Who Decides What Counts as 'Deviant'?

In the early 1950s, a young sociologist named Howard Becker spent his evenings playing piano in Chicago dance bands and his days studying the people he played alongside. He noticed something the criminology textbooks of the era could not explain. The jazz musicians he knew thought of themselves as a world apart, set against the conventional "squares" who hired them, and they smoked marijuana as a matter of routine. By the letter of the law, every one of them was a criminal, yet almost none were ever arrested, labeled, or treated as deviant by the wider society. The behavior was constant; the label was not. Becker began to suspect that the label, not the behavior, was the thing worth studying.

That suspicion overturned a question people had been asking the wrong way for a long time. The older approach assumed that some acts are simply deviant by their nature, and that the social scientist's job was to find out what was wrong with the people who committed them. Becker's insight was that this gets the problem backward: if the same act can be a crime in one setting and unremarkable in another, then the deviance cannot live inside the act; it must live in the social process that responds to it. So who, exactly, decides what counts as deviant, and how does that decision come to stick?

The Move That Changed the Question

The conventional view treated deviance as a property of behavior, the way redness is a property of an apple, so that one could in principle line up all human acts and sort them into the deviant and the non-deviant. Becker's contribution, in his 1963 book Outsiders, was to deny that such a sorting is possible at all, because the very same act lands in different bins depending on the circumstances.

His central claim is worth stating precisely, because it is more radical than it first sounds. Deviance, he argued, is not a quality of the act a person commits. It is the outcome of a social process in which people who have the power to apply labels successfully call a behavior deviant and make that label stick. This is the heart of what came to be called labeling theory. The deviant is not someone who has done a particular thing but someone to whom the label has been applied. Becker put it sharply: deviant behavior is behavior that people so label.

Notice what this does to the analytical object. The interesting variable is no longer the conduct of the rule-breaker but the variation in how identical conduct is treated. A man drinking heavily at a wedding and a man drinking heavily on a park bench are doing the same thing to their bodies, yet one is celebrating and the other is a problem to be managed. That gap, between the act and its social reception, is what labeling theory asks us to explain. It does not deny that people break rules; it insists that breaking a rule and being made into a deviant are two different events, and the second does not follow automatically from the first.

How a Label Builds a Career

If labeling theory only said that society reacts unevenly to misbehavior, it would be a modest observation. Its sharper edge comes from a claim about consequences, worked out by the sociologist Edwin Lemert, who drew a distinction between two kinds of rule-breaking that are easy to confuse and important to keep apart.

The first he called primary deviance, the initial act of breaking a rule. Primary deviance is often fleeting and frequently goes unremarked. A teenager shoplifts once, a respectable person cheats on their taxes, a student tries a drug at a party. The act happens, and in most cases nothing follows from it; the person continues to think of themselves as fundamentally law-abiding, and so does everyone else. The second kind, secondary deviance, is the behavior that develops after a person has been caught, labeled, and had the label stick. Once someone is publicly marked as a thief, an addict, or a delinquent, that identity begins to organize how others treat them and, in time, how they see themselves.

The argument is that the label does real work. A young person tagged as a delinquent may find that schools, employers, and neighbors now respond through that category, closing off the paths back to a conventional life and leaving the deviant role as one of the few available. The person settles into that role, and what looked like proof that the label was accurate is in fact partly an effect of the label itself. Lemert's point is that the deviant career is produced by the labeling, not the original act alone. The act may have been trivial; the machinery set in motion by the response to it is what makes a deviant.

A Simple Table That Pulls Two Things Apart

Becker captured the logic in a small two-by-two table worth keeping in mind, because it forces a distinction everyday language constantly blurs. One axis asks whether a person has actually broken a rule. The other asks whether the person has been labeled as deviant. Crossing these two yes-or-no questions gives four cells.

Someone who has neither broken a rule nor been labeled is simply conforming. Someone who has broken a rule and been labeled is what Becker called the pure deviant, the case ordinary thinking treats as the only kind there is. But the two remaining cells are where the framework earns its keep. The secret deviant has broken a rule yet escaped the label entirely, like Becker's marijuana-smoking musicians or the countless people who quietly break norms without consequence. The falsely accused has broken no rule yet has been labeled anyway, the innocent person convicted, the dissident smeared, the patient misdiagnosed.

The power of the table is that the secret deviant and the falsely accused are impossible to describe at all if you assume behavior and label are the same thing. By laying them out as separate axes, Becker makes visible that the two are analytically distinct, that one can occur without the other, and that the connection between them is not a fact of nature but a social outcome that has to be explained.

Four Ways Sociology Asks the Same Question

Labeling theory is one of several major frameworks sociologists use to analyze deviance, and it helps to see it against the others, because each begins from a different question and arrives at a different account. The point is not that one is correct and the rest mistaken; each illuminates a different facet of the same phenomenon.

Robert Merton's strain theory starts from the gap between the goals a society tells people to pursue, such as wealth and success, and the legitimate means it makes available to reach them. When the means are blocked, some people innovate by reaching approved goals through disapproved channels, and deviance becomes a response to structural pressure. Travis Hirschi's control theory inverts the question, asking not why people break rules but why most do not, and answers that strong bonds to family, school, and community hold conformity in place, so deviance appears where those bonds are weak. Edwin Sutherland's differential association locates the source in learning, holding that deviant behavior, including its techniques and justifying attitudes, is learned in intimate groups much as any other behavior is. Labeling theory sits alongside these three, asking not what drives a person to break a rule but what happens afterward and who has the power to make it count.

When the Label Makes the Patient

The most provocative application of labeling theory reached into psychiatry. In 1966, the sociologist Thomas Scheff argued that the diagnostic label of mental illness often produces the patient's career rather than simply reflecting an underlying condition. On his account, many people display unusual behavior at one time or another, most of it transient and ignored, until someone is caught up in the psychiatric system and labeled, after which the role of mental patient begins to shape their conduct and identity in the way Lemert described for deviance generally.

This was a strong claim, and the discipline has refined it rather than swallowing it whole. The mature position holds two things at once that are easy to mistake as contradictory. Mental illness has real dimensions; serious conditions involve genuine suffering and measurable disturbance, and they are not merely inventions of the people who diagnose them. At the same time, diagnostic labels still do social work, changing how a person is treated and how they come to understand their own experience. Labeling theory was wrong to suggest the condition is nothing but the label, but right that the label is never socially weightless.

A Single Generation, a Reversed Verdict

Nowhere is labeling theory's central prediction tested more clearly than in how quickly the catalog of deviance can change. Consider same-sex relationships. The philosopher Michel Foucault traced how, in the nineteenth century, the homosexual was constituted as a distinct kind of person and a deviant category, written into law and medicine alike. The same conduct that was a criminal offense in many jurisdictions within living memory is now, in most of those same places, the recognized basis of marriage. The behavior did not change; the labeling did.

That reversal is not an isolated case. Cannabis use, visible tattoos, and same-sex marriage have all moved from the deviant column toward the unremarkable one within a single generation across much of the Western world. Labeling theory predicts exactly this kind of transition, because it locates deviance in the labeling rather than the behavior, and labels are the work of particular people in particular institutions and can be withdrawn as readily as imposed. The same logic explains the selective application Becker noticed among the jazz musicians: identical behavior generates different labels depending on who is doing it, where, and in what context. Contemporary drug-enforcement data make the pattern stark, with arrest rates for the same conduct falling unevenly across neighborhoods and groups, just as a theory centered on the labeler, not the act, would lead you to expect.

What the Framework Sees and What It Misses

Labeling theory opened analytical space that the older individual-pathology model had foreclosed. By moving the question from the rule-breaker to the rule-enforcer, it made it possible to study power, selective enforcement, and the manufacture of deviant identities, all of which the earlier framework could not see.

Its limits become visible, though, when the framework is pressed against behaviors that produce serious harm regardless of how they are labeled. A killing remains a catastrophe for its victim whether or not the killer is ever named a murderer, and a tradition that locates all the action in labeling can slide toward implying that the harm is secondary to the social reaction. It is not. The mature discipline holds both insights together, recognizing that some conduct genuinely wounds while the route from conduct to the deviant label still runs through institutions and power. The most useful way to see this is to put several framings on a single case at once, since a labeling theorist, a Mertonian strain theorist, and a critical-conflict theorist will each frame the question differently, and each lights up an aspect the others leave in the dark.

Key Takeaways

Howard Becker's labeling theory relocates deviance from the act to the social process that names and sanctions it, arguing that deviance is not a quality of behavior but the outcome of people with the power to apply labels making them stick, which turns the uneven treatment of identical conduct into the central thing to be explained; Edwin Lemert's distinction between primary deviance, the often unremarked initial act, and secondary deviance, the role that develops once a label has taken hold, shows how the labeling itself produces the deviant career, while Becker's four-cell table, crossing rule-breaking with labeling to yield the conforming, the pure deviant, the secret deviant, and the falsely accused, pulls behavior and label apart so they can be studied separately; applied to Thomas Scheff's account of mental illness and to the rapid destigmatization of cannabis, tattoos, and same-sex marriage within a single generation, the framework predicts these reversals because labels can be withdrawn as readily as imposed, yet its limits surface against conduct that harms regardless of labeling, so the mature discipline keeps it alongside Merton's strain theory, Hirschi's control theory, and Sutherland's differential association, holding the insight that power shapes the label together with the insight that some acts truly cause harm.