Tarde's Theory of Imitation and Differential Identification Theory

Gabriel Tarde was a French magistrate and sociologist writing in the last decade of the nineteenth century, at a time when the dominant criminological explanation for offending was biological. Cesare Lombroso's argument that the criminal was an atavistic throwback with identifiable physical stigmata had wide currency across Europe, and Tarde, who had spent years presiding over criminal cases before turning to sociology, rejected it on both empirical and theoretical grounds.

His counter-claim, set out fully in Les lois de l'imitation (The Laws of Imitation), published in 1890, was that society itself is nothing more than a web of imitations, and that crime is simply one more social product transmitted the way fashion, religion, or language is transmitted.

For Tarde, every act a person performs that is not a pure invention is a copy of something someone else did first. Invention is rare; imitation is the ordinary mechanism by which any invented practice, useful or harmful, spreads through a population.

A new way of committing fraud, a new weapon, a new criminal argot, all begin as one person's innovation and then diffuse outward as others copy it, exactly as a new dance step or a new manner of dress diffuses. Crime therefore requires no separate biological or psychological theory of its own; it is subject to the same laws that govern the transmission of any other social practice.

This move had two consequences for criminology. First, it shifted attention away from the individual constitution of the offender and toward the social channels through which criminal practices travel, anticipating the ecological and cultural-transmission theories that followed in the twentieth century.

Second, it made crime statistics interpretable as evidence of imitation in action: Tarde pointed to the way particular methods of theft or particular styles of violence would appear in one region and then reappear in adjoining regions in a pattern that tracked communication and travel routes rather than any shared biology among the populations involved.

Tarde's first law states that imitation is more intense and more frequent where social contact is close, whether that closeness is physical proximity or the density of social interaction. People copy those they see and interact with regularly far more readily than they copy strangers, so imitation follows the actual geography and social structure of contact rather than spreading evenly and randomly through a population.

Tarde used this law to explain the marked urban-rural pattern in nineteenth-century French crime statistics. Cities, with their dense concentration of people and constant novel encounters, generated far more new criminal methods, and those methods then radiated outward along roads, railways, and trade routes into surrounding rural areas at a rate roughly proportional to the traffic between city and countryside.

He also distinguished contact that is genuinely close, meaning frequent and psychologically significant, from contact that is merely physical proximity without real interaction, since imitation depends on the former rather than the latter.

Writing at a time when the railway and the newspaper were rapidly multiplying the points of contact between people who would otherwise never have encountered one another's habits, Tarde expected any future improvement in communication to accelerate the same imitative process rather than to change its underlying logic.

The law also carries an implicit claim about density: a crowded prison wing, a tightly knit gang, or a small trading town where everyone knows everyone else is a higher-imitation environment than a sparse and anonymous one, because the opportunities to observe and copy another person's conduct are simply more numerous.

This is the same intuition that later underpins ecological and subcultural accounts of crime concentration in specific neighbourhoods, though those later theories add mechanisms Tarde did not specify, such as differential opportunity and social disorganisation.

Tarde's third law describes what happens once a new imitation meets an existing custom. A newly introduced practice does not simply appear alongside the old one indefinitely; it is inserted into the existing fabric of habits, and the two either merge into a hybrid practice or compete until one displaces the other. Tarde described this as imitation proceeding logically, since a population tends to resolve the contradiction between an old custom and a new alternative rather than tolerating it forever.

Applied to crime, the law of insertion explains why a new criminal technique, once introduced into a region or a criminal milieu, rarely stays confined to a curiosity practiced by its originator. It is inserted into the existing repertoire of criminal methods, and if it proves more effective, more profitable, or harder to detect than the older method it resembles, it gradually replaces that older method rather than merely adding to the total volume of crime.

A new method of counterfeiting a currency, for instance, does not simply sit beside the old method; forgers converge on whichever technique works better, and the older technique falls out of use. Tarde treated this displacement as a general logical process governed by consistency rather than by any single cause, meaning a population will tend to resolve a contradiction between two competing practices even when no authority intervenes to settle it, simply because holding both indefinitely is socially inefficient.

The mechanism also captures how legitimate and illegitimate practices interact. A new item of legitimate technology, once it diffuses widely enough, creates the opportunity for a corresponding criminal adaptation that is then inserted into existing criminal practice; the older method adapted to the previous technology fades as the newer opportunity becomes the more efficient one to imitate. Tarde treated this displacement dynamic as a general logical process, not one confined to any one type of offence.

Tarde's second law holds that imitation flows downward through the social hierarchy: people copy those they regard as socially superior, whether the superiority is one of class, wealth, urban sophistication, or visible status, rather than the reverse. Fashion in dress, in speech, and in manners spreads from the socially prominent down to those who look up to them, and Tarde argued that vice and crime follow exactly the same downward path.

On this account, practices associated with the upper classes or with fashionable city life, including forms of fraud, gambling, and violent affrays, would first appear among the socially prominent and then be copied by those of lower status who imitated their superiors' visible conduct without necessarily sharing their resources or their impunity.

Tarde also noted a geographic version of the same law, since the capital and its provincial imitators followed the same downward diffusion pattern he attributed to social class, with the metropolis setting fashions, criminal and otherwise, that provincial centres subsequently adopted.

This is the aspect of Tarde's theory most clearly overtaken by later research. Twentieth-century criminology, particularly through subcultural and strain traditions, documented crime patterns that move in the opposite direction, upward and laterally through peer networks rather than only downward from a socially superior source, and showed that criminal innovation is at least as often a lower-status adaptation to blocked opportunity as it is an imitation of elite vice.

The law of imitation of superiors by inferiors is best read today as identifying one real diffusion channel among several, rather than as a complete account of how criminal practices move through a status hierarchy. It remains useful mainly as a reminder that status and visibility, not just proximity, shape which behaviours get copied at all, a point later picked up by theories of media influence and celebrity modelling that Tarde could only sketch in outline.

Edwin Sutherland, developing differential association theory through successive editions of his criminology textbook from 1939 onward, accepted Tarde's central claim that criminal behaviour is learned rather than inherited, but rejected imitation as the operative mechanism because he judged it too vague to explain the specific pattern of who becomes criminal and who does not.

Imitation, in Tarde's broad sociological sense, covers almost any instance of one person's conduct influencing another's, which makes it difficult to specify in advance which imitations will produce criminal conduct and which will not.

Sutherland wanted a mechanism precise enough to generate testable predictions, so he replaced imitation with the claim that a person becomes delinquent when they are exposed, through intimate personal groups, to an excess of definitions favourable to violation of the law over definitions unfavourable to it, a process he called differential association. The learning that occurs includes both the specific techniques of committing an offence and the attitudes, rationalisations, and motives that make the offence seem justified.

Sutherland set out this claim as a small number of formal propositions, moving criminology toward a testable, falsifiable statement of the learning process rather than a loose analogy to fashion, and that formal structure is a large part of why differential association, rather than imitation, became the theory taught as the sociological baseline for a further generation of researchers.

Differential association is deliberately narrower than Tarde's laws of imitation. It specifies the content that is learned, defines learning as occurring through communication within intimate groups rather than through generalised social copying, and adds the idea of ratio, that what matters is the balance of favourable to unfavourable definitions a person is exposed to over time, not merely whether contact with a criminal model occurred at all.

This precision is exactly what let differential association become the dominant learning theory of crime for the following decades, and it is also the gap that Daniel Glaser identified when he proposed differential identification in 1956.

Sutherland himself acknowledged that differential association was a general statement rather than a fully specified psychological mechanism, and later theorists, Glaser among them, treated that gap as an invitation rather than a flaw, each proposing a more specific account of how favourable definitions actually take hold in a given individual.

Daniel Glaser set out differential identification theory in his 1956 article "Criminality Theories and Behavioral Images", published in the American Journal of Sociology. Glaser accepted Sutherland's central claim that criminal conduct is learned through exposure to favourable definitions, but he argued that differential association, by tying the mechanism to associations with actual people in intimate groups, could not account for offenders who had little sustained contact with other criminals yet still adopted criminal definitions.

Glaser's own formulation states that a person pursues criminal behaviour to the extent that they identify themselves with real or imaginary persons from whose perspective their criminal behaviour seems acceptable. The operative word is identification, not association or contact. A person does not need to interact with a criminal model at all; they need only to see themselves through that figure's eyes, adopting a perspective from which the criminal act looks justified, admirable, or simply unremarkable.

Identification can attach to a parent, an older sibling, a peer, but it can equally attach to a person known only by reputation, a historical figure, or a completely fictional character. Glaser also distinguished degrees of identification, noting that a person may identify strongly with one figure and only weakly or ambivalently with another present in the same environment, which is why two people raised in materially similar circumstances can diverge sharply in whether they adopt criminal definitions.

This reframing solved the specific problem Glaser had identified in Sutherland's account. It explained why an only child with no delinquent peers could still become delinquent by identifying with a media figure or a distant relative's reputation, and it explained why exposure to the same criminal peer group produces different outcomes in different individuals, since not everyone in a group identifies with the same reference figure inside it or to the same degree.

Differential identification therefore sits inside the same learning tradition as Tarde and Sutherland while shifting the operative mechanism a second time, from imitation, to associational learning, to an internal act of identification that does not require the model's physical presence.

Glaser built the absence of required contact into the theory from the outset, which is why differential identification adapted easily to mass communication in a way Tarde's imitation and Sutherland's differential association did not as readily anticipate.

A reader of a crime novel, a viewer of a film, or a listener to a broadcast can identify with a character or a public figure they will never meet, and that identification can supply the same favourable definitions an intimate association would supply under Sutherland's model.

The clearest contemporary extension is parasocial identification through online media, where an audience member forms a one-sided sense of relationship with a content creator or public figure they follow but have never met, and can absorb that figure's stated justifications as though they came from an intimate associate; research on online radicalisation has documented individuals adopting extremist definitions of acceptable violence after identifying with figures encountered only through video or livestreams, with no offline association at all.