Table of Contents


Title page

Copyright page



Part I: State, Social Order, and “the Criminal Question” in Modern Europe

1 Leviathan’s Subjects: From the Social Contract to Cesare Beccaria

From the Constitution of Subjects to the Contract to the “Penitentiary Invention”

Freedom, Self-Control, Discipline

Cesare Beccaria’s Version of Enlightenment

On Crimes and Punishments

The Realism of Beccaria’s Concept of “Free Will”

The Dialectic of Enlightenment

2 The “Positive School,” Urban Crowds, and the Social Question

The City Crowd

The Emergence of Positivism

Quételet and Guerry’s Sociological Positivism

The Italian Positive School

“Scientific” Criticism of the Positive School

“Contextual” Criticism of the Positive School: The “Southern Question”

The Positive School and the Crisis at the Turn of the Century

Freud’s Psychoanalysis and Crime

3 Emile Durkheim’s Sociology of Deviance

The Division of Labor in Society



The Sociology of Deviance

Durkheim’s Concept of State and Democracy

Two Laws of Penal Evolution

Part II: Democracy, Social Control, and Deviance in America

4 Social Control and Deviance in the New Republic

A Prologue: “America”

Protestantism, Republic, Social Control

5 Social Control and Deviance in Chicago

State and Social Control

Chicago: An Attitude of “Appreciation”

Robert Park: “The Public” and Social Control

Chicago Sociology of Deviance between “Ecological” Theory and “Social Disorganization”

Democracy and Social Control

Language and Social Control

Sister Carrie’s Less Eligibility

Louis Wirth: Urbanism as a Way of Life

6 The 1930s: Between Differential Association and Anomie

Crime and the New Deal

The Theory of Differential Association

White-Collar Crime

Talcott Parsons’s “Problem of Order”

Robert K. Merton’s Theory of Anomie

Crime, Democracy, and “the American Dream”

7 From the “Neo-Chicagoans” to Labeling Theory

The Tradition of Chicago: Social Control and the “Power Elite”

The Dialectic of Social Control: The Mass Media

Labeling Theory

David Matza: From “the Subterraneans” to “the Ban”

The Sixties: “Sympathy for the Devil”

8 From “Labeling” to a “Critical” Kind of Criminology

Zookeepers of Deviance

A “Critical” Kind of Criminology: The “Criminal Question”

The Critique of “Total Institutions” from Goffman to Foucault

Michel Foucault Cutting Off the Head of the King


The Contribution of Feminism: Social Control and the Constitution of a Gendered Subject

Part III: The “Crisis Decades”: “State,” Social Control, and Deviance Today

9 The End of “the Short Century” between Inequality and Fear

The Social Crisis of the 1970s and Mass Imprisonment in the United States

“Realist” Criminologies

Control Theory

Criminologies of the Self and the Other

Criminologies of the Self

Victimization and Self-Report Studies


Broken Windows and Zero Tolerance

The Criminology of “the Other” and the “Underclass”

The Turn of the 1990s

Automated Control

10 The Cycle of the Canaille

A Theory of Long Cycles

The Perennial Reproduction of the Canaille

Post-Fordist Corrections

What Does “Rehabilitation” Mean?

Penality and Government of Populations

Post-Fordism, Globalization, and Social Control

Conclusions: Crime, Punishment, and Social Structure



Title page


The pages that follow are a reflection of my own personal journey between the United States (more precisely California, where I lived between 1977 and 1993) and Italy, where I lived before and after those dates. The work is devoted to reconstructing the development of theories about social control (and the State), deviance, and crime in Europe and America. In so doing, I was led to revisit some of the motifs originally developed in The Prison and the Factory, written with Massimo Pavarini in Italian in 1977 (1981), and then subsequently my own The State of Social Control: A Sociological Study of Concepts of State and Social Control in the Making of Democracy (1990). The first result of this work of reconstruction was a volume in Italian, Stato, controllo sociale, devianza: Teorie criminologiche e società tra Europa e Stati Uniti, published by Paravia Bruno Mondadori Editore, Milan, in 2002. This was the work that is the ancestor, so to speak, of the present volume that I have now written in English.

Since 1993, when I returned to the School of Law of the University of Bologna, I have been teaching courses in criminology (understood as sociology of deviance and social control), and it is on the basis of this experience that I have now written this volume. From the two main concepts which oriented The State of Social Control – the all-European idea of “the State” and the all-American idea of “social control” – I have now moved to stressing a third orienting concept, that of “crime,” or “deviance.” From the political and social theory I have therefore moved to a more straightforward criminological interest, while at the same time remaining true to David Matza’s intimation of never separating “the study of crime from the workings and theory of the state,” the original sin, so to speak, of “criminological positivism” (Matza 1969: 143).

The present text is divided into three parts. The first, dedicated to “European” theories from the Classical to the Positive School to Emile Durkheim’s pioneering sociology, is a reconstruction of the relationship between the development of concepts of the “State,” the idea of deviance (or to be more philologically correct, “crime”), and the broader framework of the political organization of society and social reaction to crime in nineteenth-century Europe. The second part focuses on the analysis of the emergence of a concept of social control at the dawn of the first “mass democratic” society, i.e. the United States of America, and the reconstruction of the subsequent sociological theories of deviance between the beginning of the twentieth century and the 1970s. The third, more complex and open part retraces the events of the “current” period since the early 1970s, a period characterized by an unprecedented rise in the volume of penality in the United States and a somewhat limited “export” of American policies to Europe. This was a period when the social form of mass democracy was first extended from the United States to Europe, before entering a deep crisis, which was also marked by the emergence of what I have termed “automated control” (as opposed to “social control”).

The object of this work is therefore the reconstruction of the ways of thinking control, broadly speaking, in relation to the different modes of social organization and the prevailing concepts of “deviance” and “crime” therein. In fact, social organization and concepts of deviance imply each other in a stricter way than is usually thought – as Emile Durkheim pointed out early on.

In a sense, I started thinking about this work almost 20 years ago, after the publication of The State of Social Control. It is therefore practically impossible to thank all those who, through exchange of ideas, collaborations, and all kinds of human and social relations, have contributed to what follows. I shall therefore limit myself to remembering those who have been my mentors, none of whom is unfortunately still with us. They are Franco Bricola and Alessandro Baratta in Italy – to whose names is forever linked the impulse toward a deep renewal of critical thinking in Italian criminology – and “Don” Cressey and “Ed” Lemert in California. I had in fact the honor of being Don’s last PhD student at the University of California, Santa Barbara, from 1979 to 1986, and then colleague to Ed at the University of California, Davis, between 1986 and 1993. They were wonderful mentors, both to the young PhD student and to the subsequent junior colleague, and exceptional representatives of that high tradition of American sociological criminology to which I will be forever indebted.

My most heartfelt thanks also go to the hundreds and hundreds of students who, in courses and seminars on criminology that I have held since 1993 at the School of Law of the University of Bologna, have taught me and led me in the writing of the pages that follow, through their questions, their comments, their understandings, and oftentimes even their misunderstandings. These pages undoubtedly owe a lot to them. Special thanks go also to my friends and colleagues Malcolm M. Feeley, Rosann Greenspan, and Jonathan Simon, of the Center for the Study of Law and Society at the Boalt Hall School of Law of the University of California, Berkeley, for very generously tolerating my presence at the Center for several summers, a presence that has helped me complete several stages of this long journey (and that was often made possible thanks to the EAP Faculty Exchange Program between the University of California and the University of Bologna). A heartfelt “thank you” also goes to my Argentinean colleague and friend, Maximo Sozzo, who has painstakingly read the entire manuscript and given me invaluable help, also with an eye to the future Spanish translation. Emma Longstaff, Jonathan Skerrett, and Fiona Sewell, of Polity, have crucially helped me with preparing this volume in its final form, showing great generosity in dealing with an author for whom the English language is still, after all, a “second” language! I would also like to thank the Italian publisher Paravia Bruno Mondadori Editore for generously allowing me and Polity to freely use the Italian manuscript of Stato, controllo sociale, devianza as the initial building blocks for this work. And, last but not least, thank you so much, Peggy and Emilia, for your love, support, and patience!


When, between the 1960s and 1970s, a number of young European and American scholars from legal and social studies started looking into the apparently novel ideas of “social control” and “deviance,” within such notable organizations as the European Group for the Study of Deviance and Social Control or the “School of Criminology” of the University of California at Berkeley, it was rather typical of such intellectual efforts to consider the two concepts of “the State” and “social control” as strictly linked when not almost synonymous. The State was often seen as the “author” of social control. It reminded one of that kind of anthropomorphic conception that legal theorist Hans Kelsen had dubbed the Makroanthropos (Kelsen 1922: 3), the State as a very large human being, who “does” this and that, “organizes,” “imposes,” “prohibits,” and sometimes even assigns rewards and punishments of various sorts. The State was therein seen as a kind of “great father,” which Freud, unsurprisingly, had made the object of some of his “metapsychological” studies (Freud 1913, 1921).

Under the impact of crucial contributions by Michel Foucault (1975, 1978), it became necessary, for those young scholars, to start questioning this authoritarian, paternalist, and essentially phantasmagoric idea of the social order. In a previous volume (Melossi 1990), I tried to reconnect to such an intellectual heritage, albeit trying to show, at the same time, that the essentials of such questioning had already developed to some degree within the very origins of the social sciences in the United States. In one of its most characteristic and original statements, for instance, Arthur Bentley, the founder of North American political science, in his pioneering The Process of Government (1908), had contemptuously set aside the concepts of “State” and “sovereignty” as mere survivals of a mythical past. So much was this so that a great political philosopher of the twentieth century, Alessandro Passerin d’Entrèves, in reconstructing “the notion of the State,” was brought to observe that in contemporary political science, especially in its American version, a true “dissolution” of the State concept had taken place, such that “[t]he disruption of the notion of the State in modern political science is such a challenging and portentous event that it is surprising no detailed study should yet have been made to account for it and to explain it” (Passerin d’Entrèves 1967: 60).

In place of “the State” another key concept emerged. The idea of social control took shape in the intellectual laboratory that characterized the Progressive Era in the United States in the early decades of the twentieth century. The notion of social control certainly had a much lesser pedigree than that of the State. It belonged decisively to the social sciences rather than to political philosophy. It belonged, indeed, within the “social engineering” of the early decades of last century. It was a concept designed to capture the idea of an “intelligent” government of social change rather than the metaphysical “essence” of social order – an essence that had, when the occasion demanded, to be wrenched from the recalcitrant social body by force. This emergent view of social order was no longer a vision descending from the heavens of political philosophy in a still pre-democratic society. It was instead a perspective according to which the social sciences were coming to terms with the ongoing processes of construction of consensus among the masses. It was in fact the cooperation and collaboration of the latter that were at stake. Even in this case, social intervention descended from on high, but instead of imposing its “sovereignty” on civil society, its function was to capture, channel, and guide the deepest currents of social change. In short, the crux of the matter was a shift from the imposition of coercion to the organization of consensus. That shift was fundamental to the emergence of “democratic” societies as they came to be understood.

At the same time, however, as we shall see in this volume, every theory of the social order incorporates a theory of deviance implicitly or explicitly (which amounts to saying that every theory of deviance can be traced back to the general theory of social order that to some degree supports it). For instance, in European contractualist theories, centered on the concepts of the State and the individual, the source of deviance was essentially rooted in some kind of individual failure – whether such failure be located in an anti-juridical rationality, as in Enlightenment-based theories, or in some kind of constitutional defect, as in positivism at the end of the nineteenth century.

Already in the emerging criminological interest of the early social sciences, however – for example in the work of Guerry, Quételet, and later of course Durkheim – the main emphasis was placed on the processes that produce a criminality understood as a social fact. Each sociological theory of the social order will then produce a specific theory of deviance. Indeed, as we shall see, there seem to be two distinct contenders. On the one hand, a line of thought from Durkheim to Parsons to Merton conceives of crime as a product of structural factors. On the other hand, instead, an alternative line of thought considers crime as a kind of behavior grounded in culture, and which can be transmitted through social learning (from the Chicago School to the various strains of interactionism).

In the former, structuralist tradition, a monistic concept of social control coincided with an essentially individualist vision of the emergence of deviance. The latter, interactionist tradition, originally inspired by Georg Simmel’s theory, instead saw behaviors socially labeled as deviant as the outcome of cultural and/or normative conflict, thus linking the idea of deviance to a view of normative pluralism. This was a sociology of deviance that had already answered the objection that critics such as Colin Sumner (1994) would bring much later, according to which the relativity of a concept of deviance would inevitably bring forth the dissolution of that very concept (this is an objection that is better brought against a structural type of theory, because interactionist and relativist theories in reality started from the point made by such objections). Finally, the current period – moving from the deep crisis of sociological theories of deviance after the 1970s – witnessed a curious divarication of disciplinary orientations. There was, especially in the United States, a reawakening of nineteenth-century ideological inspirations, whether of the neo-classical or neo-positivist variety, that accompanied the resurrection, in the 1960s, of Nietzsche’s “pale monster,” the State. At the same time, however, a “culture of control” emerged that aspired to make deviance impossible by technological means, through an intervention in the “environment” of deviance and crime (Garland 2001a; Marx 2005). This was a novel orientation, in the sense that it seemed to break with the very twentieth-century notion of a relationship between control, consent, and democracy.

The Penal System between “Exclusion” and “Inclusion”

It is customary today to think of the penal system as one of “social exclusion” (Steinert and Pilgram 2003; Young 1999). This may be perceived as the penal system’s “real” function and outcome but certainly, especially in its very beginnings, it was not imagined as pursuing such a goal. This emphasis on “exclusion” – typical of a society permeated by a rhetoric of “democratic participation” – tends to obfuscate the extent to which penitentiary institutions especially were originally conceived as mechanisms of in-clusion, or in-corporation, I would say, into a social contract. This was particularly the case in republican or proto-democratic societies, such as the United States, in their beginnings (Dumm 1987 on de Tocqueville 1835, 1840, following Foucault 1975).

Post-structuralist and 1970s “critical” thinking pointed out that prisons and “ideological state apparatuses” in general had been “invented” within a broader societal effort of “making up subjects” (Matza 1969; Althusser 1970; Foucault 1975; Hacking 1986) or, in the more direct and transparent language of North American reformer Benjamin Rush (Dumm 1987: 88), “Republican machines”: citizens, that is, who will know how to govern themselves, this being a necessary prerequisite for a system based on self-government. In the pages that follow, we will devote quite a bit of reflection to such statements. What I would like to point out, however, is that the terms of any such “incorporation” into the social contract, into the social body, clearly respond to the conditions and conflicts most characteristic of that society and to the way in which social order is therein framed and conceived of.

A number of “classic” commentators, from de Beaumont and de Tocqueville (1833) to Rusche and Kirchheimer (1939), have pointed out the affinity between the main features of the penal system in a given society in a given period and the consideration that society gives its members, and especially its laboring members. According to Rusche and Kirchheimer, the valorization of labor would typically be connected with an attitude of inclusion and incorporation of the lowest strata of the working class into the productive process and society more generally. This would also be the main orientation of the penal system. The de-valorization of labor instead – in a situation, for instance, of high unemployment – would be usually connected with a concept of the penal system as exclusion, as a system at most of “warehousing” inmates.

This way of thinking about the relationships between the social structure and the penal system carries certainly more than a grain of truth. It is, however, at the same time, still quite mechanistic, because the definition of a given situation and of the policies required therein is never something objectively “given” according to strict economic standards, but is the discursive product of hegemonic processes in which political and economic elites’ “definitions of the situation” have a very important say. What is a “social crisis,” for instance, depends a great deal on the perspective of the one defining it (O’Connor 1987; Hall et al. 1978; Sparks 1992: 55–77). And from the perspective of social elites, a crisis is first and foremost an assault on their power, whether political or economic.

I would therefore submit that two situations might obtain from the standpoint of penality. In the “exclusionary” penal mode, society is (successfully) described as being in a state of “crisis,” where order needs to be re-established and the social fabric mended and brought back to unity after having been lacerated and torn apart. Here, it is often the metaphor of the State that appears: Leviathan as a purveyor of order and unity or, better, of unification (reductio ad unum) and hierarchy – as David Matza explained powerfully (1969). Because one of the main powers of the State is the power to punish (Beccaria 1764), penality is particularly apt to be used to define the powers and boundaries of sovereignty. In such a situation, the task characteristic of the system of criminal justice is one of bringing society to unity by eliminating fragmentation and anarchy.

In the situation instead where a tendency toward inclusiveness develops, this is because the social order is perceived as suffocating and unfair, and social change as necessary. The task characteristic of the system of criminal justice becomes then one of allowing for experimentation and “innovation” (in the Durkheimian sense of a type of deviance that triggers social and normative change).

How can we sociologically explore such oscillations? From a quantitative perspective, one could show, for instance, that the “productivity” of a penal system increases in situations of moral panic and crises (particularly when such crises are perceived by elites as threatening the dominant form of social relations, i.e. their power). We can see that incarceration rates tend to increase in situations of crisis (economic and/or political). Qualitatively, however, we can observe that the representations of the criminal offender change too, i.e. the representations of the criminal that society produces, and in which criminologists play a part (Melossi 2000a). By “representation” I mean the descriptive portrayal of criminals, in criminological discourse, in the public opinion, or in aesthetic discourse, as a distinctive “type” presenting identifiable moral, physical, and social characteristics according to specific locales of time and space (Leps 1992; Rafter 1997; Sparks 1992; Fritzsche 1998).

These two perspectives, quantitative and qualitative, are indeed related: the devaluation of the person who is at the center of the penal system, either as a criminal or as an inmate (usually seen by the public as synonyms), is related to a rise in the number of such persons. There is an affinity between those social processes that increase the number of inmates and those that change – for the worse – the representation of the criminal. Or rather: it is the same social process, in which the changed representations – in orienting social action – make it possible for the numbers to go up or down, and the numbers’ seesawing in turn feeds back on the quality of the representation (because if many members of a given category of people go to prison, this is taken socially to mean that those who belong to that category are indeed very bad, or inherently dangerous).

I have therefore advanced the idea that the sphere of penality, in its quantitative and qualitative variability, constitutes a sort of “gazette of morality” (Melossi 1993) through which a varying pressure is exercised on the generality of the public (given that, as we shall see in more detail, I follow Durkheim – and, for that matter, the classical theorists – in conceiving the main function of the penal system as being one of controlling society more than the criminals, who should actually be regarded as the “useful” “bearers” of such control; Foucault 1975).

In the situation characterized by a tendency to exclusion, we may observe in fact that criminologists (as well as public opinion and “aesthetic” productions, not to mention politics) assume an attitude of distance or antipathy toward the criminal: the deviant is the producer of evil (whether he or she wants it or not), social order is represented as a given order that is to be established or re-established, and the representations of the criminal are under the constellation of the monstruum, the monstrosity, far removed from any common experience and hence from the possibility of empathy.

In the situation characterized instead by a drive to include, criminologists (as well as public opinion and fictional accounts) tend to assume an attitude of vicinity or sympathy toward the criminal: the deviant is seen as in some sense a victim of society, social order is represented as justly or at least reasonably contested, and the representations of the criminal are located under the constellation of “innovation,” when not of a heroic and quasi-tragic striving against the tyranny of fate or social conditions.

Transformations in Concepts of Social Order and Representations of Crime

Essentially, what I intend to do in this volume is to develop, for the case of criminological theories, a conception inspired by Georg Rusche and Otto Kirchheimer’s views – views that, I would add, may be related to a “long cycle” or “long wave” perspective. This is an outlook according to which socio-economic development happens in long cycles where the “peak” and the “trough” are separated by about 25 years. This would be a concept more easily employed to understand what is an essentially cultural process of change, like that of penality, characterized by long-term movements instead of short “business cycles.” This approach conceives of cyclical movement as acted upon by economic and political actors. Innovation would then appear to be, for instance, a crucial instrument used by entrepreneurs in order to diffuse the power of labor in conditions of prosperity and profoundly change the type of economy in which the working class has reached a high level of power. Rusche and Kirchheimer’s concept of a linkage between punishment and the labor market could then be seen as one of the aspects of this broader picture. Imprisonment increases, and the situation within prisons becomes less tolerable, when economic and political elites are on the attack in order to hit the high levels of power reached by the working class. Then, after the re-establishment of their hegemony, imprisonment would again become “softer” and penality more bent on “reform.” This is what, in the last chapter of this volume, I call “the cycle of the canaille.”

All of this ought not to be conceived as an “understructure” “determining” a “superstructure” but, in Max Weber’s way, as relations of “elective affinity,” with long-cyclical movements produced by the discursive interaction of all the actors involved. Such movements are in fact strictly related to, and crucially facilitated by, changing representations. Particular representations tend to dominate within given cyclical periods. For instance, during prosperous and innovative periods of the cycle, a fragmented and rapidly changing society conceives of itself, through its production of social and criminological theory, as a plural and conflictual entity, where the concept of deviance, or indeed of crime, is relative to the standpoint of the one who is talking, and the representation of the criminal is a fundamentally controversial and contested representation. Some criminals at least get a “fair break” and are cast in the role of innovators and heroes rather than in that of villains. Examples of this kind of society are Europe and North America around 1900 and then again in the 1920s and in the “extended” 1960s. It is then no surprise that these periods are also characterized by declining imprisonment rates and by a public discursive rhetoric centered in social innovation, experimentation, and change.

In the opposite ideal-typical situation instead, there are long-cyclical periods when the fragmentation has reached “intolerable limits,” at least from the perspective of elites, and the need for a re-establishment of unity, authority, and hierarchy (even if under a somewhat changed balance of power) is of paramount importance on the political and cultural agenda. These periods suggest theories of social order characterized by an idea of unity and cohesion (I would call them “monist” theories) where the basic normative frame is consensually shared and where the image of the criminal is one of “the public enemy,” to use the FBI’s telling 1930s label. The criminal is now a morally repugnant individual (as described by criminologists as well as in “the public opinion” or in fiction), the one who brings a deadly threat to society’s moral order. The causes of such a threat, if at all relevant, are to be found within the criminal himself or herself, and not in social relationships (as was instead the case in the other ideal-typical situation). Examples are the periods when nation-states were first established in the nineteenth century, the world-wide Depression in the 1930s, and the period from 1973 till today, which Eric Hobsbawm has very aptly dubbed “the crisis decades” (Hobsbawm 1994: 403–32). This theoretical approach should then be particularly useful – as I will try to show in the last chapter – in order to understand the relationships between mass imprisonment after 1973 and the profound transformations of the American economy in the same period (usually described as “post-Fordist”), in particular the connections between the disciplining of the heavily racialized, deskilled manual labor force and the process of “McDonaldization” of American society (Ritzer 1993). Once again, it is no surprise that during these periods imprisonment rates tend to increase, and the public rhetoric is one re-emphasizing the value of the collectivity around concepts such as “the State,” “the Nation,” “the Community.”

I do not claim that the chronology and the impact of events are the same in all societies and cultures, or even all societies and cultures of the same type. There are time lags and there are specific cultural traditions that should be emphasized and taken into account (Nelken 1994; Savelsberg 1994; Melossi 2001; Melossi forthcoming). For instance, there is the issue of why some of the same processes in “post-Fordist” Europe could occur without mass imprisonment as the goad (the 2006 book by Michael Cavadino and James Dignan, Penal Systems: A Comparative Approach, offers relevant material to that end).

Even if it was not entirely original to him, Michel Foucault was probably the most effective proponent of the idea that, after all, imprisonment is a sort of instantiation of the criminal. “The penitentiary technique and the delinquent are in a sense twin brothers,” Foucault wrote (1975: 255). The “production” and “presentation” of criminals within prison is then applied to the criminal in general, to the criminal sans phrase, as the Italian Positive School especially made clear. Representations of crime and criminals follow the destiny of imprisonment in their uneasy relationship with economic, social, and labor conditions. Further, one has to remember Kai Erikson’s insight that representations of crime and punishment in the public arena are projections of deeper social and cultural concerns (Erikson 1966). Criminologists usually refer to social control as a “response” to deviance, and particularly to legally sanctioned deviance, “crime.” I believe any such connection between social control and crime to be a contingent matter, which depends on the specific nature of what is meant by “social control” and “crime” in a given socio-historical situation (Melossi 1994; Black 1984). During most of the twentieth century, the control of crime was subaltern to social practices largely based on the construction of consensus (Melossi 1990) and directed at controlling the generality of the public rather than the few who are responsible for what is officially defined as “crime.” In this sense, controlling crime competed with, and often took back seat to, controlling society. Indeed, controlling crime has often been but an instrument used in order to control society. This was also the view of many “classical” social theorists, most notably as clearly expressed by Durkheim:

[Punishment] does not serve, or serves only very incidentally, to correct the guilty person or to scare off any possible imitators. From this dual viewpoint, its effectiveness may rightly be questioned; in any case it is mediocre. Its real function is to maintain inviolate the cohesion of society by sustaining the common consciousness in all its vigour … Thus, without being paradoxical, we may state that punishment is above all intended to have its effect upon honest people. Since it serves to heal the wounds inflicted upon the collective sentiments, it can only fulfil this role where such sentiments exist, and in so far as they are active. (Durkheim 1893: 62–3)

Words not very different can be found in George H. Mead (1917–18) and, of course, more recently, in Foucault (1975). “Representations” of deviance, and deviants, perform the work, in a sense, of reconnecting “structural” change and change in penal policies in a way that the more economistic view of Georg Rusche and Otto Kirchheimer had left to some degree unexplored. Marie-Christine Leps (1992) effectively showed, for instance, in her reconstruction of the cultural climate in the nineteenth-century fin de siècle years, that each cultural environment in fact produces a given “knowledge” of the criminal that spans different discursive forms, from scientific tracts to newspapers to fictional accounts. Such representations perform a work in society, which consists, among other things, in orienting public morality. This work of representing and giving moral orientation can be accounted for according to a concept of the relationship between social structure and culture, which is one of “elective affinity.” Social practices that are usually ascribed to “structure” show affinity for social practices usually ascribed to “culture” (Weber 1904–5; Howe 1978). If rates of punishment, for instance, seem to “respond” to change in the economy, at least in given societies and periods (Chiricos and DeLone 1992; Melossi 1998a), this is not because of some kind of homeostatic “blind” mechanism – even if social theorists may end up so describing it – but because ideas expressed in a publicly available language about where “the economy” is going, what a “social crisis” is, what causes it and who is to blame for it, and so on, appear to change, in society, together with and in a tight cultural exchange with publicly available ideas about crime, punishment, and responsibility. Even if the social theorist’s description is simply a reduced form description (that is, a description that takes into account only exogenous and endogenous variables, without indulging in reflections about “intervening variables”; Hanushek and Jackson 1977: 227) ultimately the connection between such disparate aspects of social life is a work performed by discourse, i.e. by that discursive interaction that makes social co-ordination and therefore social practices possible. Representations are in a sense the answer to Marx’s proverbial riddle of the relationship between “structure” and “superstructure” (Marx and Engels 1845–6; Mead 1934; Smith 1976; Melossi 1985).

This kind of approach is bound to be eminently “reflexive.” Whereas what is usually called a “positivistic” attitude has powerfully contributed, especially in the last 20 or 30 years, to obscuring a “reflexive” view in criminology, sociological theory has instead tried to come to terms, for instance in the work of authors such as Anthony Giddens (1984), with the concept of reflexivity – with the challenge, that is, brought to a science of society by the scientist’s awareness of the necessary implication of his or her ideas, with the material that he or she strives to describe by means of those ideas. Not so in criminology. Or rather: the pronounced reflexivity of key works that appeared in the late 1960s and early 1970s, such as Matza’s (1969) or Taylor, Walton, and Young’s (1973), has brought the field too close to what one might call an act of euthanasia to be tolerated within the boundaries of a social science whose fate has always been intimately linked with the issue of political legitimation.