuyama, F. (2011) The Origins of Political Order: From Prehuman Times to the French Revolution. New York: Farrar, Straus and Giroux.

Gowans, C. (2021) ‘Moral Relativism’, in Zalta, E.N. (ed.) The Stanford Encyclopedia of Philosophy (Spring 2021 Edition). Stanford, CA: Metaphysics Research Lab, Stanford University (Accessed: 3 January 2026).

Harman, G. (1975) ‘Moral Relativism Defended’, The Philosophical Review, 84(1), pp. 3–22.

Holland, T. (2019) Dominion: The Making of the Western Mind. London: Little, Brown.

Hrdy, S.B. (1999) Mother Nature: A History of Mothers, Infants, and Natural Selection. New York: Knopf.

Hume, D. (1751) An Enquiry Concerning the Principles of Morals. London: A. Millar.

International Criminal Court (1998) Rome Statute of the International Criminal Court. Rome, 17 July 1998 (entered into force 1 July 2002).

League of Nations (1926) Slavery Convention. Geneva, 25 September 1926 (entered into force 9 March 1927).

Mackie, J.L. (1977) Ethics: Inventing Right and Wrong. Harmondsworth: Penguin Books.

MacCulloch, D. (2009) A History of Christianity: The First Three Thousand Years. London: Allen Lane.

Melanchthon, P. (1549) Initia doctrinae physicae dictata in academia Vuitebergensi. Wittenberg: Johann Lufft.

Pritchett, L. and Woolcock, M. (2004) ‘Solutions when the solution is the problem: Arraying the disarray in development’, World Development, 32(2), pp. 191–212.

Russell, B. (1959) ‘Face to Face’ (television interview with John Freeman), BBC Television, 4 March.

UNESCO (1982) Mexico City Declaration on Cultural Policies (World Conference on Cultural Policies, Mexico City, 26 July–6 August 1982). Paris: UNESCO.

United Nations (1948) Convention on the Prevention and Punishment of the Crime of Genocide. New York: United Nations, 9 December 1948 (entered into force 12 January 1951).

United Nations (1956) Supplementary Convention on the Abolition of Slavery, the Slave Trade, and Institutions and Practices Similar to Slavery. Geneva: United Nations, 7 September 1956 (entered into force 30 April 1957).

United Nations (1989) Convention on the Rights of the Child. New York: United Nations, 20 November 1989 (entered into force 2 September 1990).

Williams, E. (1944) Capitalism and Slavery. Chapel Hill, NC: University of North Carolina Press.

A poll of modern philosophers suggests that a majority of them are moral realists.1 They assume that morality is objective in the sense that there are moral truths that exist independently of what people might think about them. This view is, interestingly, at odds with the origin of the term “morality” which points to social conventions.

[E]tymologically, the term “moral” comes from the Latin mos, which means custom or habit, and it is a translation of the Greek ethos, which means roughly the same thing, and is the origin of the term “ethics”. — Stanford Encyclopedia of Philosophy (2019)

In line with this etymological origin, I argue that the best explanation of morality is that it emerged to regulate cooperation and coordination in social interactions, and that moral systems are social conventions. This naturalistic approach starts from the factual observation that people have partially but not fully aligned interests, leading to social interactions that mix cooperation and conflict.

This view eliminates the need to posit unobserved moral truths for moral rules to mean something. Moral rules can be understood as commonly agreed rules of social interaction among members of society. These rules feel deep and fundamental because they govern the most important game we play: the Game of Life. The idea that moral rules in a society are agreed upon is often reflected in the use of the term “social contract” to describe them. People learn these rules through education and socialisation.

In this post, I address criticisms levelled at this view of morality as the reflection of a social contract.

The guilt trap

One often hears the suggestion that if morality is only conventional, it is meaningless.2 We are then told that if moral claims have no meaning, anything goes. Some of the worst things we can think of—like killing innocent people—must be “fine” and nothing compels criminals to behave well.

The leap from “conventional” to “anything goes” is a mistake, and I’ll explain why in the next section. But I’ll start by criticising this style of counterargument, which I’ll label a guilt trap. It is a trap because it is another version of the Santa Claus fallacy. It is saying something like: “do you realise the awful implications of your beliefs”?

In fact, even if a theory were to lead us to very distasteful conclusions, that would not be a reason to reject it. Suppose that—ghastly horror—we were to find that our home, Earth, is not the centre of the universe, but just a tiny dot rotating around an average star, somewhat on the skirt of a galaxy comprised of about 100 billion stars, itself one among billions of galaxies. Too bad. Reality is not determined by our preferences. A Humean constructivist can therefore stand firm on rejecting the guilt trap because it is not a good argument to begin with.

But because the human brain is designed to argue in favour of our preferences, it might nonetheless be a useful rhetorical move for the Humean constructivist to point out that nothing fundamentally ghastly should be expected from this approach. The goal of Humean constructivism is to explain morality—practices and beliefs—as it is observed in real life. Hence, a Humean constructivist is more likely to explain why our current moral practices often make sense than to advocate throwing them away in favour of views we find abhorrent.3

Morality as the respect of a social contract

Moral systems as social contracts

The first point to make is that saying there are no absolute moral truths is not the same as saying moral claims are meaningless. A contract sets rules defining rights and duties between those who sign it. The rules created by a contract do not come from “out there”. They arise from an agreement between different parties. Saying that morality emerges from commonly agreed rules—conventions—about how to behave in society does not imply that these rules are meaningless or can be ignored at will.

Imagine Alice and Bob start playing tennis. At some point, Alice hits a perfectly good forehand that lands 1 m within the court. Bob yells “out” and claims the point. Alice is entitled, by the rules of the game, to say “you are wrong, Bob, you lost the point”. This claim is not “meaningless”. It is true, given the rules of tennis.

We can agree that the rules of tennis are conventional. They are neither written into the fabric of the universe nor implied by the rules of pure logic. Does it follow that, because they are conventional, Bob can simply ignore them? The answer is simple and non-mysterious: it is a matter of fact that Bob might choose to ignore the rules and claim he won the point even though he did not. But if he does so, he violates the implicit understanding he had with Alice that he would play fair, that is, play by the commonly agreed rules. Faced with an outright disrespect of the rules, Alice is likely to fire back: “fine, I am not playing with you anymore”. The rules of tennis are a social contract between players. Bob cannot violate this contract and expect Alice to continue playing while ignoring his violations.

The word social contract can seem grand. Past philosophers sometimes implied, like Hobbes, that there might have been a time when people literally sat down to agree on a social contract. These stories are clearly fictitious. In his discussion of the origin of state institutions, Francis Fukuyama points out that early states pretty much never formed that way:

We might label this the Hobbesean fallacy: the idea that human beings were primordially individualistic and that they entered into society at a later stage in their development only as a result of a rational calculation that social cooperation was the best way for them to achieve their individual ends. — Fukuyama (2011)

Such fairy tales about social contracts are not necessary. A social contract can be understood as a commonly agreed understanding of how things should be done, and what would count as a breach of acceptable behaviour in society. To be clear about how he uses this expression, Binmore writes:

This section has identified a social contract with an implicit self-policing agreement between members of society to coordinate on a particular equilibrium in the game of life. Some apology is appropriate for those who find this terminology misleading. […] Words other than contract—such as compact, covenant, concordat, custom, or convention—might better convey the intention that nobody is to be imagined to have signed a binding document or to be subject to pre-existing moral commitments. Perhaps the best alternative term would be “social consensus”. This does not even carry the connotation that those party to it are necessarily aware of the fact. — Binmore (1994, emphasis mine)

If morality emerges from conventions, in what sense “ought” we to abide by moral rules? The answer is that we ought to abide by moral rules to be able to continue to play in social games with others, and to remain accepted as a player by them. If we violate that social contract, we lose our right—as defined by the social contract—to be treated as a rule-abiding player, and we risk a range of sanctions. In the extreme, we might be thrown out of the social game entirely, ending up in prison or worse.4

People sometimes react to this explanation by saying: this is not a proper “ought”. This claim tends to take two forms. First, there are rules you really ought to follow, even if you would face no social sanctions for not doing so. Second, there are things that are really bad, whatever people think about them, like “killing innocent people”.

The ring of Gyges

Let’s first address the idea that morality must imply a duty to do something beyond the costs and benefits set out in the social contract. This question was already discussed in Plato’s Republic, where Glaucon counters Socrates’ idealism with a story about what happens when we gain impunity. In the story, a shepherd named Gyges finds a ring that lets him become invisible at will. Once he realises he can act without being seen, he uses the ring to enter the palace, seduce the queen, kill the king, and take the throne.

Glaucon concludes:

[E]very man, when he supposes himself to have the power to do wrong, does wrong.

For Glaucon, we follow moral rules only because we have an interest in doing so. Remove the incentives to behave well, and we would stop doing so. This view is not necessarily in line with our intuitions. Plato uses Glaucon’s view to criticise it and argue that there is more to being moral: morality comes from within us, and being moral is necessary to be at peace with oneself and happy.

From a Humean constructivist perspective, Glaucon’s argument seems, unfortunately, somewhat sound. In the end, a social contract works only if it creates incentives for people to follow it. Binmore even points out that assuming we have a moral sense that reliably drives us to behave against our interest would conflict with the logic of evolution:

[I]f the morality of everyday life required people to take actions that made them less fit, how could it have evolved? […] one of the achievements of modern social psychology has been to expose the gap that lies between how we like to think of ourselves and the reality of what we actually are. […] realism requires that we take the same attitude to how society actually works as Oliver Cromwell took to his appearance. He too would have liked to look better, but he nevertheless insisted on being painted “warts and all”. — Binmore (1998)

What is the meaning of words like right or wrong and ought

How should we think of Gyges’ actions, if morality is seen as the reflection of a social contract?

First, it is perfectly coherent to say that what Gyges is doing is “wrong”, given the social contract of his society. Rules of social contracts do not say “don’t do that except if nobody is watching.” Rules are meant to be obeyed, whether there are observers or not. It is unlikely that the excuse “Sorry, officer, I thought nobody was watching” has ever been successful in evading a speeding ticket.

Second, it is coherent to say that Gyges “ought” to avoid doing all the bad things he does. He ought to play by the rules of the social contract if he wants to remain entitled to be treated as a good social partner by other people. If he behaves badly and he is found out, for instance, if he drops his ring, he would face the consequences.

But wait, you might say: “what if there is a 0% chance that he would be found out, and therefore no reason for him not to behave badly?” Here, the Humean constructivist resists the pressure to invent a skyhook: a moral duty that exists above and beyond the social contract and somehow binds Gyges to behave well in all circumstances. If there are absolutely no penalties for violating the rules of the social contract, there is no additional force, external to the social contract, that would give Gyges an injunction to respect it.

Note that this does not mean Gyges is permitted by moral rules to behave badly. The observation that motorists can (and sometimes do) speed when there is no police around does not mean that this action is condoned by the driving code.

Explaining the role of moral intuitions in social interactions

These answers, while very clear and coherent, face the challenge of our moral intuitions: we feel that cheating when unobserved is bad. Here, however, we should not treat these intuitions as a window into moral truth. A naturalistic approach should instead explain why these intuitions exist and why they take the form they do.5 The answer is that evolution endowed us with a moral sense that helps us play social games well. As I said in the previous post:

Our moral sense, shaped by eons of evolution, is designed to provide us with emotions that help us play the Game of Life successfully. Any deviation from the moral code puts us at risk of future sanctions: blame, ostracisation, and even, in some cases, violence. Our moral emotions track the costs of deviations and bring them back into our present. They embed this shadow of the future in our daily experience.6

These moral intuitions are also designed to help us assess others’ behaviour and detect violations, subtle or blatant, of the social contract. In Gyges’ case, our intuitions tell us that what he did is wrong. A social contract works to the extent that deviations are sanctioned. Members of society are required by the social contract to sanction violations of that social contract. Not doing so is a breach of the contract in the same way as not reporting a crime can be itself a crime (“misprision of felony”). It is therefore clear that upon hearing Gyges’ story, listeners like us ought to form the kind of negative moral judgement that would justify punishing him if he were found out.

What about Gyges’ feelings? Even though Gyges might face 0% risk of being found, he might nonetheless experience moral feelings about his actions: a sense of duty not to behave badly, and guilt for thinking about or doing bad things. Here, again, the answer is not mysterious. These feelings evolved to track the possible risk of being found out and the future costs involved. Even if we face a situation of total impunity, it is reasonable for such feelings to be triggered, because situations with truly 0% risk do not really exist in real life. Even the culprit in Hitchcock’s story of a “perfect murder” in Dial M for Murder ends up being arrested. Hence, we should not be surprised that our moral emotions ring alarm bells, even when it seems that no risk of being found out exists.7

Being mindful of the deceptive appeal of the term social contract

Another problem generated by our moral sense is that the term “social contract” is itself loaded. We have a duty to respect a contract in society, because every specific contract is embedded in the overarching rules of the social contract.8 There is, however, nothing above society that sustains the social contract.

The term social contract might therefore be deceptively appealing. Our intuitions may find it more palatable than the term convention. I would, however, be slightly disingenuous if I leveraged those positive intuitions to make the idea of morality as a social contract sound more convincing.

I shall emphatically not argue that members of society have an a priori obligation or duty to honor the social contract. On the contrary, it will be argued that the only viable candidates for a social contract are those agreements, implicit or explicit, that police themselves. […] — Binmore (1994, emphasis in the original)

The role of trade-offs in morality

Note that these emotions do not need to feel to us like they are tracking risk return trade-offs. We can feel strongly motivated by such feelings, and some people experience them more than others.9 In other words, the fact that our moral sense tracks risk-return trade-offs is an ultimate explanation of moral behaviour, but we often experience moral judgements as driven by strong feelings and principles that seem devoid of calculation. These feelings are the proximate explanation of our moral behaviour.10

Having said that, it would be naive to ignore trade-offs. We would like to think we would act morally in all circumstances, but the statement “The surprising thing is not that every man has his price, but how low it is,” commonly attributed to Napoleon, bites because for every Schindler willing to sacrifice his interests for moral principles, there were many more Eichmann who followed wherever institutional carrots and sticks led them.11

my hero in the Republic [is] Plato’s brother, Glaucon—who gets short shrift from Socrates for putting a view close to the social contract ideas defended in this book. I think that a modern Glaucon would be agreeing with me that fairness norms evolved to select one of the many ways to balance power in a group. — Binmore (2005)

The Humean constructivist answer to the Ring of Gyges challenge

We can therefore summarise a Humean constructivist answer to the challenge of the ring of Gyges as follows:

1. We can say that Gyges was “wrong” given the social contract.

2. According to that social contract, Gyges “ought” not to violate its rules. But a social contract that does not attach appropriate incentives to its rules will fail to be followed. If a social contract has appropriate incentives, then the rules of the social contract align with Gyges’ interests: not only should he follow them from the social contract perspective, but he also should follow them from the perspective of his own interests.

3. If Gyges violates the social contract thanks to his impunity, other members of society will have strong moral feelings condemning it.

4. Gyges might also have negative moral feelings about it (though these feelings might erode with repeated experience of impunity).

Not “anything goes”

It is easy to misread the claim that moral norms are man-made (or socially constructed) as implying that anything goes. To start with, moral norms need to be respected for society to work. A society where civility and trust disappear breaks down.

In addition, moral norms cannot take any possible shape or form. The logic of strategic interactions imposes constraints on them.

Not everything can work as a social contract

Saying that because moral norms are socially constructed, they can be anything is a bit like saying that because buildings are man-made, they can take any shape. Buildings can take many shapes, but to withstand the test of time they must respect the laws of physics. It is exactly the same for moral norms. It is obvious that they differ widely across times and places, but for them to work over long periods, they need to respect the “laws” of social interaction, so to speak. In other words, they must be compatible with the patterns of behaviour that emerge in human societies, given that individuals have partially but incompletely aligned incentives.

The idea that moral and fairness norms are equilibria of social games has a key implication. If they remain stable and respected, it is not because people are brainwashed and mindlessly obey imposed rules. Instead, social norms are stable because they are self-enforcing: everybody has an interest in following them when everybody else follows them.

Consider driving on the road in a country like the USA. Why is the rule “everybody drives on the right side of the road” respected by nearly everybody in society? One might be tempted to say it is because it is illegal not to do so. But there are many countries where it is illegal to do various things, and many people nonetheless treat the law as a flexible instruction rather than a strict rule. The real reason almost everybody drives on the right in the US is that it would not pay to drive on the left. You would most often not arrive earlier, you would risk an accident, and because nobody else does it, you would likely attract negative judgment from people around you.

Among all the possible moral codes one could conceive, only a tiny proportion are equilibria. The rule “everybody drives on the right” is not the only solution. In several countries, like the UK or Australia, the rule “everybody drives on the left” is used instead. But no country has adopted in its official driving code rules like “drive on the side you fancy” or “cars drive on the left and trucks on the right”. There are good reasons for these absences. These alternatives are not solutions to the driving game.12 Any country with such rules would find not only that they lead to bad outcomes but also that people end up not respecting them. Conventions to drive on the left or on the right would emerge anyway. Historically, for example, the convention to keep left on the road emerged in Britain as a practice long before it was formally written into national traffic law.13

Moral, legal and ethical systems are conventions, but they are not arbitrary: they are constrained by the structure of conflict and cooperation in society.

A lot of the things we find abhorrent generate such feelings for good reasons

We can expect that moral systems that survive the test of time will encode concerns for participants’ lives and physical integrity. An argument of the type “killing random people is wrong, therefore there are absolute moral truths” mistakes our proximate moral feelings—attuned to the fact that modern societies ban random killings—for evidence that there is an absolute truth “out there” about it.

A Humean constructivist, and in particular Binmore, points out that if our current rules are the way they are, it is typically because they help sustain the way our society works. In liberal democratic countries, moral systems form a social contract well-suited to support large-scale cooperation. Therefore, in addition to the fact that Humean constructivists typically share the same negative moral feelings as everybody else about ideas such as allowing random killings, they also have an additional reason to reject such radical suggestions: they would likely tear the fabric of society.14

Mainstream ethical philosophy is lost in conceptual mazes because it takes our moral intuitions at face value, as if they pointed to something “out there” that justifies our moral stances. We “ought” to do this or that because it is right or good in an absolute sense. Moral intuitions are, however, the tail that wags the dog. Starting from them to build an understanding of morality treats the proximate psychological mechanisms that help humans navigate cooperative social interactions as a primary source of insight into morality, instead of looking at the ultimate causes of morality: the logic of human cooperation, which shaped these intuitions.

Saying morality arises from social conventions, and that we can think of moral systems as a social contract, often raises the concern that nihilism would engulf society if such paltry views were to take hold. All the types of gruesome abuse we might think of would be acceptable because humans would be free to do anything.

As I have put here, it is preposterous. For sure, the readers of this post have countless times followed social conventions when playing board games, sporting matches and everyday interactions, without assuming that absolute truths reside behind the rules they followed. Similarly, moral rules must be followed for society to work. Because there is no enforcer outside society, moral rules need to be self-enforcing to be sustainable: moral systems must come with a range of formal and informal sanctions that ensure that nobody has an interest in stepping outside the rules of appropriate behaviour. Our moral intuitions are then a reflection of the fact that our cognition is well designed for us to navigate these situations, anticipate what is appropriate, be wary of what might not be, and so on.

Finally, the idea that moral systems are social contracts in no way implies that anything goes. On the contrary, it is reasonable to assume that among all the social rules we could imagine, only a tiny subset can be part of a sustainable social contract.

I expect that some readers will point out a gaping hole in my argument. Even if not all rules are permitted, there is no guarantee that some very unpleasant social contract cannot work as an equilibrium in a society. What are we to say about it then? Should we say that slavery or the subjugation of women is fine if another society seems to be stable with such institutions? This question is an important one, and it will be the topic of my next post.

References

Acton, J. (1887) ‘Letter to Archbishop Mandell Creighton (5 April 1887)’, in Acton–Creighton Correspondence.

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: The MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: The MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Fukuyama, F. (2011) The Origins of Political Order: From Prehuman Times to the French Revolution. New York, NY: Farrar, Straus and Giroux.

Hare, J. (2019) ‘Religion and Morality’, in Zalta, E.N. (ed.) The Stanford Encyclopedia of Philosophy (Fall 2019 edn). Metaphysics Research Lab, Stanford University.

Lie-Panis, J. and André, J.-B. (2022) ‘Cooperation as a signal of time preferences’, Proceedings of the Royal Society B: Biological Sciences, 289(1973), 20212266.

In a note last year, Richard Dawkins asked a simple question: is morality objective? Professional philosophers often put the question this way: are there statements that are truth-apt (they can be right or wrong) and whose truth is stance-independent (whether they are right or wrong does not depend on people’s thoughts or preferences about it).

In the previous post, I argued that the answer to that question is negative: we don’t have good reasons to assume that there are moral truths “out there”. Such an answer is sometimes met with shock and possibly horror. The philosopher Rudolf Carnap reports an anecdote about how people reacted to his claim that value statements (including moral ones) do not describe facts and cannot be proved true or false. One of his colleagues even pondered whether he should call the police to put him in jail.

According to these critics, to deny to value statements the status of theoretical assertions and thereby the status of demonstrating their validity must necessarily lead to immorality and nihilism. In Prague I found a striking example of this view in Oskar Kraus, the leading representative of the philosophy of Franz Brentano. I heard from the students that in one of his seminars he characterized my thesis of the nature of value statements as so dangerous for the morality of youth that he had seriously pondered the question whether it was not his duty to call on the state authorities to put me in jail. — Carnap (1963)1

Carnap’s story captures a familiar fear: once we give up on absolute moral truths, we are left with moral chaos.

Until now, my posts on the topic have been mostly about pushing back against these widespread intuitions and against established philosophies building on them. Now is the time to flesh out more how we can think of morality from a naturalistic perspective, once we abandon the idea that there are absolute moral truths “out there”.

Locating the naturalistic position in the space of moral philosophy

For most people, ethical discussions often seem a bit mysterious. What is right? What is wrong? And how do we know how to answer these questions?

Metaethics—the branch of philosophy that looks at what morality is—can seem daunting. It is filled with abstract notions and technical terms. Early in his career, prior to producing the work that would end in his landmark book on the topic, On What Matters, Derek Parfit had purposely stayed away from that field. He is said to have told a colleague about it:

I don’t do metaethics. I find it much too hard. — Parfit2

Here, I provide a quick introduction to metaethics that breaks down the different philosophical positions based on answers to a few key questions. This makes it easier to locate where the naturalistic position I am developing here stands.3

Moral realism

The first question is the one asked by Dawkins: is morality objective? Those answering yes are “moral realists”. As I indicated in my last post, an international poll of philosophers suggests that a majority of them endorse this view. The bulk of my post was dedicated to arguing against it.

One can think of two types of justifications for such a position. The first aims to ground an objective morality in natural facts. This “moral naturalism” is the position adopted by thinkers like Peter Singer and Sam Harris.4 It starts from factual descriptions of human experience (for example, that pain is undesirable and happiness desirable). It then argues that we can take an evaluative stance on this basis (for example, that pain is “bad” not just for the person, but in some higher sense that gives us a moral duty to reduce it).

Criticism of moral realism. This introduction of an evaluative stance, not grounded in natural facts, does all the work here: it smuggles an ought into the theory. It does not show that morality is objective. It assumes it. It is a logical leap that the reader is free to make, but it is not warranted by a naturalistic approach.

The second approach is to say that reasoning itself can help us identify moral truths. This approach is labelled “robust realism”, it is the one held by philosophers like Immanuel Kant and Derek Parfit.

Criticism of robust realism. I previously discussed Parfit’s argument.5 His claim to only use reason to find moral truths is a bit incorrect. In the end, he relies on our moral intuitions as a kind of primary ground upon which to build his theory: he investigates our intuitions with thought experiments and tries to build a theory which rationalises our “reasoned intuitions” (i.e. intuitions that resist some kind of rational investigation about how right they are). The problem here is that we have a very good naturalistic account of why we would have moral intuitions, based on evolutionary explanations. There is therefore no reason to assume that the existence of moral intuitions is evidence of objective moral truths “out there”.

Moral anti-realism

Let’s assume we abandon moral realism. A question we can ask next is: ”OK, morality is not objective in an absolute way, but can moral statements nonetheless be considered true or false in another way?”

Here, one option is to say no. Moral statements are neither true nor false, because moral statements do not refer to a moral reality that can make them true or false. Instead, moral claims reflect our preferences. Hence, when we say “murder is wrong” we just mean “I really don’t like murder”. This approach, proposed by philosopher A.J. Ayer,6 is called emotivism. It is also nicknamed the boo-hurrah theory of morality. A statement like “murder is wrong” just means “boo murder” and “charity is good” just means “hurrah charity”. Our moral statements, therefore, reflect our feelings about things.

Criticism of emotivism. I see at least two issues with emotivism. First, identifying moral statements with moral feelings raises the question of why we have them and why they take the particular form they do. Could we have any type of moral feelings? Could we expect “hurrah murder” to be as likely as “boo murder” as a feeling? If not, why? Emotivism leaves this question unresolved. A second issue is that moral statements seem to have a logical structure (this is called the Frege–Geach problem).7 We can make apparently valid inferences such as “murder is wrong, so intending to murder is wrong” or use moral claims inside conditionals: “If murder is wrong, then paying someone to murder is wrong.” But emotivism rejects the idea that moral claims can be true or false, so it is not clear why they should behave in arguments as if they were the kind of sentences that can be true or false. If moral language is only a matter of expressing feelings, why couldn’t someone consistently say “boo murder” and “hurrah for murder intentions,” in the same way that they can dislike vanilla ice cream but like chocolate ice cream?

Another possible answer is to say: yes, moral statements can be true or false, but ordinary positive moral judgements are all false! There is no realm of objective moral facts “out there”, so when we say “murder is wrong” we are systematically mistaken about what exists. This view, famously defended by John Mackie and known as error theory, holds that moral judgements are always in error. From that perspective, much traditional moral debate on moral truths looks a bit like the supposed Byzantine debates on whether angels are male or female: it discusses claims about presupposed entities that do not exist.8

Criticism of error theory. Saying that moral claims are wrong when they assume absolute moral truths is fine. But it doesn’t follow that moral claims can’t be true or false in a more limited way. The sentence “You can’t do that”, frequently uttered when playing board games, can be true or false even though it does not rely on any absolute truths “out there”, only on the existence of commonly agreed rules of play. In fact, Mackie invites us to look at empirical studies of how morality actually works in human societies, and Binmore saw his own work as following in Mackie’s footsteps.

Mackie tells us to look […] for a framework within which to make sense of […] anthropological data, he directs our attention to Von Neumann’s theory of games. — Binmore (2005)

The naturalistic approach I develop here answers yes to the question of whether moral statements can be true or false. However, if they are not true or false in the sense of moral realism, in what sense can they be true or false? The answer is that different social groups can have different moral rules. Within each group, a statement such as “X is wrong” can be true or false given these rules.

If this is the case, it is natural to ask where these rules come from: how are they constructed? One possible answer is radical cultural relativism. This position takes the view that different moral systems happen to be what they are in a given society for historically contingent reasons that might be arbitrary. These moral systems are not comparable across times and places. They are “incommensurable” (there is no measure we could use to compare them). As a result, we are never able to compare moral rules across cultures because we do not have a valid external standpoint from which to do so.9

Criticism of radical cultural relativism. Radical cultural relativism has too many unanswered questions. It does not offer a positive theory of how moral systems emerge and evolve. Why do societies have the moral systems they have? Are they arbitrary? Can they prescribe any kind of action at all? If so, why should we care about them? Why do people follow them?10

The view I develop here does not pursue this route. It posits that moral systems do not simply happen to be what they are for purely historically contingent reasons. They emerged from cultural evolution to serve a function: to help foster and regulate social cooperation. This approach is often referred to as Humean constructivism because philosopher David Hume was perhaps the earliest proponent of this view:

The rules of equity or justice depend entirely on the particular state and condition in which men are placed, and owe their origin to that utility which results to the public from their strict observance. — Hume

Surprisingly, this approach remains a blind spot in mainstream academic metaethics. The view I am developing here builds on important contributions by authors such as Robert Sugden, Brian Skyrms, Cristina Bicchieri, and Ken Binmore, whose work is largely ignored in standard metaethical discussions. A quick look at the main textbooks is revealing. Neither these authors, nor their method of inquiry—game theory—is mentioned in van Roojen’s Metaethics (325 pages), Miller’s Contemporary Metaethics (570 pages), or Chakraborti’s Introduction to Ethics (768 pages).

Binmore noted this somewhat peculiar absence of engagement with game theory in the philosophical literature on morality. When philosophers have to look into the possible insights game theory can bring to our understanding of morality:

It is at this point that scientific philosophers tend to falter. Game theory books are mostly written by economists, who use an exotic vocabulary and complicated mathematical equations. — Binmore (2005)

I strongly agree with Binmore that this is a pity. In a way, it is not warranted, as many important findings from game theory can be explained simply. Indeed, it is the standpoint of this Substack to bring some of its key insights in a faithful but approachable way.

[…] there is nothing arcane about game theory. On the contrary, what has to be explained in this chapter is embarrassingly easy. The exotic vocabulary and the fancy equations of game theory books are just more of the dust that scholars always kick up lest it be found out that what they have to say isn’t very profound. — Binmore (2005)11

As a result of this relative absence in metaethics, many moral philosophers are likely less familiar with the conceptual stance I present and may, on the basis of my previous post against moral realism, have been unsure where to locate it: emotivism? cultural relativism? error theory? This post (and the next) places this perspective in an approach that stands mostly outside the bulk of the discussion in metaethics, and yet is a position that is conceptually solidly grounded in our understanding of human interactions.

Another way to represent what distinguishes this position from most metaethical positions is to map these positions in a diagram distinguishing both realism from anti-realism and game-theoretical/evolutionary foundations from an absence of such foundations. The graph below does this for a non-exhaustive list of thinkers, blending classical philosophers and more modern thinkers.12

A moral theory of the Seemly

Having located this approach in the intellectual space of moral philosophy, it is now time to flesh it out and describe what vision it offers about morality. It is a big challenge to explain it simply, while doing it justice within the space of a Substack post. Here I will describe what it stands for. In the next post, I’ll discuss the implications of this position and compare these implications to other ethical positions.

The Good, the Right and the Seemly

Binmore distinguishes two main ways of thinking of morality as objective “out there”. The first is to assume that some things have an objective value and we ought to act to foster the presence of this value in the world. Such theories are called consequentialist. We can call them theories of the Good. The second is to assume that there are some fundamental moral principles out there that we ought to follow. Such theories are called deontological. We can call them theories of the Right.

A Humean constructivist approach rejects the claim that there are absolute moral truths out there—be they values or principles—that should guide our actions. Instead, it posits that our moral code is a human creation that serves to foster social cooperation.13

Having rejected the theories of the Good and the theories of the Right, Binmore labels his approach a theory of the Seemly, that is, a theory of morality as what is appropriate given social conventions about morality.

I try to capture this down-to-earth attitude to morality by calling it a theory of the seemly. Although theories of the seemly aren’t currently popular among moral philosophers, they have a long and respectable intellectual history. Aristotle, Epicurus, and Hume are perhaps the most famous Western exponents of the approach, but if one could mention only one name it would have to be that of Confucius.14

Within a theory of the seemly, things aren’t good or right in themselves; they are good or right because they are generally held to be good or right in a particular society. — Binmore (2005)

Thinkers like Robert Sugden, Brian Skyrms, Cristina Bicchieri and Ken Binmore develop such a theory of the Seemly by using game theory to explain the content and inner workings of moral systems. In short, this view relies on two key ideas. First, moral rules are equilibria of social games: they regulate stable patterns of social interactions. Second, these rules have emerged and persisted because they help solve recurrent coordination and conflict problems in ways that are, on balance, advantageous for enough members of the society to be stable.15

Moral rules are rules of the Game of Life

From this perspective, if there is an “ought” to follow these rules, it is not an unconditional mysterious ought coming from intangible moral values and principles that apply universally. We ought to respect moral rules because they are the rules of the social games we play. If you start playing Association Football (soccer), you should not grab the ball and run towards the opposite goal. Why? Because “handball is wrong” in football. And if you want to play football, you should respect that rule.

You might still ask, but in what sense ought we to abide by moral rules? Why shouldn’t we be happy instead just doing what we want? Here again, the explanation is no longer mysterious. You ought to abide by moral rules because they are an equilibrium of the social games you play. As a consequence, they are self-enforced by sanctions built into the game for those who do not respect them. It is perfectly meaningful to say something like “If you don’t want to be penalised, you ought to respect the rules”. This is a conditional ought, which is rational for you to follow, given the rules of the game. It is not an unconditional ought based on some assumed external moral truth.

I expect that this deflationary account of moral claims—i.e. an account which reduces the meaning of moral claims—will feel unsatisfactory to many. Isn’t morality radically more important than the rules of a sport? Some may point out that we feel more strongly about statements like “murder is wrong” than about “handball is wrong in football”.

The answer here is that yes, moral rules are much more important to us than sporting rules because moral rules are the rules of the Game of Life, the one that determines our success and setbacks in life, at work, with friends or romantic partners, with family members, and so on. It makes perfect sense for us to have a moral sense that makes us care greatly about how the Game of Life is played. There is no reboot and no respawn in this game, and it determines everything that matters to us.

Our moral sense, shaped by eons of evolution, is designed to provide us with emotions that help us play the Game of Life successfully. Any deviation from the moral code puts us at risk of future sanctions: blame, ostracisation, and even, in some cases, violence. Our moral emotions track the costs of deviations and bring them back into our present. They embed this shadow of the future in our daily experience. Talking about our moral cognition, cognitive scientists André et al. (2022) write:

[…] morality is not a set of rigid and stereotyped principles, but a flexible context-dependent calculator of what one must do to secure a good reputation and attract future cooperative investment from others.

Such an explanation might seem to conflict with the phenomenology of our experience of moral behaviour—i.e. how we experience engaging in making moral judgements. Moral feelings are typically experienced as primary feelings, not the reflections of some practical trade-offs about future payoffs. Isn’t this discussion of “calculations” misguided?16

It is important, here, to stress the two different levels of explanation of behaviour in evolutionary theory: ultimate explanations are about why a type of behaviour was selected by evolution, while proximate explanations are about how such types of behaviour are generated by our cognition. Saying that our moral sense is designed to track the payoffs of respecting or deviating from moral rules is an ultimate explanation. But, in our everyday lives, we typically choose to behave morally because we experience moral emotions, this is the proximate explanation of our behaviour. And our moral feelings can be strong because the stakes of the Game of Life are high.

Humean constructivism, in spite of a coherent conceptual framework grounded in ideas and results from game theory, is marginal in mainstream ethical discussions. The reason might come from the technicality of game theory, but also possibly from the seemingly less appealing view that morality is not grounded in absolute principles but in social conventions.

Aside from intellectual considerations, theories of the Seemly do not have the same social cachet as the grandiose theories of the Good and the Right. But perhaps the fact that people like me are moved to write books like this is a sign that the time for Seemliness has come. — Binmore (1998)

However grandiose the theories of the Right and the Good might be, they will only lead us to err endlessly in conceptual mazes if they are misguided. From that perspective, having the epistemic courage to abandon these absolute views and accept the social nature of morality opens the way for us to attain a greater clarity on what morality is and how it works.

Derek Parfit started his highly influential book Reasons and Persons with a quote from Nietzsche that reflects his view that abandoning religious foundations for morality opens us to a radically new understanding of morality, something real but secular. In the same way, the Humean constructivist perspective on morality opens a radical perspective, different from much that has preceded it in moral and political philosophy, enlightening even if a bit daunting in parts.

At last the horizon appears free to us again, even granted that it is not bright; at last our ships may venture out again, venture out to face any danger; all the daring of the lover of knowledge is permitted again; the sea, our sea, lies open again; perhaps there has never yet been such an “open sea”. — Nietzsche

References

André, J.B., Fitouchi, L., Debove, S. and Baumard, N., 2022. An evolutionary contractualist theory of morality. PsyArXiv. May, 24.

Ayer, A.J. (1936) Language, Truth and Logic. London: Gollancz.

Bicchieri, C. (2006) The Grammar of Society: The Nature and Dynamics of Social Norms. Cambridge: Cambridge University Press.

Binmore, K. (1994) Game Theory and the Social Contract, Volume 1: Playing Fair. Cambridge, MA: MIT Press.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Binmore, K. (2012) ‘Kitcher on Natural Morality’, Analyse & Kritik, 34(1), pp. 129–140.

Carnap, R. (1963) ‘Intellectual Autobiography’, in Schilpp, P.A. (ed.) The Philosophy of Rudolf Carnap. La Salle, IL: Open Court, pp. 3–84.

Chakraborti, C. (2023) Introduction to Ethics: Concepts, Theories, and Contemporary Issues. Singapore: Springer.

Edmonds, D. (2023) Parfit: A Philosopher and His Mission to Save Morality. Princeton, NJ: Princeton University Press.

Hume, D. (1751) An Enquiry Concerning the Principles of Morals. London: A. Millar.

Mackie, J.L. (1977) Ethics: Inventing Right and Wrong. London: Penguin.

Miller, A. (2003) An Introduction to Contemporary Metaethics. Cambridge: Polity Press.

Nietzsche, F. (1882) The Gay Science. (Various editions; citation here refers to the work from which the closing passage is taken.)

Parfit, D. (1984) Reasons and Persons. Oxford: Clarendon Press.

Parfit, D. (2011) On What Matters. Oxford: Oxford University Press.

Skyrms, B. (1996) Evolution of the Social Contract. Cambridge: Cambridge University Press.

Sugden, R. (1986) The Economics of Rights, Cooperation and Welfare. Oxford: Basil Blackwell.

van Roojen, M. (2015) Metaethics: A Contemporary Introduction. New York: Routledge.

von Neumann, J. and Morgenstern, O. (1944) Theory of Games and Economic Behavior. Princeton, NJ: Princeton University Press.

Westermarck, E. (1906–1908) The Origin and Development of the Moral Ideas. London: Macmillan.

This Substack is not for the epistemically faint-hearted (i.e. not for those afraid of what the truth might be), so let’s get something out of the way. I believe it does not make sense to talk of morality as being objective, in the sense that there would be absolute moral truths.

The reason is that, if you take a naturalistic approach without skyhooks, I do not see in what sense there could be moral truths external to us. Where would they come from? How would we access them? I share the incredulity Binmore expresses here:

Diogenes remarked that he had seen Plato’s cups and table, but had yet to see his cupness and tableness. I feel much the same about the Platonism of traditional theories of the Good and the Right. — Binmore (1998)1

The philosopher most well-known for rejecting the absolute reality of morality is Nietzsche. In Beyond Good and Evil, he stated:

There are no moral phenomena, only a moral interpretation of phenomena.

For him, moral statements do not reflect some objective moral truths. They are social and historical narratives, human inventions. But wait, isn’t this paving the way to moral relativism, and the justification of any horrible behaviour we can think of? Here, I discuss why there is no reason to believe that there are absolute moral truths. In the next post, I’ll discuss why this conclusion does not lead to the justification of asocial behaviour.

Preferences over ice cream and Nazis

One specific aspect of morality is that we tend to experience moral norms as objective and externally imposed, rather than as mere personal preferences. Philosopher Kyle Stanford illustrates this with the contrast between saying “I like chocolate ice cream” and “The Nazis were bad.” The first is clearly a matter of taste, but the second feels different, something we take to be true regardless of anyone’s opinions.

Humans experience the demands of morality as somehow imposed on us externally: we do not simply enjoy or prefer to act in ways that satisfy the demands of morality; we see ourselves as obligated to do so regardless of our subjective preferences and desires, and we regard such demands as imposing unconditional obligations not only on ourselves, but also on any and all agents whatsoever, regardless of their preferences and desires. — Stanford (2018)2

The notion that moral injunctions have some objective nature is associated with moral realism, the idea that moral rules are real laws, in the same sense as the laws of physics.3

The philosophers’ view

Modern philosophy has become mostly secular, but it is striking that the majority of philosophers seem to continue to support the view that there are absolute truths, independently of the question of the existence of a God. A recent 2020 PhilPapers Survey indicated a clear majority supporting moral realism among philosophers across the world.

On Substack, philosophers like Peter Singer, , , , , , are in favour of moral realism while , , and are opposed to it.

The argument for moral realism

Moral realists argue that we can build a theory of morality. There are many different justifications for moral realism, but they often share similar features, such as the observations that we have strong moral feelings and experience them as objective, which might vindicate looking for the objective reasons behind these feelings.

To engage seriously with moral realism, I’ll focus on the arguments of two influential thinkers: Derek Parfit, a leading moral philosopher, and Sam Harris, a well-known public intellectual with a wide popular impact.4

Derek Parfit’s non-naturalist moral realism

Derek Parfit is commonly seen as one of the most important moral philosophers of recent times, if not the most important. His views have had a major influence in academic circles and beyond. It therefore makes sense to turn to him when looking for one of the best possible advocates of moral realism. He defends the objectivity of morality in his book On What Matters (2011).

At the heart of Parfit’s view is the claim that, in addition to “internal” reasons we have for action (those that come from our personal desires), there are objective “external” reasons to act that apply to us whether or not we are motivated by them. Imagine a person who enjoys causing pain and feels no sympathy or guilt. Given his present desires, he would be happier if he tortured a defenceless prisoner. On a purely “internal reasons” view, this person has no reason not to torture; in fact, he has an internal reason to torture, because it satisfies his strongest desires. Parfit wants to say that, even if this person has no such internal motives, he still has a powerful reason not to torture. That reason does not depend on what he now wants; it is grounded in the suffering he would inflict. This is what Parfit means by an external reason, and he argues that such external reasons are real.

We all have reasons to regret anyone’s suffering, and to prevent or relieve this person’s suffering if we can, whoever this person may be, and whatever this person’s relation to us. We have such reasons to prevent or to regret the suffering of any sentient or conscious being. — Parfit (2011)

If there are external reasons, can we systematise them and build a moral theory that makes sense of them? Parfit thinks so, and he claims that, in the process, this theory unifies the main moral approaches, Kantianism, rule consequentialism, and contractualism, which, on his view, have been climbing the same mountain from different sides.5

This systematisation relies on making sure we uncover a moral theory that is internally coherent: it does not lead to contradictory injunctions. It also relies on checking that its conclusions align with our reasoned moral intuitions (those moral intuitions that withstand some scrutiny) about what the right thing to do is.

We might question what the epistemic status of such a systematisation is. Let us say Parfit is successful at building such a theory. What does that mean?

From a naturalistic point of view, there are basically two ways in which a theory can earn its keep. First, it should be logically coherent. Second, it should be answerable to experience in some way. Hume makes this point bluntly:

If we take in our hand any volume; of divinity or school metaphysics, for instance; let us ask, Does it contain any abstract reasoning concerning quantity or number? No. Does it contain any experimental reasoning concerning matter of fact and existence? No. Commit it then to the flames: for it can contain nothing but sophistry and illusion. — Hume

Parfit claims that his reasoning does not fit within this dichotomy, that it is possible to find irreducibly normative truths by rational reflection. In fact, I think he is in practice only doing a variation on it. To start with, he clearly aims to produce a consistent theory. That is central to his project. Then, how does he “test” whether this theory is acceptable? He says that he reasons with cases and principles. But in practice, on what grounds does a case lead him to accept or reject a conclusion? It is whether the injunction to act in a certain way in that case fits with our considered moral intuitions.6

In other words, Parfit uses moral intuitions as the raw material against which principles are tested, using clever thought experiments that set up unusual situations as stress tests for the general application of those principles. From a naturalist point of view, what he actually does is to organise our moral intuitions into a coherent system, rather than to open an independent window onto a further layer of moral reality.

A problem with this approach is that it gives a lot of weight to our moral intuitions as the primary substrate on which to build a theory. The fact that we have intuitions is obvious, but why should we treat our intuitions as pointing to some moral truth “out there” to be discovered?7

One reason given by Michael Huemer, who also defends moral realism, is that

It is reasonable to assume that things are as they appear, in the absence of grounds for doubting this. — Huemer8

Applied to morality, the idea is: moral truths appear to us to be objective, and unless we have strong reasons to distrust this appearance, we should take it at face value. Our intuitions about right and wrong are then treated as evidence for objective moral facts.

With all due respect to Huemer, this is a weak argument once we take seriously what we know about how our moral psychology evolved. We have very good independent explanations of why we have moral intuitions, without positing a realm of moral truths “out there”. Our moral sense appears to be an evolved tool for navigating social life: coordinating cooperation, managing conflicts of interest, punishing betrayal, and so on. Seen in that light, there is no particular reason to expect our intuitions to track some separate, non-natural moral reality. We should expect them instead to track what tends to work well for creatures like us in the kinds of social games we actually play.

Put differently: even if you can build a neat and coherent theory that organises these intuitions, that does not by itself show that the theory describes objective moral facts. It may simply be a good higher-level description of the strategies that help us do well in social interaction. If our shared moral intuitions are there to guide us towards sensible behaviour in repeated social interactions, they will need to be broadly compatible with equilibrium strategies of social games: strategies that it is in our interest to follow when others follow them too.

For instance, to the extent that it is good to cooperate with people who are willing to cooperate with us, and not to cooperate with people who just want to take advantage of us, we should have clear moral intuitions about reciprocity.

In pure epistemic terms, it is not clear why the enterprise Parfit undertakes would lead to objective moral truths that are “mind-independent”, as described by moral realists. This criticism was advanced precisely by philosopher Sharon Street:

The challenge for realist theories of value is to explain the relation between these evolutionary influences on our evaluative attitudes, on the one hand, and the independent evaluative truths that realism posits, on the other. Realism, I argue, can give no satisfactory account of this relation. — Street (2006)

Even if there were objective moral truths “out there”, there would be no reason to expect our moral intuitions to be special windows onto them, since these intuitions are shaped by evolution, which is a blind process with no reason to drive our cognition towards intangible moral truths. Instead, it shapes our cognition to help us be successful in the sense of surviving and having surviving offspring.9

Sam Harris’ naturalist moral realism

Sam Harris is a neuroscientist and public intellectual, with several New York Times best-selling books. He hosts the podcast Making Sense, which has a large international audience. Harris developed his perspective on morality in his book The Moral Landscape (2010).

I chose to discuss Harris’ views because many readers are likely familiar with them, and also because his starting point is naturalistic, like the one I adopt in this Substack. Harris wants to ground morality in facts about conscious creatures and their experiences, not in a religious or metaphysical realm. This makes it easier to engage with his ideas and to identify where disagreements lie.

Harris’ starting point is that we are evolved beings with nervous systems that generate conscious experiences of pain and pleasure. From this, he says that some types of experiences are better than others.

He asks us to imagine the “worst possible misery for everyone”: a world in which every conscious being suffers as much as possible, for as long as possible, with no compensating benefits. Opposed to that, we can imagine much better possible worlds. Harris then asks whether we are prepared not to call the first world “bad” and the others “better”. If we are, we have already accepted that there are objective moral facts: some states of the world really are better than others, because they contain more well-being.

There is an obvious question though: as much as we feel that the first world is awful, why should we call it bad in an absolutely-objectively-true way, rather than in a we-really-don’t-like-it way? Harris anticipates this question:

But you haven’t said why the well-being of conscious beings ought to matter to us. If someone wants to torture all conscious beings to the point of madness, what is to say that he isn’t just as ‘moral’ as you are?

His answer has two parts. First, he claims that consciousness is “the only intelligible domain of value”:

I think we can know, through reason alone, that consciousness is the only intelligible domain of value. What is the alternative? I invite you to try to think of a source of value that has absolutely nothing to do with the (actual or potential) experience of conscious beings.

Frankly, I do not follow. With his question “what is the alternative?”, he simply introduces the assumption that there must be some “value” in the world. This is a skyhook. If you assume there needs to be such an objective value, you have pretty much provided a justification for a normative theory out of thin air.10 But what tells us that such an objective value exists? Isn’t that the first question to answer?11

Also, Harris says that “through reason alone” we can know that consciousness is the only domain of value. But, actually, he does not show this using reason. He is really only saying: “come on, you know deep down it is true”.

So really, it is a call to our intuition, and Harris joins Parfit (and many other moral realists) using our intuitions as a primary source of evidence. Having argued for such a value, it is then easy to claim that there are objective rights and wrongs.

Once we admit that the extremes of absolute misery and absolute flourishing—whatever these states amount to for each particular being in the end—are different and dependent on facts about the universe, then we have admitted that there are right and wrong answers to questions of morality.

Why should we agree with his line of argument? Because it is obvious, and, if we don’t, Harris suggests we are a kind of psychopath:

Anyone who doesn’t see that the Good Life is preferable to the Bad Life is unlikely to have anything to contribute to a discussion about human well-being. Must we really argue that beneficence, trust, creativity, etc., enjoyed in the context of a prosperous civil society are better than the horrors of civil war endured in a steaming jungle filled with aggressive insects carrying dangerous pathogens? I don’t think so. (emphasis mine)

The key here is what sense is given to better. Obviously, pretty much everybody would prefer to live in a prosperous civil society rather than in a civil-war-ridden society. But that does not mean that there is an objective value out there saying so. Indeed, the whole exercise should be to prove it, not say it is self-evident.12 Saying so is relying on our intuitions shaped by evolution.

So, Harris is a naturalist who smuggles an ought out of thin air on the ground that it is obvious if we think about it. The problem is that an immaterial ought is anything but obvious. Starting from a naturalist point of view has the merit of grounding the theory in well-known facts about human psychology, but it also makes the leap of faith to a normative standpoint easier to spot in its arbitrariness.

The motivation: appeal to consequences

Reflecting on his achievement, Parfit stated:

My life is my work. I believe I have found some good reasons for believing that values aren’t just subjective and that some things really do matter. If my arguments don’t succeed, my life has been wasted. — Edmonds (2021)

This existential motivation should, if anything, make us doubt Parfit’s intellectual impartiality. In a previous post, I described how the desire to believe something is true because we prefer it to be true, what I call the Santa Claus fallacy, often underlies technical arguments about morality.13

When looking at the motivations of moral realists, they often openly express a strong preference for the existence of moral truths, to the point of fitting the Santa Claus fallacy.14

Consider, for instance, David Enoch in Taking Morality Seriously:

In particular, under non-objectivist metaethical assumptions […] it would be morally impermissible to stand one’s moral ground in any number of conflicts or disagreements where it does seem permissible (perhaps even required) to stand one’s moral ground.15

In other words: if we want to be able to say that something is morally wrong, we need to believe there are objective moral truths.

Another moral realist, Russ Shafer-Landau, writes in Moral Realism: A Defence:

[if modesty, liberality, and tolerance were human constructs], then the illiberal and the intolerant, who have constructed things differently, would be making no error. — Shafer-Landau

In other words: if we want to be able to say that the illiberal and intolerant are morally wrong, we need to believe there are objective moral truths.

Huemer says in Ethical Intuitionism:

Lastly, moral anti-realism undermines our sense of meaning in life, and this brings me to one of the reasons why I find anti-realism unbelievable. I think anti-realism really boils down to the view that nothing matters. […] Life has no meaning, because “meaning” is one of those ‘spooky’, non-natural properties that anti-realists do not believe in.16 — Huemer

In other words: if we want our lives to have meaning, we need to believe that there are objective moral truths.

Turning back to the two thinkers I discussed earlier, Parfit wrote in On What Matters:

On subjective theories, nothing matters. We should reject the arguments for this bleak view.

And Harris wrote in The Moral Landscape:

Few people seem to recognize the dangers posed by thinking that there are no true answers to moral questions.

These appeals to consequences are flawed arguments. Whether I prefer something to be true or not should not influence my belief in it being true. I might have a very strong preference for Santa Claus to be real; it won't make it in the least more real. Similarly, the fact that I would very much like moral rules to be objective because of the judgements they would allow me to make is irrelevant to assessing whether these rules exist objectively.17 That such brilliant minds resort to such a flawed type of argument is testimony to the fact that our reason is a tool selected for us to convince ourselves and others, not primarily to find the truth.

In the end, the search for a justification for an objective theory of moral truth might be seen as an attempt to retain, within a secular framework, the type of absolute moral mindset that Judeo-Christian traditions had until now provided. This criticism was explicitly made by Sartre

Towards 1880, when the French professors endeavoured to formulate a secular morality, they said something like this: God is a useless and costly hypothesis, so we will do without it. However, if we are to have morality, a society and a law-abiding world, it is essential that certain values should be taken seriously; they must have an a priori existence ascribed to them. It must be considered obligatory a priori to be honest, not to lie, not to beat one’s wife, to bring up children and so forth; so we are going to do a little work on this subject, which will enable us to show that these values exist all the same, inscribed in an intelligible heaven although, of course, there is no God. […] nothing will be changed if God does not exist; we shall rediscover the same norms of honesty, progress and humanity, and we shall have disposed of God as an out-of-date hypothesis which will die away quietly of itself. — Sartre (1946)

William James suggested that philosophers often fall into two temperamental camps: the tender-minded, who are drawn to idealistic, and often more comforting, views, and the tough-minded, who insist on sticking to experience, facts, and logic, even when the resulting conclusions are not especially appealing.

This Substack definitely stands on the tough-minded side. There is no point in trying to argue that Santa exists just because we would like it to be the case. If we agree there is a reality out there, it does not depend on what we would like it to be. Understanding it as it is requires us to follow where logic and facts guide us. Furthermore, given the design of human reason—filled with reasoning biases that conveniently land us on our preferred conclusions—we should lean towards being suspicious of, rather than drawn to, our own preferences. In the domain of morality, this means that while we may find the idea of absolute moral truths appealing, we should be particularly vigilant about the reasons we accept in support of this view. The view I present in this post is that, when we do so, there are no good reasons to believe that it makes sense to talk of an objective moral reality out there.18

This conclusion leads to a natural question. Does the rejection of moral realism imply a radical relativism that takes the form of subjectivism (moral judgements are nothing more than personal preferences)? Does it mean that anything goes and that we have to agree that things like torture, slavery, and genocide are fine?

The answer here is negative. If morality is understood as being about social rules of cooperation, moral judgments are not just meaningless drivel. In that sense, the naturalistic perspective I develop here is different from the conclusions Nietzsche draws about morality. Discussing this, Binmore states:

[…] the classic error that Nietzsche makes on behalf of all the other thinkers who contrive to muddle the distinction between relativism and subjectivism is to confuse moral values and personal tastes. Morality evolved to regulate interactive human behavior. All the members of a particular society need to be playing the same morality game for its rules to serve any purpose. What would be the point of morality if it differed not only from one society to another, but also from one individual to another? — Binmore (1998, emphasis in the original)

A natural question is: what exactly is morality in this naturalist perspective? This will be the topic of my next post.

References

Binmore, KG (1998) Game Theory and the Social Contract, Volume 2: Just Playing. MIT Press, Cambridge, MA.

Bourget, D & Chalmers, DJ (2023) ‘Philosophers on Philosophy: The 2020 PhilPapers Survey’, Philosophers’ Imprint, 23(11).

Edmonds, D (2023) Parfit: A Philosopher and His Mission to Save Morality. Princeton University Press, Princeton, NJ.

Enoch, D (2011) Taking Morality Seriously: A Defence of Robust Realism. Oxford University Press, Oxford.

Golub, C (2021) ‘Is there a Good Moral Argument against Moral Realism?’, Ethical Theory and Moral Practice, 24(1), pp. 151–164.

Harris, S (2010) The Moral Landscape: How Science Can Determine Human Values. Free Press, New York.

Huemer, M (2005) Ethical Intuitionism. Palgrave Macmillan, Basingstoke.

Hume, D (1748) Philosophical Essays Concerning Human Understanding. A. Millar, London.

Hume, D (1755) ‘Of Suicide’, in EF Miller (ed.) Essays: Moral, Political, and Literary. Liberty Fund, Indianapolis.

James, W (1907) Pragmatism: A New Name for Some Old Ways of Thinking. Longmans, Green & Co., New York.

Mackie, J.L. (1977) Ethics: Inventing Right and Wrong. Harmondsworth: Penguin Books.

Nietzsche, F (1886) Beyond Good and Evil: Prelude to a Philosophy of the Future. C. G. Naumann, Leipzig.

Parfit, D (2011) On What Matters. 2 vols. Oxford University Press, Oxford.

Pölzler, T & Wright, JC (2019) ‘Empirical research on folk moral objectivism’, Philosophy Compass, 14(5), e12589.

Sartre, J-P (1946) L’existentialisme est un humanisme. Les Éditions Nagel, Paris.

Shafer-Landau, R (2003) Moral Realism: A Defence. Clarendon Press, Oxford.

Stanford, PK (2018) ‘The difference between ice cream and Nazis: Moral externalization and the evolution of human cooperation’, Behavioral and Brain Sciences, 41, e95.

Street, S (2006) ‘A Darwinian Dilemma for Realist Theories of Value’, Philosophical Studies, 127(1), pp. 109–166.

13

One of the commenters to my post made exactly this point

Religion is one of the most natural justifications for the existence of objective moral truths. Sacred texts handed out to us by a superior divine being provide a seemingly natural origin to morality.

This idea is so prevalent that the absence of belief in God is associated with an anxiety about the absence of morality. Sartre, an atheist, pointed out this concern:

[It is] extremely embarrassing that God does not exist, for there disappears with Him all possibility of finding values in an intelligible heaven. There can no longer be any good a priori, since there is no infinite and perfect consciousness to think it. It is nowhere written that ‘the good’ exists, that one must be honest or must not lie, since we are now upon the plane where there are only men. — Sartre

This belief that God is necessary to hold together our desire for morality to mean something is often used to justify explicitly or implicitly the existence of God, in what I call a Santa Claus fallacy—”Santa must exist because it would be too horrible if he didn’t”.

Here I explain why the idea that morality requires religion gets things backward, not only does religion not clearly underpin our moral sense, it is most likely our moral sense which has shaped religions.

The Euthyphro dilemma

Sartre’s quote reflects a widely shared belief that it is the existence of God which justifies the existence of moral truths. But is it really the case?

A problem with this idea was pointed out more than 2,000 years ago by Plato. In his dialogue Euthyphro, he presents a dilemma about morality and God. Is something moral because God said it is? Or does God give us injunctions to do something because it is moral?

If we accept the idea that something is moral because God said so, it makes morality arbitrary in the sense that we would have to accept as moral whatever God told us was moral. What if God told us that stealing or killing our neighbour is moral? To many, it would seem wrong for God to say that.

If, on the contrary, we say that God gives us injunctions because they are moral, then morality must be defined outside of God. In that case, we have not solved the question of the origin of morality.1

When religion conflicts with our moral intuitions

Evidence from psychology seems to suggest that believers behave as if they assume morality exists outside of God. This is illustrated by how they react to scenarios where God would recommend that we do things we unambiguously consider bad. Studying morality in children, psychologist Larry Nucci found that:

[…] the majority of children at all ages and from both religious groups rejected the notion that God’s command would make it right for a person to steal. […] Instead, children attempt to coordinate their notions of God with what they know to be morally right. — Nucci (1991)

He cites this interview with a conservative Jewish boy as an illustration:

Question. Let’s suppose that God had written in the Torah that Jews should steal, would it then be right for Jews to steal?

Answer. No.

Question. Why not?

Answer. Even if God says it, we know he can’t mean it, because we know it is a very bad thing to steal. We know he can’t mean it. Maybe it’s a test, but we just know he can’t mean it.

Question. Why wouldn’t God mean it?

Answer. Because we think of God as very good—absolutely perfect person.

Question. And because He’s perfect, He wouldn’t say to steal? Why not?

Answer. Well—because we people are not perfect, but we still understand. We are not dumb either. We still understand that stealing is a bad thing.2

A recent psychological study found that when asked whether it would be right for God to support a range of behaviour we typically reject, most children stuck to their moral norms instead of accepting the idea that they should simply follow a divine injunction.

Children of varying religious backgrounds identified moral norms… as authority-independent… even if God had made a commandment requiring people to steal, stealing would still be immoral. — Reinecke and Solomon (2023)

Another element in support of the primacy of our moral intuitions is that critics of religion often point to elements in them that conflict with our modern moral sense—like the justification of mass killings or of the inferior status of women.3 What would be the point of these criticisms if believers were just willing to accept that whatever God prescribes is the right thing to do? If critics resort to such criticisms, it is because they believe it will appeal to a moral sense that exists outside of religion and can be used to convince believers to disagree with religious teaching.

The cultural evolution of religions and how our moral sense shaped it

There are indeed good reasons to think that our morality precedes religion: our moral sense is a tool to help us navigate social interactions and cooperate successfully with others without being pushovers. Nicolas Baumard developed precisely this view in his book The Origins of Fairness:

Our moral disposition is specialized, autonomous, universal, and innate because it was selected by evolution. — Baumard (2010)

It is then not religion that gave us our moral intuitions, but our moral intuitions that helped shape religions. In support of this view, one can first observe that religions are incredibly diverse and that many forms of religion observed across times and places differ radically from the large monotheist religions we are accustomed to.

The anthropologist Stewart Guthrie discusses this diversity in his book Faces in the Clouds. This kind of comparison is useful to rule out some simple explanations of why religion arose in the first place.

Religion does not simply emerge to offer us a belief in an afterlife, because many religions do not place much weight on a consoling afterlife, and, in others, the afterlife is barely articulated or quite bleak.

Religion cannot have simply emerged to make sense of the world in some grand or ultimate way, because many religions are largely this-worldly and pragmatic, concerned with local misfortunes rather than with providing a comprehensive cosmology.

Religion cannot be simply about making us moral either, because many religions do not prescribe systematic moral behaviour, and some famous gods (such as the Greek gods) show little intrinsic interest in morality.

What Guthrie suggests is that religions arise, at least in part, from our evolved propensity to over-detect agency in the environment: we quite literally tend to see faces in the clouds. From an evolutionary point of view, this bias is useful because the cost of a false positive (thinking there is an agent when there is none) is usually small, while the cost of a false negative (failing to notice a predator, rival or ally) can be fatal. In ambiguous situations, it is therefore adaptive to be overly susceptible to assuming that some agent might be present. That is why we find creaking noises in old houses creepy and why we might jolt at the sound of leaves rustling in the bushes.

Guthrie’s claim is that religious concepts build on this general tendency: once the mind is primed to interpret ambiguous patterns as agents, it is a short step to populating the world with invisible persons, spirits and gods.

The anthropologist Pascal Boyer builds on Guthrie’s idea and argues that this propensity to over-detect agency helps explain the appeal of spirits and gods as minimally counterintuitive agents: agents that respect most of our intuitions about behaviour and even physics, with a few notable exceptions such as being invisible or able to perform miracles. Boyer suggests that religious traditions naturally blend such agents with our moral and social intuitions, so that gods and spirits have desires, likes and dislikes, form bonds with us, and take revenge if we break these bonds.

As such stories evolve, some themes are likely to act as cultural attractors4 because of how they resonate with our existing intuitions and preferences. This seems to be the case for moralising gods,5 who justify and enforce our moral sense. Moralising gods might be particularly appealing psychologically because they provide a justification for our feelings that good deeds should be rewarded and bad deeds punished. They satisfy our desire for a just world where no crime goes unpunished. Baumard and Boyer have argued that it is precisely this intuitive appeal that made such religious narratives historically successful.

Despite their differences, religious and non-religious movements owe their cultural success to the fact that these explicit, coherent accounts of moral prescriptions are congruent with universal, and much older, evolved moral intuitions. — Baumard and Boyer (2013)

In addition, moralising gods can help make societies more cohesive and cooperative, and therefore more successful. By creating beliefs that justify cooperative behaviour and increase the perceived costs of anti-social behaviour, moralising gods might help societies sustain a higher degree of social trust and a higher level of cooperation.6 In his sociology of religion, Max Weber has emphasised that one of the reasons for the economic success of Protestants was the effect of an ever-judging God (with the absence of the regular process of forgiveness of confession present in Catholicism):

[The puritan Protestant was characterised by] trust, especially his economic trust in the absolutely unshakable and religiously determined righteousness of his brother in faith. — Weber 1915

The game-theoretic foundations of our morality

So instead of religions being the foundation of our morality, it is our moral sense that likely led large world religions to evolve progressively into their moralising form. In support of this view, we have very good ideas about why our moral sense has precisely the features it has.

One of the most striking findings of the research in game theory in the second part of the 20th century was that it is rational to cooperate in repeated interactions7 and that very simple strategies like tit-for-tat (being kind to people who have been kind to you and not being kind to those who have not).8 The reason is simple. When people interact over time, being uncooperative now risks losing the gains from cooperation later. In his book The Evolution of Cooperation, Robert Axelrod stressed how the shadow of the future can make honesty and cooperation the best policy.

These findings left, however, one issue unresolved: there are many ways to cooperate and each way is associated with different splits from the gains of cooperation. Alice and Bob might be happy to form a family, but to have a happy and functional family they need to split the house chores, the childcare duties, the income-generating activities, and so on. There are many ways to split these things, and 50-50 everywhere is not necessarily a feasible or preferred solution. To start with, Alice cannot share the long and difficult process of making children with her partner, and she might expect him to pull his weight in other domains.

A key insight from game theorist Ken Binmore is that agreeing on how to organise cooperation, and split the gains it generates, involves some bargaining. He developed a theory where a key cognitive tool to solve this problem is for us to put ourselves in the shoes of other people, to take their side to be able to propose solutions that would be agreeable to all.

This cognitive tool is nothing else than the Golden Rule: “Do unto others as you would have them do unto you.”9 A striking aspect of this rule is how it can be found in pretty much the same form across widely different cultures and religions.10

It can also be found in secular ethical philosophy. Kant’s categorical imperative can be understood as a sophisticated elaboration of the Golden Rule:

Act only in accordance with that maxim through which you can at the same time will that it become a universal law. — Kant (1785)11

And John Rawls presents the original position with its veil of ignorance as a procedural way of applying Kant’s categorical imperative to the choice of social rules.

The original position may be viewed, then, as a procedural interpretation of Kant’s conception of autonomy and the categorical imperative within the framework of an empirical theory. — Rawls (1971)

Binmore thinks that such philosophies have been attractive because the Golden Rule, which he sees as a “simplified version of the original position,”12 is a fundamental principle of our psychology to navigate social interactions and negotiate with others. In short, we put ourselves in other people’s shoes to find a solution that works for all.13

For me, the original position is merely a stylized representation of a fairness norm that evolved along with the human race for the purpose of resolving small-scale coordination problems. — Binmore (1998)

In that perspective, it is not just reciprocity that is one of the foundations of our psychology; it is also about how we organise this reciprocity to coordinate on mutually agreed solutions. These deep intuitions exist universally because they are solutions to problems of cooperation faced for eons by our ancestors. Our moral sense is based on these deep intuitions. They do not derive from religion; instead, they shaped world religions.

In his article on Morality and religion in the Encyclopedia of Religion, Ronald Green observed the uncanny commonalities of religions’ moral principles in spite of their large diversity in traditions and beliefs.

One of the most striking impressions produced by comparative study of religious ethics is the similarity in basic moral codes and teachings. […] It is sometimes assumed, because religious traditions hold widely different religious beliefs, that their ethics must correspondingly differ; what is remarkable, however, is that these great differences in beliefs apparently do not affect adherence to at least the fundamental moral rules. — Green (1987)

This striking commonality makes sense once we understand that it reflects deep moral intuitions that precede the existence of large world religions, because they are part of the cognitive tools evolution has handed us to be successful as part of a highly social species.

It is common to hear the concern that without religion, morality is impossible. This perspective gets things backward. It is not religions that induce moral motives in humans. Instead, our moral motives pre-existed religions and world religions built on these and gave them a supernatural justification.

References

Atkinson, Q.D. and Bourrat, P. (2011) ‘Beliefs about God, the afterlife and morality support the role of supernatural policing in human cooperation’, Evolution and Human Behavior, 32(1), pp. 41–49.

Axelrod, R. (1984) The Evolution of Cooperation. New York: Basic Books.

Baumard, N. (2010) Comment nous sommes devenus moraux: Une histoire naturelle du bien et du mal [English translation: The Origins of Fairness]. Paris: Odile Jacob.

Baumard, N. and Boyer, P. (2013) ‘Explaining moral religions’, Trends in Cognitive Sciences, 17(6), pp. 272–280.

Binmore, K. (1998) Game Theory and the Social Contract, Volume 2: Just Playing. Cambridge, MA: MIT Press.

Binmore, K. (2005) Natural Justice. Oxford: Oxford University Press.

Boyer, P. (2001) Religion Explained: The Evolutionary Origins of Religious Thought. New York: Basic Books.

Darwin, C. (1871) The Descent of Man, and Selection in Relation to Sex. London: John Murray.

Dawkins, R. (2006) The God Delusion. London: Bantam Press.

Green, R.M. (1987) ‘Morality and religion’, in Eliade, M. (ed.) Encyclopedia of Religion, Vol. 10. New York: Macmillan, pp. 92–106.

Guthrie, S. (1993) Faces in the Clouds: A New Theory of Religion. New York: Oxford University Press.

Kant, I. (1785) Groundwork of the Metaphysics of Morals. Riga: Johann Friedrich Hartknoch.

Norenzayan, A. (2013) Big Gods: How Religion Transformed Cooperation and Conflict. Princeton, NJ: Princeton University Press.

Nucci, L.P. (1991) ‘Doing justice to morality in contemporary values education’, in Benninga, J.S. (ed.) Moral, Character, and Civic Education in the Elementary School. New York: Teachers College Press, pp. 21–39.

Plato (c. 399 BCE) Euthyphro. In: Cooper, J.M. (ed.) (1997) Plato: Complete Works. Indianapolis, IN: Hackett Publishing.

Rawls, J. (1971) A Theory of Justice. Cambridge, MA: Harvard University Press.

Reinecke, M.G. and Solomon, L.H. (2023) ‘Children deny that God could change morality’, Cognitive Development, 68, Article 101393.

Roes, F.L. and Raymond, M. (2003) ‘Belief in moralizing gods’, Evolution and Human Behavior, 24(2), pp. 126–135.

Russell, B. (1927) Why I Am Not a Christian. London: Watts & Co.

Sartre, J.-P. (1946) L’existentialisme est un humanisme. Paris: Éditions Nagel.

Sperber, D. (1996) Explaining Culture: A Naturalistic Approach. Oxford: Blackwell.

Trivers, R.L. (1971) ‘The evolution of reciprocal altruism’, The Quarterly Review of Biology, 46(1), pp. 35–57.

Weber, M. (1904–1905) ‘Die protestantische Ethik und der Geist des Kapitalismus’, Archiv für Sozialwissenschaft und Sozialpolitik, 20–21.

Soon after they learn to talk, children seem to grasp one somewhat abstract notion: fairness. They are quick to decry the unfairness of many situations: how food is shared at the dinner table, how their parents’ attention is divided among siblings, or how their teacher treats them compared with other pupils. This early disposition to identify what is morally wrong1 and what should be done to make it right further develops as they grow up. In the adult world, normative statements are commonplace. Turn on the TV, and an intellectual will tell you that society should be like this or like that; a politician will say that it is not right for the government to act that way; and an activist will state that social justice requires a change in policy to improve society.

But wait a minute, what do all these statements mean? What does it mean to say we ought to do that, or that it is right to do this? Asking these questions might surprisingly reveal that we typically don’t have very good answers! And, even more surprisingly, experienced thinkers’ answers often rely on intellectual tricks, skyhooks, whose weakness is overlooked only because they deliver the kind of answer our intuitions find comforting.

In this post, and upcoming ones, I argue that we can build a theory of morality without such tricks, a theory consistent with our understanding of natural and behavioural sciences. This theory provides strikingly simple answers to the deep questions that have been asked in philosophy about morality.

Morality put to the toddler’s acid test

Because moral statements are so commonplace, it is easy to fail to appreciate how shallow our thinking about their meaning is. To see this, let’s put them to what I’d call the toddler’s acid test. Toddlers don’t know much, but they are often able to unsettle our confidence in our own “knowledge” with a simple trick: asking questions like “why?”, not just once, but repeatedly, until it becomes evident that we do not actually have all the answers.

Take a simple interaction from the 1990 children's picture book Why? In that book, a little girl, Lily, keeps asking “why?” to her father. Such an unrelenting line of inquiry quickly unveils the dangerously thin layer of knowledge on which adults’ daily confidence is actually based. Consider this possible exchange:

—Don’t sit down there on the grass, Lily.

—Why?

—Because it is not good to have stains on your dress.

—Why is it not good?

—Because you would not look respectable when our friends arrive.

—Why should I look respectable?

[Most dads who are not experts in the philosophy of social norms would start struggling here]

Adults typically end such discussions with “because that’s the way it is”, which is asking for faith in a pragmatic solution that works instead of providing a deep understanding.2

Now imagine a situation where the toddler asks a question about fairness. Let's say she observes on the news that a policy is implemented to increase the income of a category of the population. She inquires

—Why this policy?

—Because they face some hardship and the government is helping them.

—Why is the government helping them?

—Because it is good to help people who have difficulties.

—Why is it good to help people with difficulties?

In truth, most of us would struggle to articulate a meaningful answer to such a disarming toddler. In fact, many professional philosophers would too.3

The physicist Richard Feynman made a very interesting observation about “why?” questions: they are potentially endless.

When you explain a “why,” you have to be inside some framework that allows something to be taken as true. Otherwise, you’re perpetually asking “why.” — Feynman in an interview

One way to end the cycle of “why” questions about morality is to reach a solid and clear foundation. In the same way as mathematics can be grounded in initial statements, axioms, that are accepted as self-evidently true, one hope is that we can do the same about morality. Unfortunately, finding such principles, if ever they exist, has not proven easy.4 De facto, instead of relying on such foundations, theories of morality are often held by skyhooks.

Skyhooks

Skyhooks are explanations that hold our thesis in the air without any further grounding.5 The clearest example of skyhooks is the assumption that morality comes literally from the sky: its rules are handed to us by divine beings.

Examples of divine codes of morality are the Law Code given by the Sun-god Shamash to King Hammurabi, or the Ten Commandments given to Moses on Mount Sinai. Such stories seemingly solve the problem of the foundations of morality. You just have to look at the sacred book of your religion to find out the rules of morality.

A secular alternative is the idea that there are moral truths independently of religious beliefs. It is a widely held intuition that morality is somehow objective, as if there were absolute laws of morality, woven into the fabric of the universe, and which only need to be unveiled. The philosopher Grotius proposed a formal version of this intuition. He argued that humans have natural rights that emerge from natural laws that exist out there, so to speak.6 It is as if there were an invisible book of morality out there in the sky, only requiring us to find ways to recover its content.