Book review: The Rewards of Punishment

Mathieu Deflem
www.mathieudeflem.net

This is a copy of a book review in the International Review of Modern Sociology 35(2):333-335, Autumn 2009.
Also aavailable as pdf file.

Please cite as: Deflem, Mathieu. 2009. The Rewards of Punishment: A Relational Theory of Norm Enforcement, by Christine Horne. International Review of Modern Sociology 35(2):333-335.



This book presents a theory of norm enforcement on the basis of experimental means of research in social psychology. The problem of norm enforcement is approached on the basis of the dilemma that norms are not always and not equally enforced and that some norms may persist even when they have a strong negative impact on communities and their members. Norms are hereby defined as rules that have a modicum on consensus (legitimacy) and that are socially enforced. The relational theory that author Christine Horne adopts in the book suggests that people will more likely enforce norms when they sanction someone’s behavior because of the relative degree of relevance they attribute to the social relations that are maintained with norm violators. In other words, people are more likely to punish when they conceive that certain forms of behavior have negative consequences and when the members of a group are more rather than less interdependent.

The relational theory of norm enforcement is tested in a series of experiments (which are explained in detail in an appendix to the book), involving various situations in which laboratory subjects can decide to enforce norms and punish people. First, a norms game is developed in which subjects interact with one another through networked computers and have the choice of giving punishments. Results of experimental research show that punishments increase as the consequences become larger. Additionally, sanctions were also found to be a function of the degree of interaction among subjects.

Second, Horne experimentally researches a metanorms situation. Metanorms are defined as norms that regulate sanctioning (the enforcement of norms), rewarding the behavior of those who enforce norms. The theory adopted in the book again suggests that metanorms are especially likely to be developed and enforced when people are engaged in interdependent relationships. To test this proposition, a metanorms game is developed whereby subjects can decide about punishing wrongdoing and rewarding those who punish. The results of Horne’s experimental research show that support to sanctioners is higher as subjects are more interdependent with one another. Besides interdependence strengthening metanorms, both interdependence and metanorms are also found to affect norm enforcement as subjects were more likely to sanction under conditions of high interdependence.

Third, the author develops a final set of experiments to address the enforcement of norms of expectations on issues that have no intrinsic negative consequences, such as fashion. Relying on the relational theory of metanorms, it is suggested that people will sanction behavior simply because it is atypical or out of the norm. Additionally, the punishment of atypical behaviors will be more likely when these behaviors are more meaningful. To test these propositions, Horne constructed an experimental design that involves an expectations game in which subjects are given the choice between two entirely meaningless options. Their choices turn out to be influenced by the behavior of others in that subjects conforming with majority choices more likely punish those that did not. Additionally, the sanctioning of atypical behavior is found to be more likely when that behavior has an a priori socially desirable meaning.

In the book’s final chapters, Horne takes the insights derived from her experimental research to the world outside the laboratory. For example, she applies the theory to norms regulating the sexual activities of males across various cultures. The prediction is made that sex norms for males will be more permissive in matrilineal societies because women are there less dependent on their male partners. Anthropological data support the proposition. Horne formulates other such applications of the relational theory to situations of informal crime control and the enforcement of international human rights. Oriented at informing policy, Horne finally speculates that strong legal systems weaken the cohesion or interdependence of communities because they provide an alternative to informal controls. Law would does not merely respond to weak normative controls, but contribute to weaken those controls.

This work makes an interesting contribution to the study of informal and formal sanctioning and reward mechanisms. At the level of interacting individuals, this study is informative on some of the dynamics of sanctioning behaviors. An important question, however, is not whether laboratory conditions can be applied to real-life situations, but which level of social conduct they can be applied to. Surely, the abstractness of theories developed and tested in the artificial settings of a university laboratory does have the advantage of such theories being applicable to a wide variety of social settings. However, the level at which the theories can be transposed is necessarily social-psychological and cannot readily extend, at least not without additional speculation and risk, to the level of groups and social structures.

“Norms are difficult to study in naturally occurring settings,” the author writes (p. 16). Given my own interests in the sociology of law as a social institution (and my resulting lack of knowledge in experimental social psychology), I must in this respect observe that such a statement makes sense only from a social-psychological orientation that adopts an understanding of norms in terms of individuals following, breaking, and enforcing rules of behavior. Yet, such a conception is baffling for sociologists interested in social and cultural structures, who have, at least since Durkheim, routinely studied norms at the level of the structures and processes of social institutions.

To be sure, the real-life settings Horne mentions where her relational theory can be applied are not without interest and significance. She mentions the example of the abusive faculty member who seeks to undermine colleagues in a desperate attempt to control the social environment and cover up a lack of accomplishments, alcohol addiction, or other weaknesses. Those of us who work in academic departments know the situation sadly all too well. But the relational theory does not bring to light the broader institutional framework, especially the cultural and legal contexts, in which such behavior is tolerated and/or enforceable to varying degrees.

Related to the book’s social-psychological orientation, moreover, I suspect that the opinions on the value of this book, from the viewpoint of its methodological orientation, will squarely fall in two camps. Those that like experimental research in the social sciences will like this book, and those that do not favor such methods will not like it much. The author addresses the usual concerns that are raised against experimental social-science research and also argues that the theories that are developed in laboratory research still need to be applied in other settings in the naturally occurring social world. Yet, while the abstractness of the relational theory is a clear strength, there remains an important unresolved issue of transposibility, not so much from the laboratory to the ‘real’ world, as from the social-psychological level of interaction to the structural level of social institutions. Rather than a defense of the use of experiments in social-science research, this work may instead be received as an affirmation that such methods are restricted to social-psychological questions. The formal requirements for theories to be abstract may not convince sociologists to abandon the study of the behavior of institutions in concrete socio-historical contexts, and it does not have to be otherwise. There are after all more than formal requirements for scholarship to be valuable as well as valid.