Social norms, the informal rules that govern behavior in groups andsocieties, have been extensively studied in the social sciences.Anthropologists have described how social norms function in differentcultures (Geertz 1973), sociologists have focused on their socialfunctions and how they motivate people to act (Durkheim 1895 [1982],1950 [1957]; Parsons 1937; Parsons & Shils 1951; James Coleman1990; Hechter & Opp 2001), and economists have explored howadherence to norms influences market behavior (Akerlof 1976; Young1998a). More recently, also legal scholars have touted social norms asefficient alternatives to legal rules, as they may internalizenegative externalities and provide signaling mechanisms at little orno cost (Ellickson 1991; Posner 2000).

With a few exceptions, the social science literature conceives ofnorms as exogenous variables. Since norms are mainly seen asconstraining behavior, some of the key differences between moral,social, and legal norms—as well as differences between norms andconventions—have been blurred. Much attention has instead beenpaid to the conditions under which norms will be obeyed. Because ofthat, the issue ofsanctions has been paramount in the socialscience literature. Moreover, since social norms are seen as centralto the production of social order or social coordination, research onnorms has been focused on thefunctions they perform. Yeteven if a norm may fulfill important social functions (such as welfaremaximization or the elimination of externalities), it cannot beexplained solely on the basis of the functions it performs. Thesimplistic functionalist perspective has been rejected on severalaccounts; in fact, even though a given norm can be conceived as ameans to achieve some goal, this is usually not the reason why itemerged in the first place (Elster 1989a, 1989b). Moreover, although aparticular norm may persist (as opposed to emerge) because of somepositive social function it fulfills, there are many others that areinefficient and even widely unpopular.

Philosophers have taken a different approach to norms. In theliterature on norms and conventions, both social constructs are seenas the endogenous product of individuals’ interactions (Lewis1969; Ullmann-Margalit 1977; Vandershraaf 1995; Bicchieri 2006). Normsare represented as equilibria of games of strategy, and as such theyare supported by a cluster of self-fulfilling expectations. Beliefs,expectations, group knowledge and common knowledge have thus becomecentral concepts in the development of a philosophical view of socialnorms. Paying attention to the role played by expectations insupporting social norms has helped differentiate between social norms,conventions, and descriptive norms: an important distinction oftenoverlooked in the social science accounts, but crucial when we need todiagnose the nature of a pattern of behavior in order to intervene onit.

• 1. General Issues
• 2. Early Theories: Socialization
• 3. Early Theories: Social Identity
• 4. Early Theories: Cost-Benefit Models
• 5. Game-Theoretic Accounts
• 6. Experimental Evidence
• 7. Evolutionary Models
• 8. Conclusion
• Bibliography
• Academic Tools
• Other Internet Resources
• Related Entries

1. General Issues

Social norms, like many other social phenomena, are the unplannedresult of individuals’ interaction. It has been argued thatsocial norms ought to be understood as a kind of grammar of socialinteractions. Like a grammar, a system of norms specifies what isacceptable and what is not in a society or group. And, analogously toa grammar, it is not the product of human design. This view suggeststhat a study of the conditions under which norms come intobeing—as opposed to one stressing the functions fulfilled bysocial norms—is important to understand the differences betweensocial norms and other types of injunction (such as hypotheticalimperatives, moral codes, or legal rules).

Another important issue often blurred in the literature on norms isthe relationship between normative beliefs and behavior. Some authorsidentify norms with observable, recurrent patterns of behavior. Othersonly focus on normative beliefs and expectations. Such accounts findit difficult to explain the complexity and heterogeneity ofnorm-driven behaviors, as they offer an explanation of conformity thatis at best partial.

Some popular accounts of why social norms exist are the following.Norms are efficient means to achieve social welfare (Arrow 1971;Akerlof 1976), prevent market failures (Jules Coleman 1989), or cutsocial costs (Thibaut & Kelley 1959; Homans 1961); norms areeither Nash equilibria of coordination games or cooperative equilibriaof prisoner’s dilemma-type games (Lewis 1969; Ullmann-Margalit1977), and as such they solve collective action problems.

Akerlof’s (1976) analysis of the norms that regulate landsystems is a good example of the tenet that “norms are efficientmeans to achieve social welfare”. Since the worker is muchpoorer and less liquid than the landlord, it would be more natural forthe landlord rather than the tenant to bear the risk of crop failure.This would be the case if the landlord kept all the crops, and paidthe worker a wage (i.e., the case of a “wage system”).Since the wage would not directly depend on the worker’s effort,this system leaves no incentive to the worker for any effort beyondthe minimum necessary. In sharecropping, on the contrary, the workeris paid both for the effort and the time he puts in: a more efficientarrangement in that it increases production.

Thibaut and Kelley’s (1959) view of norms as substitutes forinformal influence has a similar functionalist flavor. As an example,they consider a repeated battle of the sexes game. In this game, somebargaining is necessary for each party to obtain, at leastoccasionally, the preferred outcome. The parties can engage in acostly sequence of threats and promises, but it seems better to agreebeforehand on a rule of behavior, such as alternating between therespectively preferred outcomes. Rules emerge because they reduce thecosts involved in face-to-face personal influence.

Likewise, Ullman-Margalit (1977) uses game theory to show that normssolve collective action problems, such as prisoner’sdilemma-type situations; in her own words, “… a normsolving the problem inherent in a situation of this type is generatedby it” (1977: 22). In a collective action problem, self-centeredrational choices produce a Pareto-inefficient outcome.Pareto-efficiency is restored by means of norms backed by sanctions.James Coleman (1990), too, believes that norms emerge in situations inwhich there are externalities, that is, in all those cases in which anactivity produces negative (positive) effects on other parties,without this being reflected in direct compensation; thus the producerof the externality pays no cost for (reaps no benefit from) theunintended effect of their activity. A norm solves the problem byregulating the externality-producing activity, introducing a system ofsanctions (rewards).

Also Brennan, Eriksson, Goodin, and Southwood (2013) argue that normshave a function. Norms function to hold us accountable to each otherfor adherence to the principles that they cover. This may or may notcreate effective coordination over any given principle, but they placeus in positions where we may praise and blame people for theirbehaviors and attitudes. This function of accountability, they argue,can help create another role for norms, which is imbuing practiceswith social meaning. This social meaning arises from the expectationsthat we can place on each other for compliance, and the fact thatthose behaviors can come to represent shared values, and even a senseof shared identity. This functional role of norms separates it frombare social practices or even common sets of desires, as thosenon-normative behaviors don’t carry with them the socialaccountability that is inherent in norms. The distinctive feature ofthe Brennan et al. account of norms is the centrality ofaccountability: this feature is what distinguishes norms from othersocial practices.

All of the above are examples of afunctionalist explanationof norms. Functionalist accounts are sometimes criticized for offeringa post hoc justification for the existence of norms (i.e., the merepresence of a norm does not justify inferring that that norm exists toaccomplish some social function). Indeed, a purely functionalist viewmay not account for the fact that many social norms are harmful orinefficient (e.g., discriminatory norms against women and minorities),or are so rigid as to prevent the fine-tuning that would be necessaryto accommodate new cases. There, one would expect increasing socialpressure to abandon such norms.

According to some authors, we can explain the emergence of normswithout any reference to the functions they eventually come toperform. Since the norms that are most interesting to study are thosethat emerge naturally from individuals’ interactions (Schelling1978), an important theoretical task is to analyze the conditionsunder which such norms come into being. Because norms often provide asolution to the problem of maintaining social order—and socialorder requires cooperation—many studies on theemergence anddynamics of norms have focused on cooperation. Norms of honesty,loyalty, reciprocity and promise-keeping are indeed important to thesmooth functioning of social groups. One hypothesis is that suchcooperative norms emerge in close-knit groups where people haveongoing interactions with each other (Hardin 1982). Evolutionary gametheory provides a useful framework for investigating this hypothesis,since repeated games serve as a simple approximation of life in aclose-knit group (Axelrod 1984, 1986; Skyrms 1996; Gintis 2000). Inrepeated encounters people have an opportunity to learn from eachother’s behavior, and to secure a pattern of reciprocity thatminimizes the likelihood of misperception. In this regard, it has beenargued that the cooperative norms likely to develop in close-knitgroups are simple ones (Alexander 2000, 2005, 2007); in fact, delayedand disproportionate punishment, as well as belated rewards, are oftendifficult to understand and hence ineffective. Although normsoriginate in small, close-knit groups, they often spread well beyondthe narrow boundaries of the original group. The challenge thusbecomes one of explaining the dynamics of the norm propagation fromsmall groups to large populations.

If norms can thrive and spread, they can also die out. A poorlyunderstood phenomenon is the sudden and unexpectedchange ofwell-established patterns of behavior. For example, smoking inpublic without asking for permission has become unacceptable, and onlya few years ago nobody would have worried about using gender-ladenlanguage. One would expect inefficient norms (such as discriminatorynorms against women and minorities) to disappear more rapidly and withgreater frequency than more efficient norms. However, Bicchieri (2016)points out that inefficiency is not a sufficient condition for anorm’s demise. This can be seen by the study of crime andcorruption: corruption results in huge social costs, but suchcosts—even when they take a society to the brink ofcollapse—are not enough to generate an overhaul of thesystem.

An influential view of norms considers them as clusters ofself-fulfilling expectations (Schelling 1960), in that someexpectations often result in behavior that reinforces them. A relatedview emphasizes the importance ofconditional preferences insupporting social norms (Sugden 2000). In particular, according toBicchieri’s (2006) account, preferences for conformity to socialnorms are conditional on “empirical expectations” (i.e.,first-order beliefs that a certain behavior will be followed) as wellas “normative expectations” (i.e., second-order beliefsthat a certain behavior ought to be followed). Thus, norm complianceresults from the joint presence of a conditional preference forconformity and the belief that other people will conform as well asapprove of conformity.

Note that characterizing norms simply as clusters of expectationsmight be misleading; similarly, a norm cannot simply be identifiedwith a recurrent behavioral pattern either. If we were to adopt apurely behavioral account of norms there would be no way todistinguish shared rules of fairness from, say, the collective morninghabit of tooth brushing. After all, such a practice does not depend onwhether one expects others to do the same; however, one would not eventry to ask for a salary proportionate to one’s education, if oneexpected compensation to merely follow a seniority rule. In fact,there are behavioral patterns that can only be explained by theexistence of norms, even if the behavior prescribed by the norm inquestion is currently unobserved. For example, in a study of the Ikpeople, Turnbull (1972) reported that starved hunters-gatherers triedhard to avoid situations where their compliance with norms ofreciprocity was expected. Thus they would go out of their way not tobe in the position of gift-taker, and hunted alone so that they wouldnot be forced to share their prey with anyone else. Much of theIk’s behavior could be explained as a way of eluding existingreciprocity norms.

There are many other instances ofdiscrepancies betweenexpectations and behavior. For example, it is remarkable toobserve how often people expect others to act selfishly, even whenthey are prepared to act altruistically themselves (Miller &Ratner 1996). Studies have shown that people’s willingness togive blood is not altered by monetary incentives, but typically thosevery people who are willing to donate blood for free expect others todonate blood only in the presence of monetary rewards. Similarly, allthe interviewed landlords answered positively to a question aboutwhether they would rent an apartment to an unmarried couple; however,they estimated that only 50% of other landlords would accept unmarriedcouples as tenants (Dawes 1972). Such cases of pluralistic ignoranceare rather common; what is puzzling is that people may expect a givennorm to be upheld in the face of personal evidence to the contrary(Bicchieri & Fukui 1999). Furthermore, there is evidencesuggesting that people who donate blood, tip on a foreign trip, givemoney to beggars or return a lost wallet often attempt to underplaytheir altruistic behavior (by supplying selfish motives that seeminglyalign their actions with a norm of self-interest; Wuthnow 1991).

In a nutshell, norms refer to actions over which people have control,and are supported by shared expectations about what should or shouldnot be done in different types of social situations. However, normscannot be identified just with observable behavior, nor can theymerely be equated with normative beliefs.

The varying degrees of correlation between normative beliefs andactions are an important factor researchers can use to differentiateamong various types of norms. Such a correlation is also a key elementto consider when critically assessing competing theories of norms: webegin by surveying the socialized actor theory, the social identitytheory, and some early rational choice (cost-benefit) models ofconformity.

2. Early Theories: Socialization

In the theory of the socialized actor (Parsons 1951), individualaction is intended as a choice among alternatives. Human action isunderstood within a utilitarian framework as instrumentally orientedand utility maximizing. Although a utilitarian setting does notnecessarily imply a view of human motives as essentially egoistic,this is the preferred interpretation of utilitarianism adopted byTalcott Parsons and much contemporary sociology. In this context, itbecomes crucial to explain through which mechanisms social order andstability are attained in a society that would otherwise be in apermanent Hobbesian state of nature. In short, order and stability areessentially socially derived phenomena, brought about by acommonvalue system—the “cement” of society. Thecommon values of a society are embodied in norms that, when conformedto, guarantee the orderly functioning and reproduction of the socialsystem. In the Parsonian framework norms are exogenous: how such acommon value system is created and how it may change are issues leftunexplored. The most important question is rather how norms get to befollowed, and what prompts rational egoists to abide by them. Theanswer given by the theory of the socialized actor is that peoplevoluntarily adhere to the shared value system, because it isintrojected to form a constitutive element of the personality itself(Parsons 1951).

In Parsons’ own words, a norm is

a verbal description of a concrete course of action, … ,regarded as desirable, combined with an injunction to make certainfuture actions conform to this course. (1937: 75)

Norms play a crucial role in individual choice since—by shapingindividual needs and preferences—they serve as criteria forselecting among alternatives. Such criteria are shared by a givencommunity and embody a common value system. People may choose whatthey prefer, but what they prefer in turn conforms to socialexpectations: norms influence behavior because, through a process ofsocialization that starts in infancy, they become part of one’smotives for action. Conformity to standing norms is a stable, acquireddisposition that is independent of the consequences of conforming.Such lasting dispositions are formed by long-term interactions withsignificant others (e.g., one’s parents): through repeatedsocialization, individuals come to learn and internalize the commonvalues embodied in the norms. Internalization is conceived as theprocess by which people develop a psychological need or motive toconform to a set of shared norms. When norms are internalizednorm-abiding behavior will be perceived as good or appropriate, andpeople will typically feel guilt or shame at the prospect of behavingin a deviant way. If internalization is successful external sanctionswill play no role in eliciting conformity and, since individuals aremotivated to conform, it follows that normative beliefs and actionswill be consistent.

Although Parsons’ analysis of social systems starts with atheory of individual action, he views social actors as behavingaccording to roles that define their identities and actions (throughsocialization and internalization). The goal of individual action isto maximize satisfaction. The potential conflict between individualdesires and collective goals is resolved by characterizing the commonvalue system as one that precedes and constrains the social actor. Theprice of this solution is the disappearance of the individual actor asthe basic unit of analysis. Insofar as individuals are role-bearers,in Parsons’ theory it is social entities that act: entities thatare completely detached from the individual actions that created them.This consideration forms the basis for most of the criticisms raisedagainst the theory of the socialized actor (Wrong 1961); suchcriticisms are typically somewhat abstract as they are cast in theframework of the holism/individualism controversy.

On the other hand, one may easily verify whether empirical predictionsdrawn from the socialized actor theory are supported by experimentalevidence. For instance, the following predictions can be derived fromthe theory and easily put to test. (a) Norms will change very slowlyand only through intensive social interaction. (b) Normative beliefsare positively correlated to actions; whenever such beliefs change,behavior will follow. (c) If a norm is successfully internalized,expectations of others’ conformity will have no effect on anindividual’s choice to conform.

Some of the above statements are not supported by empirical evidencefrom social psychology. For example, it has been shown that there maynot be a relation between people’s normative beliefs (orattitudes) and what people in fact do. In this respect, it should benoted that experimental psychologists have generally focused on“attitudes”, that is, “evaluative feelings of pro orcon, favorable or unfavorable, with regard to particularobjects” (where the objects may be “concreterepresentations of things or actions, or abstract concepts”;Insko & Schopler 1967: 361–362). As such, the concept ofattitude is quite broad: it includes normative beliefs, as well aspersonal opinions and preferences. That said, a series of fieldexperiments has provided evidence contrary to the assumption thatattitudes and behaviors are closely related. LaPiere (1934) famouslyreported a sharp divergence between the widespread anti-Chineseattitudes in the United States and the tolerant behavior he witnessed.Other studies have pointed to inconsistencies between anindividual’s stated normative beliefs and her actions (Wicker1969): several reasons may account for such a discrepancy. Forexample, studies of racial prejudice indicate that normative beliefsare more likely to determine behavior in long-lasting relationships,and least likely to determine behavior in the transient situationstypical of experimental studies (Harding et al. 1954 [1969]; Gaertner& Dovidio 1986). Warner and DeFleur (1969) reported that the mainvariable affecting discriminatory behavior is one’s belief aboutwhatsociety (e.g., most other people) says one should do, asopposed to what onepersonally thinks one should do.

In brief, the social psychology literature provides mixed evidence insupport of the claim that an individual’s normative beliefs andattitudes influence her actions. Such studies, however, do notcarefully discriminate among various types of normative beliefs. Inparticular, one should distinguish between “personal normativebeliefs” (i.e., beliefs that a certain behavior ought to befollowed) and “normative expectations” (i.e., what onebelieves others believe ought to be done, that is, a second-orderbelief): it then becomes apparent that oftentimes only suchsecond-order beliefs affect behavior.

The above constitutes an important criticism of the socialized actortheory. According to Parsons, once a norm is internalized, members ofsociety are motivated to conform by an internal sanctioning system;therefore, one should observe a high correlation among all orders ofnormative beliefsand behavior. However, experimentalevidence does not support such a view (see also: Fishbein 1967;Cialdini et al. 1991). Another indication that the socialized actortheory lacks generality is the observation that norms can changerather quickly, and that new norms often emerge in a short period oftime among complete strangers (Mackie 1996). Long-term or closeinteractions do not seem to be necessary for someone to acquire agiven normative disposition, as is testified by the relative ease withwhich individuals learn new norms when they change status or group(e.g., from single to married, from student to faculty, etc.).Moreover, studies of emergent social and political groups have shownthat new norms may form rather rapidly, and that the demise of oldpatterns of behavior is often abrupt (Robinson 1932; Klassen et al.1989; Prentice & Miller 1993; Matza 1964). Given theaforementioned limitations, Parsons’ theory might perhaps betaken as an explanation of a particular conception of moral norms (inthe sense of internalized, unconditional imperatives), but it cannotbe viewed as a general theory ofsocial norms.

3. Early Theories: Social Identity

It has been argued that behavior is often closely embedded in anetwork of personal relations, and that a theory of norms should notleave the specific social context out of consideration (Granovetter1985). Critics of the socialized actor theory have called for analternative conception of norms that may account for the often weakrelation between beliefs and behavior (Deutscher 1973). Thisalternative approach takes social relations to be crucial inexplaining social action, and considers social identity as a keymotivating factor. (A strong support for this view amonganthropologists is to be found in the work of Cancian 1975.)

Since the notion of social identity is inextricably linked to that ofgroup behavior, it is important to clarify the relation between theseconcepts. By “social identity” we refer, in Tajfel’sown words, to

that part of an individual’s self-concept which derives from hisknowledge of his membership of a social group (or groups) togetherwith the value and emotional significance attached to that membership.(Tajfel 1981: 255)

Note that a crucial feature of social identity is that one’sidentification with the group is in some sense a conscious choice: onemay accidentally belong to a group, but we can meaningfully talk ofsocial identification only when being a group-member becomes (at leastin part) constitutive of who one is. According to Tajfel’stheory, when we categorize ourselves as belonging to a particulargroup, the perception and definition of the self—as well as ourmotives—change. That is, we start perceiving ourselves and ourfellow group-members along impersonal, “typical”dimensions that characterize the group to which we belong. Suchdimensions include specific roles and the beliefs (or actions) thataccompany them.

Turner et al.’s (1987) “self-categorization theory” provides a more specificcharacterization of self-perception, or self-definition, as a systemof cognitive self-schemata that filter and process information. Suchschemata result in a representation of the social situation thatguides the choice of appropriate action. This system has at least twomajor components, i.e., social and personal identity. Social identityrefers to self-descriptions related to group memberships. Personalidentity refers to self-descriptions such as individual charactertraits, abilities, and tastes. Although personal and social identitiesare mutually exclusive levels of self-definition, this distinctionmust be taken as an approximation (in that there are manyinterconnections between social and personal identities). It is,however, important to recognize that we often perceive ourselvesprimarily in terms of our relevant group memberships rather than asdifferentiated, unique individuals. So—depending on thesituation—personal or group identity will become salient (Brewer1991).

For example, when one makes interpersonal comparisons between oneselfand other group-members, personal identity will become salient;instead, group identity will become salient in situations in whichone’s group is compared to another group. Within a group, allthose factors that lead members to categorize themselves as different(or endowed with special characteristics and traits) will enhancepersonal identity. If a group has to solve a common task, but eachmember is to be rewarded according to her contribution, personalabilities are highlighted and individuals will perceive themselves asunique and different from the rest of the group. Conversely, if allgroup-members are to equally share the reward for a jointly performedtask, group identification will be enhanced. When the differencebetween self and fellow group-members is accentuated, we are likely toobserve selfish motives and self-favoritism against othergroup-members. When instead group identification is enhanced, in-groupfavoritism against out-group members will be activated, as well asbehavior contrary to self-interest.

According to Turner, social identity is basically a cognitivemechanism whose adaptive function is to make “groupbehavior” possible. Whenever social identification becomessalient, a cognitive mechanism of categorization is activated in sucha way to produce perceptual and behavioral changes. Suchcategorization is called a stereotype, the prototypical description ofwhat members of a given category are (or are believed to be). It is acluster of physical, mental and psychological characteristicsattributed to a “typical” member of a given group.Stereotyping, like any other categorization process, activates scriptsor schemata, and what we call group behavior is nothing but scriptedbehavior. For example, the category “Asian student” isassociated with a cluster of behaviors, personality traits, andvalues: we often think of Asian students as respectful, diligent,disciplined, and especially good with technical subjects. Whenthinking of an Asian student solely in terms of group membership, weattribute her the stereotypical characteristics associated with hergroup, so she becomes interchangeable with other group-members. Whenwe perceive people in terms of stereotypes, we depersonalize them andsee them as “typical” members of their group. The sameprocess is at work when we perceive ourselves as group-members:self-stereotyping is a cognitive shift from “perceiving oneselfas unique” to “perceiving oneself in terms of theattributes that characterize the group”. It is this cognitiveshift that mediates group behavior.

Group behavior (as opposed to individual behavior) is characterized byfeatures such as a perceived similarity between group-members,cohesiveness, a tendency to cooperate to achieve common goals, sharedattitudes or beliefs, and conformity to group norms. Once anindividual self-categorizes as member of a group, she will perceiveherself as “depersonalized” and similar to othergroup-members in the relevant stereotypical dimensions. Insofar asgroup-members perceive their interests and goals asidentical—because such interests and goals are stereotypicalattributes of the group—self-stereotyping will induce agroup-member to embrace such interests and goals as her own. It isthus predicted that pro-social behavior will be enhanced by groupmembership, and diluted when people act in an individualistic mode(Brewer 1979).

The groups with which we happen to identify ourselves may be verylarge (as in the case in which one self-defines as Muslim or French),or as small as a friends’ group. Some general group identitiesmay not involve specific norms, but there are many cases in whichgroup identification and social norms are inextricably connected. Inthat case group-members believe that certain patterns of behavior areunique to them, and use their distinctive norms todefinegroup membership. Many close-knit groups (such as the Amish or theHasidic Jews) enforce norms of separation proscribing marriage withoutsiders, as well as specific dress codes and a host of otherprescriptive and proscriptive norms. There, once an individualperceives herself as a group-member, she will adhere to the groupprototype and behave in accordance with it. Hogg and Turner (1987)have called the process through which individuals come to conform togroup norms “referent informational influence”.

Group-specific norms have (among other things) the twofold function ofminimizing perceived differences among group-members and maximizingdifferences between the group and outsiders. Once formed, such normsbecome stable cognitive representations of appropriate behavior as agroup-member. Social identity is built around group characteristicsand behavioral standards, and hence any perceived lack of conformityto group norms is seen as a threat to the legitimacy of the group.Self-categorization accentuates the similarities between one’sbehavior and that prescribed by the group norm, thus causingconformity as well as the disposition to control and punishtransgressors. In the social identity framework, group norms areobeyed because one identifies with the group, and conformity ismediated by self-categorization as an in-group member. A tellinghistorical example of the relationship between norms and groupmembership was the division of England into the two parties of theRoundheads and Cavaliers. Charles Mackay reports that

in those days every species of vice and iniquity was thought by thePuritans to lurk in the long curly tresses of the monarchists, whilethe latter imagined that their opponents were as destitute of wit, ofwisdom, and of virtue, as they were of hair. A man’s locks werea symbol of his creed, both in politics and religion. The moreabundant the hair, the more scant the faith; and the balder the head,the more sincere the piety. (Mackay 1841: 351)

It should be noted that in this framework social norms are defined bycollective—as opposed to personal—beliefs aboutappropriate behaviors (Homans 1950, 1961). To a certain extent, thischaracterization of social norms is closer to recent accounts than itis to Parsons’ socialized actor theory. On the other hand, adistinct feature of the social identity framework is thatpeople’s motivation to conform comes from their desire tovalidate their identity as group-members. In short, there are severalempirical predictions one can draw from such a framework. Given thetheory’s emphasis on identity as a motivating factor, conformityto a norm is not assumed to depend on an individual’sinternalization of that norm; in fact, a change in social status orgroup membership will bring about a change in the norms relevant tothe new status/group. Thus a new norm can be quickly adopted withoutmuch interaction, and beliefs about identity validation may changevery rapidly under the pressure of external circumstances. In thiscase, not just norm compliance, but norms themselves are potentiallyunstable.

The experimental literature on social dilemmas has utilized the“priming of group identity” as a mechanism for promotingcooperative behavior (Dawes 1980; Brewer & Schneider 1990). Thetypical hypothesis is that a pre-play, face-to-face communicationstage may induce identification with the group, and thus promotecooperative behavior among group-members. In effect, rates ofcooperation have been shown to be generally higher in social dilemmaexperiments preceded by a pre-play communication stage (Dawes 1991).However, it has been argued that face-to-face communication mayactually help group-members gather relevant information about oneanother: such information may therefore induce subjects to trust eachother’s promises and act cooperatively, regardless of any groupidentification. In this respect, it has been shown that communicationper se does not foster cooperation, unless subjects are allowed totalk about relevant topics (Bicchieri & Lev-On 2007). Thisprovides support for the view that communication does not enhancecohesion but rather focuses subjects on relevant rules of behavior,which do not necessarily depend on group identification.

Cooperative outcomes can thus be explained without resorting to theconcept of social identity. A social identity explanation appears tobe more appropriate in the context of a relatively stable environment,where individuals have had time to make emotional investments (or atleast can expect repeated future interactions within the same group).In artificial lab settings, where there are no expectations of futureinteractions, the concept of social identity seems less persuasive asan explanation of the observed rates of cooperation. On the otherhand, we note that social identity does appear to play a role inexperimental settings in which participants are divided into separategroups. (In that case, it has been shown that participants categorizethe situation as “we versus them”, activating in-grouployalty and trust, and an equal degree of mistrust toward theout-group; Kramer & Brewer 1984; Bornstein & Ben-Yossef1994.)

Even with stable environments and repeated interactions, however, atheory of norm compliance in terms of social identity cannot avoid thedifficulty of making predictions when one is simultaneously committedto different identities. We may concurrently be workers, parents,spouses, friends, club members, and party affiliates, to name but afew of the possible identities we embrace. For each of them there arerules that define what is appropriate, acceptable, or good behavior.In the social identity framework, however, it is not clear whathappens when one is committed to different identities that may involveconflicting behaviors.

Finally, there is ample evidence that people’s perceptions maychange very rapidly. Since in this framework norms are defined asshared perceptions about group beliefs, one would expectthat—whenever all members of a group happen to believe thatothers have changed their beliefs about core membershiprules—the very norms that define membership will change. Thestudy of fashion, fads and speculative bubbles clearly shows thatthere are some domains in which rapid (and possibly disruptive)changes of collective expectations may occur; it is, however, muchless clear what sort of norms are more likely to be subject to rapidchanges (think of dress codes rather than codes of honor). The socialidentity view does not offer a theoretical framework fordifferentiating these cases: although some norms are indeed related togroup membership, and thus compliance may be explained throughidentity-validation mechanisms, there appear to be limits to thesocial identity explanation.

4. Early Theories: Cost-Benefit Models

Early rational choice models of conformity maintained that, sincenorms are upheld by sanctions, compliance is merely apayoff-maximizing strategy (Rommetveit 1955; Thibaut & Kelley1959): when others’ approval and disapproval act as externalsanctions, we have a “cost-benefit model” of compliance(Axelrod 1986; James Coleman 1990). Rule-complying strategies arerationally chosen in order to avoid negative sanctions or to attractpositive sanctions. This class of rational choice models defines normsbehaviorally, equating them with patterns of behavior (whiledisregarding expectations or values). Such approach relies heavily onsanctions as a motivating factor. According to Axelrod (1986), forexample, if we observe individuals to follow a regular pattern ofbehavior and to be punished if they act otherwise, then we have anorm. Similarly, Coleman (1990) argues that a norm coincides with aset of sanctions that act to direct a given behavior.

However, it has been shown that not all social norms involve sanctions(Diamond 1935; Hoebel 1954). Moreover, sanctioning works generallywell in small groups and in the context of repeated interactions,where the identity of participants is known and monitoring isrelatively easy. Still, even in such cases there may be a so-calledsecond-order public goods problem. That is, imposing negativesanctions on transgressors is in everybody’s interest, but theindividual who observes a transgression faces a dilemma: she is todecide whether or not to punish the transgressor, where punishingtypically involves costs; besides, there is no guarantee that otherindividuals will also impose a penalty on transgressors when facedwith the same dilemma. An answer to this problem has been to assumethat there exist “meta-norms” that tell people to punishtransgressors of lower-level norms (Axelrod 1986). This solution,however, only shifts the problem one level up: upholding the meta-normitself requires the existence of a higher-level sanctioningsystem.

Another problem with sanctions is the following: a sanction, to beeffective, must be recognized as such. Coleman and Axelrod typicallytake the repeated prisoner’s dilemma game as an example of theworking of sanctions. However, in a repeated prisoner’s dilemmathe same action (“C” or “D”) must serve asboth the sanctioning action and the target action. By simply lookingat behavior, it is unclear whether the action is a function of asanction or a sanction itself. It thus becomes difficult to determinethe presence of a norm, or to assess its effect on choice as distinctfrom the individual strategies of players.

A further consideration weakens the credibility of the view that normsare upheld only because of external sanctions. Often we keepconforming to a norm even in situations of complete anonymity, wherethe probability of being caught transgressing is almost zero. In thiscase fear of sanctions cannot be a motivating force. As a consequence,it is often argued that cases of “spontaneous” complianceare the result of internalization (Scott 1971): people who havedeveloped an internal sanctioning system feel guilt and shame atbehaving in a deviant way. Yet, we have seen that the Parsonian viewof internalization and socialization is inadequate, as it leads topredictions about compliance that often run counter to empiricalevidence.

In particular, James Coleman (1990) has argued in favor of reducinginternalization to rational choice, insofar as it is in the interestof a group to get another group to internalize certain norms. In thiscase internalization would still be the result of some form ofsocialization. This theory faces some of the same objections raisedagainst Parsons’ theory: norms that are passed on from parentsto children, for example, should be extremely resistant to change;hence, one should expect a high degree of correlation between suchnorms and behavior, especially in those cases where norms prescribespecific kinds of actions. However, studies of normative beliefs abouthonesty—which one typically acquires during childhood—showthat such beliefs are often uncorrelated with behavior (Freeman &Ataöv 1960).

Bicchieri (1990, 1997) has presented a third, alternative view aboutinternalization. This view of internalization is cognitive, and isgrounded on the assumption that social norms develop in small,close-knit groups where ongoing interactions are the rule. Once anindividual has learned to behave in a way consistent with thegroup’s interests, she will tend to persist in the learnedbehavior unless it becomes clear that—on average—the costof upholding the norm significantly outweighs the benefits. Smallgroups can typically monitor their members’ behavior andsuccessfully employ retaliation whenever free-riding is observed. Insuch groups an individual will learn, maybe at some personal cost, tocooperate; she will then uphold the cooperative norm as a“default rule” in any new encounter, unless it becomesevident that the cost of conformity has become excessive. The ideathat norms may be “sluggish” is in line with well-knownresults from cognitive psychology showing that, once a norm hasemerged in a group, it will tend to guide the behavior of its memberseven when they face a new situation (or are isolated from the originalgroup; Sherif 1936).

Empirical evidence shows that norm-abiding behavior is not, as theearly rational choice models would have it, a matter of cost/benefitcalculation. Upholding a norm that has led one to fare reasonably wellin the past is a way of economizing on the effort one would have toexert to devise a strategywhen facing a new situation. Thiskind of “bounded rationality” approach explains why peopletend to obey norms that sometimes put them at a disadvantage, as isthe case with norms of honesty. This does not mean, however, thatexternal sanctions never play a role in compliance: for example, inthe initial development of a norm sanctions may indeed play animportant role. Yet, once a norm is established, there are severalmechanisms that may account for conformity.

Finally, the view that one conformsonly because of thethreat of negative sanctions does not distinguish norm-abidingbehavior from an obsession or an entrenched habit; nor does that viewdistinguish social norms from hypothetical imperatives enforced bysanctions (such as the rule that prohibits naked sunbathing on publicbeaches). In these cases avoidance of the sanctions associated withtransgressions constitutes a decisive reason to conform, independentlyof what others do. In fact, in the traditional rational choiceperspective, the only expectations that matter are those about thesanctions that follow compliance or non-compliance. In thoseframeworks, beliefs about how other people will act—as opposedto what they expect us to do—are not a relevant explanatoryvariable: however, this leads to predictions about norm compliancethat often run counter to empirical evidence.

5. Game-Theoretic Accounts

The traditional rational choice model of compliance depicts theindividual as facing a decision problem in isolation: if there aresanctions for non-compliance, the individual will calculate thebenefit of transgression against the cost of norm compliance, andeventually choose so as to maximize her expected utility. Individuals,however, seldom choose in isolation: they know the outcome of theirchoice will depend on the actions and beliefs of other individuals.Game theory provides a formal framework for modeling strategicinteractions.

Thomas Schelling (1960), David Lewis (1969), Edna Ullmann-Margalit(1977), Robert Sugden (1986) and, more recently, Peyton Young (1993),Cristina Bicchieri (1993), and Peter Vanderschraaf (1995) haveproposed a game-theoretic account according to which a norm is broadlydefined as anequilibrium of a strategic interaction. Inparticular, a Nash equilibrium is a combination of strategies (one foreach individual), such that each individual’s strategy is a bestreply to the others’ strategies. Since it is an equilibrium, anorm is supported by self-fulfilling expectations in the sense thatplayers’ beliefs are consistent, and thus the actions thatfollow from players’ beliefs will validate those very beliefs.Characterizing social norms as equilibria has the advantage ofemphasizing the role that expectations play in upholding norms. On theother hand, this interpretation of social norms does notprimafacie explain why people prefer to conform if they expect othersto conform.

Take for exampleconventions such as putting the fork to theleft of the plate, adopting a dress code, or using a particular signlanguage. In all these cases, my choice to follow a certain rule isconditional upon expecting most other people to follow it. Once myexpectation is met, I have every reason to adopt the rule in question.In fact, if I do not use the sign language everybody else uses, I willnot be able to communicate. It is in my immediate interest to followthe convention, since my main goal is to coordinate with other people.In the case of conventions, there is a continuity between theindividual’s self-interest and the interests of the communitythat supports the convention. This is the reason why David Lewismodels conventions as equilibria ofcoordination games. Suchgames have multiple equilibria, but once one of them has beenestablished, players will have every incentive to keep playing it (asany deviation will be costly).

Take instead a norm of cooperation. In this case, the expectation thatalmost everyone abides by it may not be sufficient to inducecompliance. If everyone is expected to cooperate one may be tempted,if unmonitored, to behave in the opposite way. The point is thatconforming tosocial norms, as opposed to conventions, isalmost never in the immediate interest of the individual. Often thereis a discontinuity between the individual’s self-interest andthe interests of the community that supports the social norm.

The typical game in which following a norm would provide a bettersolution (than the one attained by self-centered agents) is amixed-motive game such as the prisoner’s dilemma or thetrust game. In such games the unique Nash equilibrium represents asuboptimal outcome. It should be stressed that—whereas aconvention is one among several equilibria of a coordinationgame—a social norm can never be an equilibrium of a mixed-motivegame. However, Bicchieri (2006) has argued that when a norm exists ittransforms the original mixed-motive game into a coordinationone. As an example, consider the following prisoner’s dilemmagame (Figure 1), where the payoffs are B=Best, S=Second, T=Third, and W=Worst. Clearlythe only Nash equilibrium is to defect (D), in which case both playersget (T,T), a suboptimal outcome. Suppose, however, that society hasdeveloped a norm of cooperation; that is, whenever a social dilemmaoccurs, it is commonly understood that the parties should privilege acooperative attitude. Should, however, does not imply“will”, therefore the new game generated by the existenceof the cooperative norm has two equilibria: either both players defector both cooperate.

Figure 1

Note that, in the new coordination game (which was created by theexistence of the cooperative norm), the payoffs are quite differentfrom those of the original prisoner’s dilemma. Thus there aretwo equilibria: if both players follow the cooperative norm they willplay an optimal equilibrium and get (B,B), whereas if they both chooseto defect they will get the suboptimal outcome (S,S). Players’payoffs in the new coordination game differ from the original payoffsbecause their preferences and beliefs will reflect the existence ofthe norm. More specifically, if a player knows that a cooperative normexists and has the right kind of expectations, then she will have apreference to conform to the norm in a situation in which she canchoose to cooperate or to defect. In the new game generated by thenorm’s existence, choosing to defect when others cooperate isnot a good choice anymore (T,W). To understand why, let us look moreclosely to the preferences and expectations that underlie theconditional choice to conform to a social norm.

Bicchieri (2006) defines the expectations that underlie normcompliance, as follows:

• (a)Empiricalexpectations: individuals believe that a sufficiently largesubset of the relevant group/population conforms to the norm insituations of typeS and either
• (b)Normativeexpectations: individuals believe that a sufficiently largesubset of the relevant group/population expects them to conform to thenorm in situations of typeS;

or

• (b′)Normativeexpectations with sanctions: individuals believe that asufficiently large subset of the relevant group/population expectsthem to conform to the norm in situations of typeS, prefersthem to conform and may sanction behavior.

Note that universal compliance is not usually needed for a norm toexist. However, how much deviance is socially tolerable will depend onthe norm in question. Group norms and well-entrenched social normswill typically be followed by almost all members of a group orpopulation, whereas greater deviance is usually accepted when normsare new or they are not deemed to be socially important. Furthermore,as it is usually unclear how many people follow a norm, differentindividuals may have different beliefs about the size of the group offollowers, and may also have different thresholds for what“sufficiently large” means. What matters to conformity isthat an individual believes that her threshold has been reached orsurpassed. For a critical assessment of the above definition ofnorm-driven preferences, see Hausman (2008).

Brennan et al. (2013) also argue that norms of all kinds share in anessential structure. Norms are clusters of normative attitudes in agroup, combined with the knowledge that such a cluster of attitudesexists. On their account, “A normative principleP is anorm within a groupG if and only if:

1. A significant proportion of the members ofGhaveP-corresponding normative attitudes; and
2. A significant proportion of the members ofGknow that a significant proportion of the members ofG haveP-corresponding attitudes” (Brennan et al. 2013: 29)

On this account, a “P-corresponding normativeattitude” is understood to be a judgment, emotional state,expectation, or other properly first personal normative belief thatsupports the principleP (e.g., Alice thinking most peopleshouldP would count as a normative attitude).Condition (i) is meant to reflect genuine first personal normative commitments,attitudes or beliefs.Condition (ii) is meant to capture those cases where individuals know that a largepart of their group also shares in those attitudes. Putting conditions(i) and (ii) together offers a picture that the authors argue allowsfor explanatory work to be done on a social-level normative conceptwhile remaining grounded in individual-level attitudes.

Consider again the new coordination game ofFigure 1: for players to obey the norm, and thus choose C, it must be the casethat each expects the other to follow it. In the originalprisoner’s dilemma, empirical beliefs would not be sufficient toinduce cooperative behavior. When a norm exists, however, players alsobelieve that others believe they should obey the norm, and may evenpunish them if they do not. The combined force of empirical andnormative expectations makes norm conformity a compelling choice, beit because punishment may follow or just because one recognizes thelegitimacy of others’ expectations (Sugden 2000).

It is important to understand that conformity to a social norm isalways conditional on the expectations of what the relevant other/swill do. Weprefer to comply with the norm as we have certainexpectations. To make this point clear, think of the player who isfacing a typical one-shot prisoner’s dilemma with an unknownopponent. Suppose the player knows a norm of cooperation exists and isgenerally followed, but she is uncertain as to whether the opponent isa norm-follower. In this case the player is facing the followingsituation (Figure 2).

Figure 2

With probabilityp, the opponent is a norm-following type, andwith probability \(1 - p\) she is not. According to Bicchieri,conditional preferences imply that having a reason to be fair,reciprocate or cooperatein a given situation does not entailhaving any general motive or disposition to be fair, reciprocate orcooperate as such. Having conditional preferences means that one mayfollow a norm in the presence of the relevant expectations, butdisregard it in its absence. Whether a norm is followed at a giventime depends on the actual proportion of followers, on theexpectations of conditional followers about such proportion, and onthe combination of individual thresholds.

As an example, consider a community that abides by strict norms ofhonesty. A person who, upon entering the community, systematicallyviolates these norms will certainly be met with hostility, if notutterly excluded from the group. But suppose that a large group ofthieves makes its way into this community. In due time, people wouldcease to expect honesty on the part of others, and would find noreason to be honest themselves in a world overtaken by crime. In thiscase, probably norms of honesty would cease to exist, as the strengthof a norm lies in its being followed by many of the members of therelevant group (which in turn reinforces people’s expectationsof conformity).

What we have discussed is a “rational reconstruction” ofwhat a social norm is. Such a reconstruction is meant to capture someessential features of norm-driven behavior; also, this analysis helpsus distinguish social norms from other constructs such as conventionsor personal norms. A limit of this account, however, is that it doesnot indicate how such equilibria are attained or, in other terms, howexpectations become self-fulfilling.

While neoclassical economics and game theory traditionally conceivedof institutions as exogenous constraints, research in politicaleconomy has generated new insights into the study ofendogenousinstitutions. Specifically, endogenous norms have been shown torestrict the individual’s action set and drive preferences overaction profiles (Bowles 1998; Ostrom 2000). As a result, the“standard” economic framework positing exogenous (and inparticular self-centered) preferences has come under scrutiny. Widelydocumented deviations from the predictions of models withself-centered agents have informed alternative accounts of individualchoice (for one of the first models of “interdependentpreferences”, see Stigler & Becker 1977).

Some alternative accounts have helped reconcile insights aboutnorm-driven behavior with instrumental rationality (Elster 1989b).Moreover, they have contributed to informing the design of laboratoryexperiments on non-standard preferences (for a survey of earlyexperiments, see Ledyard 1995; more recent experiments are reviewed byFehr & Schmidt 2006 and Kagel & Roth 2016). In turn,experimental findings have inspired the formulation of a wide range ofmodels aiming to rationalize the behavior observed in the lab (Camerer2003; Dhami 2016).

It has been argued that the upholding of social norms could simply bemodeled as the optimization of a utility function that includes theothers’ welfare as an argument. For instance, consider some ofthe early “social preference” theories, such as Bolton andOckenfels’ (2000) or Fehr and Schmidt’s (1999) models ofinequity aversion. These frameworks can explain a good wealth ofevidence on preferences for equitable income distributions; theycannot however account forconditional preferences like thosereflecting principles of reciprocity (e.g., I will keep the commonbathroom clean, if I believe my roommates do the same). As notedabove, the approach to social norms taken by philosophically-inclinedscholars has emphasized the importance of conditional preferences insupporting social norms. In this connection, we note that some of thesocial preference theories do account for motivations conditional onempirical beliefs, whereby a player upholds a principle of“fair” behavior if she believes her co-players will upholdit too (Rabin 1993; Dufwenberg & Kirchsteiger 2004; Falk &Fischbacher 2006; Charness & Rabin 2002). These theoriespresuppose that players are hardwired with a notion of fair or kindbehavior, as exogenously defined by the theorist. Since theyimplicitly assume that all players have internalized aunique—exogenous—normative standpoint (as reflected insome notion of fairness or kindness), these theories do not explicitlymodelnormative expectations. Hence, players’preferences are assumed to be conditional solely on their empiricalbeliefs; that is, preferences are conditional on whether others willbehave fairly (according to an exogenous principle) or not.

That said, we stress that social preferences should not be conflatedwith social norms. Social preferences capturestabledispositions toward an exogenously defined principle of conduct(Binmore 2010). By contrast, social norms are better studied asgroup-specific solutions to strategic problems (Sugden 1986;Bicchieri 1993; Young 1998b). Such solutions are brought about by aparticular class of preferences (“norm-drivenpreferences”), conditional on the relevant set of empiricalbeliefs and normative expectations. In fact, we stress that“what constitutes fair or appropriate behavior” oftenvaries with cultural or situational factors (Henrich et al. 2001;Cappelen et al. 2007; Ellingsen et al. 2012). Accounting forendogenous expectations is therefore key to a full understanding ofsocial norms.

Relatedly, Guala (2016) offers a game-theoretic account ofinstitutions, arguing that institutions are sets of rules inequilibrium. Guala’s view incorporates insights from twocompeting accounts of institutions: institutions-as-rules (perhapsbest rendered by North 1990), and institutions-as-equilibria. From thefirst account, he captures the idea that institutions create rulesthat help to guide our behaviors and reduce uncertainty. With rules inplace, we more or less know what to do, even in new situations. Fromthe second, he captures the idea that institutions are solutions tocoordination problems that arise from our normal interactions. Theinstitutions give us reasons to follow them. The function of therules, then, is to point to actions that promote coordination andcooperation. Because of the equilibrium nature of the rules, eachindividual has an incentive to choose those actions, provided othersdo too. Guala relies on a correlated equilibrium concept to unite therules and equilibria accounts. On this picture, an institution issimply a correlated equilibrium in a game, where other correlatedequilibria would have been possible.

6. Experimental Evidence

In what follows we focus on lab experiments that identify social normsby explicitly measuring both empirical and normative expectations.

Xiao and Bicchieri (2010) designed an experiment to investigate theimpact on trust games of two potentially applicable—butconflicting—principles of conduct, namely,equality andreciprocity. Note that the former can be broadly defined as arule that recommends minimizing payoff differences, whereas the latterrecommends taking a similar action as others (regardless of payoffconsiderations). The experimental design involved two trust gamevariants: in the first one, players started with equal endowments; inthe second one, the investor was endowed with twice the money that thetrustee was given. In both cases, the investor could choose totransfer a preset amount of money to the trustee or keep it all. Uponreceiving the money, the trustee could in turn keep it or elsetransfer back some of it to the investor: in the equal endowmentcondition (“baseline treatment”), both equality andreciprocity dictate that the trustee transfer some money back to theinvestor; by contrast, in the unequal endowment condition(“asymmetry treatment”), equality and reciprocity dictatedifferent actions as the trustee could guarantee payoff equality onlyby making a zero back-transfer. Xiao and Bicchieri elicitedsubjects’ first- and second-orderempirical beliefs(“how much do you think other participants in your role willtransfer to their counterpart?”; “what does yourcounterpart think you will do?”) andnormativeexpectations (“how much do you think your counterpartbelieves you should transfer to her?”). The experimental resultsshow that a majority of trustees returned a positive amount wheneverreciprocity would reduce payoff inequality (in the baselinetreatment); by contrast, a majority of trustees did not reciprocatethe investors’ transfer when doing so would increase payoffinequality (in the asymmetry treatment). Moreover, investors correctlybelieved that less money would be returned in the asymmetry treatmentthan in the baseline treatment, and most trustees correctly estimatedinvestors’ beliefs in both treatments. However, in the asymmetrytreatment empirical beliefs and normative expectations conflicted:this highlights that, when there is ambiguity as to which principle ofconduct is in place, each subject will support the rule of behaviorthat favors her most.

Reuben and Riedl (2013) examine the enforcement of norms ofcontribution to public goods in homogeneous and heterogeneous groups,such as groups whose members vary in their endowment, contributioncapacity, or marginal benefits. In particular, Reuben and Riedl areinterested in the normative appeal of two potentially applicablerules: theefficiency rule (prescribing maximal contributionsby all) and the class ofrelative contribution rules(prescribing a contribution that is “fair” relative to thecontributions of others; e.g., equality and equity rules). Reuben andRiedl’s results show that, in the absence of punishment, nopositive contribution norm emerged and all groups converged towardfree-riding. By contrast, with punishment, contributions wereconsistent with the prescriptions of the efficiency rule in asignificant subset of groups (irrespective of the type of groupheterogeneity); in other groups, contributions were consistent withrelative contribution rules. These results suggest that even inheterogeneous groups individuals can successfully enforce acontribution norm. Most notably, survey data involving third partiesconfirmed well-defined yet conflicting normative views about theaforementioned contribution rules; in other words, both efficiency andrelative contribution rules are normatively appealing, and are indeedpotential candidates for emerging contribution norms in differentgroups.

Bicchieri and Chavez (2010) designed an experiment to investigate normcompliance in ultimatum games. Specifically, their experiment involveda variant of the ultimatum game whereby the proposer could choose oneof the following three options:($5, $5),($8, $2),orCoin (in which case one of the other two allocations wouldbe selected at random). This design allows for two plausible notionsof fairness: as anequal outcome ($5, $5) or as afairprocedure (Coin). The experimenters elicited subjects’normative expectations about the actions they thought would beconsidered fair by most participants: proposers and responders showeda remarkable degree of agreement in their notions of fairness, as mostsubjects believed that a majority of participants deemed both ($5, $5)and Coin to be appropriate. Further, the experimenters had subjectsplay three instances of the above ultimatum game under differentinformation conditions. In the “full information”condition, all participants knew that the Coin option was available,and that responders would know if their respective proposer had chosenCoin. In the “private information” condition, respondersdid not know that Coin was available to proposers, and proposers wereaware of responders’ ignorance. In the “limitedinformation” condition, participants knew that the Coin optionwas available, but responders would not be able to distinguish whethertheir respective proposer had implemented one of the two allocationsdirectly or had chosen Coin instead. The experimental results showthat when normative expectations supporting the Coin option wereeither absent (in theprivate condition) or could be defiedwithout consequence (in thelimited condition), the frequencyof choice of ($5, $5) and ($8, $2), respectively, were considerablyhigher than those of Coin. Moreover, the frequency of Coin choices washighest in thepublic information condition, where suchoption was common knowledge and its outcome transparent: this showsthat there proposers followed the rule of behavior that favored themmost, and that such a rule was effectively a social norm. On the otherhand, substantial norm evasion characterized proposers’ behaviorin the limited information condition, where ($8, $2) was the mostfrequent choice.

In a subsequent study, Chavez and Bicchieri (2013) measured empiricaland normative expectations (as well as behavior) of third parties whowere given the opportunity to add to or deduct from the payoffs ofsubjects who had participated in an ultimatum game. Third partiestended to reward subjects involved in equal allocations and tocompensate victims of unfair allocations (rather than punish unfairbehavior); on the other hand, third parties were willing to punishwhen compensation was not an available option. The experimentalresults further show that third parties shared a notion of fairness(as indicated by their normative expectations), and that such notionwas sensitive to contextual differences.

Krupka and Weber (2013) introduced an interesting procedure foridentifying social norms by means of pre-play coordination games. Inbrief, using alternative (between-subjects) variants of the dictatorgame, Krupka and Weber had participants assess the extent to whichdifferent actions were collectively perceived as socially appropriate:subjects providing these ratings effectively faced a coordinationgame, as they were incentivized to match the modal response given byothers in the same situation (such a pre-play coordination game wasintended to verify the presence of shared normative expectations).Krupka and Weber went on to use these elicited assessments to predictother subjects’ compliance with the relevant social norm in eachdictator game variant (for another application of the same elicitationprocedure, see Gächter et al. 2013).

Similarly, Schram and Charness’ (2015) proposed a procedure forinducing a shared understanding of the relevant rule of behavior, inthe lab. In short, Schram and Charness had participants in dictatorgames receive advice from a group of third parties. The informationreceived simply revealed what a group of uninvolved subjects thoughtdictatorsought to do: as such, the information receivedgenerated an exogenous variation in the dictators’ normativeexpectations. Schram and Charness’ results show that choices areindeed affected by this information.

Bicchieri and Xiao (2009) designed an experiment to investigate whathappens when empirical and normative expectations conflict. To thatend, participants in a dictator game were exposed to different piecesof information. Specifically, two groups of dictators were given some“descriptive information”; that is, they were told whatother subjects haddone in another session (i.e., one groupwas told that previous participants had made for the most part agenerous offer, while the other group was told that most participantshad made a selfish offer). Further, another two groups of dictatorswere given some “normative information”; that is, theywere told what previous subjects saidought to be done (i.e.,one group was told that most previous participants thought that oneshould make a generous offer, while the other group was told that mostparticipants thought that one should make a selfish offer). Othergroups were given both descriptive and normative information. Theexperimental results show that—whenever such information did notconflict—both descriptive and normative messages had asignificant influence on dictators’ own expectations andsubsequent choices. When messages conflicted in that one indicatedgenerosity and the other indicated selfishness, only the descriptiveinformation affected dictators’ behavior. This suggests that ifpeople recognize that others are breaching the norm, then they will nolonger feel compelled to follow the relevant rule of behaviorthemselves.

To conclude, the studies surveyed here provide evidence of the roleplayed by expectations in affecting behavior in a variety of socialdilemmas. In this regard, we note that in contrast to the vastliterature on empirical beliefs, the number of lab studies thatdirectly measure normative expectations is relatively limited: moreresearch is clearly needed to investigate the interplay of empiricaland normative information about applicable rules of behavior.

7. Evolutionary Models

Thus far we have examined accounts of social norms that take forgranted that a particular norm exists in a population. However, for afull account of social norms, we must answer two questions related tothe dynamics of norms. First, we must ask how a norm can emerge. Normsrequire a set of corresponding beliefs and expectations to supportthem, and so there must be an account of how these arise. Second, wemust investigate the conditions under which a norm is stable undersome competitive pressure from other norms. Sometimes, multiplecandidate norms vie for dominance in a population. Even if one normhas come to dominate the population, new norms can try to“invade” the existing norm’s population ofadherents.

Let us now turn to the question of norm emergence. Here we can seethree classes of models: first, a purely biological approach, second,a more cognitive approach, and third, a structured interactionsapproach. The most famous of the biological approaches to norms seekto explain cooperative behavior. The simplest models are kin selectionmodels (Hamilton 1964). These models seek to explain altruistictendencies in animals by claiming that, as selection acts on genes,those genes have an incentive to promote the reproductive success ofother identical sets of genes found in other animals. This mode ofexplanation can provide an account of why we see cooperative behaviorswithin families, but being gene-centered, cannot explain cooperativebehavior toward strangers (as strangers should not be sufficientlygenetically related to merit altruistic behavior).

Models of “reciprocal altruism” (Trivers 1971, 1985), onthe other hand, tell us that cooperative behavior has no chance ofevolving in random pairings, but will evolve in a social framework inwhich individuals can benefit from building reputations for being niceguys. Reciprocal altruism, however, does not require an evolutionaryargument; a simple model of learning in ongoing close-knit groups willdo, and has the further advantage of explaining why certain types ofcooperative behavior are more likely to emerge than others. All thatmatters in these models is that agents can properly identify otheragents, such that they can maintain a record of their past behavior.This allows for the possibility of reputations: people who have thereputation of being cooperative will be treated cooperatively, andthose who have a reputation of being unfair will be treatedunfairly.

A variation on the idea of reciprocal altruism can be seen in Axelrod(1986). Axelrod presents a “norms game” in which agentsprobabilistically choose to comply with the norm, or deviate from it,and then other agents can probabilistically choose to punish anydeviations at some cost to them. Agents can choose over time to bemore or less “bold”, which determines the rate at whichthey attempt defections, and they can likewise choose to be more orless “vengeful”, which determines how often they punish.Axelrod noted that if the game is left like this, we find that thestable state is constant defection and no punishment. However, if weintroduce a meta-norm—one that punishes people who fail topunish defectors—then we arrive at a stable norm in which thereis no boldness, but very high levels of vengefulness. It is underthese conditions that we find a norm emerge and remain stable.Axelrod’s model aims to illustrate that norms requiremeta-norms. That is, failure to retaliate against a defection must beseen as equivalent to a defection itself. What Axelrod does notanalyze is whether there is some cost to being vigilant. Namely,watching both defectors and non-punishers may have a cost that, thoughnominal, might encourage some to abandon vigilance once there has beenno punishment for some time.

Bicchieri, Duffy and Tolle (2004) present an alternative model of normemergence to explain how a norm of impersonal trust/reciprocity canemerge and survive in a heterogeneous population. This model does notrely on a meta-norm of punishment; instead, it is purely driven byrepeated interactions of conditional strategies. In their model,agents play anywhere from 1 to 30 rounds of a trust game for 1,000iterations, relying on the 4 unconditional strategies, and the 16conditional strategies that are standard for the trust game. Aftereach round, agents update their strategies based on the replicatordynamic. As the number of rounds grows, a norm of impersonaltrust/reciprocity emerges in the population. Most interestingly,however, the norm is not associated with a single strategy, but it issupported by several strategies behaving in similar ways. This modelsuggests that Trivers’ basic model works well in normal socialcontexts, but we can further enrich the story by allowing a socialnorm to supervene on several behavioral strategies.

The third prominent model of norm emergence comes from Brian Skyrms(1996, 2004) and Jason Alexander (2007). In this approach, twodifferent features are emphasized: relatively simple cognitiveprocesses and structured interactions. Both have explored a variety ofgames (such as the prisoner’s dilemma, the stag hunt, divide thedollar, and the ultimatum game) as exemplars of situations that offerthe possibility of the emergence of a moral norm. Though Skyrmsoccasionally uses the replicator dynamic, both tend to emphasizesimpler mechanisms in an agent-based learning context. In particular,learning rules like “imitate the best” or best responseare used, as they are much less cognitively demanding. Alexanderjustifies the use of these simpler rules on the grounds that, ratherthan fully rational agents, we are cognitively limited beings who relyon fairly simple heuristics for our decision-making. Rules likeimitation are extremely simple to follow. Best response requires a bitmore cognitive sophistication, but is still simpler than a fullyBayesian model with unlimited memory and computational power. Thesesimpler learning rules provide the same function as the replicatordynamic: in between rounds of play, agents rely on their learning ruleto decide what strategy to employ. Note that both Skyrms and Alexandertend to treat norms as single strategies.

The largest contribution of this strain of modeling comes not from theassumption of boundedly rational agents, but rather the carefulinvestigation of the effects of particular social structures on theequilibrium outcomes of various games. Much of the previous literatureon evolutionary games has focused on the assumptions of infinitepopulations of agents playing games against randomly-assignedpartners. Skyrms and Alexander both rightly emphasize the importanceof structured interaction. As it is difficult to uncover and representreal-world network structures, both tend to rely on examiningdifferent classes of networks that have different properties, and fromthere investigate the robustness of particular norms against thesealternative network structures. Alexander (2007) in particular hasdone a very careful study of the different classical networkstructures, where he examines lattices, small world networks, boundeddegree networks, and dynamic networks for each game and learning rulehe considers. A final feature of Skyrms and Alexander’s work isa refinement on this structural approach: they separate out twodifferent kinds of networks. First, there is theinteractionnetwork, which represents the set of agents that any given agentcan actively play a game with. Second is theupdate network,which is the set of agents that an agent can “see” whenapplying her learning rule. The interaction network is thusone’s immediate community, whereas the update network is allthat the agent can see. To see why this is useful, we can imagine acase not too different from how we live, in which there is a fairlylimited set of other people we may interact with, but thanks to aplethora of media options, we can see much more widely how othersmight act. This kind of situation can only be represented by clearlyseparating the two networks.

Thus, what makes the theory of norm emergence of Skyrms and Alexanderso interesting is its enriching the set of idealizations that one mustmake in building a model. The addition of structured interaction andstructured updates to a model of norm emergence can help make clearhow certain kinds of norms tend to emerge in certain kinds ofsituation and not others, which is difficult or impossible to capturein random interaction models.

Now that we have examined norm emergence, we must examine what happenswhen a population is exposed to more than one social norm. In thisinstance, social norms must compete with each other for adherents.This lends itself to investigations about the competitive dynamics ofnorms over long time horizons. In particular, we can investigate thefeatures of norms and of their environments, such as the populationsthemselves, which help facilitate one norm becoming dominant overothers, or becoming prone to elimination by its competitors. Anevolutionary model provides a description of the conditions underwhich social norms may spread. One may think of several environmentsto start with. A population can be represented as entirelyhomogeneous, in the sense that everybody is adopting the same type ofbehavior, or heterogeneous to various degrees. In the former case, itis important to know whether the commonly adopted behavior is stableagainst mutations. The relevant concept here is that of anevolutionarily stable strategy (ESS; Maynard Smith &Price 1973; Taylor & Jonker 1978): when a population ofindividuals adopts such a strategy, it cannot be successfully invadedby isolated mutants, since the mutants will be at a disadvantage withrespect to reproductive success. An evolutionarily stable strategy isa refinement of the Nash equilibrium in game theory. Unlike standardNash equilibria, evolutionarily stable strategies must either bestrict equilibria, or have an advantage when playing againstmutant strategies. Since strict equilibria are always superior to anyunilateral deviations, and the second condition requires that the ESShave an advantage in playing against mutants, the strategy will remainresistant to any mutant invasion. This is a difficult criterion tomeet, however. For example, a classic Tit-For-Tat strategy in theprisoner’s dilemma is not an ESS. Many strategies performequally well against it, including the very simple “AlwaysCooperate” strategy, let alone Tit-For-Two-Tats, and any numberof variations. Tit-For-Tat is merely an evolutionarily neutralstrategy relative to these others. If we only consider strategies thatare defection-oriented, then Tit-For-Tat is an ESS, since it will dobetter against itself, and no worse than defection strategies whenpaired with them.

A more interesting case, and one relevant to a study of thereproduction of norms of cooperation, is that of a population in whichseveral competing strategies are present at any given time. What wewant to know is whether the strategy frequencies that exist at a timeare stable, or if there is a tendency for one strategy to becomedominant over time. If we continue to rely on the ESS solutionconcept, we see a classic example in the hawk-dove game. If we assumethat there is no uncorrelated asymmetry between the players, then themixed Nash equilibrium is the ESS. If we further assume that there isno structure to how agents interact with each other, this can beinterpreted in two ways: either each player randomizes her strategy ineach round of play, or we have a stable polymorphism in thepopulation, in which the proportion of each strategy in the populationcorresponds to the frequency with which each strategy would be playedin a randomizing approach. So, in those cases where we can assume thatplayers randomly encounter each other, whenever there is a mixedsolution ESS we can expect to find polymorphic populations.

If we wish to avoid the interpretive challenge of a mixed solutionESS, there is an alternative analytic solution concept that we canemploy: the evolutionarily stable state. An evolutionarily stablestate is a distribution of (one or more) strategies that is robustagainst perturbations, whether they are exogenous shocks or mutantinvasions, provided the perturbations are not overly large.Evolutionarily stable states are solutions to a replicator dynamic.Since evolutionarily stable states are naturally able to describepolymorphic or monomorphic populations, there is no difficulty withintroducing population-oriented interpretations of mixed strategies.This is particularly important when random matching does not occur, asunder those conditions, the mixed strategy can no longer be thought ofas a description of population polymorphism.

Now that we have seen the prominent approaches to both norm emergenceand norm stability, we can turn to some general interpretiveconsiderations of evolutionary models. An evolutionary approach isbased on the principle that strategies with higher current payoffswill be retained, while strategies that lead to failure will beabandoned. The success of a strategy is measured by its relativefrequency in the population at any given time. This is most easilyseen in a game theoretic framework. A game is repeated a finite numberof times with randomly selected opponents. After each round of thegame, the actual payoffs and strategies of the players become publicknowledge; on the basis of this information, each player adjusts herstrategy for the next round. The payoff to an individual playerdepends on her choice as well as on the choices of the other playersin the game, and players are rational in the sense that they arepayoff-maximizers. In an evolutionary model, however, players learnand adapt in a non-Bayesian way, that is, they do not condition onpast experience using Bayes’ Rule. In this sense, they are nottypical rational learners (Nachbar 1990; Binmore & Samuelson1992).

In an evolutionary approach behavior is adaptive, so that a strategythat did work well in the past is retained, and one that fared poorlywill be changed. This can be interpreted in two ways: either theevolution of strategies is the consequence of adaptation by individualagents, or the evolution of strategies is understood as thedifferential reproduction of agents based on their success rates intheir interactions. The former interpretation assumes short timescalesfor interactions: many iterations of the game over time thus representno more than a few decades in time in total. The latter interpretationassumes rather longer timescales: each instance of strategy adjustmentrepresents a new generation of agents coming into the population, withthe old generation dying simultaneously. Let us consider theramifications of each interpretation in turn.

In the first interpretation, we have agents who employ learning rulesthat are less than fully rational, as defined by what a Bayesian agentwould have, both in terms of computational ability and memory. Assuch, these rules tend to be classified as adaptive strategies: theyare reacting to a more limited set of data, with lower cognitiveresources than what a fully rational learner would possess. However,there are many different adaptive mechanisms we may attribute to theplayers. One realistic adaptive mechanism is learning by trial anderror; another plausible mechanism is imitation: those who do best areobserved by others who subsequently emulate their behavior (Hardin1982). Reinforcement learning is another class of adaptive behavior,in which agents tweak their probabilities of choosing one strategyover another based on the payoffs they just received.

In the second interpretation, agents themselves do not learn, butrather the strategies grow or shrink in the population according tothe reproductive advantages that they bestow upon the agents thatadhere to them. This interpretation requires very long timescales, asit requires many generations of agents before equilibrium is reached.The typical dynamics that are considered in such circumstances comefrom biology. A standard approach is something like the replicatordynamic. Norms grow or shrink in proportion to both how many agentsadhere to them at a given time, and their relative payoffs. Moresuccessful strategies gain adherents at the expense of less-successfulones. This evolutionary process assumes a constant-sized (or infinite)population over time. This interpretation of an evolutionary dynamic,which requires long timescales, raises the question of whether normsthemselves evolve slowly. Norms can rapidly collapse in a very shortamount of time. This phenomenon could not be represented within amodel whose interpretation is generational in nature. It remains anopen question, however, as to whether such timescales can beappropriate for examining the emergence of certain kinds of norms.While it is known that many norms can quickly come into being, it isnot clear if this is true of all norms.

Another challenge in using evolutionary models to study social normsis that there is a potential problem of representation. Inevolutionary models, there is no rigorous way to represent innovationor novelty. Whether we look at an agent-based simulation approach, ora straightforward game-theoretic approach, the strategy set open tothe players, as well as their payoffs, must be defined in advance. Butmany social norms rely on innovations, whether they are technologicalor social. Wearing mini-skirts was not an option until they wereinvented. Marxist attitudes were largely not possible until Marx. Theage at which one gets married and how many children one has are highlylinked to availability of and education about birth controltechnologies. While much of the study of norms has focused on moregeneric concepts such as fairness, trust, or cooperation, the fullbreadth of social norms covers many of these more specific norms thatrequire some account of social innovation.

This representational challenge has broad implications. Even when wecan analytically identify evolutionarily stable states in a particulargame, which is suggestive of norms that will be converged upon, we nowhave a problem of claiming that this norm has prospects for long-termstability. Events like the publication of the Kinsey report candramatically shift seemingly stable norms quite rapidly. As theunderlying game changes in the representation, our previous results nolonger apply. In the face of this representational problem, we caneither attempt to develop some metric of the robustness of a givennorm in the space of similar games, or more carefully scope the claimsthat we can make about the social norms that we study with thismethodology.

Although some questions of interpretation and challenges ofrepresentation exist, an important advantage of the evolutionaryapproach is that it does not require sophisticated strategic reasoningin circumstances, such as large-group interactions, in which it wouldbe unrealistic to assume it. People are very unlikely to engage infull Bayesian calculations in making decisions about norm adherence.Agents often rely on cognitive shortcuts to determine when norms oughtto be in effect given a certain context, and whether or not theyshould adhere to them. Evolutionary models that employ adaptivelearning strategies capture these kinds of cognitive constraints, andallow the theorist to explore how these constraints influence theemergence and stability of norms.

8. Conclusion

The study of social norms can help us understand a wide variety ofseemingly puzzling behaviors. According to some accounts, a socialnorm results from conditional preferences for conforming to a relevantbehavioral rule. Such preferences are conditional on two differentkinds of beliefs: empirical and normative expectations.

This and other accounts of social norms still leave much to beinvestigated. Explaining how normative expectations come to existremains an open question. Another open question to consider is how onecould intervene to change socially harmful norms.

Finally, we stress that differentcontextual factors (such asthe framing and characteristics of the strategic problem, the role oneis assigned, the social category with which one identifies, as well ashistorical and chance events) often come to be associated withdifferent notions of “appropriate behavior”. Accountingfor endogenous expectations is therefore key to a full understandingof norm-driven behavior. More research—both theoretical andexperimental—is needed to further illuminate the impact ofexpectations on strategic decisions.

Bibliography

• Akerlof, George A., 1976, “The Economics of Caste and of theRat Race and Other Woeful Tales”,Quarterly Journal ofEconomics, 90(4): 599–617. doi:10.2307/1885324
• Alexander, Jason McKenzie, 2000, “Evolutionary Explanationsof Distributive Justice”,Philosophy of Science, 67(3):490–516. doi:10.1086/392792
• –––, 2005, “The Evolutionary Foundationsof Human Altruism”,Analyse & Kritik, 27(1):106–113. [Alexander 2005 available online]
• –––, 2007,The Structural Evolution ofMorality, Cambridge: Cambridge University Press.doi:10.1017/CBO9780511550997
• Andreoni, James and B. Douglas Bernheim, 2009, “Social Imageand The 50–50 Norm: A Theoretical and Experimental Analysis ofAudience Effects”,Econometrica, 77(5):1607–1636. doi:10.3982/ECTA7384
• Arrow, Kenneth J., 1971, “A Utilitarian Approach to theConcept of Equality in Public Expenditure”,QuarterlyJournal of Economics, 85(3): 409–15.doi:10.2307/1885930
• Axelrod, Robert, 1984,The Evolution of Cooperation, NewYork: Basic Books.
• –––, 1986, “An Evolutionary Approach toNorms”,American Political Science Review, 80(4):1095–1111. doi:10.1017/S0003055400185016
• Battigalli, Pierpaolo and Martin Dufwenberg, 2007, “Guilt inGames”,American Economic Review: Papers andProceedings, 97(2): 170–176. doi:10.1257/aer.97.2.170
• –––, 2009, “Dynamic PsychologicalGames”,Journal of Economic Theory, 144(1): 1–35.doi:10.1016/j.jet.2008.01.004
• Bénabou, Roland and Jean Tirole, 2006, “Incentives andProsocial Behavior”,American Economic Review, 96(5):1652–1678. doi:10.1257/aer.96.5.1652
• Bicchieri, Cristina, 1990, “Norms of Cooperation”,Ethics, 100(4): 838–861. doi:10.1086/293237
• –––, 1993,Rationality andCoordination, Cambridge: Cambridge University Press. SecondEdition, 1996.
• –––, 1997, “Learning to Cooperate”,in Cristina Bicchieri, Richard C. Jeffrey, and Brian Skyrms,TheDynamics of Norms, Cambridge: Cambridge University Press.
• –––, 2002, “Covenants Without Swords:Group Identity, Norms, and Communication in Social Dilemmas”,Rationality and Society, 14(2): 192–228.doi:10.1177/1043463102014002003
• –––, 2006,The Grammar of Society: TheNature and Dynamics of Social Norms, Cambridge: CambridgeUniversity Press. doi:10.1017/CBO9780511616037
• –––, 2016,Norms in the Wild: How toDiagnose, Measure, and Change Social Norms, New York: OxfordUniversity Press. doi:10.1093/acprof:oso/9780190622046.001.0001
• Bicchieri, Cristina and Alex Chavez, 2010, “Behaving asExpected: Public Information and Fairness Norms”,Journal ofBehavioral Decision Making, 23(2): 161–178.doi:10.1002/bdm.648
• Bicchieri, Cristina, John Duffy, and Gil Tolle, 2004, “TrustAmong Strangers”,Philosophy of Science, 71(3):286–319. doi:10.1086/381411
• Bicchieri, Cristina and Yoshitaka Fukui, 1999, “The GreatIllusion: Ignorance, Informational Cascades and the Persistence ofUnpopular Norms”,Business Ethics Quarterly, 9(1):127–155. doi:10.2307/3857639
• Bicchieri, Cristina and Azi Lev-On, 2007, “Computer-MediatedCommunication and Cooperation in Social Dilemmas: An ExperimentalAnalysis”,Politics, Philosophy and Economics, 6(2):139–168. doi:10.1177/1470594X07077267
• Bicchieri, Cristina and Alessandro Sontuoso, 2015, “I CannotCheat on You after We Talk”, inThe Prisoner’sDilemma, Martin Peterson (ed.), Cambridge: Cambridge UniversityPress, pp. 101–114. doi:10.1017/CBO9781107360174.007
• –––, forthcoming, “Game-Theoretic Accountsof Social Norms”, in M. Capra, R. Croson, T. Rosenblatt, and M.Rigdon (eds.),The Handbook of Experimental Game Theory,Cheltenham: Edward Elgar Publishing.
• Bicchieri, Cristina and Erte Xiao, 2009, “Do the RightThing: But Only If Others Do So”,Journal of BehavioralDecision Making, 22(2): 191–208. doi:10.1002/bdm.621
• Bicchieri, Cristina and Jiji Zhang, 2012, “An Embarrassmentof Riches: Modeling Social Preferences in Ultimatum Games”, inPhilosophy of Economics, (Handbook of the Philosophy ofScience, Volume 13), Uskali Mäki (ed.), Amsterdam:Elsevier.
• Binmore, Ken, 2010, “Social Norms or SocialPreferences?”,Mind & Society, 9(2): 139–157.doi:10.1007/s11299-010-0073-2
• Binmore, Kenneth G. and Larry Samuelson, 1992, “EvolutionaryStability in Repeated Games Played by Finite Automata”,Journal of Economic Theory, 57(2): 278–305.doI:10.1016/0022-0531(92)90037-I
• Bolton, Gary E. and Axel Ockenfels, 2000, “ERC: A Theory ofEquity, Reciprocity, and Competition”,American EconomicReview, 90(1): 166–193. doi:10.1257/aer.90.1.166
• Bornstein, Gary and Meyrav Ben-Yossef, 1994, “Cooperation inIntergroup and Single-Group Social Dilemmas”,Journal ofExperimental Social Psychology, 30(1): 52–67.doi:10.1006/jesp.1994.1003
• Bowles, Samuel, 1998, “Endogenous Preferences: The CulturalConsequences of Markets and Other Economic Institutions”,Journal of Economic Literature, 36(1): 75–111.
• Brennan, Geoffrey, Lina Eriksson, Robert E. Goodin, and NicholasSouthwood, 2013,Explaining Norms, New York: OxfordUniversity Press. doi:10.1093/acprof:oso/9780199654680.001.0001
• Brewer, Marilynn B., 1979, “In-group Bias in the MinimalIntergroup Situation: A Cognitive-Motivational Analysis”,Psychological Bulletin, 86(2): 307–324.doi:10.1037/0033-2909.86.2.307
• –––, 1991, “The Social Self: On Being theSame and Different at the Same Time”,Personality and SocialPsychology Bulletin, 17(5): 475–482.doi:10.1177/0146167291175001
• Brewer, Marilynn B. and Sherry K. Schneider, 1990, “SocialIdentity and Social Dilemmas: A Double-Edged Sword”, in DominicAbrams and Michael A. Hogg (eds.),Social Identity Theory:Constructive and Critical Advances, Wheatsheaf, NY:Harvester.
• Camerer, Colin F., 2003,Behavioral Game Theory. Experimentsin Strategic Interaction, Princeton, NJ: Princeton UniversityPress.
• Cancian, Francesca M, 1975,What are Norms? A Study of Beliefsand Action in a Maya Community, Cambridge: Cambridge UniversityPress.
• Cappelen, Alexander W., Astri Drange Hole, Erik Ø Sørensen andBertil Tungodden, 2007, “The Pluralism of Fairness Ideals: AnExperimental Approach”,American Economic Review,97(3): 818–827. doi:10.1257/aer.97.3.818
• Charness, Gary and Matthew Rabin, 2002, “UnderstandingSocial Preferences with Simple Tests”,Quarterly Journal ofEconomics, 117(3): 817–869.doi:10.1162/003355302760193904
• Chavez, Alex K. and Cristina Bicchieri, 2013, “Third-PartySanctioning and Compensation Behavior: Findings from the UltimatumGame”,Journal of Economic Psychology, 39:268–277. doi:10.1016/j.joep.2013.09.004
• Cialdini, Robert B. and Noah J. Goldstein, 2004, “SocialInfluence: Compliance and Conformity”,Annual Review ofPsychology, 55: 591–621.doi:10.1146/annurev.psych.55.090902.142015
• Cialdini, Robert B., Carl A. Kallgren, and Raymond R. Reno, 1991,“A Focus Theory of Normative Conduct”, inAdvances inExperimental Social Psychology, volume 24, Mark P. Zanna (ed.),New York: Academic Press, pp. 201–234.doi:10.1016/S0065-2601(08)60330-5
• Coleman, James S., 1990,Foundations of Social Theory,Cambridge, MA: Belknap.
• Coleman, Jules L., 1989, “Afterword: Rational ChoiceApproach to Legal Rules”,Chicago-Kent Law Review,65(1): 177–191. [Jules Coleman 1989 available online]
• Dawes, Robyn M., 1972,Fundamentals of Attitude Measurement, New York: Wiley.
• –––, 1980, “Social Dilemmas”,Annual Review of Psychology, 31: 169–193.doi:10.1146/annurev.ps.31.020180.001125
• –––, 1991, “Social Dilemmas, EconomicSelf-Interest, and Evolutionary Theory”, in Donald R. Brown andJ.E. Keith Smith (eds.),Recent Research in Psychology: Frontiersof Mathematical Psychology: Essays in Honor of Clyde Coombs, NewYork: Springer-Verlag. doi:10.1007/978-1-4612-3088-5_2
• Deutscher, Irwin, 1973, “What”, in hisWhat WeSay/What We Do: Sentiments & Acts, Glenview, IL: ScottForesman.
• Dhami, Sanjit S., 2016,The Foundations of Behavioral EconomicAnalysis, New York: Oxford University Press.
• Diamond, A.S., 1935,Primitive Law, London: Watts.
• Dufwenberg, Martin and Georg Kirchsteiger, 2004, “A Theoryof Sequential Reciprocity”,Games and EconomicBehavior, 47(2): 268–298.doi:10.1016/j.geb.2003.06.003
• Durkheim, Émile, 1895 [1982],Les Règles de la méthodesociologique, Paris. Translated asThe Rules of SociologicalMethod, W. D. Hall (trans.), Glencoe, IL: The Free Press.
• –––, 1950 [1957],Leçons de Sociologie:Physique des Moeurs et du Droit, Istanbul: Cituri BiraderletBasimevi. Translated asProfessional Ethics and Civic Morals,Cornelia Brookfield (trans.), Glencoe, IL: The Free Press.
• Ellickson, Robert C., 1991,Order Without Law: How NeighborsSettle Disputes, Cambridge, MA: Harvard University Press.
• Ellingsen, Tore, Magnus Johannesson, Johanna Mollerstrom, and SaraMunkhammar, 2012, “Social Framing Effects: Preferences orBeliefs?”,Games and Economic Behavior, 76(1):117–130. doi:10.1016/j.geb.2012.05.007
• Elster, Jon, 1989a,The Cement of Society: A Study of SocialOrder, Cambridge: Cambridge University Press.doi:10.1017/CBO9780511624995
• –––, 1989b, “Social Norms and EconomicTheory”Journal of Economic Perspectives, 3(4):99–117. doi:10.1257/jep.3.4.99
• Falk, Armin and Urs Fischbacher, 2006, “A Theory ofReciprocity”,Games and Economic Behavior, 54(2):293–315. doi:10.1016/j.geb.2005.03.001
• Fehr, Ernst and Klaus M. Schmidt, 1999, “A Theory ofFairness, Competition, and Cooperation”,Quarterly Journalof Economics, 114(3): 817–868.doi:10.1162/003355399556151
• –––, 2006, “The Economics of Fairness,Reciprocity and Altruism—Experimental Evidence and NewTheories”, in Serge-Christophe Kolm and Jean Mercier Ythier(eds.),Handbook of The Economics of Giving, Altruism andReciprocity (Volume 1), Amsterdam: North-Holland/Elsevier, pp.615–691. doi:10.1016/S1574-0714(06)01008-6
• Fishbein, Martin E., 1967, “A Consideration of Beliefs andTheir Role in Attitude Measurement”, inReadings in AttitudeTheory and Measurement, Martin E. Fishbein (ed.), New York:Wiley.
• Freeman, Linton C. and Türröz Ataöv, 1960, “Invalidity ofIndirect and Direct Measures of Attitude Toward Cheating”,Journal of Personality, 28(4): 443–447.doi:10.1111/j.1467-6494.1960.tb01631.x
• Gächter, Simon, Daniele Nosenzo, and Martin Sefton, 2013,“Peer Effects in Pro-Social Behavior: Social Norms or SocialPreferences?”,Journal of The European EconomicAssociation, 11(3): 548–573. doi:10.1111/jeea.12015
• Gaertner, Samuel L. and John F. Dovidio, 1986, “The AversiveForm of Racism”, in John F. Dovidio and Samuel L. Gaertner(eds.),Prejudice, Discrimination, and Racism: Theory andResearch, Orlando, FL: Academic Press, pp. 61–89.
• Geanakoplos, John, David Pearce, and Ennio Stacchetti, 1989,“Psychological Games and Sequential Rationality”,Games and Economic Behavior, 1(1): 60–79.doi:10.1016/0899-8256(89)90005-5
• Geertz, Clifford, 1973, “Thick Description: Toward anInterpretive Theory of Culture”, inThe Interpretation ofCultures: Selected Essays, New York: Basic Books, pp.3–30.
• Gintis, Herbert, 2000,Game Theory Evolving, Princeton:Princeton University Press.
• Granovetter, Mark, 1985, “Economic Action and SocialStructure: the Problem of Embeddedness”,American Journal ofSociology, 91(3): 481–510. doi:10.1086/228311
• Guala, Francesco, 2016,Understanding Institutions: TheScience and Philosophy of Living Together, Princeton, NJ:Princeton University Press.
• Hamilton W.D., 1964, “The Genetical Evolution of SocialBehaviour I and II”,Journal of Theoretical Biology, 7:1–16 and 17–52. doi:10.1016/0022-5193(64)90038-4 and10.1016/0022-5193(64)90039-6
• Hardin, Russell, 1982,Collective Action, New York:Resources for the Future.
• Harding, John, Harold Proshansky, Bernard Kutner, and IsidorChein, 1954 [1969], “Prejudice and Ethnic Relations”, inGardner Lindzey (ed.),Handbook of Social Psychology, volume2, Reading, MA: Addison-Wesley, pp. 1021–1061. Reprinted inGardner Lindzey and Elliot Aronson (eds.),Handbook of SocialPsychology, second edition, volume 5, Reading, MA: AddisonWesley, 1969, pp. 1–76.
• Hausman, Daniel M., 2008, “Fairness and Social Norms”Philosophy of Science, 75(5): 850–860.doi:10.1086/594529
• Hechter, Michael and Karl-Dieter Opp, 2001,Social Norms,New York: Russel Sage Foundation.
• Henrich, Joseph and Robert Boyd, 2001, “Why People PunishDefectors: Weak Conformist Transmission Can Stabilize CostlyEnforcement of Norms in Cooperative Dilemmas”,Journal ofTheoretical Biology, 208(1): 79–89.doi:10.1006/jtbi.2000.2202
• Henrich, Joseph, Robert Boyd, Samuel Bowles, Colin Camerer, ErnstFehr, Herbert Gintis, and Richard McElreath, 2001, “In Search ofHomo Economicus: Behavioral Experiments in 15 Small-ScaleSocieties”,American Economic Review, 91(2):73–78. doi:10.1257/aer.91.2.73
• Hirshmann, Albert O., 1982,Shifting Involvements: PrivateInterest and Public Action, Princeton: Princeton UniversityPress.
• Hoebel, Adamson E., 1954,The Law of Primitive Man,Cambridge, MA: Atheneum.
• Hogg, Michael A. and John C. Turner, 1987, “Social Identityand Conformity: A Theory of Referent Informational Influence”,in William Doise and Serge Moscovici (eds.),Current Issues inEuropean Social Psychology, Volume 2, Cambridge: CambridgeUniversity Press, pp. 139–182.
• Homans, George Caspar, 1950,The Human Group, New York:Harcourt, Brace & Company.
• –––, 1961,Social Behavior, New York:Harcourt Brace and World.
• Insko, Chester A. and John Schopler, 1967, “TriadicConsistency: A Statement of Affective-Cognitive-ConativeConsistency”,Psychological Review, 74(5):361–376. doi:10.1037/h0020278
• Kagel, John H. and Alvin E. Roth, 2016,Handbook ofExperimental Economics, Volume 2, Princeton, NJ: PrincetonUniversity Press.
• Klassen, Albert D., Colin J. Williams, and Eugene E. Levitt, 1989,Sex and Morality in the U.S.: An Empirical Enquiry Under theAuspices of the Kinsey Institute, Middletown, CT: WesleyanUniversity Press.
• Kramer, Roderick M. and Marilynn B. Brewer, 1984, “Effectsof Group Identity on Resource Use in a Simulated CommonsDilemma”,Journal of Personality and Social Psychology,46(5): 1044–1057. doi:10.1037/0022-3514.46.5.1044
• Krupka, Erin L. and Roberto A. Weber, 2013, “IdentifyingSocial Norms Using Coordination Games: Why Does Dictator Game SharingVary?”,Journal of The European Economic Association,11(3): 495–524. doi:10.1111/jeea.12006
• LaPiere, Richard T., 1934, “Attitudes vs. Actions”,Social Forces, 13(2): 230–237. doi:10.2307/2570339
• Ledyard, John, 1995, “Public Goods Experiments”, inJohn H. Kagel and Alvin E. Roth (eds.),The Handbook ofExperimental Economics, Princeton, NJ: Princeton UniversityPress.
• Lewis, David, 1969,Convention: A Philosophical Study,Cambridge, MA: Harvard University Press.doi:10.1002/9780470693711
• –––, 1975, “Languages and Language”,inLanguage, Mind, and Knowledge, (Minnesota Studies inthe Philosophy of Science, 6), Keith Gunderson (ed.),Minneapolis: University of Minnesota Press, pp. 3–35. [Lewis 1975 available online]
• López-Pérez, Raúl, 2008, “Aversion to Norm-Breaking: AModel”,Games and Economic Behavior, 64(1):237–267. doi:10.1016/j.geb.2007.10.009
• Mackay, Charles, 1841,Extraordinary Popular Delusions and theMadness of Crowds, N. Stone (ed.), Hertfordshire: Wordsworth,1995.
• Mackie, G., 1996, “Ending Footbinding and Infibulation: AConvention Account”,American Sociological Review,61(6): 999–1017. doi:10.2307/2096305
• Matza, David, 1964,Delinquency and Drift, New York: Wiley.
• Maynard Smith, J. and G. R. Price, 1973, “The Logic ofAnimal Conflict”,Nature, 246(5427): 15–18.doi:10.1038/246015a0
• Miller, Dale T. and Rebecca K. Ratner, 1996, “The Power ofthe Myth of Self-Interest”, in Leo Montada and Melvin J. Lerner(eds.),Current Societal Concerns About Justice, New York:Plenum Press. doi:10.1007/978-1-4757-9927-9_3
• Nachbar, J.H., 1990, “‘Evolutionary’ SelectionDynamics in Games: Convergence and Limit Properties”,International Journal of Game Theory, 19(1): 59–89.doi:10.1007/BF01753708
• North, Douglass C., 1990, “A Transaction Cost Theory ofPolitics”Journal of Theoretical Politics, 2(4):355–367. doi:10.1177/0951692890002004001
• O’Gorman, Hubert J., 1975, “Pluralistic Ignorance andWhite Estimates of White Support for Racial Segregation”,Public Opinion Quarterly, 39(3): 313–330.doi:10.1086/268231
• Olson, Mancur, 1965 [1971],The Logic of Collective Action:Public Goods and the Theory of Groups, revised edition,Cambridge, MA: Harvard University Press.
• Ostrom, Elinor, 2000, “Collective Action and The Evolutionof Social Norms”,Journal of Economic Perspectives,14(3): 137–158. doi:10.1257/jep.14.3.137
• Pagel, Elaine, 2003,Beyond Belief: The Secret Gospel ofThomas, New York: Vintage Books.
• Parsons, Talcott, 1951,The Social System, New York:Routledge.
• –––, 1937 [1968],The Structure of SocialAction. A Study in Social Theory with Special Reference to a Group ofRecent European Writers, New York, London: Free Press.
• Parsons, Talcott and Edward A. Shils, 1951,Towards a GeneralTheory of Action, Cambridge, MA: Harvard University Press.
• Posner, Eric A., 2000,Law and Social Norms, Cambridge,MA: Harvard University Press.
• Prentice, Deborah A. and Dale T. Miller, 1993, “PluralisticIgnorance and Alcohol Use on Campus: Some Consequences ofMisperceiving the Social Norm”,Journal of Personality andSocial Psychology, 64(2): 243–56.doi:10.1037/0022-3514.64.2.243
• Rabin, Matthew, 1993, “Incorporating Fairness into GameTheory and Economics”,American Economic Review, 83(5):1281–1302.
• Reuben, Ernesto and Arno Riedl, 2013, “Enforcement ofContribution Norms in Public Good Games with HeterogeneousPopulations”,Games and Economic Behavior, 77(1):122–137. doi:10.1016/j.geb.2012.10.001
• Robinson, Claude E., 1932,Straw Votes: A Study of PoliticalPrediction, New York: Columbia University Press.
• Rommetveit, Ragnar, 1955,Social Norms and Roles: Explorationsin the Psychology of Enduring Social Pressures with EmpiricalContributions from Inquiries into Religious Attitudes and Sex Roles ofAdolescents from Some Districts in Western Norway, Oslo:Akedemisk Forlag.
• Schelling, Thomas C., 1960,The Strategy of Conflict,Cambridge, MA: Harvard University Press.
• –––, 1978,Micromotives andMacrobehavior, New York: Norton.
• Schram, Arthur and Gary Charness, 2015, “Inducing SocialNorms in Laboratory Allocation Choices”,ManagementScience, 61(7): 1531–1546. doi:10.1287/mnsc.2014.2073
• Scott, John Finley, 1971,Internalization of Norms: ASociological Theory of Moral Commitment, Englewood Cliffs, NJ:Prentice Hall.
• Sherif, Muzafer, 1936,The Psychology of Social Norms,New York: Harper.
• Skyrms, Brian, 1996,Evolution of the Social Contract,Cambridge: Cambridge University Press.doi:10.1017/CBO9780511806308
• –––, 2004,The Stag Hunt and the Evolutionof Social Structure, Cambridge: Cambridge University Press.doi:10.1017/CBO9781139165228
• Stigler, George J. and Gary S. Becker, 1977, “De GustibusNon Est Disputandum”,American Economic Review, 67(2):76–90.
• Sugden, Robert, 1986 [2004],The Economics of Rights,Co-operation and Welfare, second edition, Basingstoke: PalgraveMacmillan, 2004.
• –––, 2000, “The Motivating Power ofExpectations”, in Julian Nida-Rümelin and Wolfgang Spohn(eds.),Practical Rationality, Rules, and Structure,Dordrecht: Kluwer. doi:10.1007/978-94-015-9616-9_7
• Tajfel, Henri, 1973, “The Roots of Prejudice: CognitiveAspects”, inPsychology and Race, Peter Watson (ed.),Chicago: Aldine.
• –––, 1981,Human Groups and SocialCategories: Studies in Social Psychology, Cambridge: CambridgeUniversity Press.
• Taylor, Peter D. and Leo B. Jonker, 1978, “EvolutionaryStable Strategies and Game Dynamics”,MathematicalBiosciences, 40(1–2): 145–156.doi:10.1016/0025-5564(78)90077-9
• Thibaut, John W. and Harold H. Kelley, 1959,The SocialPsychology of Groups, New York: Wiley.
• Trivers, Robert L., 1971, “The Evolution of ReciprocalAltruism”,Quarterly Review of Biology, 46(1):35–57. doi:10.1086/406755
• –––, 1985,Social Evolution, MenloPark, CA: Benjamin/Cummings.
• Turnbull, C. M., 1972,The Mountain People, New York:Touchstone.
• Turner, John C., Michael A. Hogg, Penelope J. Oakes, Stephen D.Reicher, and Margaret S. Wetherell, 1987,Rediscovering the SocialGroup: A Self-Categorization Theory, Oxford: Blackwell.
• Ullmann-Margalit, Edna, 1977,The Emergence of Norms,Oxford: Clarendon Press.
• Vanderschraaf, Peter, 1995, “Convention as CorrelatedEquilibrium”,Erkenntnis, 42(1): 65–87.doi:10.1007/BF01666812
• Warner, Lyle G. and Melvin L. DeFleur, 1969, “Attitude as AnInteractional Concept: Social Constraint and Social Distance asIntervening Variables Between Attitudes and Action”,American Sociological Review, 34(2): 153–169.doi:10.2307/2092174
• Wicker, Allan W., 1969, “Attitude versus Actions: TheRelationship of Verbal and Overt Behavioral Responses to AttitudeObjects”,A Journal of Social Issues, 25(4):41–78. doi:10.1111/j.1540-4560.1969.tb00619.x
• Wrong, Dennis H., 1961, “The Oversocialized Conception ofMan in Modern Sociology”,American Sociological Review,26(2): 183–193. doi:10.2307/2089854
• Wuthnow, Robert, 1991,Acts of Compassion: Caring for Othersand Helping Ourselves, Princeton: Princeton UniversityPress.
• Xiao, Erte and Cristina Bicchieri, 2010, “When EqualityTrumps Reciprocity”,Journal of Economic Psychology,31(3): 456–470. doi:10.1016/j.joep.2010.02.001
• Young, H. Peyton, 1993, “The Evolution ofConventions”,Econometrica, 61(1): 57–84.doi:10.2307/2951778
• –––, 1998a, “Social Norms and EconomicWelfare”,European Economic Review, 42(3–5):821–830. doi:10.1016/S0014-2921(97)00138-4
• –––, 1998b,Individual Strategy and SocialStructure: An Evolutionary Theory of Institutions, Princeton, NJ:Princeton University Press.

Academic Tools

	How to cite this entry.
	Preview the PDF version of this entry at theFriends of the SEP Society.
	Look up this entry topic at theIndiana Philosophy Ontology Project (InPhO).
	Enhanced bibliography for this entryatPhilPapers, with links to its database.

Other Internet Resources

• Arló-Costa, Horacio and Arthur Paul Pedersen, “Social Norms, Rational Choice and Belief Change”, manuscript, Carnegie Mellon Research Showcase.
• Axelrod, Robert, 1992, “How to Promote Cooperation”,Current Contents—Social and Behavioral Sciences, 44:10 (November 2, 1992); this is a scan of the version that appeared inCurrent Contents—Arts and Humanities, 23: 16 (November9, 1992).
• Selected Papers on Social Norms atEconPapers.
• Social Norm, entry inWikipedia.

Related Entries

belief |common knowledge |convention |evolution: the concept of, to 1872 |game theory |game theory: evolutionary

Acknowledgments

Section 6 of this entry has been adapted with minor changes from“Game-Theoretic Accounts of Social Norms”, by CristinaBicchieri and Alessandro Sontuoso, inThe Handbook of ExperimentalGame Theory, Mónica Capra, Rachel Croson, Tanya Rosenblatt, andMary Rigdon (eds.), Cheltenham: Edward Elgar Publishing,forthcoming.