In this book John Davis explores the different conceptions of the human individual in recent economics. Among the approaches he covers are behavioural economics, experimental economics, social preferences approaches, game theory, neuroeconomics, evolutionary economics and Amartya Sen's capability approach. The key question addressed by his critical survey is how these approaches revise or replace the atomist conception of the individual that is traditional in economics. I will begin this review by outlining Davis's motivation and theoretical approach, proceed then to present his main theses about various theoretical approaches in economics, and conclude by presenting some general critical comments.

The starting point of Davis’ approach is normative political theory. Much of political thinking employs concepts and language borrowed from economics. Davis argues that while economists themselves are not strongly motivated by philosophical concerns, from the point of view of political theory it is imperative to reconstruct and evaluate the conceptions of individual that underlie economic theorizing. The basic argument is: if individuals are important in political theory, they should also be important in economic theory. This agenda creates the basic tension for the book: while most of it is engaged with theoretical debates in economics, the ultimate criteria come from outside economic research. Consequently, if the reader forgets, or disagrees with Davis's starting point, he or she is bound to feel in many places that Davis’ desiderata are artificial philosophical constraints that may not do much for the improvement of economic theory. I wish Davis had argued more extensively for the legitimacy of the challenge he poses for economic theories. It is not obvious that the conceptualization of individual that is developed for the purposes of economics should also serve as a sufficient basis for political theorizing. Davis does not really address the possibility that the concepts required for political thinking are incommensurable with those that serve best the epistemic aims of economics. It is possible that the conception of individual that we consider mandatory in normative contexts is a hindrance for the scientific development of economics.

Be this as it may, according to Davis any satisfactory conception of the individual has to meet two criteria. First, according to the individuation criterion, individuals have to be successfully represented as distinct and independent beings. Second, according to the re-identification criterion, individuals have to be re-identifiable in those same terms across processes of change. The latter presupposes the first criterion: a conception of the individual can pass the requirements of the individuation criterion and fail with the re-identification criterion, but not the other way around. One of the main claims by Davis is that some of the new theoretical approaches in economics do succeed in showing individuals to be distinct and independent but fail to show how they are enduringly so. Thus these approaches are not able to handle intertemporal choice or economic activities that presuppose enduring individuals.

According to Davis, the traditional homo economicus conception of the individual does not really explain individuality but merely stipulates a certain use of the term ‘individual’. Its individuation procedure is circular as it presupposes the individual – characterized by its preferences – as exogenously given. The postwar neoclassicism continued the commitment to this stipulative approach to individuals as it co-defined individuality and rationality by axiomatically requiring that preferences are ‘well-behaved’. Thus people are identified as single individuals formally rather than empirically. The new approaches to economics covered by Davis attempt to replace this conception, but they face two challenges. The first is preventing the individuals from fragmenting into collections of sub-personal multiple selves. The second is the ability to tell how individuals can be distinguished from the supra-personal social world. The main claim of the book is that most of the new approaches ultimately fail either in one or both of these challenges. While there is clear progress in economic understanding of individuality, the continuing attachment to the old homo economicus conception of the individual prevents these approaches to reach their full potential.

The first part of the book begins by considering two strategies of behavioural decision research: Kahneman and Tversky's prospect theory and the preference construction approach of Lichtenstein and Slovic. Prospect theory represents a rather conservative approach to revising expected utility theory. By increasing the realism of the psychological foundations of choice theory, this version of behavioural economics seeks to improve economics essentially on its own terms. However, Davis argues, this attempt is compromised by an attachment to atomism. Consequently, prospect theory faces a severe multiple-selves problem, particularly in the context of intertemporal choice. The preference construction approach, in turn, is more radical – and unsurprisingly less influential in economics. Nonetheless, it faces similar problems. According to Davis, it fragments the individual into multiple selves to the degree that individuality ceases to exist altogether. But the approach faces an additional problem: an ambiguity in ‘preference construction’ leaves it open as to what or who constructs the preferences. Thus the preference construction approach allows no obvious basis for talking about individuals as distinct beings, or for identifying them. It fails to tell how individuals can be distinguished from the supra-personal world of social interaction.

Chapters 3 and 4 discuss two further developments in behavioural economics. Chapter 3 starts with Bénabou and Tirole's attempt to solve the multiple selves problem within the framework of homo economicus. Davis argues that this approach ultimately fails to account for such things as self-control and social interaction. While Bénabou and Tirole make arguments for the unity of the individual, they undermine these arguments when they extend their game-theoretic analysis to social interaction. Similarly, the ‘libertarian paternalism’ advocated by Sunstein and Thaler fails to adequately integrate interaction into the analysis of individuals with multiple selves.

Chapter 4 takes on the strategy of introducing interaction directly into the individual utility function as social preferences. Davis first discusses Akerlof and Kranton's idea of identity as self-image that can be incorporated within the utility function via cognitive dissonance minimization. He argues that the notion of social identity employed by Akerlof and Kranton has its origins in psychological theory. It focuses on the given individual and asks how individuals align themselves through self-categorization with given social categories. This approach does not give an account of personal identity and it leaves open how individuals and social structures are mutually determining. Similarly, it recognizes that individuals can have multiple identities, but it does not have resources to deal with their possible interconnections. Thus, according to Davis, Akerlof and Kranton are not able to prevent individuals from fragmenting into multiple selves or to explain how they manage to maintain a sense of unity as single individuals. The social preference approach that incorporates non-egoistic preferences into the agent's utility function faces similar problems. Apart from reducing reasons to mere preferences and missing the impersonal standpoint that is crucial in philosophical accounts of altruism, these approaches also create a novel version of the problem of multiple selves. The individuals would have both the utility function (or self) that applies to the preferences specific to them and utility functions (or selves) that apply to the preferences of others.

The second part of the book turns to accounts that make social interaction central to the explanation of the individual and thus promise to escape the limits of atomism. In Chapter 5 Davis argues that game theory in principle involves the radical idea that individuals interact directly, but that classical game theory is unable to develop this idea satisfactorily because of its sustained commitment to the atomistic conception of individuals. On the other hand, experimental game theory, especially in the context of indefinitely repeated games, seems to threaten to dissolve individuality into sociality altogether. Davis also argues that experimental game theory raises doubts about external validity as the experimental subject might take into account the goals and expectations of the experimenter. These problems imply, in Davis’ judgement, that a more radical departure from the received view of individuals is needed.

Chapter 6 discusses two such game-theory based approaches to individuality. The first is Michael Bacharach's account of team reasoning where individuals are seen as having multiple selves (qua members of different groups) as their supra-personal social identities. The second is Don Ross's neuroeconomics approach in which people's different neural selves are relatively independent agents that play intra-personal games with one another that produce single unified individuals that play inter-personal games with one another. Davis argues that Bacharach fails to show that his multiple selves do not dissolve into the many different social groups with which they identify. In contrast, Ross needs to show that his sub-personal agents do not fragment into their many competing sub-personal selves. Davis concludes that both of these approaches satisfy the individuation criterion, but fail to show how individuals can be re-identified as enduringly distinct and independent, in other words they do not answer the second, supra-personal, challenge.

Chapter 7 goes back to Herbert Simon's attempts to explain ecologically rational individuals as adaptive and self-organizing. Davis goes on to develop a relational-evolutionary conception of individual that builds on Vernon Smith's market experimentalism and Ken Binmore's evolutionary game theory. According to Davis, these conceptions, finally, successfully address both the individuation and re-identification issues, at least in principle. When combined with agent-based modelling, they provide a basis for economics that could avoid both the problem of multiple selves and the problem of dissolution of self to social interaction.

However, Davis argues that there is still something missing from the relational–evolutionary conception of the individual. People's human capacities should also be seen as evolving, in other words, they ought to be seen as capabilities people develop in social interaction. These missing pieces are articulated in part 3, chapter 8, which introduces the capabilities approach and argues that it should be combined with a social psychological account of self-identity as self-narrative. Davis concludes the chapter by arguing that the capability approach gives a much better handle of human heterogeneity that characterizes modern societies than a traditional preferences-based account of mainstream economics.

Chapter 9 takes finally a more systematic look at notions of personal and social identity. The chapter begins with a discussion of Horst, Kirman and Teschl's model of endogenous identity determination and Darity, Mason and Stewart's evolutionary game theory model of racial identity norms. Davis argues that while both of these models are promising, they fail to give a full account of the dynamics of social identity. As a basis for a more comprehensive account, Davis introduces a distinction between relational and collective social identity. The first refers to role-based identities that identify the self as a certain kind of person, while the second refers to identifications of the self with a group or category as a whole. The relation identity emphasizes differences from others, whereas the collective identity focuses on membership and thus commonalities with others. Furthermore, while collective identity is based on third-person representations, the relational identity ultimately derives its meaning from reflexive first-person representations. Davies argues that only by taking into account both of these notions, and the specific processes involved in their construction, can dynamics of personal identity be captured as an evolving narrative.

The last chapter returns to themes that motivate the book and sketches how political theory would change if the capability-based approach to individuality was taken seriously. Among other things, Davis argues that if economic policy is ultimately designed to promote the aims of individuals, it needs to be reconfigured to accommodate the capability-based rather than a preference-based conception of the individual. One of the implications of this is that economists have to take more seriously things like human rights, individual dignity and value plurality.

As a whole, Davis's book provides a tour of quite diverse ways in which mainstream economics is changing and incorporating influences from the outside. This is one of the strengths of the book as Davis also manages to give an interesting testimony about the peculiar way in which economics interacts with other disciplines. As he observes, economics is much more open to outside influence than it used to be. However, it is still highly selective in appropriating and accommodating developments from elsewhere. In my mind, Davis's book is at its best in documenting this selectiveness. For example, the comparison between prospect theory and the construction of preferences approach is very illuminating. Similarly, Davis's discussion of different notions of social identity helps to show how economists’ disciplinary preferences affect what they adopt from other disciplines.

The main goal of the book is to make a case for the importance of the issue of individuality. In my judgement, the success on this front is variable. While in some cases he convincingly shows that something essential is missing, in some other cases the challenge seem to be out of place. For example, while Davis's criticism of prospect theory raises important issues, I found his criticism of the construction of preferences approach overstretched. The main focus of this line of research is on the subpersonal mechanisms at work in preference articulation and it does not make sense to ask for agency in these processes. It is not obvious that a psychological theory like this is meant to answer Davis's challenges about individuality at all. More generally, Davis is very eager to present quite strong conclusions about the abilities of various approaches to deal with the issue of individuality. However, in most cases his treatment of the approaches is not thorough enough to support such strong claims. It might have been advisable to cut down the number of approaches to be discussed and give more attention to those remaining.

The defence of the capability-based approach in political theory can be seen both as the book's motivation and as the third main goal. As I have already pointed out, this motivation could have been presented more clearly. Most of the book is engaged in methodological and theoretical issues without any direct connection to political theory and it is difficult to imagine what the connection could be. As a reflection of this, the last chapter does not really connect with the rest of the book. While it reminds the reader of the ultimate motivation of the author, it covers too many issues to really provide new arguments in support of the capability approach.