Deviance is behavior that violates the social norms of a group, institution, or society. In sociology, the term also encompasses the processes through which conduct, identities, and personal attributes come to be defined and treated as unacceptable. It is an analytical category rather than a synonym for wrongdoing: an action may be considered deviant without causing harm, while harmful practices may be customary, legally permitted, or institutionally protected.
Deviance is relational and historically variable. Its identification depends on which expectations apply, who interprets them, and what authority those interpreters possess. The same conduct can be celebrated in one setting, tolerated in another, and punished in a third. Sociological explanations therefore examine both departures from expectations and the creation, enforcement, and transformation of those expectations.
The study of deviance connects questions about social order with questions about inequality and change. Its principal subjects include rule-breaking, punishment, stigma, informal sanctions, institutional misconduct, and the emergence of alternative ways of life. Its methods range from observation of everyday encounters to longitudinal studies of offending and comparative analysis of legal and cultural change.
Boundaries of the concept
Deviance is distinct from statistical rarity. A rare linguistic ability or an unusually high income need not violate any norm, whereas widespread practices such as minor traffic violations may remain normatively prohibited. Nor is deviance identical to personal eccentricity. An unusual preference becomes sociologically relevant as deviance when it conflicts with expectations and attracts, or risks attracting, social disapproval. Identifying a norm requires evidence about expectations, not merely the observation that most people behave in a particular way.
Deviance and crime overlap but are not interchangeable. Crime consists of conduct prohibited by criminal law within a particular jurisdiction. Deviance includes many noncriminal departures from conventions, such as breaches of etiquette or occupational expectations. Conversely, a prohibited act may receive little informal condemnation within a particular group. Workplace practices, for example, can violate external regulations while being treated internally as ordinary or even necessary. The relationship between legality, acceptability, and harm must therefore be investigated rather than assumed.
Norms operate at several levels. A profession may impose obligations more demanding than those of the wider society, while a peer group may reward conduct that schools or employers condemn. A participant can consequently be conformist in one reference group and deviant in another. Subcultures are especially important because they may provide alternative standards, vocabularies, and sources of recognition. They are not inherently deviant, however; many organize distinctive tastes without substantially conflicting with wider rules.
Researchers also distinguish the behavior from the response to it. A concealed violation may occur without public labeling, while a false accusation may produce sanctions despite the absence of the alleged act. Some approaches define deviance primarily through norm violation; others emphasize successful designation by an audience. These approaches direct attention to different objects and can yield different findings. A study counting prohibited acts is not necessarily measuring the same phenomenon as one counting disciplinary cases.
“Positive deviance” describes departures from expectations that are regarded as admirable or that produce unusually successful outcomes under comparable constraints. The term is used in organizational and applied research, although its meaning differs from the traditional emphasis on condemnation. Exceptional generosity can depart from an expected level of obligation without becoming a stigmatized identity. Such cases reveal why rarity, norm violation, moral judgment, and social reaction are best treated as distinguishable dimensions.
Social order and the functionalist tradition
Early sociological accounts challenged the assumption that deviance was simply a defect in otherwise harmonious societies. In The Rules of Sociological Method (1895), Émile Durkheim argued that crime was normal in the sociological sense: it appeared across societies and was connected to the conditions of collective life. “Normal” did not mean morally desirable, nor did it imply that every level or form of crime was harmless. It meant that the complete elimination of norm violation was not a realistic description of social organization.
Durkheim’s reasoning rested partly on the variability of moral boundaries. Even a community composed of exceptionally conscientious people would identify comparatively minor departures as serious offenses. Shared expectations could become more exacting rather than eliminate transgression. Punitive responses, in his account, expressed and renewed collective sentiments. Public condemnation showed participants where a moral boundary lay and reaffirmed their membership in a community organized around it.
This argument became central to structural functionalism. Deviance could clarify rules, strengthen solidarity among those condemning it, and reveal pressures for institutional change. A challenge initially treated as unacceptable might expose a mismatch between established expectations and changing conditions. These are possible social consequences, not evidence that every deviant act benefits society. Violence, exploitation, and exclusion can produce lasting damage, while condemnation can unite one group through the mistreatment of another.
Kai Erikson developed the boundary-maintenance approach in Wayward Puritans (1966), a historical study of Massachusetts Bay. He examined how public episodes of deviance helped a community articulate its identity. The analytical focus shifted from the characteristics of isolated offenders to the relationship between collective crises, enforcement, and the symbolic boundaries of membership. This perspective also encouraged attention to ceremonies of accusation, expulsion, and punishment.
Functional explanations face a methodological difficulty: a consequence does not automatically explain why a practice originated or persisted. Showing that punishment produces solidarity does not establish that solidarity caused the punishment. Communities also contain competing moral positions rather than a single consensus. Later work has therefore retained Durkheim’s insight that deviance and social order are interconnected while questioning explanations that presume a unified moral community or infer social necessity from institutional survival.
Anomie, strain, and unequal opportunities
Durkheim used anomie to describe conditions in which social regulation becomes weakened or disrupted. In Suicide (1897), he connected certain forms of suicide to disturbances in the regulation of aspirations, including disturbances associated with rapid economic change. Anomie was not simply individual confusion or an absence of all rules. It concerned the relationship between socially organized expectations, limits, and desires.
Robert K. Merton reformulated the idea in “Social Structure and Anomie” (1938). His argument concerned a discrepancy between culturally emphasized goals and unequal access to legitimate means of achieving them. In a society that strongly prizes monetary success, people may encounter pressure to attain that goal while facing restricted educational or occupational opportunities. The resulting strain can make prohibited means more attractive. This account locates pressure toward deviance within social structure, rather than attributing it solely to defective personalities.
Merton distinguished five adaptations. Conformity accepts both approved goals and institutionalized means. Innovation retains the goals while adopting prohibited means; fraud is a common illustration. Ritualism abandons or reduces commitment to dominant success goals while continuing to follow approved procedures. Retreatism rejects both goals and means, whereas rebellion seeks to replace them with alternatives. These are analytical patterns, not fixed personality types, and individuals may move among them or display different adaptations in different areas of life.
Subsequent theories addressed matters Merton left underdeveloped. Albert Cohen’s Delinquent Boys (1955) examined status frustration and the formation of delinquent group standards. Richard Cloward and Lloyd Ohlin’s Delinquency and Opportunity (1960) argued that access to illegitimate opportunities, like access to legitimate ones, is unequally distributed. Learning profitable theft or entering an organized illegal market requires contacts, skills, and opportunities; disadvantage alone does not supply them.
Robert Agnew’s general strain theory, introduced in a major formulation in 1992, broadened strain beyond blocked economic success. It included the loss of positively valued experiences and exposure to negative treatment. Emotions, especially anger, could connect strain to offending, but the outcome depended on coping resources, social relationships, and available alternatives. The theory does not predict that adversity invariably produces deviance.
Strain approaches are strongest when they specify mechanisms rather than equate poverty with criminality. Most disadvantaged people do not commit serious offenses, and privileged actors also engage in misconduct. Organizational competition, threatened status, or pressure to meet financial targets can generate strains among relatively advantaged groups. Empirical tests must therefore distinguish material deprivation, perceived injustice, aspirations, emotional responses, and actual opportunities.
Learning, social bonds, and opportunities for action
Learning theories explain deviance through processes that also produce conventional conduct. Edwin H. Sutherland’s differential association theory, formulated in Principles of Criminology in 1939 and revised in 1947, proposed that criminal behavior is learned through interaction, especially within intimate groups. Learning includes techniques and motives, as well as definitions favorable or unfavorable to violating the law. Associations differ in frequency, duration, priority, and intensity, so exposure is not reducible to a simple count of delinquent acquaintances.
Later social learning approaches, notably those associated with Ronald Akers, incorporated reinforcement and imitation. A behavior may become more likely when it produces approval, money, status, or relief from an unpleasant situation. Rewards and punishments can be anticipated as well as experienced. This perspective explains why deviance can be highly disciplined: participants may learn specialized skills, comply with internal rules, and punish unreliable associates even while violating external expectations.
Control theories reverse the question. Instead of asking why people acquire deviant motivations, they ask what ordinarily restrains behavior that offers an immediate advantage. Travis Hirschi’s Causes of Delinquency (1969) identified attachment, commitment, involvement, and belief as elements of a social bond. Concern for others’ opinions, investment in conventional achievements, participation in activities, and acceptance of rules can inhibit delinquency. Socialization matters here because it develops relationships and commitments, not merely knowledge of prohibitions.
These approaches also examine how people suspend moral restraints. Gresham Sykes and David Matza’s “Techniques of Neutralization” (1957) described justifications such as denying responsibility, denying injury, denying the victim, condemning condemners, and appealing to higher loyalties. A person may recognize a general rule yet claim that a particular violation is excusable. Such accounts help explain movement between conventional and prohibited conduct without assuming a complete rejection of prevailing morality.
Opportunity-centered theories focus more directly on situations. Lawrence Cohen and Marcus Felson’s routine activity approach (1979) linked predatory offenses to the convergence of a motivated offender, a suitable target, and the absence of capable guardianship. Changes in everyday routines can alter exposure even when individual motivations remain stable. This framework applies most directly to offenses requiring particular encounters, not to every form of deviance.
A recurring research problem is separating influence from selection. Similar friends may become more alike through interaction, but individuals also choose companions with similar dispositions. Likewise, weak school attachment may precede misconduct or result from exclusion following it. Longitudinal evidence is needed to distinguish these reciprocal processes.
Labeling, stigma, and deviant careers
Labeling theory places the social reaction to conduct at the center of analysis. Rooted in symbolic interactionism, it examines how meanings emerge through encounters and institutional decisions. Howard S. Becker’s Outsiders (1963) argued that deviance cannot be understood solely as a property of an act: rules are created and applied by social groups, and their application helps constitute the category of the outsider. The approach asks who can impose a definition, on whom, and with what consequences.
Edwin Lemert distinguished primary from secondary deviance in Social Pathology (1951). Primary deviance refers to departures that need not become central to a person’s identity or social position. Secondary deviance develops when reactions to those departures reorganize the person’s circumstances and self-understanding. Repeated exclusion, for instance, may reduce access to conventional employment or social networks and increase dependence on groups organized around the stigmatized activity. This is a possible sequence, not an inevitable progression.
Erving Goffman’s Stigma (1963) examined the management of socially discrediting attributes. Stigma is relational: an attribute becomes discrediting within particular expectations about identity. Goffman distinguished situations in which a stigmatized attribute is known or immediately apparent from those in which it can be concealed. The latter can involve decisions about disclosure, passing, and the management of information. Stigma therefore extends beyond behavioral violations to characteristics that others interpret as signs of an unacceptable identity.
The concept of a deviant career describes changing participation, identity, and social position over time. Becker’s studies of marijuana use showed how learning to interpret an experience and participating in a social setting could be integral to continued use. Career analysis does not assume steady escalation. Entry, intermittent participation, withdrawal, and re-entry can all occur, and institutional encounters may redirect rather than determine a trajectory.
Labeling theory is sometimes criticized for underexplaining initial behavior or minimizing serious harm. Its strongest formulations do not require denying that acts occur before official intervention. They distinguish the causes of an act from the causes of its detection, classification, and consequences. Research can investigate whether a sanction deters later misconduct, strengthens a stigmatized identity, restricts opportunities, or combines these effects.
Empirical analysis must also avoid assuming that every label is equally powerful. A private insult, an occupational blacklist, and a criminal conviction have different institutional reach. The durability of labeling depends on records, audiences, material dependence, and opportunities to contest or escape the designation.
Power, inequality, and the production of rules
Conflict approaches question the assumption that rules simply express a shared moral consensus. They examine how groups with unequal resources influence legislation, enforcement priorities, and public definitions of threat. This does not imply that every rule serves only a dominant group or that all prohibited acts are harmless. It directs attention to whose interests receive protection, whose injuries become visible, and whose conduct attracts surveillance.
Sutherland’s work on white-collar crime, presented in his 1939 presidential address to the American Sociological Society and developed in his 1949 book, challenged explanations centered on poverty and street offending. His definition emphasized crimes committed by persons of respectability and high social status in the course of their occupations. The concept helped expose how occupational misconduct could be handled through regulatory or civil processes and consequently remain less visible in conventional criminal statistics.
The category of white-collar crime is not identical to corporate crime. An employee may offend for personal benefit against an employer, while corporate wrongdoing may be undertaken to advance organizational objectives. Collective decision-making, divided responsibilities, and technical complexity can make accountability difficult to establish. Internal language may recast a violation as routine practice or a commercial necessity, reducing its visibility to participants as well as outsiders.
Feminist scholarship expanded the field by examining gendered expectations, the neglect of victimization, and the treatment of women within criminal justice and other institutions. It questioned theories derived mainly from male samples and investigated why similar conduct could carry different consequences depending on gender. Later work emphasized the interaction of gender with class, racial classification, sexuality, and other social positions, rather than treating women or men as homogeneous categories.
Power also shapes the distinction between misconduct and legitimate dissent. Civil disobedience deliberately challenges legal rules while appealing to principles presented as more authoritative. The sociological classification of such conduct as norm-breaking does not resolve its legal or ethical justification. Similarly, describing a prohibition as socially constructed does not establish that it is arbitrary: rules may respond to demonstrable injuries while remaining historically produced and unevenly enforced.
Comparative research therefore separates several questions: what conduct causes harm, what conduct is prohibited, which violations are detected, and which actors are punished. These distributions need not coincide. Commitments to human rights, due process, and equality before the law provide institutional standards against which the treatment of alleged violators can be assessed without assuming that accusation establishes culpability.
Institutions, sanctions, and the organization of conformity
Social control includes the processes through which conduct is regulated and expectations are maintained. Informal controls include ridicule, gossip, withdrawal of trust, praise, and exclusion from social interaction. Formal controls involve authorized procedures such as school discipline, professional licensing, employment sanctions, and criminal punishment. The distinction concerns institutional organization, not necessarily severity: informal exclusion can have profound consequences even without an official judgment.
Control is frequently anticipatory. People adjust their behavior because they expect scrutiny, wish to retain valued relationships, or have internalized standards. Families, schools, religious organizations, workplaces, and peer groups communicate expectations through everyday routines. These institutions may reinforce one another, but they may also issue conflicting demands. An employee expected to achieve an unrealistic sales target, for example, may face tension between organizational rewards and professional obligations.
Michel Foucault’s Discipline and Punish (1975) examined the historical development of disciplinary power. Its central subjects included surveillance, examination, documentation, and the organization of bodies in institutions. Drawing on Jeremy Bentham’s panopticon design, Foucault described how the possibility of observation could encourage self-monitoring. His account concerns more than punishment after an offense: it asks how institutional arrangements produce standards of normal performance and make individuals comparable.
Organizations also generate distinctive forms of deviance. Diane Vaughan’s The Challenger Launch Decision (1996) analyzed how the acceptance of anomalies could become routine within organizational decision-making. The associated concept of normalization of deviance describes a process in which departures from expected safety margins become accepted because earlier departures did not produce an immediately visible disaster. It need not involve a conscious intention to cause harm or a simple decision to ignore a known, fully understood danger.
Sanctions can have multiple and conflicting effects. Exclusion may protect a group from an immediate threat while disrupting education, employment, or supportive relationships. Highly punitive environments may deter some acts while encouraging concealment or discouraging the reporting of mistakes. Conversely, tolerance of misconduct can expose others to continuing harm. Institutional analysis examines these mechanisms rather than treating either punishment or leniency as uniformly effective.
The legitimacy and procedural form of control also matter. Clear rules, consistent application, opportunities to be heard, and review mechanisms affect how disciplinary decisions are experienced and contested. A sanction is consequently not just a quantity of deprivation; it is an event embedded in relationships of authority, recognition, and institutional trust.
Measurement, evidence, and research ethics
Measuring deviance requires an explicit definition of the norm, the relevant population, the time period, and the behavior or reaction being counted. Police records measure incidents and persons entering administrative systems, not all prohibited conduct. School suspension figures measure disciplinary decisions as well as student behavior. Differences across places or periods may reflect altered reporting practices, enforcement priorities, legal classifications, or recording systems rather than equivalent changes in underlying conduct.
Researchers commonly distinguish incidence from prevalence. Incidence concerns the number of events occurring during a specified period; prevalence concerns the proportion of people or other units involved at least once. A small number of repeat participants can account for many incidents, so these measures answer different questions. Studies must also distinguish allegations, arrests, charges, convictions, and sanctions. These are stages of institutional processing rather than interchangeable indicators.
Victimization surveys reveal some experiences that never reach official records. They are especially useful for offenses for which respondents can identify an event and report it meaningfully. Their limits include recall error, nonresponse, survey coverage, and difficulties measuring conduct without an identifiable individual victim. Self-report studies can capture concealed behavior, including minor violations, but are affected by question wording, willingness to disclose, memory, and the sampling of populations that may exclude highly marginalized groups.
Ethnography, interviews, and participant observation examine meanings and practices that standardized measures can miss. They can show how participants learn rules, negotiate identities, and interpret enforcement. Such methods are particularly useful for understanding hidden settings and the difference between official accounts and everyday routines. They do not automatically establish how common a practice is across a large population, making the relationship between detailed description and generalization an important methodological question.
Causal inference presents additional difficulties. Observed associations between punishment and later offending can reflect differences that existed before punishment. Associations between neighborhood disadvantage and crime may involve residential selection, institutional resources, exposure to violence, or policing patterns. Longitudinal studies, natural experiments, and carefully designed comparisons can help distinguish mechanisms, although each depends on assumptions that require scrutiny.
Research ethics are especially demanding where disclosure can expose participants to prosecution, dismissal, retaliation, or stigma. Consent, confidentiality, secure data handling, and careful presentation of identifying details are central concerns. Researchers must also avoid treating official categories as neutral descriptions or reproducing stereotypes through selective sampling. The same dataset can illuminate misconduct and reveal how institutions make some misconduct more visible than others.
Historical change, public controversies, and digital settings
Definitions of deviance change through legal reform, social movements, scientific reclassification, and everyday shifts in practice. These processes do not necessarily move together. Conduct may become legal while remaining stigmatized, or become socially tolerated before its legal status changes. Public acceptance also varies across institutions and groups. Change therefore involves multiple boundaries rather than a single transition from deviant to normal.
The treatment of homosexuality illustrates the distinction between psychiatric classification, legality, and social acceptance. In 1973, the American Psychiatric Association voted to remove homosexuality as a diagnostic category from its classification system; a residual category concerning distress about sexual orientation remained until 1987. The World Health Assembly adopted the tenth revision of the International Classification of Diseases in 1990, in which homosexuality itself was not classified as a disorder. These institutional changes did not by themselves eliminate discrimination or produce uniform legal change. Historical analysis concerns the classification process, not an assumption that sexual orientation is intrinsically deviant.
Moral panic describes an intensified public response in which a condition or group is presented as a threat to social values. Stanley Cohen’s Folk Devils and Moral Panics (1972) developed the concept through the treatment of British youth groups known as Mods and Rockers. The approach examines interactions among news coverage, public concern, political claims, and enforcement. Its use requires evidence: the existence of strong concern does not alone establish exaggeration, and a socially amplified threat may still involve real injuries.
Digital environments alter both norm formation and enforcement. Online communities develop local expectations about speech, attribution, participation, and identity. Platform rules operate alongside national law and informal user sanctions. Moderation can involve automated detection, user reports, human review, and appeals, producing a layered system in which the same content may receive different treatment across services or jurisdictions. Algorithms can influence visibility without making a formal legal judgment.
Digital traces also change the duration and audience of labeling. Material created for one setting can circulate into another, detached from its original context. Searchable records and copied content may preserve an accusation or violation long after the original interaction. At the same time, online networks can enable people previously stigmatized in their local surroundings to find recognition, organize collectively, and challenge prevailing classifications.
These settings create methodological problems as well as new evidence. Removed content is often inaccessible to independent researchers, platform users are not representative of whole populations, and recorded engagement does not directly measure agreement. Research on digital deviance must distinguish prohibited behavior, harmful effects, moderation decisions, and public visibility rather than treating them as a single phenomenon.