Reviewed by
Ryan Muldoon
University of Pennsylvania
Modelling Norms opens with an intriguing question: what is the right question to ask when we wish to understand social behavior? Is it why people cooperate? Or why people break the rules of their society? Each question strikes us as important, but as authors Corinna Elsenbroich and Nigel Gilbert convincingly argue, each question is only partial. Instead, they argue, we should investigate these questions together, by developing our understanding of social norms. Social norms offer each basic question a vantage point that unifies their domain of inquiry, while at the same time allowing social science to focus on the problem of going from micro scale preferences and attitudes to macro scale patterns of behavior. The authors go on to argue that when we study social norms, one of the most powerful tools that we can deploy is agent-based modeling. This book can be broken into three basic parts. Chapters 2 and 3 offer the reader a systematic look at the state of the literature on social norms and criminal behavior, respectively. These chapters provide the theoretical motivation for modeling, and are paired with a very nice chapter 4, which introduces the reader to the basic concepts behind agent-based modeling. These initial chapters offer the reader a clear picture of the framework within which the book proceeds. The next five chapters are all short reviews of existing social norms modeling literature, organized around particular themes - environmental factors, the role of punishment, imitation dynamics, social settings of norms, and internalization. Each chapter reviews a small selection of papers, and concludes with a common "Achievements and Shortcomings" section which aims to draw out some lessons learned from each modeling approach. The final 6 chapter present the authors’ own views on how social norms ought to be modeled, starting with a very necessary chapter 10, which aims to offer a systematic critique of the social norm modeling literature, and offers a positive account of how norms ought to be modeled. The next four chapters present models that aim to rectify what they authors take to be shortcomings in previous work. In each chapter, the authors walk through both the theory presented in the first part of the book, and how to compose ideas from several models into a single model to better capture complex phenomena. In this ambitious last section, we see topics ranging from juvenile delinquency to the social construction of knowledge, to morality and we-intentionality. Each chapter discusses building up a model out of components, offers a summary of results and observations, concluding remarks.
This book broadly succeeds in its aims. It serves as a very nice introduction to the literature on modeling social norms - one I could easily see being productively deployed in a course on the subject. Of particular value is that the authors devote ample time to detailing the relevant accounts of social norms, so the reader can better see how and why some modeling choices were made in the models that were reviewed. The book, especially given its short length, does a nice job balancing between discussing the theory of norms and the particular challenges brought on by modeling social phenomena.
Though I think this book is a valuable resource, I do think it has some important flaws. I will highlight two. From a social norms perspective, I am puzzled as to why the dichotomy that the book opens with is then elaborated by using crime as the paradigm of norm-breaking. Criminal behavior in some contexts may well be norm-following. The authors distinguish between social and legal norms in the introduction, but then ignore it in the text. This narrows our thinking, by implicitly painting social norms as pro-social, and norm breaking to be anti-social (and illegal). Many norms are harmful. Many pro-social norms are not laws. That’s part of what makes social norms important to study - they guide our behavior in ways that aren’t codified in our formal institutions. By conflating social norm violation with legal norm violation, this important aspect of norms is washed away.
From a modeling perspective, I would have liked to see more detail in the presentation of the authors’ models. Insofar as they are to be exemplars of what chapter 10 argued for, I wanted more detail on the models themselves, their results, and even the process used to develop them. Chapter 10 highlights the authors’ preference for KIDS instead of KISS, but I would have liked to see some nuance to the discussion, particularly when considering differing goals of modelers. I was also dismayed to see no real discussion of multi-model approaches, or robustness analysis, given we are dealing with complex adaptive systems.
Despite these criticisms, the book is both a pleasant read and a very nice introduction to the literature.
Return to Contents
of this issue