The Social Learning Tradition

Interpersonal Relationships and Deviant Behavior


Edwin H. Sutherland

Sutherland's differential association theory

Although Sutherland began work on a general explanation of criminal behavior in the 1920s, his first formal statement of differential association theory appeared in the 1939 edition of his textbook, Principles of Criminology. Sutherland's subsequent revision of the theory in the 1947 edition of his textbook continues to influence contemporary theoretical and empirical work on the social learning of deviant behavior.

Sutherland's theory consists of nine statements that specify various elements of the interpersonal process through which individuals learn to engage in law-violating behavior. The central explanatory principle of differential association--which is contained in statement 6 of the theory--is simple and straightforward: a person engages in criminal behavior "because of an excess of definitions favorable to violation of law over definitions unfavorable to violation of law" (Sutherland and Cressey, 1974: 75). All of us are members of intimate personal groups where we are exposed to definitions unfavorable to violation of law. In the family environment, for instance, we learn from our parents that we should respect and obey the law. However, most of us have also at some time been involved in groups where the law is defined in unfavorable terms. For example, some of your close friends might define laws prohibiting the use of certain drugs as unfair and might assure you that the chances of being caught for drug use are small. But, as Sutherland points out above, the mere fact that interpersonal communication in some groups exposes a person to unfavorable definitions of the law is not sufficient to produce criminal behavior. Criminal behavior results only when such definitions exceed noncriminal definitions that support the law. Thus, it is not the absolute amount of exposure to criminal patterns that is important; the differential or ratio of associations with criminal and noncriminal patterns is what provides the theoretical key to Sutherland's explanation of criminal behavior.

However, it is clear from other statements in Sutherland's theory and from his comments on those statements that the content of the criminal learning process includes much more than just definitions that oppose or support legal norms. More broadly, people learn various criminal or anticriminal patterns that are differentially present in their social environment (1974: 76). A criminal pattern refers to a complex but organized set of behaviors and cultural meanings that support violation of the law. The behavioral component of a criminal pattern "includes. . . techniques of committing the crime, which are sometimes very complicated, sometimes very simple" (1974: 75). But, merely learning how to commit a crime is not sufficient to cause a person actually to engage in criminal behavior. Criminal patterns also include a subjective component--cultural meanings attached to criminal behavior that make a person willing or motivated to violate the law. In addition to unfavorable definitions or evaluations of the law, these cultural meanings also include various "motives, drives, rationalizations, and attitudes" that channel a person's behavior in a criminal direction (1974: 75). Although Sutherland does not devote detailed attention to anti-criminal patterns, we can reasonably assume that this concept refers to patterns of law-abiding behavior and conventional cultural meanings. Most of us are not criminals, Sutherland would argue, because our associations with anticriminal patterns predominate over our exposure to criminal patterns. However, this ratio of associations is reversed in the social environment of criminals. Reiterating the central principle of differential association theory, Sutherland states that "(w)hen persons become criminal, they do so because of contacts with criminal patterns and also because of isolation from anticriminal patterns" (1974: 76).

What is the nature of the learning process by which criminal patterns are acquired by the individual? Statements 2, 3, and 8 in Sutherland's theory give concise answers to this question (1974: 75-76):

2. Criminal behavior is learned in interaction with other persons in a process of communication.
3. The principal part of the learning of criminal behavior occurs within intimate personal groups.
8. The process of learning criminal behavior by association with criminal and anticriminal patterns involves all of the mechanisms that are involved in any other learning.

In short, criminal behavior is a product of normal social learning through interaction in primary groups, such as friends or family.

Sutherland’s argument that law-violating behavior is socially learned from an excess of interpersonal associations with criminal patterns is appealing for its clarity and simplicity. But, as statement 7 in his theory indicates, he was cautious not to oversimplify his analysis of this micro-level process (1974: 76): “Differential associations may vary in frequency, duration, priority, and intensity.” Here, he specifies a number of analytical variables that figure into the misleadingly simple conception of an "excess" of associations. This excess may come about through more frequent exposure to criminal patterns than to anticriminal patterns, or it may be a function of greater duration  of contact with criminal patterns over time. With regard to the variable of priority, Sutherland suggests that criminal learning early in life may be especially influential on the individual. Finally, Sutherland does not precisely define the meaning of intensity, but he indicates that “it has to do with such things as the prestige of the source of a criminal or anticriminal pattern and with emotional reactions related to the associations” (1974: 76). All of these contingencies in the social learning process can affect the relative impact of criminal patterns or anticriminal patterns on the individual’s behavior.

Criticisms and revisions of Sutherland’s theory

In the decades following his 1947 reformulation of differential association theory, Sutherland’s influence on micro-level theorizing in the sociology of deviance was perhaps even more pervasive than was Merton’s influence on macro-level approaches. Theoretical work in the social learning tradition has primarily consisted of critical debates over the merits of Sutherland’s explanation of individual criminality and of attempts to revise or extend differential association theory (see Cressey, 1960).

The qualities that made Sutherland a leading social scientist show through in the fact that he, himself, was one of the most astute critics of differential association theory. One of his major reservations about the theory was its failure to incorporate criminal opportunities as an essential factor for the occurrence of criminal behavior (1956b).  Cloward and Ohlin (1960) later benefited from Sutherland’s critical scrutiny of his own theory by building their subcultural theory of delinquency around the concept of illegitimate opportunity. Sutherland was also concerned about the part played by variations in individual personality in the differential association process (1956a). His sociological explanation of crime focused attention on variations in the social environment in the form of interpersonal contacts with criminal or anticriminal patterns. However, he felt that this theory should be modified in some way to take into account personality traits, such as bashfulness, that might affect a person’s susceptibility to these environmental influences. 

Picking up on the latter criticism, Daniel Glaser (1956) proposed a revision of differential association theory that attempts to focus greater attention on individual factors that intervene between environmental contacts and criminal behavior. Glaser states his theory of differential identification as follows (1956: 440, emphasis deleted):

[A] person pursues criminal behavior to the extent that he identifies himself with real or imaginary persons from whose perspective his criminal behavior seems acceptable.

Glaser argues that this formulation improves upon Sutherland’s mechanistic image of individuals being pushed toward criminal or noncriminal behavior by opposing environmental forces. The theory of differential identification views the individual as a more active, voluntary participant in the criminal learning process. The key question in explaining criminal behavior, according to Glaser, is with whom does the individual subjectively choose to identify, criminals or noncriminals? Criminal associations and other environmental factors external to the individual are primarily relevant only “to the extent that they can be shown to affect the choice of the other from whose perspective the individual views his own behavior” (1956: 440).

It is not clear whether Glaser intended to revise differential association theory completely or merely to substitute his brief statement of differential identification for some of the statements in Sutherland’s theory. As it stands, differential identification theory is less specific than differential association theory regarding the nature and content of the social learning process. While Glaser’s explanation may be less deterministic than Sutherland’s, it is also less sociological. In contrast to Sutherland’s focus on variations in the social relationships and cultural meanings to which an individual is actually exposed, Glaser’s explanation shifts attention to psychological phenomena—variations in the individual’s internal identification with other persons, whether those persons are real or imaginary.

The issues of determinism and psychological explanation also become crucial in the assessment of several more recent attempts to revise Sutherland’s theory in accordance with the concepts and principles of psychological theories of behavioral learning (Jeffery, 1965; Burgess and Akers, 1966; Akers, 1977, 1998; Adams, 1973). These revisions of differential association theory are inspired by the fact that Sutherland did not present an intensive analysis of the kind of learning involved in individual criminality. Although Sutherland emphasized the social character of criminal learning and its fundamental similarity to non-criminal learning, he did not attempt to specify the psychological details of this learning process.

In the most ambitious attempt to fill in these details, Burgess and Akers (1966) used general principles of learning developed by the psychologist B. F. Skinner (1953) as the basis for a complete reformulation of Sutherland’s theory. Terming their revision differential association-reinforcement theory, Burgess and Akers argue that criminal behavior (or any other learned behavior) can be explained as a function of the reinforcement (i.e., rewards) or punishment a person receives from the environment. Put simply, behavior that produces a reward will tend to be performed more frequently in the future, while behavior that is followed by punishment will tend to decrease in frequency. Differential reinforcement occurs when under certain environmental conditions one particular behavior leads to greater or more frequent rewards than does another alternative behavior. The behavior that receives greater environmental reinforcement—for instance, deviant behavior—becomes dominant over the alternative behavior, such as conforming behavior. Therefore, through a process of differential reinforcement, environmental conditions can increase the probability of deviant acts by individuals.

A number of other concepts and general principles of behavior learning are used by Burgess and Akers (1966) to revise or delete each of the original statements in Sutherland’s theory. Even more than Sutherland, Burgess and Akers are able to drive home the point that criminal behavior is “normal” in the sense that it is learned in the same way as any other behavior. In later work, Akers (1977; 1998) has attempted to show how differential association-reinforcement theory might be used to explain other forms of deviant behavior, such as mental illness, alcoholism, and sexual deviance. Thus, an important contribution of this work has been to extend the social learning tradition to a number of forms of deviant behavior in addition to criminality.

Since differential association-reinforcement theory is phrased in the language of psychological learning theory, it might seem obvious that this is a nonsociological explanation of deviant behavior. However, this question is not as clear-cut as it might seem. Unlike some psychological perspectives that explain individual behavior as a function of variations in personality and other internal factors, the body of behavioristic theory on which Burgess and Akers base their work stresses conditions in the environment as the key determinants of human behavior. Burgess and Akers, following Sutherland, are also careful to point out that primary groups are the major source of reinforcement involved in the social learning of deviant behavior. However, as opposed to Sutherland and most other sociologists, Burgess and Akers do not limit their attention to the social environment. Differential association-reinforcement theory explicitly includes nonsocial reinforcement and learning without direct contact with other persons as possible sources of deviant behavior (1966; also see Adams, 1973). Only empirical research can determine how essential interpersonal relationships and other social factors are for the learning of deviant behavior, but the inclusion of nonsocial factors in Burgess and Akers’ theory carries it beyond the boundaries of a strictly sociological explanation (see Akers et al., 1979). 

Despite the criticisms that have been directed at it, Sutherland’s differential association theory still represents the best example of sociological analysis within the social learning tradition. The special emphasis Sutherland placed on social relationships and shared cultural meanings has been lost to varying degrees in later revisions of his theory. Yet, as recent developments in this micro-normative tradition would suggest, attention to some nonsociological elements in deviant behavior may provide some useful insights for research.

Control theory: an alternative to social learning theory

Like other normative theories, all versions of social learning theory are addressed to the question: "Why do they do it?" This question treats deviance as problematic. That is, it focuses analytical attention on special pressures or motivations that cause some people (they) to engage in norm-violating behavior. On the other hand, control theories raise a rather different question for sociological analysis: "Why don't we do it?" Here, conformity rather than deviance is treated as problematic, for the latter question shifts attention to social and personal controls that prevent most people (we) from engaging in norm-violating acts. From the viewpoint of control theorists, deviant behavior is "caused" only in the sense that norm-violating acts are more likely to occur when the controls that cause conformity are weak or absent.

Control theory has its origins in the early works of Durkheim and the Chicago sociologists, but the most influential proponent of this alternative to social learning theory is Travis Hirschi. In his book, Causes of Delinquency (1969), Hirschi clarifies the basic themes of control theory and provides a convincing demonstration of how an analysis of the problem of conformity can yield important insights into the phenomenon of deviance. Hirschi's own version of control theory focuses on several elements of the "social bond" that ties individuals to conventional society and explains their conforming behavior. Like so many other contemporary ideas in the field of deviance, the notion of a social bond goes back to Durkheim's analysis of different types of suicide (1951). In addition to his examination of anomie suicide, Durkheim distinguished another type of suicide that results from a lack of social integration—the "relaxation of social bonds" between individuals and society (1951: 214). This type of suicide, which Durkheim termed egoistic suicide, occurs under conditions where members of society share relatively few common beliefs or where group relationships are weak or lacking. Durkheim describes the egoistic or individualistic way of life encouraged by such conditions as follows (1951: 209):

The more weakened the groups to which (the individual) belongs, the less he depends on them, the more he consequently depends only on himself and recognizes no other rules of conduct than what are founded on his private interests.

Although individuals are in a sense freer from society under egoistic conditions, they are also less protected from the impact of personal troubles and misfortunes. In the absence of supportive social bonds, suicide rates increase as the isolated "individual yields to the slightest shock of circumstance" (1951: 215). The concept of a social bond was used by Durkheim for a macro-level explanation of how egoistic conditions contributed to high rates of suicide in certain societies or segments of society. Hirschi, on the other hand, applies this concept in a micro-level analysis of individual conformity and deviance.

Hirschi (1969: 16-34) distinguishes four elements of the social bond to conventional society that controls or prevents individual deviant behavior: attachment, commitment, involvement, and belief. The element of attachment comes closest to Durkheim's conception of social or group integration. The greater the extent to which the individual is attached to conventional groups, such as the family, the less likely he or she will be free to deviate. Commitment is a more rational element of the social bond that refers to the investments a person has in conventional forms of behavior. People committed to conventional educational or occupational careers are unlikely to risk their loss by engaging in deviant behavior. The element of involvement is based on the commonsense idea that "idle hands are the devil's workshop." That is, people who are highly involved in conventional activities have neither the time nor the energy required for planning and carrying out deviant activities. Whereas these first three elements refer to relationships and activities in the social environment that tie the individual to the conventional social order, Hirschi's final element, belief, refers to a more subjective aspect of the social bond. Hirschi assumes that there is a common set of conventional values and norms in society, but he points out that "there is variation in the extent to which people believe they should obey the rules of society, and, furthermore, that the less a person believes he should obey the rules, the more likely he is to violate them" (1969: 26).

Hirschi proposes that these four elements can be combined into an interrelated system of variables that together determine the likelihood of conforming or deviant behavior: "... the more closely a person is tied to conventional society in any of these ways, the more closely he is likely to be tied in the other ways" (1969: 27). However, it is clear that Hirschi considers attachment to be the most important element or variable in his theory of the social bond. He devotes special attention to how close relationships and emotional attachment with conventional groups not only affect one's beliefs in the moral rules of society but also determine one's commitment to conventional educational and occupational pursuits. Commitment to such conventional activities, in turn, determines one's involvement or allocation of time and resources to conforming behavior. Therefore, Hirschi implies that a lack of attachment to the family, to teachers, or to other conventional social relationships is particularly likely to free the individual from the social bond and to increase the probability of deviant behavior.

Like so many other normative theorists, Hirschi applies his theory to juvenile delinquency, although it has implications for other forms of deviant behavior as well. In Causes of Delinquency, Hirschi presents results from a questionnaire survey of a large sample of secondary school students that provide considerable support for predictions from social bonding theory. One finding from this study deserves comment because of its theoretical relevance to the social learning tradition. Hirschi reports that when a boy has strong bonds to conventional society (as measured by questions relating to family and school), his tendency to engage in delinquent acts is not substantially increased by association with delinquent friends (1969: 145-158). This and other findings in Hirschi's study suggest that social bonding theory might provide a more adequate account of interpersonal sources of delinquency than does social learning theory. At the very least, Hirschi's work indicates that the social learning tradition has some serious competition from control theory as a micro-normative explanation of adolescent deviance.

Adapted from pp. 14-15, 153-163 of James D. Orcutt, Analyzing Deviance, Dorsey Press, 1983.