The nine propositions of differential association constitute Edwin Sutherland’s final and most precisely stated account of how criminal behavior is acquired, published in the 1947 fourth edition of Principles of Criminology as the culmination of more than a decade of theoretical refinement within the Criminology Theories silo. Rather than offering a loose set of correlates associated with crime, Sutherland set out a chain of interconnected causal statements specifying that criminal behavior is learned, that it is learned through communication within intimate personal groups, and that the direction of a person’s motives and drives depends on whether the definitions to which that person is exposed favor or oppose violation of the law. The nine propositions were designed to function as a unified logical system, each proposition building on the ones preceding it, so that the theory could explain both why particular individuals become criminal and why crime rates vary systematically across groups and communities. This article works through each proposition in turn, examines the logical architecture connecting them, and assesses the interpretive debates that have surrounded the theory’s most technically demanding claims, particularly the ratio-based formulation captured in the seventh proposition.
Introduction
Sutherland intended the nine propositions to be read as a connected sequence rather than as nine independent claims, since later propositions depend logically on assumptions established by earlier ones; proposition one establishes that criminal behavior is learned rather than inherited or independently invented, a claim that sets the stage for propositions two through five, which specify the mechanism, context, and content of that learning process. Propositions six and seven then introduce the theory’s most distinctive and most frequently debated claim, that criminality results from an excess of definitions favorable to violation of law over definitions unfavorable to violation, modified by the frequency, duration, priority, and intensity of the associations through which those definitions are transmitted. Propositions eight and nine close the system by specifying that the learning mechanism involved in acquiring criminal behavior is identical to the mechanism involved in acquiring any other behavior, and by explicitly ruling out general needs and values as adequate explanations of crime, since these needs and values are equally characteristic of non-criminal behavior.
This article is organized around the Edwin Sutherland and Differential Association Theory biographical and historical treatment already available in this silo, and it assumes that background while focusing narrowly on the internal logic of the nine propositions themselves. Each section below groups the propositions that function together as a logical unit, following the grouping structure most commonly used in criminological pedagogy, and closes with an assessment of how each proposition has fared under sustained empirical and philosophical scrutiny in the decades since 1947.
Proposition One: Criminal Behavior Is Learned
Rejecting Innate and Independent Invention Explanations
The first proposition states simply that criminal behavior is learned, a claim whose apparent simplicity conceals a sweeping rejection of two alternative explanations that Sutherland considered inadequate: that criminal behavior is innate, a product of inherited constitution, and that criminal behavior might arise through independent invention, meaning that an individual could spontaneously devise criminal conduct without instruction or example from others (Sutherland, 1947). By ruling out both innateness and independent invention in the same proposition, Sutherland positioned differential association theory in direct opposition to biological determinism, which located the source of crime in hereditary constitution, and to any account that treated criminal ingenuity as a spontaneous individual achievement rather than a socially transmitted skill.
Ronald Akers’s later synthesis of the social learning tradition credited proposition one with anticipating a premise that behavioral psychology would later establish on independent empirical grounds: that complex behavior patterns, criminal or otherwise, are acquired through experience rather than expressed directly from inherited predisposition (Akers, 2009). Sutherland’s rejection of independent invention carried a specific evidentiary implication: since most complex behavior, and virtually all technically demanding criminal behavior, requires some degree of instruction, observation, or apprenticeship, the theory predicted that careful investigation of any individual criminal career would reveal identifiable social sources of criminal technique and motivation rather than spontaneous emergence (Sutherland & Cressey, 1978). This prediction proved consistent with Sutherland’s own earlier fieldwork in The Professional Thief, which documented an extended apprenticeship process through which a novice acquired both technical skill and normative orientation from established professional criminals (Sutherland, 1937). The proposition thus functioned not merely as an abstract theoretical claim but as a methodological directive: investigators should look for the social origins of any given criminal behavior rather than treating it as self-evidently emergent from individual character.
The Learning Analogy to Legitimate Behavior
Proposition one’s deeper theoretical significance lies in its implicit analogy between criminal and legitimate behavior: if criminal behavior is learned in the same general sense that legitimate behavior is learned, then criminology need not develop a special, sui generis theory of criminal motivation distinct from general theories of human learning (Sutherland, 1947). This move was consequential because it allowed Sutherland to treat criminality as continuous with ordinary social behavior rather than as a categorically distinct phenomenon requiring explanation by pathology, positioning differential association theory as an application of general learning principles to a specific behavioral outcome rather than as a specialized theory of a specialized population.
Critics have noted that proposition one, taken alone, is compatible with a wide range of subsequent theoretical elaborations, since many theories agree that criminal behavior is learned while disagreeing sharply about the mechanisms, content, and context of that learning (Akers, 1998). The proposition’s importance, therefore, lies less in its standalone content than in its function as the logical premise that licenses the more specific claims contained in propositions two through nine; without establishing that criminal behavior is learned rather than innate or spontaneously invented, the subsequent propositions specifying how, where, and through what content that learning occurs would lack their foundational warrant. Later social learning theorists, including Ronald Akers, retained this foundational premise even while substantially modifying the mechanisms Sutherland specified in later propositions, illustrating how proposition one functioned as durable bedrock beneath decades of theoretical elaboration (Burgess & Akers, 1966). Daniel Glaser’s differential identification theory offered one of the earliest such modifications, retaining proposition one’s core claim that criminal behavior is learned while arguing that identification with real or symbolic others, rather than direct associational frequency alone, better captures the process by which individuals come to favor law violation (Glaser, 1956); Albert Cohen’s work on delinquent subcultures likewise accepted the learned-behavior premise while emphasizing collective, group-level status processes as the vehicle of that learning (Cohen, 1955).
Propositions Two and Three: The Process and Context of Learning
Communication with Other Persons
Proposition two specifies that criminal behavior is learned in interaction with other persons in a process of communication, a claim that anchors the theory firmly within the interactionist tradition associated with George Herbert Mead and situates learning as an inherently social and dyadic or small-group process rather than a solitary cognitive achievement (Sutherland, 1947). Sutherland’s specification of communication as the vehicle of learning excluded, at least in its original formulation, learning through impersonal or mass-mediated sources such as newspapers, films, or later broadcast media, a limitation that Sutherland defended on the grounds that such impersonal communication rarely carried the intensity and personal significance necessary to shape an individual’s motives and self-concept (Sutherland & Cressey, 1978).
This restriction to interpersonal communication has generated one of the most persistent lines of criticism directed at the theory, since it was formulated decades before electronic and networked communication became a primary vector for the transmission of both technical criminal knowledge and normative definitions favorable to crime, a limitation examined further in the dedicated treatment of the theory’s application to online and cyber-enabled offending elsewhere in this silo. Even accepting this limitation, proposition two’s core insight, that criminal learning is a communicative and relational process rather than a matter of individual discovery, has remained broadly durable, informing decades of research on peer influence, gang recruitment, and family transmission of criminal norms (Warr, 2002).
Intimate Personal Groups as the Primary Learning Context
Proposition three narrows the communicative context specified in proposition two by asserting that the principal part of the learning of criminal behavior occurs within intimate personal groups, a claim that privileges family, close friendship networks, and small peer groups over more diffuse or impersonal social contexts as the primary sites of criminal socialization (Sutherland, 1947). Sutherland’s emphasis on intimacy reflected his conviction that the transmission of definitions favorable to law violation requires a degree of trust, emotional significance, and sustained contact that only close personal relationships typically provide, a claim consistent with subsequent research on the outsized influence of close friends, relative to more casual acquaintances, on adolescent delinquent involvement (Warr, 2002).
The intimacy requirement also carries an important negative implication: proposition three suggests that more impersonal or fleeting associations, even if numerically frequent, should exert comparatively weak influence on criminal learning relative to associations characterized by closeness and emotional significance. This claim has been operationalized in subsequent empirical research through measures distinguishing close friends from mere acquaintances and through attention to family relationships as a distinct and often earlier-occurring context of associational influence, examined more fully in the dedicated treatment of the theory’s application to family relationships elsewhere in this silo. Sutherland’s insistence on intimacy as a qualifying condition, rather than treating all associations as interchangeable, represents one of the theory’s more precise and empirically productive specifications, even as later researchers have debated exactly how intimacy should be measured and whether its influence is genuinely distinct from sheer frequency of contact (Matsueda, 1988).
Propositions Four and Five: Content of Learning and Definitions
Techniques, Motives, and Rationalizations
Proposition four specifies the content of what is learned within criminal associations: when criminal behavior is learned, the learning includes techniques of committing the crime, which are sometimes very complicated and sometimes very simple, and the specific direction of motives, drives, rationalizations, and attitudes (Sutherland, 1947). This proposition distinguishes differential association theory from purely technical or opportunity-based accounts of crime by insisting that criminal learning involves not merely the acquisition of skill but the acquisition of a normative and motivational orientation that makes the use of that skill for criminal purposes psychologically available and justified to the individual.
The inclusion of rationalizations within the content of learning proved especially generative for later scholarship, most directly informing Sykes and Matza’s subsequent theory of techniques of neutralization, which elaborated in much greater empirical detail the specific verbal and cognitive strategies, including denial of responsibility and denial of injury, through which offenders neutralize the moral constraints that would otherwise inhibit criminal conduct (Sykes & Matza, 1957). Cressey’s own embezzlement research similarly extended proposition four by showing how trusted employees required a specific verbalization, or rationalization, permitting them to redefine embezzlement as something other than theft before they could act on an otherwise unshared financial problem, directly illustrating how techniques and rationalizations operate jointly rather than independently in the commission of crime (Cressey, 1953).
Definitions Favorable and Unfavorable to Violation of Law
Proposition five specifies that the direction of motives and drives is learned from definitions of the legal codes as favorable or unfavorable, and Sutherland was careful to clarify that in some societies an individual is surrounded by persons who unanimously define the legal codes as rules to be observed, while in other social settings an individual is surrounded by persons whose definitions are conflicting, so that the individual is often exposed to both kinds of definitions simultaneously (Sutherland, 1947). This proposition introduces the theory’s central conceptual unit, the definition, understood as an orientation toward legal codes rather than as a single discrete belief, and it is this concept of competing definitions, examined in dedicated depth in the companion article on definitions favorable and unfavorable to crime elsewhere in this silo, that supplies the raw material for the ratio specified in proposition six.
Sutherland’s formulation of proposition five deliberately avoided reducing definitions to simple approval or disapproval of specific criminal acts, instead framing them as broader orientations toward the legitimacy and bindingness of law itself, an orientation that could vary across social groups, subcultures, and occupational settings in ways that a narrower act-specific formulation would not capture (Sutherland & Cressey, 1978). This breadth allowed the theory to explain not only the acquisition of specific criminal techniques but the acquisition of a general disposition toward law violation, a disposition capable of generalizing across offense types in ways that more narrowly technique-focused theories could not readily account for, a generalization capacity that later researchers have both praised for its explanatory reach and criticized for its resistance to precise empirical operationalization (Matsueda, 1988).
Propositions Six and Seven: The Principle of Differential Association
Excess of Definitions Favorable to Violation
Proposition six contains what Sutherland and subsequent commentators have generally regarded as the theoretical heart of the entire system: a person becomes delinquent because of an excess of definitions favorable to violation of law over definitions unfavorable to violation of law (Sutherland, 1947). This is the proposition from which the theory takes its name, since it specifies that criminality results not from mere exposure to any criminal definitions whatsoever, an exposure that is nearly universal in modern society, but from a differential balance in which favorable definitions numerically or functionally outweigh unfavorable ones within an individual’s associational history. The proposition is explicitly relativistic and comparative rather than absolute, meaning that the theory predicts criminal behavior not from the presence of criminal associations alone but from their weight relative to conventional, law-abiding associations.
This comparative framing carried an important theoretical corollary that Sutherland made explicit: differential associations may vary in frequency, duration, priority, and intensity, a specification that anticipates proposition seven and that prevents proposition six from being read as a crude numerical head-count of criminal versus non-criminal acquaintances (Sutherland, 1947). Sutherland intended the ratio to capture the qualitative weight of associations rather than their sheer number, a distinction that later empirical researchers have found difficult to operationalize with precision, since survey and interview methods more readily capture the number and frequency of associations than their qualitative priority or intensity, a methodological difficulty discussed further in the dedicated review of empirical tests of the theory elsewhere in this silo (Pratt et al., 2010).
Frequency, Duration, Priority, and Intensity
Proposition seven supplies the four modalities by which the differential associations specified in proposition six are to be weighted: frequency, referring to how often an individual is exposed to a given set of definitions; duration, referring to the length of time over which exposure occurs; priority, referring to the developmental timing of exposure, with earlier-formed definitions generally presumed more influential; and intensity, referring to the prestige and emotional significance of the source of a given definition (Sutherland, 1947). This four-part specification, examined at much greater length in the dedicated companion article on frequency, duration, priority, and intensity elsewhere in this silo, represents Sutherland’s attempt to render the abstract ratio of proposition six operational by identifying the specific dimensions along which associations vary in their capacity to shape an individual’s balance of definitions.
The frequency-duration-priority-intensity framework has proven simultaneously the theory’s most conceptually rich contribution and its most persistent empirical liability, since Sutherland never specified a precise algorithm by which these four dimensions should be combined or weighted relative to one another, leaving researchers to develop their own, sometimes inconsistent, operationalizations (Cressey, 1960). Cressey himself, in defending the theory against charges of untestability, argued that this apparent imprecision reflected the genuine complexity of the phenomenon being modeled rather than a fatal theoretical flaw, and he maintained that careful qualitative and quantitative attention to all four modalities, rather than reliance on frequency alone, would yield a more accurate application of the theory than the frequency-only operationalizations common in early empirical tests (Cressey, 1960). This defense has been only partially persuasive to later methodologists, who continue to debate whether the ratio specified in propositions six and seven can be rendered genuinely falsifiable given the absence of a specified combinatory rule.
Propositions Eight and Nine: Learning Mechanisms and the Limits of Need Explanations
Learning by All the Mechanisms Involved in Other Learning
Proposition eight states that 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, explicitly rejecting the notion that criminal learning requires a distinct or specialized psychological process, such as imitation alone, separate from the general mechanisms of human learning (Sutherland, 1947). This proposition reinforced the theory’s continuity with general behavioral and social-psychological science, positioning differential association theory as an application of established learning principles rather than as a freestanding, criminology-specific psychological theory, a positioning that later proved directly useful to Ronald Akers and Robert Burgess when they reformulated the theory in explicitly operant-conditioning terms.
Burgess and Akers’s 1966 reformulation took proposition eight as its point of departure, arguing that Sutherland had correctly identified the need for a general learning-theoretic mechanism but had failed to specify that mechanism with the precision available through the behavioral psychology of B. F. Skinner, and their differential association-reinforcement theory, examined in dedicated depth elsewhere in this silo, sought to supply exactly the mechanistic precision that proposition eight gestured toward without fully achieving (Burgess & Akers, 1966). This reformulation illustrates how proposition eight functioned less as a completed theoretical claim than as an invitation for subsequent scholars to specify the general learning mechanisms Sutherland referenced but did not himself elaborate using the technical vocabulary of learning psychology.
Why General Needs and Values Do Not Explain Crime
Proposition nine closes the theoretical system with a deliberately negative claim: while criminal behavior is an expression of general needs and values, it is not explained by those general needs and values, since noncriminal behavior is an expression of the same needs and values (Sutherland, 1947). Sutherland offered this proposition specifically to foreclose explanations of crime that appealed to the desire for money, status, or excitement, arguing that such motives are equally characteristic of law-abiding individuals and therefore cannot, by themselves, distinguish criminal from noncriminal conduct; what distinguishes the criminal, in Sutherland’s account, is not the presence of these general motives but the specific associational history that channels those motives toward criminal rather than legitimate means of satisfaction.
This proposition has been credited by sympathetic commentators as one of Sutherland’s most incisive theoretical moves, since it anticipated later strain-theoretic debates about the relationship between generalized aspirations and criminal means by insisting that generalized aspirations alone possess no explanatory power absent an account of how those aspirations become linked to criminal rather than conventional means (Vold, Bernard, & Snipes, 2002). Critics, meanwhile, have noted that proposition nine, in ruling out general needs and values as explanatory, arguably narrows the theory’s explanatory reach precisely at the moment when a fuller account of individual differences in criminal propensity might be most useful, a tension addressed at greater length in the dedicated treatment of criticisms of the theory elsewhere in this silo (Akers, 1998).
Interpretive Debates and the Logical Architecture of the Nine Propositions
The Ratio Problem and Operationalization Challenges
The most persistent interpretive difficulty surrounding the nine propositions concerns the precise meaning and measurability of the ratio specified in proposition six and elaborated in proposition seven, since Sutherland never provided a formula specifying how frequency, duration, priority, and intensity should be combined into a single measure of the net balance of definitions favorable and unfavorable to law violation (Matsueda, 1988). Early empirical researchers, including James Short, largely operationalized the ratio through simplified measures of frequency of association with delinquent peers, a simplification that critics argue understates the theory’s original complexity while nonetheless yielding empirical results broadly consistent with the theory’s core prediction that delinquent associations predict delinquent behavior (Short, 1957).
This operationalization gap has generated an extended methodological literature attempting to develop more faithful measures of the full ratio, including efforts to incorporate priority and intensity through retrospective measures of when associations began and how much respect or admiration the individual held for the associates in question, though such measures remain more difficult to collect reliably than simple frequency counts (Pratt et al., 2010). The persistence of this measurement challenge, more than seventy years after the propositions were first stated in their canonical form, illustrates both the theory’s continued centrality to empirical criminology and the genuine difficulty of translating a qualitatively rich theoretical construct into a fully specified quantitative model, a difficulty that has motivated much of the subsequent theoretical elaboration examined in the dedicated treatment of the theory’s principal criticisms elsewhere in this silo. Travis Hirschi’s rival social bond theory, developed explicitly as an alternative to the ratio-based logic of propositions six and seven, argued that a simpler measure of an individual’s attachment, commitment, involvement, and belief in conventional institutions predicted delinquency at least as well as measures of differential peer association (Hirschi, 1969), a challenge that prompted Barbara Costello and Paul Vowell’s later reanalysis of the original Richmond Youth Project data, which found that carefully specified differential association measures retained independent predictive power even after controlling for Hirschi’s bonding variables (Costello & Vowell, 1999).
Criticism and Defense of the Propositions as a System
Viewed as an integrated logical system rather than as nine isolated claims, the propositions have drawn both praise for their internal coherence and criticism for specific points of ambiguity that ripple through the system as a whole; because proposition six depends on the definitional concept introduced in proposition five, and proposition seven depends on the ratio concept introduced in proposition six, any imprecision at an early stage of the chain propagates into later, more empirically consequential propositions (Cressey, 1960). Defenders of the system, including Cressey, have argued that this interdependence is a strength rather than a weakness, since it forces careful theoretical work to proceed systematically rather than piecemeal, while critics have argued that the same interdependence makes the theory difficult to test in isolated components, since a failure to confirm one proposition may reflect measurement failure at an earlier stage of the causal chain rather than genuine theoretical inadequacy (Akers, 1998).
Melvin DeFleur and Richard Quinney undertook one of the earliest attempts to reconstruct the nine propositions as a formal axiomatic system, translating Sutherland’s prose statements into a set of interrelated postulates and theorems with the explicit goal of clarifying which empirical predictions the theory actually entailed and exposing points at which the original propositions left key terms undefined (DeFleur & Quinney, 1966). Their reconstruction confirmed that propositions six and seven carried the theory’s principal empirical burden while also revealing that several auxiliary terms, including the precise threshold at which an excess of favorable definitions becomes causally sufficient, remained formally unspecified in Sutherland’s original text, a gap that subsequent empirical researchers, including Gary Jensen in an early test using survey data on delinquent attitudes and behavior, attempted to close through explicit measurement strategies (Jensen, 1972).
Contemporary criminological theory continues to draw on the nine propositions both as a historical touchstone and as an active analytical framework, with scholars in the social learning and life-course traditions regularly returning to the original propositions to clarify ambiguities in their own reformulated models, a pattern that testifies to the propositions’ continued generative capacity even amid seventy years of critique (Warr, 2002). The nine propositions thus occupy an unusual position in the history of criminological theory: simultaneously a fixed historical text, unrevised since Cressey’s careful preservation of Sutherland’s original wording across later textbook editions, and a living analytical resource that each new generation of criminologists reinterprets in light of its own methodological and substantive concerns.
Conclusion
The nine propositions of differential association form a deliberately sequential logical system in which each claim depends on the ones preceding it, moving from the foundational assertion that criminal behavior is learned, through the specification of communicative context and intimate association, to the definitional content of that learning, and finally to the ratio-based principle that criminality results from an excess of definitions favorable to law violation weighted by frequency, duration, priority, and intensity. Propositions eight and nine close the system by situating criminal learning within general learning mechanisms and by explicitly denying that generalized needs and values can explain criminal behavior on their own, a move that sharpened the theory’s causal claims considerably relative to the looser multiple-factor approaches Sutherland sought to displace.
Seventy years of subsequent scholarship have both extended and complicated this system, generating reformulations such as Akers’s differential association-reinforcement theory that supply mechanistic precision the original propositions gestured toward but did not fully specify, and generating a substantial methodological literature grappling with the operationalization of the ratio concept at the system’s core. That this seventy-year engagement continues to generate active research, rather than settling into purely historical interest, testifies to the enduring analytical power of Sutherland’s original architecture, even as its most technically demanding claims remain the subject of ongoing empirical and theoretical debate.
Related Articles
- Edwin Sutherland and Differential Association Theory
- Learning Criminal Behavior Through Social Interaction
- Definitions Favorable and Unfavorable to Crime
- Frequency, Duration, Priority, and Intensity
- Differential Association-Reinforcement Theory
References
- Akers, R. L. (1998). Social learning and social structure: A general theory of crime and deviance. Northeastern University Press.
- Akers, R. L. (2009). Social learning and social structure: A general theory of crime and deviance (2nd ed.). Transaction Publishers.
- Burgess, R. L., & Akers, R. L. (1966). A differential association-reinforcement theory of criminal behavior. Social Problems, 14(2), 128–147.
- Cohen, A. K. (1955). Delinquent boys: The culture of the gang. Free Press.
- Costello, B. J., & Vowell, P. R. (1999). Testing control theory and differential association: A reanalysis of the Richmond Youth Project data. Criminology, 37(4), 815–842.
- Cressey, D. R. (1953). Other people’s money: A study in the social psychology of embezzlement. Free Press.
- Cressey, D. R. (1960). Epidemiology and individual conduct: A case from criminology. Pacific Sociological Review, 3(2), 47–58.
- DeFleur, M. L., & Quinney, R. (1966). A reformulation of Sutherland’s differential association theory and a strategy for empirical verification. Journal of Research in Crime and Delinquency, 3(1), 1–22.
- Glaser, D. (1956). Criminality theories and behavioral images. American Journal of Sociology, 61(5), 433–444.
- Hirschi, T. (1969). Causes of delinquency. University of California Press.
- Jensen, G. F. (1972). Parents, peers, and delinquent action: A test of the differential association perspective. American Journal of Sociology, 78(3), 562–575.
- Matsueda, R. L. (1988). The current state of differential association theory. Crime & Delinquency, 34(3), 277–306.
- Pratt, T. C., Cullen, F. T., Sellers, C. S., Winfree, L. T., Madensen, T. D., Daigle, L. E., Fearn, N. E., & Gau, J. M. (2010). The empirical status of social learning theory: A meta-analysis. Justice Quarterly, 27(6), 765–802.
- Short, J. F. (1957). Differential association and delinquency. Social Problems, 4(3), 233–239.
- Sutherland, E. H. (1937). The professional thief. University of Chicago Press.
- Sutherland, E. H. (1947). Principles of criminology (4th ed.). J. B. Lippincott.
- Sutherland, E. H., & Cressey, D. R. (1978). Criminology (10th ed.). J. B. Lippincott.
- Sykes, G. M., & Matza, D. (1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22(6), 664–670.
- Vold, G. B., Bernard, T. J., & Snipes, J. B. (2002). Theoretical criminology (5th ed.). Oxford University Press.
- Warr, M. (2002). Companions in crime: The social aspects of criminal conduct. Cambridge University Press.