21
The collective consciousness
Sociologists sometimes refer to the “collective consciousness” of a community, the set of values and beliefs shared (and known or believed to be shared) by its members. On the value side, the collective consciousness includes moral and social norms, religion, and political ideologies. On the belief side, it includes opinions about factual matters as well as about causal relations, ranging from rumors about the white slave trade to beliefs about the perverse effects of unemployment benefits. In this chapter I consider social norms and their operation. In the next chapter, I consider modes of collective or, better, interactive belief formation. There is a double asymmetry in my treatment of values and beliefs. On the one hand, I have little to say about the emergence of social norms, not because the question is uninteresting but because I find it too hard. On the other hand, I have little to say about the substance of popular or collective beliefs. Their content varies greatly in time and space, whereas the mechanisms of emergence, propagation, change, and collapse of beliefs are more invariant.
The operation of social norms
Consider two statements:
Always wear black clothes in strong sunshine.
Always wear black clothes at a funeral.
The first injunction is a matter of instrumental rationality, since the air between the body and the clothes circulates more rapidly when the garments are black. The second expresses a social norm, which has no obvious instrumental significance. The existence and importance of social norms cannot be doubted. The proximate causes involved in their operation are reasonably well understood. Yet their ultimate origin and function (if any) remain controversial.
A social norm is an injunction to act or to abstain from acting. Some norms are unconditional: “Do X; do not do Y.”1 They include the norms not to eat human flesh, not to have sexual intercourse with a sibling, not to break into the queue, never to wear red clothes (as some mothers tell their daughters), to wear black clothes at a funeral, to begin with the outermost knife and fork and work inward toward the plate, to treat the sickest patient first (even if the chances of curing him are worse than those for other patients). Other norms are conditional: “If you do X, then do Y,” or “If others do X, then do X.” In many groups, there is a norm that the person who first suggests that some action be taken is then charged with carrying it out;2 as a result, many good suggestions are never made. A childless couple may feel subject to a norm that whoever first suggests they have a child will have a larger share in raising it; as a result some couples who would like to have a child may remain childless.3 There may not be a norm telling me to send Christmas cards to my cousins, but once I begin there is a norm to continue and another norm telling my cousins to reciprocate. Yet although conditional, these norms are not conditional on any outcome to be realized by the action, as is the injunction to wear black in strong sunshine.
Revolutionary action by crowds can also be subject to social norms: it is legitimate to destroy, but not to steal; to kill, but not to rape. Tocqueville commented on the efficacy of such norms in the revolutions of 1789 and 1848. Concerning the first, he wrote that “if among the armed men someone committed a base action, he was immediately jailed by his comrades. This is a peculiar feature of our French people.” Concerning the 1848 Revolution, of which he had been an eyewitness, he wrote that the ban on theft did “not prevent a lot of robbery on such days, for … there are always rascals everywhere who jeer at the morality of the main body and are very contemptuous of its conception of honor when nobody is looking.” In other words, those motivated by moral norms can act as enforcers of a social norm with the same content.
More examples will be offered later. First, however, I need to say something about what lends causal efficacy to social norms and how they differ from other norms. A simple response to the first question is that social norms operate through informal sanctions directed at norm violators. Typically, sanctions affect the material situation of the offender, either by the mechanism of direct punishment or by the loss of opportunities caused by social ostracism. A farmer who violates community norms may see his barn burned down and his sheep disemboweled. Alternatively, he may find his neighbor denying his request for help with the harvest. The mechanism of gossip can act as a multiplier on these sanctions, by adding third-party sanctions to the original second-party punishment.
Consider what a cattle farmer might do when a neighbor's cattle repeatedly trespass on her land. She may seize the cattle, at a benefit to herself and at some cost to the neighbor. She may destroy the cattle or reduce their value (e.g. by castrating a bull), at no benefit to herself and some cost to the neighbor. She may herd the offending livestock to some distant place, at some cost to herself and to the neighbor. Or she might cut off all relations with the neighbor (ostracism). The last response could be inefficient, however, in that it might not deter future trespasses. The first response might be seen as an aggressive taking rather than as a punishment. The second and especially the third responses are more adequate, in that they clearly indicate an intention to punish, if need be at some cost to the punisher.
In general, however, I believe that ostracism or avoidance is the most important reaction to norm violations. If instead of repeated trespass the neighbor had engaged in a one-shot break of promise, cutting off relations would have been the more natural reaction. This claim is supported by the general idea that social norms operate through the emotions of shame in the norm violator and of contempt in the observer of the violation (Chapter 8). Because the action tendency of contempt is avoidance, which often causes material losses for the ostracized person, there is a link between the emotional response and the imposition of sanctions. Yet the sanctions are often more important as a vehicle for the communication of emotion than they are in their own right. Moreover, the cost of sanctioning to the sanctioner may be especially important in communicating the strength of his emotion.
The sanction theory of social norms runs into an obvious problem: what motivates the sanctioners to punish? What is in it for them? Typically, sanctioning is costly or risky for the sanctioner. Even if he does not give up an opportunity for mutually profitable interaction, the expression of disapproval might trigger an angry and even violent reaction in the target. There is an important distinction here between spontaneous disapproval and deliberate shaming. The latter can easily backfire, causing the target to be angry rather than ashamed. Even when the disapproval is in fact spontaneous, the target may, perhaps self-servingly, interpret it as intentional shaming and react accordingly. For this reason, sanctioning is a risky business. Why, then, do people engage in it?
One answer might be that non-punishers themselves risk punishment. This no doubt happens. In a society with strong norms of revenge one might expect that a person who fails to shun someone who fails to take revenge would herself be shunned. Among schoolchildren, a child might be more willing to interact with a “nerd” when not observed by classmates. Yet a child who abstains from joining the mob in harassing a child who is friendly to the nerd is unlikely himself to be harassed. Hence the third-party harassers are not likely to be motivated by the fear of punishment. Experimentally, the question might be examined by seeing whether third parties would punish Responders who, by accepting very low offers in the Ultimatum Game, fail to punish ungenerous Proposers. I would be surprised if they did, and even more surprised if fourth-party observers punished non-punishing third parties. At a few removes from the original violation, this mechanism ceases to be plausible.4
A more parsimonious and adequate explanation of sanctioning relies on the spontaneous triggering of contempt and the associated action tendency. Anger, too, may be involved, because of the fluid distinction between social and moral norms. Also, flaunting one's violation of social norms is likely to trigger anger rather than contempt because it tells other people that one does not care about their reactions. Although these spontaneous action tendencies may be kept in check by the costs and risks of sanctioning, they may be capable of overriding the latter. Ostracizing the nerd who could help his classmates with homework is costly, as was the refusal of aristocrats under the ancien régime to let their daughters marry wealthy commoners. When a “taste” for discrimination takes the form of refusing to employ or buy from members of despised minority groups or women, economic efficiency may suffer. Often, such behavior reflects the operation of social norms rather than of idiosyncratic individual preferences, as shown by phrases such as “Jew-lover” or “nigger-lover” used to condemn those who go against the norm.
Sanctioning requires information about the norm violation, either by direct observation or by hearsay, notably gossip. In large communities such information may be hard to come by. For this reason, it may be hard to enforce the social norm of voting in national elections. To overcome this difficulty, one might publish the names of non-voters. In a large-scale field experiment, a substantially higher turnout was observed among those who received mailings promising to publicize their turnout to their household or their neighbors. In this study, the citizens automatically received information about who voted and who did not. Alternatively, one might leave it to the citizens whether to seek out this information, by posting the names of non-voters on the internet. This practice already exists in Argentina, where it is combined with a fine for non-voting. The “naming, shaming and blaming” produced by publicity could also be a substitute for fines, and perhaps a more effective method.
What social norms are not
Social norms need to be distinguished from a number of related phenomena: moral norms, quasi-moral norms, legal norms, and conventions. Although the dividing lines may be fluid, there are clear-cut cases in each category. Both moral and quasi-moral norms (Chapter 5) are capable of shaping behavior even when the agent believes herself to be unobserved by others. By contrast, the shame that sustains social norms is triggered by the perceived contempt of others. The corresponding action tendency is to escape from their accusing stares: to hide, run, and even kill oneself.
Legal norms differ from social norms in that they are enforced by specialized agents who typically impose direct punishment rather than ostracism, experiments with legal “shaming” notwithstanding. Legal and social norms interact in numerous ways. In 1990, for instance, some state legislators in Louisiana pushed for reduction of criminal sanctions applicable to informal punishers of flag burners. Even after an edict of 1701 allowed the French nobility to engage in commerce (only wholesale, not retail), it was more than fifty years before they overcame the social norms prohibiting the practice. In some communities, there are social norms against appealing to legal norms, whereas in others people litigate at the drop of a hat.
Conventions, or convention equilibria, can in principle be enforced through the sheer self-interest of the agent, without any action by others. As noted in Chapter 18, they are often quite arbitrary. At the first day of a conference, each participant may find his or her seat more or less randomly. On the second day, a convention has been created: people converge to their chosen seats because doing so is the obvious (focal-point) allocative mechanism. On the third day, the convention has hardened into an entitlement: I get angry if another participant has taken “my” seat. Yet although the social norm cements the arbitrary convention and makes it more likely to be respected, it is not indispensable. Among New Yorkers, there is a convention to celebrate New Year's Eve in Times Square, but since few people would know whether a given person showed up or not, there is little opportunity for sanctioning. Even if the norm of driving on the right side of the road were not reinforced by social norms and legal norms, the dangers to the driver of switching into the left lane would be a strong deterrent.
A complex category is that of unwritten legal and political norms such as constitutional conventions.5 These are usually not legally enforceable, although courts may take account of them in decisions. Instead, they are enforced by political sanctions, or the fear of such sanctions. Until 1940, for instance, the American constitutional convention that nobody could serve as president more than twice was enforced by the belief that anyone who tried to do so would be defeated. This was why Ulysses Grant did not stand for a third term. Such norms, of which there are many, have some of the flavor of social norms, since they are enforced by the diffuse force of public opinion rather than by specialized agencies.6 Other political conventions are better seen as equilibria in repeated games. In many parliamentary systems there is, for instance, a convention that when an administration leaves office its internal documents are sealed and become available (to historians) only after several decades. Although any given administration might be tempted to open the archives of its predecessors and use them as political ammunition, the knowledge that this would set a precedent for its successor to do the same is sufficient to deter it from doing so. This is not a convention in the sense of Chapter 18, since each administration would prefer to deviate from it as long as others do not.
Norms and externalities
There are norms against those who impose small negative externalities on many others (Chapter 17). When people litter in the park, spit in the street, urinate in the lake, or drink from the office coffee pot without dropping a quarter into the cup, they usually try to do so unseen. Even when they do not actually fear sanctions, the mere thought that others might think badly of them may deter them from performing these actions when observed. Norms of this kind are socially useful in the strong sense that they make everybody better off. The norm against spitting in public places is an especially good example. Before one knew how contagious diseases were spread, spitting was a perfectly acceptable practice and widely catered to by spittoons. Once the mechanism of contagion was understood, “No spitting” signs appeared in many public places. Today, the norm is so entrenched (in some countries at least) that the signs have been taken down.
In this example, we can observe the norm emerging and claim with some confidence that it came about because it was in the public interest. The danger was perceived, a legal norm was created, and the social norm followed. Whether the perception of negative externalities can create social norms without the intermediary step of public intervention is more questionable. The mere fact that a norm is needed, and perceived to be needed, does not automatically bring it about. In developing countries, there is no social norm to limit family size. Social norms against overgrazing and overfishing have not emerged spontaneously to prevent the tragedy of the commons. There is no norm regulating the use of antibiotics, although their excessive use imposes externalities on others through the development of more resistant micro-organisms. Norms against playing music on the public beach and against using cell phones in the concert hall also owe (I conjecture) their origin to action by the relevant authorities. Over and over again, we find that outside intervention is necessary to stop people from imposing these negative externalities on each other. In some cases, as in the norm against spitting, people may refrain even when the legal norm disappears or ceases to be enforced. In others, such as China's “one-child” policy, it seems unlikely (but perhaps not impossible) that the behavior would persist if the regulation were to be lifted.
Smaller groups may be able to impose these norms without external intervention. In the workplace, there is often a strong norm against rate busters, because it is believed that their efforts might cause the management to lower the piece rate. (In this case the externality takes the form of an increase in the probability of a rate cut.) Although the management might want to commit itself to a policy of fixed piece rates, to induce workers to make a stronger effort, it may not be able to make a credible promise to this effect. Strikebreakers, too, are often heavily sanctioned by their fellow workers. It is perhaps significant that these two cases involve common opposition to an adversary. In a “game against nature” such as overgrazing, solidarity does not seem to emerge as easily, because free riding is not seen as betrayal. In firms that pay workers using individual piece rates, the norm against rate busting may emerge because it is seen as benefiting the “enemy.” It would be less likely to arise (except if fueled by envy) in a workers’ cooperative.
Other social norms target negative externalities that one group of people imposes on another. The norm against smoking, even in places where it is still legally allowed, is an example.7 In many Western societies today, guests who smoke often abstain without even asking the host whether they would be allowed to. What one might call “noise externalities” underlie the norm “Children should be seen but not heard.” There are two ways in which this injunction could be a social norm and not merely a form of parental punishment. First, children might ostracize other children who violate the norm. Second, parents might ostracize other parents whose children violate it. In train compartments, those who want to impose a “fresh air externality” on others usually lose the contest with those who impose a “stuffy air externality.” (On buses in Paris, this norm is posted as an obligation.) The reason may be that closed windows are perceived as the default option and hence as a normative baseline.
Norms and conformism
Some social norms are little but injunctions not to stick one's neck out. Inhabitants of small towns everywhere will recognize the “Law of Jante” written down (in 1933) by one who got away:
Thou shalt not believe thou art something.
Thou shalt not believe thou art as good as we.
Thou shalt not believe thou art more wise than we.
Thou shalt not fancy thyself better than we.
Thou shalt not believe thou knowest more than we.
Thou shalt not believe thou art greater than we.
Thou shalt not believe thou amountest to anything.
Thou shalt not laugh at us.
Thou shalt not believe that anyone is concerned with thee.
Thou shalt not believe thou canst teach us anything.
These norms can have very bad social consequences. They can discourage the gifted from using their talents and may lead to their being branded as witches if nevertheless they use them. Luck, too, is frowned upon. Among the Bemba of Northern Rhodesia, it is said that to find a beehive with honey in the woods is good luck; to find two beehives is very good luck; to find three is witchcraft.
Codes of honor
Strong and often subtle norms can regulate behavior in feuds, vendettas, duels, and revenge more generally. The norms define the actions that call for a retaliation or a challenge, the conditions under which and the means by which it can or must be carried out, and the fate of someone who fails to live up to the primary norm. Beginning with the last, the failure to take revenge often causes a kind of civic death, in which the agent is completely cut off from normal social relations. Within his family, his opinion counts for nothing; if he ventures outside his home, he is met with ridicule or worse. It is a paradigmatic situation of contempt, inducing intolerable shame.
Anything that can be seen, however remotely, as an insult to the agent's honor can trigger retaliation. In prerevolutionary Paris, the Vicomte de Ségur, a prominent rake about town, amused himself by writing small epigrams in verse. A rival who was jealous of his reputation wrote a little verse himself subtly mocking Ségur's verses. As revenge, Ségur seduced the rival's mistress and then, when she announced that she was pregnant, told her that he had just been using her to get back at his rival and that now that he had attained his aim he was no longer interested in her. (She subsequently died in childbirth.) He went back to Paris and told the story to anyone who would listen, never encountering disapproval. Les liaisons dangereuses, it seems, was but a feeble imitation of reality.
In nineteenth-century Corsica, there were four circumstances that justified or required vengeance: when a woman had been dishonored, when an engagement had been broken, when a close relative had been killed, and when false testimony in court led to the conviction of a member of one's family. In one case, a notary was convicted of homicide on false testimony and subsequently died in prison. His brother became a bandit and over a period of years killed all fourteen prosecution witnesses. These are all cases of vengeance for the sake of maintaining one's honor. The system of honor also included, however, actions undertaken for the purpose of gaining honor. Montaigne refers to “what is said by the Italians when they wish to reprove that rash bravery found in younger men by calling them bisognosi d'honore, ‘needy of honour.’”
In the American South people react more strongly to perceived insults than do northerners. Homicide rates are higher in the South, and people express stronger approval of violent reactions to affronts. In an ingenious study, a confederate of the experimenter bumped into the subject, “accidentally on purpose,” and called him an “asshole.” Afterward cortisol levels (reflecting reactions to the incident) and testosterone levels (reflecting preparation for future aggression) rose dramatically more in southern than in northern subjects. In another experiment, subjects continued walking down the hallway where they had been “bumped” and saw a large football player type (a confederate) walking toward them in a determined manner. The hallway had been cluttered with tables so that there was room only for one person to pass at a time, essentially creating a game of Chicken. Southerners went much closer to the other person (three feet) before they “chickened out” than did northerners (nine feet).
Do codes of honor serve any social function? If they do, can the function explain why they exist? The idea that the practice of revenge is a useful form of population control is too arbitrary to be taken seriously. An alternative view, that norms of revenge provide a functional equivalent of organized law enforcement in societies with a weak state, is also implausible. The Mediterranean and Middle Eastern societies that have subscribed to these norms have had levels of violence and mortality rates among young men far above what are found elsewhere.8 As suggested by the observation by Montaigne just quoted, norms of revenge and the larger code of honor in which they are embedded may light as many fires as they put out. Often, feuds create more disruption than they control.
Others have argued that norms of honor evolve in sparsely settled herding societies, in which a reputation for willingly using violence serves as a useful, even indispensable, deterrent to theft. The culture of honor in the American South has been explained in this perspective. Over and above the general problems of functional explanation, this analysis runs into the difficulty that codes of honor were equally strong in the court of the French kings in the seventeenth and eighteenth centuries, to name only one non-rural example. Some of those who focus on codes of honor in the urban aristocracy rather than among rural herders then come up with another functional explanation: in the absence of war the nobility “needed” duels to keep up their warlike spirit. If one does not provide a mechanism by which the need would generate its own satisfaction, this argument is worthless. These polemical comments do not imply that I have a better explanation to offer.
Norms of etiquette
A further set of social norms are those involved in rules of manners or etiquette. Codes of dress, language, table behavior, and the like are often relentless in their detail, condemning to ostracism those who miss the smallest nuance.9 In all societies there is a norm regulating the appropriate distance from other people to maintain on social occasions. If one moves inside the private space of a person (in the United States perhaps fifteen inches) one risks being shunned as uncouth. The norm is unusual, however, in that the individuals concerned are often unaware of its existence and operation. Most norms of etiquette are highly codified, often literally so. They are not only (for the most part) pointless but also sometimes even cruel in their consequences, as when a five-year-old girl goes home in tears because her friends ridiculed her new stroller for her baby doll on the grounds that it had no brakes. In prerevolutionary Paris, a young officer, wealthy but not noble, tried to gatecrash a ball at Versailles. “He was treated so severely that in his despair over the ridicule with which he was covered, at a time when ridicule was the worst of all evils, he killed himself when he came back to Paris.”
The puzzle is why these intrinsically trivial matters take on such importance. The disproportionate disapproval triggered by a breach of etiquette may be due to the unfounded belief that people are all of a piece (Chapter 12), so that someone who violates an unimportant norm is likely also to violate more consequential ones. Also, the violation of trivial norms of etiquette may be seen as a non-trivial show of disregard for what other people think. This leaves unexplained, however, why the unimportant norms exist at all.
Functional explanations are very common. The subtle rules of etiquette among the elite exist, allegedly, in order to make it more difficult for outsiders to “crash the party” by imitating the rule-governed behavior. There is no doubt that these rules often have the effect of keeping upstarts down, but that does not offer an explanation of why they exist. As many self-proletarianized students have discovered, it is very difficult to break into the working class for someone who was not born in it. In Norway in the 1970s, for instance, young Maoists found that making fun of the royal family was a sure way of alienating themselves from the class they were trying to join. Yet nobody has suggested that the norms of the working class exist in order to make it more difficult for outsiders to pass themselves off as workers. The argument makes no more sense for the norms of the elite.
Norms of etiquette can impose heavy costs, while benefiting nobody. In eighteenth-century Massachusetts, social norms required widows and widowers to provide mourners with rings, gloves, and scarves, all of which had to come from England. In 1741, the Massachusetts House of Representatives tried to put an end to these wasteful practices, by forbidding the distribution of scarves and rings and limiting the number of those who could receive gloves to six persons, in addition to the minister and six pallbearers. The legislation was largely ignored. When the conflict with England caused economic depression, it was welcomed by some individuals because, in the words of a contemporary, “each individual being ransomed from the tyranny of fashion, will be free to act as his circumstances may require, and such freedom can scarce be purchas'd too dear.” In contemporary China, too, many poor families are ruined by norms of highly public gift-giving at funerals and weddings. Even if they get something back when they are hosting an event, ceremonies on average cost more than twice the income from gifts received. There is evidence that the resulting “income squeeze” leads to in utero malnutrition and subsequently stunted growth.
When expensive private schools impose the use of uniforms on the students, it is probably also to reduce the cost of following social norms of dress – and to remove the visible contrast between those who can afford it and those who cannot. A more interested motive lay behind the sumptuary laws that have been enacted at many times and places, to prevent commoners from imitating the manners of the aristocracy. Here, too, however, the sheer expense mattered. In the Italian town of Lucca, the social norm requiring expensive dowries threatened to reduce the population, because fathers were unable to equip their daughters to make them eligible. The municipality enacted a limitation on the number of expensive dresses per person, but it was not enforceable. These examples serve to undermine a rosy view of social norms as providing a solution to problems that institutions cannot address. The opposite can be true: the norms can be harmful, and institutions incapable of restraining them.
Norms of drinking
If social norms were invariably geared to enhance the welfare of the individual or of society we might expect them to be directed against heavy drinking that is perceived to have harmful short-term or long-term consequences. There are indeed many norms of this kind. Some norms, usually linked to religion, demand total abstinence. Islam and some Protestant sects have absolute bans on alcohol. Secular norms, by contrast, often enjoin drinking in moderation. The Italian norm “Never drink between meals” has the dual effect of limiting total consumption and of reducing the rate of absorption of alcohol, thus buffering the short-term effect on the body. In Iceland, there are norms against drinking in the presence of the children and against drinking on fishing trips.
Alcohol-related norms do not, however, always enhance welfare. There are norms that condemn abstinence, as well as norms that enjoin people to drink heavily. Among the Mapuche Indians of Chile, drinking alone is criticized, and so is abstinence; such behavior is seen as showing lack of trust. The traditional French culture condemns both the teetotaler and the drunkard. In Italy, distrust of abstainers is expressed in a proverb, “May God protect me from those who do not drink.” In youth subcultures of many countries, abstainers are subject to heavy pressure and ridicule. Conversely, there are many societies in which heavy drinking is socially prescribed. In Mexico and Nigeria, the macho qualities shown in the ability to drink heavily are much admired. In pre-revolutionary Russia, excessive drinking was obligatory in the subculture of young officers.
When abstinence is condemned or when heavy drinking is socially mandatory, would-be abstainers may have to resort to subterfuge. In Sweden, a common question is “Do you want sherry, or are you driving?” It is so accepted that abstaining alcoholics often say they are driving because this relieves them of the social pressure that otherwise would certainly be exerted by the host to convince the guest to have a drink. The norm of drinking can only be offset by another norm (against drunk driving). Similarly, it has been argued that conversions to Protestantism provide an alternative for some Latin Americans who want to opt out of the system of community governance in which even the rituals often involve heavy drinking and drunkenness. Again, the norm of drinking can only be overridden by another norm, which in this case has the backing of religion.
These are cases of the strategic use of norms. Conversely, people can behave strategically to get around the norms. Some ancient Chinese considered alcohol itself to be sacred and drank it only in sacrificial ceremonies; eventually, they would sacrifice whenever they wanted to drink. In Spain, at certain hours, not to drink on an empty stomach is a tacit cultural proscription, so food will be included with the drinking. In both cases, we observe a reverse of the original causal link: rather than obeying the conditional norm of drinking only when they are doing X, people do X whenever they want to have a drink.
Norms of queuing
The queue is a norm-ridden social system. At the same time, the queue is a transient phenomenon, unlike the other contexts I have discussed. Before I proceed to discuss norms of queuing, let me pursue this contrast for a moment.
It is probably a common intuition that norms have less impact on behavior in communities with high turnover. According to Tocqueville, “Men who live in democracies are too mobile to allow some group of them to establish and enforce a code of etiquette. Each individual therefore behaves more or less as he pleases, and manners are always to some extent incoherent because they are shaped by each individual's feelings and ideas rather than conforming to an ideal model held up in advance for everyone to imitate.” Earlier, I referred to the small-town norm of “Don't stick your neck out,” with the implication that in more anonymous interactions deviant behavior would be less severely sanctioned. In light of this intuition, it is interesting that as communities grow larger and more mobile, we observe the emergence of norms regulating the behavior among strangers. This remark strengthens the interpretation of norms in terms of emotion rather than material sanctions. Under most circumstances, it is difficult to impose tangible sanctions on a person who violates a queue norm. There is not even much space for avoidance. People can give full rein, however, to expressions of contempt or indignation.
There is a norm, I believe, against walking up to the person at the head of a bus queue and offering him or her money in exchange for the place. This norm is obviously inefficient: if the person who is asked accepts and moves to the back of the line in exchange for the money, both persons benefit and nobody is hurt. According to Tocqueville, such norms against open display of wealth in public are specific to democratic societies: “Do you see this opulent citizen? … His dress is simple, his demeanor modest. Within the four walls of his home, luxury is adored.” There may also be an underlying idea that the use of queuing is a valuable counterweight to the pervasive use of money in allocating scarce goods. To prevent the rich from getting everything, let some goods be allocated by a mechanism that puts them at a disadvantage, because of their greater opportunity costs of queuing. In Communist Poland, where queuing was endemic, there was no norm against purchasing a place in a queue, probably because this practice was seen as one of many necessary forms of jockeying for position. Other forms included hiring people to stand in line or moving back and forth between several queues while asking people in each of them to “hold the place.” There were norms regulating these activities, and deviations were sanctioned. A surprising norm was the rule against reading while queuing. According to an observer, “women do not want to be told, even by implication, that they are actually wasting time in queues. If one reads or works in the queue, this implicitly reminds others that they are wasting time. The response is to scold the deviant, putting the reminder out of sight and mind.” In addition, people reading or working would shirk their duty of monitoring violations of queue norms.
A different kind of violation occurs when someone intrudes in a queue, whether at the head of the line or somewhere in the middle. In this case, the negative reactions of people behind the intruder in the queue might be due to considerations of cost, be it in the form of time costs or (if they are queuing for a scarce good) material costs. Alternatively, they might be due to outrage or indignation. Experiments find that although both factors may be at work, subjects usually have a stronger reaction to illegitimate intrusions than to legitimate ones that impose equal costs. (A telling fact in these studies is that the confederates of the experimenter who were asked to intrude in the queue felt the task to be highly aversive.) There is often a norm to the effect that responsibility for rejecting intruders lies with the person immediately behind him or her. There are also norms regulating place holding in queues. In my supermarket, the norm seems to be that it is acceptable to leave the shopping cart in the line to go to pick up one item from the shelves, but not to go back several times. In Australian football queues, the norm in leaving position markers is that one must not be absent for periods longer than two to three hours. Although it might seem more efficient if most people placed a marker in the queue and then went home for a while, this practice would violate equality since the people who remained in the queue to maintain it would be disfavored.
The basic principle of fairness in queuing, “First in, first out,” can be violated when there are multiple and independent queues. Thus reported customer satisfaction is higher in the single-queue Wendy's restaurants than in the multi-queue Burger King and McDonald's restaurants, although the latter average half the waiting time of Wendy's. At Houston airport, customers with checked luggage complained about the baggage delay (a one-minute walk to the carousel and a seven-minute wait at the carousel), compared to passengers with hand luggage who could proceed directly to the taxi stand. When the airport authorities changed the disembarking location so that all customers had to walk for six minutes, complaints dropped to nearly zero.
The norm of tipping
Tipping for service is not a negligible phenomenon. Estimates of tips in US restaurants range from $5 billion to $27 billion a year; adding tips to taxi drivers, hairdressers, and others would yield a larger figure. Estimates of the fraction of income that waiters derive from tips range from 8 percent (the Internal Revenue Service assumption) to 58 percent for waiters serving full-course meals. In some contexts tipping may seem puzzling, in others less so. If you go to the same hairdresser each time you need a haircut, you tip to ensure good service; the same applies for meals in your favorite restaurant. Tipping in one-shot encounters, such as a taxi ride or a meal in a restaurant you do not expect to visit again, is more puzzling. Such behavior is in fact doubly puzzling: it cannot be sustained by two-party interaction over time, nor by third-party sanctions at the time of the encounter. If you are the only passenger in the taxi, other people are rarely in a position to know whether you tip the taxi driver adequately; nor are other customers in the restaurant likely to notice how much you tip your waiter.
Tipping, it has been argued, is an efficient way of remunerating waiters. It is obviously easier for the client to monitor the quality of service than it is for the restaurant owner. Hence decentralizing the monitoring function and linking reward to observed performance are a way of overcoming the “principal-agent problem” (how to prevent workers from shirking) that besets many contractual relationships (Chapter 25). Tipping, therefore, might be part of an “implicit contract” for the purpose of enhancing efficiency. But as Sam Goldwyn said, an unwritten contract isn't worth the paper it's written on. The argument, like many other attempts to explain social norms, is merely a piece of unsupported functionalism. The idea that restaurant owners who forbid tipping are eliminated in the competition with those who allow it is entirely conjectural, and in any case would not explain why customers do tip in the latter. Also, when assessed empirically, tipping does not seem to pass the appropriate efficiency tests. It does not, for instance, appear to be more prevalent in occupations where monitoring is easier. The fact that waiters often pool their tips also undermines the efficiency argument. Indirectly, however, the pooling of tips could enhance efficiency, if waiters ostracize colleagues who give such bad service that they bring little back to the pool.
I do not know why there is a norm to tip in certain occupations and not in others, or why the same service receives a tip in some countries and not in others. Once a norm exists, however, we can understand why people tip: they simply do not like the idea that others, such as a disappointed taxi driver, might disapprove of them, even if they do not expect to meet them again. Being the object of the contemptuous stare of the other is not necessary. It may be enough simply to know or have reason to believe that the other feels contempt. To take another example, the belief that others might disapprove explains why I abstain from picking my nose on the subway platform when a train is passing by without stopping, even if there are no other people on either platform.
Why norms?
The importance of social norms for the regulation of behavior and the proximate mechanism by which they operate are, as I said, fairly well understood. I do not believe, however, that we have a good understanding of their origin. There are two separate questions. First, what is the evolutionary origin of the correlative emotions of shame and contempt that sustain social norms? In other words, why are there social norms at all? Second, why do specific norms exist in specific societies? How and when do they arise; how and when do they disappear?
A simple answer to the first question is that we care intensely about what other people think about us. We seek their approval and fear their disapproval. This answer, however, only raises the same question at one remove: why should we care about what other people think about us? In some cases, to be sure, a reputation can be useful and worth cultivating. Yet the thought that the taxi driver might think badly about us if we do not leave a tip is entirely divorced from reputational concerns. Also, since the reason others think badly about us is that we have violated a social norm, explaining norms by the desire that others not think badly about us is to some extent circular.
Concerning the second question, the most common answer is that norms emerge to regulate externalities. There is something to this idea if we add, as I argued we should, that social norms against imposing negative externalities on others are usually ushered in by an outside authority. There is a general social norm to obey the law. If fines were seen as prices, and prison as no more stigmatizing than a stay in a hospital, there would be no such norm, but in general these reactions to lawbreaking are not seen as equivalent to other, objectively equal burdens. People feel ashamed of going to jail and try to hide the fact if they can.10 When the law bans behavior that imposes negative externalities on others, the social norm of obeying the law may spill over into a norm against that behavior. The norm may persist even if the law that gave rise to it falls into disuse. This outcome may be hard to distinguish, however, from the emergence of the “good equilibrium” in an Assurance Game (Chapter 18). If the state induces cooperation by punishing defectors and then dismantles the punishment apparatus, people may continue to cooperate because each person's top-ranked situation is the one in which she and everybody else cooperate (there is no free-rider temptation).
With regard to many of the other norms I have discussed, such as the norm against offering money to buy someone's place in the bus queue, norms of etiquette, or norms of tipping, it is harder to come up with an explanation of their emergence and persistence. One line of argument, often offered by economists, is that the persistence of norms can be explained as equilibrium behavior and that their emergence is a matter of accident and history about which social science has little to say. Since an implicit premise of this book is that the dividing line between social science and history is artificial and pointless, I cannot agree with the latter claim. As to the former, I have argued that social norms typically do not exhibit the best-response logic that characterizes strategic games. When, unobserved, I observe another violating a norm, sanctioning the violator is typically not a best response.
Bibliographical note
This chapter builds on, and (I hope) improves on, the account of norms I proposed in The Cement of Society (Cambridge University Press, 1989) and, more succinctly, in “Social norms and economic theory,” Journal of Economic Perspectives 3 (1989), 99–117. Influential discussions of social norms are J. Coleman, Foundations of Social Theory (Cambridge, MA: Harvard University Press, 1990), R. Ellickson, Order Without Law (Cambridge MA: Harvard University Press, 199 ), and E. Posner, Law and Social Norms (Cambridge, MA: Harvard University Press, 2000). I learned from all of them but was not persuaded by any. For an instructive criticism of Posner, see the review by R. McAdams, Yale Law Journal 110 (2001), 625–90. Useful discussions of unwritten constitutional norms or conventions are found in two articles by J. Jaconelli, “The nature of constitutional convention,” Legal Studies 24 (1999), 24–46, and “Do constitutional conventions bind?” Cambridge Law Journal 64 (2005), 149–76. The two American examples are taken from H. Horwill, The Usages of the American Constitution (Oxford University Press, 1925). The Law of Jante is taken from A. Sandemose, A Fugitive Crosses his Trail (New York: Knopf, 1936). The role of witchcraft in sustaining norms against sticking one's neck out is discussed in K. Thomas, Religion and the Decline of Magic (Harmondsworth: Penguin, 1973). I discuss codes of honor and revenge in Chapter 3 of Alchemies of the Mind (Cambridge University Press, 1999). The story about the Vicomte de Ségur is taken from Les mémoires de la Comtesse de Boigne (Paris: Mercure de France, 1999), vol. I, pp. 73–4. These memoirs (ibid., p. 38) are also the source of the story about the young officer who killed himself out of shame for being ridiculed. Norms of etiquette are the topic of P. Bourdieu, Distinction (Cambridge MA: Harvard University Press, 1987), with a distinctly functionalist slant. The experimental studies on “the culture of honor” are reported in R. Nisbett and D. Cohen, The Culture of Honor (Boulder, CO: Westview Press, 1996). The examples of norms of drinking are taken from my Strong Feelings (Cambridge, MA: MIT Press, 1999). The misadventures of self-proletarianized students in Norway are charted in a wonderfully amusing novel, unfortunately not translated into English, by D. Solstad, Gymnaslaerer Pedersens beretning om den store politiske vekkelsen som har hjemsøkt vårt land (Oslo: Gyldendal, 1982). A movie version with English subtitles is available. The story about funeral norms in Massachusetts is told in T. H. Breen, The Marketplace of Revolution (Oxford University Press, 2005), and the story about dowries in Lucca in H. Freudenberger, “Fashion, sumptuary laws, and business,” Business History 37 (1963), 433–49. The effects of costly funerals and weddings in China are documented in X. Chen and X. Chang, “Costly posturing: relative status, ceremonies, and early child development in China,” working paper (Yale School of Public Health, 2012). For norms of queuing in Communist Poland, see J. Hrada, “Consumer shortages in Poland,” Sociological Quarterly 26 (1985), 387–404. For some of the other examples, see L. Mann, “Queue culture: the waiting line as a social system’, American Journal of Sociology 75 (1969), 340–54. The efficiency-based explanation of the norm of tipping is offered by N. Jacob and A. Page, “Production, information costs and economic organization: the buyer monitoring case,” American Economic Review 70 (1980), 476–8. It is criticized in M. Conlin, M. Lynn, and T. O'Donoghue, “The norm of restaurant tipping,” Journal of Economic Behavior & Organization 5 (2003), 297–321, which proposes an account closer to the one sketched here.
1 In the following, “conditional” and “unconditional” refer to the content of the social norm. As noted in Chapter 5, all social norms are conditional in the sense that their operation is contingent on the presence of an observer.
2 This norm may be linked to focal-point reasoning.
3 Hence the norm induces a game of Chicken.
4 In the Old South, things were organized more rigorously. “Peer pressure demanded that every young man assist the [slave] patrols and take turns in whipping the suspects.”
5 The phrase “constitutional convention” is used about two entirely different aspects of constitutions: the unwritten norms that supplement the written constitution and the constituent assemblies that are often used to adopt a written constitution.
6 Two American examples illustrate the force of these norms. When T. Roosevelt stood for a third term after a split in the Republican Party, which had failed to nominate him, feelings ran high. While on one of his speech-making tours, he was shot at by a man of unbalanced mind, who said: “I shot Theodore Roosevelt because he was a menace to the country. He should not have a third term. I shot him as a warning that men must not try to have more than two terms as President.” Regarding the norm (which remains unwritten in about half of the American states) that all members of the state electoral college must vote for the candidate who received the greatest number of votes, one observer predicted that “an Elector who failed to vote for the nominee of his party would be the object of execration, and in times of very high excitement might be the subject of a lynching.”
7 The most important externality is caused by smoke inhalation (passive smoking). It is sometimes also claimed that smokers impose a negative externality on other smokers who want to quit but cannot resist the desire to smoke triggered by the visual cue of others smoking.
8 One might object that the relevant comparison is with the level of violence that would obtain in the “state of nature.” If that state is defined by an exclusive concern with self-interest and the absence of any state-like agencies, it would not produce violence motivated by envy, spite, or anger. The anticipation of forceful appropriation by the strong of the goods produced by the weak might have a chilling effect on production, thus preventing actual violence from occurring.
9 In aristocratic societies, gross deviations are sometimes accepted, when seen as deliberate rather than as evidence of rule ignorance. Proust's Charlus is an example.
10 In Norway, there used to be a mandatory three-weeks’ prison sentence for drunk driving. Some people allegedly took a sunlamp with them into the prison cell, to acquire a tan they could use to buttress their story of having taken a holiday.