Social Norms as Rules of Social Games
Social Norms as Rules of Social Games
How game theory unveils the nature and workings of social norms
One of the most interesting aspects of human life is how our behaviour is articulated around social norms that, in many ways can seem arbitrary. People drive on the right of the road in many countries but on the left in others. They shake hands to greet each other in some places, and hug and kiss in others. In some cultures, guests finish their plates when invited for dinner to show appreciation for the meal, while in others they leave food to indicate they were served enough.
We are well aware of these differences. The saying “when in Rome, do as the Romans do” reflects our understanding that we should adapt to the local norms to fit in. But what exactly are norms? Why do they matter, how do they shape our behaviour and how do they change? Social norms are one of the most puzzling aspects of human behaviour, often seeming to defy rational explanation. In this post, I’ll show how game theory can provide clear answers to these questions.
Social norms as cultural moulds
To discuss social norms, let’s start with what sociology can tell us about them. Unfortunately, some sociologists, like Christine Horne have highlighted the many unanswered questions about social norms.
One aspect of social norms emphasised in sociology is that they act as external constraints on our behaviour, originating from society. This view was already proposed by the father of sociology, Emile Durkheim, who aimed to clearly separate the new discipline from psychology by making it clear that there are some things that influence behaviour and are social in nature and external to each individual. He defined “social facts” as:
Manners of acting, thinking and feeling external to the individual, which are invested with a coercive power by virtue of which they exercise control over him. - Durkheim (1895)
Social norms are such facts. They are rules of behaviour existing in a given social context. These rules are coercive, in part because they are enforced by others when we are found not to respect them.
Norms are ordinarily enforced by sanctions, which are either rewards for carrying out those actions regarded as correct or punishments for carrying out those actions regarded as incorrect. - Coleman (1990)
In addition to external sanctions, sociologists have argued that norms can become internalised. The sociologist Talcott Parsons is known for stating that individuals internalise the values and the social roles they must conform to in a functional society.
A stable system requires above all the internalization of value-orientations to a degree which will sufficiently integrate the goals of the person with the goals of the collectivity. - Parsons and Shils (1951)
As a consequence of this internalisation of social norms, people feel psychological pressure to conform to the social roles and expectations ascribed to them. In his discussion of social norms, Jon Elster (1994) pointed out that they are “sustained by the feelings of embarrassment, anxiety, guilt, and shame that a person suffers at the prospect of violating them.”1
In that view, individuals are receptacles of social and cultural elements that shape their personality, behaviour and feelings. There is limited focus on the agency of individuals, their personal desire, their incentives, and their strategies. Instead, the focus is on the social constraints that shape their actions and beliefs. The feeling of autonomy, the idea that we do what we want in our lives is largely an illusion.
We are the victims of an illusion which leads us to believe we have ourselves produced what has been imposed upon us externally. - Durkheim (1895)
These views have influenced the discourse on social norms. Individuals are often described either as naive social pawns following blindly arbitrary rules or as unhappy victims of the coercion that “society” puts on them. This way of seeing social norms permeates public discussions when existing norms are being criticised. One example is gender norms which are often described as external constraints imposed by society.
This vision, however, raises a range of questions. How does “society” create norms? Why are these norms here instead of others? How does society drill these norms into our heads and make us feel bad for deviating from them? Why would people be so naive and lack agency in regard to social norms? How to explain that people may be so compliant at following some norms, while they can be very apt at finding ways not to follow other norms like when they engage in tax avoidance or speed on the road?2
These questions can be answered and it is possible to make sense of the features of social norms by adopting a richer vision of social behaviour that emphasises the agency of individuals.
Social norms as equilibria of social games.
The nature and functioning of social norms are considerably clarified using insights from game theory. There is a compelling body of literature that describes and discusses social norms as equilibria of social games. As I have discussed in a previous post, an equilibrium in game theory is a situation where, given what everybody else is doing, you do not have an interest in changing your actions.
The philosopher David Lewis, was among the first scholars to approach social norms as equilibria from a game-theoretic standpoint, in his influential book, Convention (1969).
He sees social conventions as solutions to coordination games. A typical example is the choice of driving on one side of the road. Drivers are better off driving on the same side to avoid collisions. For arbitrary reasons, a country may have adopted one side instead of the other as a convention. It is then in the interest of every driver to follow this convention and not deviate. You may think that we drive on one side of the road because it is the law. But laws only work if they reflect a social convention. It is easy to establish a law that enshrines a previously existing convention. But governments who try to implement new laws deviating from existing conventions typically find it very difficult (Basu, 2018).3
Coordination games are everywhere when you think about it. This is partly because successful interactions with others require some degree of shared intentionality (O'Madagain and Tomasello, 2022). That is, we need a common understanding of the situation and a common understanding that this understanding is shared by other people. We also need an understanding of the likely intentions of other people and how other people’s actions inform us about their intentions. Similarly, we need to be confident that they understand our intentions and understand how our actions signal our intentions.
The social norms that pervade our social interactions are therefore not “arbitrary” in that sense. They are good solutions that greatly facilitate social interactions and remove doubts and uncertainties that would act as frictions limiting our ability to quickly navigate social situations. Social norms allow us to seamlessly do things that are much more complex than they look, like walking in a crowded public space, ordering a meal in a restaurant, and taking turns talking in meetings.
Seeing somebody violating a norm, breaks our expectations about this person’s behaviour. Even if the violation is small it may raise reasonable concerns: what else may this person be doing next? It is why minor deviations of social norms often generate feelings of awkwardness, sometimes creepiness.
The game theorist Ken Binmore (2005) extended this logic to norms of fairness which determine who should get what in social situations where benefits and costs have to be shared among people. A key insight from game theory is that substantial gains can often be secured from sustained cooperation in repeated interactions. However, the manner in which these gains can be shared is not uniquely determined. Technically, there are many efficient equilibria, where players could agree to share the gains from cooperation and consequently be better off than not cooperating. However, some of these agreements would lead to a very unequal division, with one player receiving a disproportionately larger share of the benefits. To successfully sustain cooperation, people must agree on how to divide the gains from such collaboration.
Binmore points out that fairness norms are helping us to do exactly this. These norms function as conventions that facilitate coordination and the selection of one of the efficient equilibria. The “fair” way to share benefits and costs is just one of many possible means of dividing gains. It is the option we commonly agree to settle on, and it is the one we expect others to consent to. Our willingness to accept it and our shared expectations that others will also comply help us to coordinate swiftly without needing to renegotiate each time.
Fairness norms are ubiquitous. They pervade our social interactions and help us solve problems like: Deciding which child should get the cherry on the top of the cake (“Your brother is the youngest, we’ll give it to him”); Deciding who should get praise in collective endeavours (“Susan should be particularly congratulated on the amount of work she put into this project”); Deciding who should have a say in collective decisions (“The inhabitants of Smallville had to be consulted as this highway project would impact their community”).
From understanding norms as equilibria of social games, we recover and explain key aspects of norms discussed in sociology:
Norms are external to individuals because equilibria emerge as social configurations that are usually beyond the control of each individual.
Norms constrain individual behaviour because they come with social rewards for compliance and penalties for deviations. We are those creating the corresponding rewards and penalties because it is in our interest to be seen as invested in shared norms and not to be seen as condoning norm violations. Norms are therefore self-enforcing because enforcement is part of the norm. 4
This perspective also helps solve the puzzles about social norms.
People are not social pawns who follow social norms blindly. Instead, they react to the social rewards and penalties they can expect from their actions.
The psychological strain felt when violating norms has a function. The feelings of shame and guilt act as the shadow costs of the risks of future social penalties (Gilbert 2003; Sznycer, et al. 2016). They help us to anticipate the external risks of being found violating social norms in terms of loss of reputation and material punishment.5
Norms (and the psychological cost of deviating from them) are not created and imposed by something called “society”. Instead, they are stable configurations of expectations and behaviours that perpetuate themselves because each individual has an interest to have these expectations and adopt these behaviours. They are de facto rules of the social games we play, and—typically—no one either created them in the past or controls them now.
Norm changes
An essential insight in considering social norms as equilibria is that they are not necessarily “good”. The origin of social norms can be accidental in nature and their effect might be collectively harmful.
Note that being an equilibrium does not make a social norm good or efficient; there are lots of bad equilibria around. It simply means that the expectations and actions of all the parties concerned are consistent, or that their expectations are self-fulfilling. - Bicchieri (2005)
In that case, society would benefit from changing norms, but such changes can be hard as people don’t have an interest in altering their behaviour until others do likewise. Understanding norms as equilibria helps us understand how norms can be changed.
Changing collective expectations
The view that people are naively driven by social norms may result in support for a paternalistic approach to behavioural alteration, guided by supposedly more informed actors like government officials. Yet evidence on interventions targeting social norm change shows that merely instructing people about the “better” behaviours they should adopt often fails (Tankard and Paluck, 2016).
Changing norms necessitates a collective shift in expectations and behaviours which can be difficult to achieve. People may reasonably balk at the idea of being the first mover as they face the risk of being penalised for deviating from the existing norm, without gaining benefits from the new norm in case it does not catch on. This aspect of norms explains the way they change: norms tend to be sticky most of the time, but they tend to change quickly when they do change.
The process will tend to exhibit long periods of stasis in which it is close to some equilibrium, punctuated by relatively brief periods in which the equilibrium shifts in response to stochastic shocks. We call this the punctuated equilibrium effect. Young (1998)
The histories of both the legalisation of marijuana in the USA and the legalisation of same-sex marriages illustrate how norms can appear almost immutable for extended periods, only to change swiftly later on.
A quick change of social norms is typically associated with a change of the beliefs and preferences of a large proportion of people, in favour of the new norm. This evolution is in line with the view that individuals’ internalised beliefs and behaviour do not exist because they are imposed by “society” for norms to persist. Instead, norms persist as social equilibria, and individuals have incentives to uphold beliefs and preferences that align with the existing norms. When social norms do change, people have an interest to align their beliefs and preferences with the new norm. People can prove very flexible and able to do so.
Healthy scepticism towards external interventions
The understanding that people are not naively following social norms adds another reason to have some scepticism toward paternalistic external interventions. Nobel Prize winner Elinor Ostrom (1990) made the point that people are not helpless at solving social problems. On the contrary, social groups can be very good at setting up effective norms through collective discussion and self-organisation. The possibility to discuss and negotiate collectively a new norm can be pivotal to facilitate a shift in social norms. Locally negotiated norms frequently produce better outcomes than externally imposed rules by government officials.
The effectiveness of leaders and role models
The nature of norms as equilibria also helps explain why group leaders, role models and people highly connected socially can have a much larger impact on the change in social norms than the average person (e.g. Paluck and Shepherd 2012). These effects are likely not primarily because people blindly follow leaders, but rather because leaders and role models can act as coordination mechanisms that help people trust that others will synchronously change their behaviour and expectations.
Whether leaders know what they are doing better than their followers or not, they can be very useful to a society as a coordinating device for solving the equilibrium selection problem. - Binmore (2001)
In summary, viewing social norms as equilibria of social games helps us understand their features—why they are stable and how they constrain individual behaviour through rewards for compliance and penalties for deviations. It also enlightens us on why society can sometimes be mired in undesirable social norms and how these norms can evolve. Most importantly, it portrays people not as mere automatons living with an illusory sense of autonomy but as individuals primarily interested in successfully navigating social interactions. When social norms change, people can adapt with notable agility. In essence, people are not social pawns; they are players of social games, though they (mostly) do not set the games’ rules.
This post is part of a series on how game theory helps us make sense of a wide range of social behaviours. In previous posts, I wrote a quick introduction to game theory, and I showed how it is strikingly effective at explaining tennis players’ strategies. You can subscribe to receive future posts.
References
Basu, K., 2018. The republic of beliefs: A new approach to law and economics. Princeton university press.
Bicchieri, C., 2005. The grammar of society: The nature and dynamics of social norms. Cambridge University Press.
Binmore, K., 2001. Natural justice and political stability. Journal of Institutional and Theoretical Economics, pp.133-151.
Binmore, K., 2005. Natural justice. Oxford university press.
Binmore, K., 2022. Thinking about social norms. National Institute Economic Review, 259, pp.20-30.
Durkheim, E., 1895/2013. The rules of sociological method, Palgrave Macmillan.
Elster, J., 1994. Rationality, emotions, and social norms. Synthese, pp.21-49.
Fisman, R. and Miguel, E., 2007. Corruption, norms, and legal enforcement: Evidence from diplomatic parking tickets. Journal of Political economy, 115(6), pp.1020-1048.
Gilbert, P., 2003. Evolution, social roles, and the differences in shame and guilt. Social Research: An International Quarterly, 70(4), pp.1205-1230.
Horne, C., 2001. Sociological perspectives on the emergence of social norms in Hechter, M. and Opp, K.D. eds., Social norms.
Lewis, D., 1969/2008. Convention: A philosophical study. John Wiley & Sons.
O'Madagain, C. and Tomasello, M., 2022. Shared intentionality, reason-giving and the evolution of human culture. Philosophical Transactions of the Royal Society B, 377(1843), p.20200320.
Ostrom, E., 1990. Governing the commons: The evolution of institutions for collective action. Cambridge university press.
Paluck, E.L. and Shepherd, H., 2012. The salience of social referents: a field experiment on collective norms and harassment behavior in a school social network. Journal of personality and social psychology, 103(6), p.899.
Parsons, T., 1954. Essays in sociological theory. The Free Press.
Parsons, T. and Shils, E.A. eds., 1951. Toward a general theory of action. Harvard University Press.
Sznycer, D., Tooby, J., Cosmides, L., Porat, R., Shalvi, S. and Halperin, E., 2016. Shame closely tracks the threat of devaluation by others, even across cultures. Proceedings of the National Academy of Sciences, 113(10), pp.2625-2630.
Tankard, M.E. and Paluck, E.L., 2016. Norm perception as a vehicle for social change. Social Issues and Policy Review, 10(1), pp.181-211.
Young, H.P., 1998. Individual strategy and social structure: An evolutionary theory of institutions. Princeton University Press.
According to Parsons and Shils, when people struggle to match the expectations from these internalised social roles, they feel psychological strain.
The different individual actors participating in the social sytem will each have different roles, and they will accordingly differ in their specific goals and cognitive orientations. Role expectations… sharpen the edges of commitments and they impose further disciplines upon the individual. They can do so only as long as the conditions are present in the personality and the social system which enable human beings to live up to these kinds of expectations, which diminish or absorb the strains to which people are subjected, including both the "internal strains" connected with difficulty in fulfilling internalized norms and the strains which are associated with divergence from expectation. - Parsons and Shils (1951)
Jon Elster is here proposing a finer description of the psychological mechanisms behind these “internal strains”.
One view in sociology, defended by Talcott Parsons has been to consider society as an organism with social norms being the rules required for society to function and perpetuate itself.
Every social system is a functioning entity. That is, it is a system of interdependent structures and processes such that it tends to maintain a relative stability and distinctiveness of pattern and behavior as an entity by contrast with its - social or other - environment, and with it a relative independence from environmental forces. It "responds", to be sure, to the environmental stimuli, but is not completely assimilated to its environment, maintaining rather an element of distinctiveness in the face of variations in environmental conditions. To this extent it is analogous to an organism. - Parsons (1954), emphasis mine.
But more often, social scientists simply posit that social systems have some dynamics of their own (as emerging properties) with social norms and their evolution resulting from these.
David Lewis discussed the situation when official enforcement conflicts with conventional expectations in the driving game:
There is a complication: if we do not drive on the right, the highway patrol will catch us and we will be punished. So we have an independent incentive to drive on the right, and this second incentive is independent of how the others drive. But it makes no important difference. If I expected the others to be on the left, I would be there too, highway patrol or no highway patrol. My preference for driving on the same side as the others outweighs any incentive the highway patrol may give me to drive on the right. And so it is for almost everyone else, I am sure. The highway patrol modifies the payoffs in favor of driving on the right; but there are still two different coordination equilibria. The punishments are superfluous if they agree with our convention, are outweighed if they go against it, are not decisive either way, and hence do not make it any less conventional to drive on the right. - Lewis (1969)
Social norms are self-enforcing because they compel you to abide by them and punish deviants, or else you would be punishable yourself. It is the case if the norm is a perfect equilibrium: a combination of behaviours that everybody would still want to follow if someone deviates. Ken Binmore (2022) compares this to the relatively surprising effectiveness of tyrannical leaders who rely on the compliance of the oppressed. He illustrates it with the example of Alice in Wonderland:
Alice in Wonderland obeys the orders of the Queen of Hearts because she anticipates that otherwise the Queen will order Bob to cut off her head. Bob would obey this order because he anticipates that otherwise the Queen will order Carol to cut off his head. Carol would obey the order to punish Bob because she fears that someone else would cut off her head. Since this someone else might be Alice, we are looking at a spiral of self-confirming beliefs, which may seem too fragile to support anything solid. It is true that the beliefs go round in a circle, but the folk theorem shows that this fragility is an illusion, since the behaviour generated by the beliefs holds together as a perfect equilibrium.
The punishments required to sustain social norms can actually be relatively subtle:
Such talk of punishment evokes visions of torture chambers and the gibbet. Countless heretics have suffered in this way, but nearly all the punishments necessary to sustain an established social norm are administered without either the punishers or the victim even being aware that a punishment has taken place. No stick is commonly flourished. What happens most of the time is that the carrot is withdrawn a tiny bit. Shoulders are turned slightly away. Greetings are imperceptibly gruffer. Eyes wander elsewhere. These are all warnings that deviants ignore at their peril.
It does not imply that the psychological feelings that sustain our respect of social norms always adapt quickly to new settings. People experiencing a change in social norms around them may still be influenced by their attitudes that were shaped when navigating different social norms. Fisman and Miguel (2007) found for instance that the number of parking tickets received by diplomats working at the UN in New York is influenced by the degree of corruption in their home countries.
This is a nice summary. The view of norms as the temporary equilibria of a *many* agent dynamic system is clearly right. But there are many factors in the system, and whether nefarious powerful groups can create new norms or not I tend to agree they can't create new ones out of whole cloth - they can and have shaded them or prolonged the acceptance of poor equilibria in ways that are bad for the majority. So I'd argue that something like the vibe of the mental model trap you call out is pointing to something real.
I really enjoyed this summary on social norms! Francesco Guala's "Understanding Institutions" fits the literature cited here well. From his book I especially like the notion "correlation device". Staying with the traffic example: a traffic light is a correlation device because it sends a reliable signal to the players of the game about which strategy to pursue. The content of the correlation device doesn't really matter as long as the signal is clear to all players of the game. We don't stop at red and go at green because something about these colours inherently makes us behave this way. We could have cats for stop and dogs for go, or any other pairs, and traffic would still be regulated. I'd say that red and green are mutually salient cues, or focal points, grounded in cultural conventions (plus the colour red has a particular psychological effect). I find society level focal points fascinating: what emerge as focal points in the first place, how they stabilise at the society level, etc.
A side note: I'm not sure what exactly Tomasello means by "shared intentionality". By and large, people are subject to the same costs of not playing the social game right (reputation costs, internal feelings of shame and guilt). Then, one may assume that, in general, people have the same intentions, which is to avoid these costs by not violating norms. Shared intentionality is not the same as having a common understanding of norms. Having a common understanding of norms can regulate social interactions: if I know what the prevailing norms are in a given social context, I know how to behave. I also know that others know what the norms are, therefore I expect them to behave in a certain way, and I know that they too expect me to behave in a certain way.