In little more than a year, three new books on Amartya Sen and his ideas have been published. After all that has been written already on Sen, one may wonder if that is not too much. As to these three books – all edited volumes containing a selection of papers of well-known researchers in the respective fields – the overall answer clearly is no. Each volume has different strengths, which I will try to highlight below. However, a review of three books necessarily involves concentrating on a limited set of contributions only, and the selection reflects in part my personal interests.

Amartya Sen, edited by Christopher W. Morris (AS from now on) presents an overview of the many different strands of thought and research areas to which Sen has contributed over the years. It contains eight chapters corresponding to eight different research areas and, still, some of Sen's interests are not represented. One important aspect of Sen's writings is his critique of the requirement of internal consistency of choice and completeness in relation to rationality. This is the subject of Dhongde and Pattanaik's article ‘Preference, choice and rationality: Amartya Sen's critique of the theory of rational choice in economics’. Famous is Sen's example of the fruit basket that is passed around at a party and in which only one apple is left, which the person, wanting to behave decently is not choosing. However, had there been another basket with two apples, the person would have chosen the apple rather than nothing. This behaviour violates consistency conditions such as WARP or Condition α and is therefore usually considered irrational. Yet Sen has always been sceptical about formulations of rational choice that are framed exclusively in terms of choice, without reference to anything ‘external’ to the choice such as underlying preferences, goals or motivations. Indeed, in this example, there is nothing irrational if we take into account that this person is not merely maximizing her self-interest, but also acting decently. One solution is to specify the space of options in a way that it captures all motives and concerns of the agent. If we do so, the inconsistency disappears. The difficulty with this of course is, as Dhongde and Pattanaik underline, that one would need to know what the agent cares about, that is her objectives and motives. But this is exactly what the revealed-preference theory of rational choice tried to avoid.

Dhongde and Pattanaik proceed to discuss Sen's critique of the preference-based theory of choice. This concerns primarily the discussion of self-interest that is supposed to underlie all individual actions. Dhongde and Pattanaik remind us, and so does Christopher Morris in his chapter on ethics and economics (in AS), that Sen usually distinguishes between three different notions of ‘privateness’: self-centered welfare, self-welfare goal and self-goal choice. The first concept means that a person's welfare only depends on her own consumption. The second one means that a person's goal is to maximize her welfare. Here, the person may feel ‘sympathy’ with others and thus the welfare of others enters indirectly as part of the welfare of the person herself. Finally, self-goal choice means that each action is guided by the person's own goals and not by those of others, for example. These concepts of privateness are generally accepted in economics. In fact, Dhongde and Pattanaik regroup self-welfare goal and self-goal choice under the heading of ‘self-welfare choice’, referring to the fact that in economics these concepts are used to mean that whatever the person does, it is to further her welfare and thus her self-interest in a very broad sense. Hence advancing one's self-interest is typically considered to be the rational thing to do. To see everything as self-welfare choice is probably also the reason why so many models in economics start with a given utility function that the agent is simply maximizing, even if this utility function contains several different goals or motivations. This presupposes that the person was able to form an all-things-considered (ATC) preference ordering, for which she was able to determine the trade-off of different motivations (Baigent Reference Baigent1995). However, whether such an underlying ATC-preference ordering exists is a big question, which generally is not discussed (Teschl Reference Teschl2010; Arlegi and Teschl Reference Arlegi and Teschl2011).

Sen thinks that self-interest, how broadly defined it may be, is not the only motivation for behaviour. This also means that he does not think that the maximizing conception of rationality is the only one available. For him, the three concepts of privateness introduced above are quite independent of each other. A person may violate the self-welfare goal (i.e. not maximize her own welfare) but still comply with self-goal choice (i.e. advance another goal of hers that differs from her own welfare). On the other hand, it may be possible to choose something that goes against self-goal choice (i.e. by choosing according to the goals of someone else, for example), but nevertheless advance her self-welfare goal. The latter possibility may sound confusing because if a person violates self-goal choice, how can she advance her self-welfare goal? But this is not so confusing if we look, as Sen does, at the Prisoner's Dilemma game: if the person did act according to her self-welfare goal, she, and her counterpart (assuming he or she has the same goal) would end up in the Pareto-dominated situation. It is well known that if they cooperated, the two players could achieve a better outcome for both. But how are they made to cooperate? It is the crux of the prisoner's dilemma that standard maximizing rationality does not bring people to cooperate. Hence a different rationality is required, which Sen calls ‘commitment’. This concept is explicitly discussed in Morris's article, but funnily enough not in Dhongde and Pattanaik's chapter (in AS), even though they discuss it conceptually without mentioning the term commitment. A typical commitment would be to act according to social norms or rules of conduct. Morris reminds us that the maximizing conception of rationality has difficulties with norms or rules. Either the norm says that one should choose the best alternative – but then it tells nothing new and goes in the sense of maximizing rationality. Or it prescribes to choose a less preferred option – but this seems to be irrational with respect to maximizing rationality. Commitment however makes norm following rational.

The idea of commitment is tricky and Morris's chapter pays tribute to this. Several issues can be associated with it. One is that acting with respect to one's commitment may sever the link between welfare and choice, i.e. it may go against self-welfare goal. This of course goes very much against the spirit of most analysis in economics, which, as mentioned above, supposes that all actions advance people's interest and thus welfare in some way or another. The question then is, as highlighted also in the chapter by Dhongde and Pattanaik, how is one supposed to assess the overall well-being of a person. One answer is to say that a person's well-being is not to be assessed in terms of ‘utilities’ or preference satisfaction, but in terms of her functionings and capabilities. If commitment reduces a person's welfare in terms of preference satisfaction, it can still contribute to achieve what the person would like to be and to do (that is, a valued functioning), and is in this sense a positive value. It can do so because the person is pursuing another of her goals (and thus satisfying self-goal choice), or it can do so because the person is following a norm or acting in accordance with some other people's goal (and thus violating self-goal choice as well). But we may raise the question whether this is a satisfactory way of assessing well-being in this particular case.

The prisoner's dilemma game has the advantage that any violation of self-goal choice is leading to a better outcome for the person. But what if interacting with others leads to outcomes that are inferior? This question needs to be asked, especially in light of Sen's important discussion of ‘adapted preferences’: where can we draw the line between a situation in which a person chooses an inferior outcome on the basis of adapted preferences and a situation in which a person is ‘committed’ to a particular norm? While there are norms, rules and behavioural obligations that facilitate cooperation and thus life with other people, we can also imagine prescribed conduct, which may not. The question is whether this distinction can always be drawn in terms of an assessment of functionings and capabilities. These issues are unfortunately not taken account of in either of the two articles discussed above.

This leads to another important question. With reference to an article written by Philip Pettit (Reference Pettit2005), Morris wonders to what extent norms alias commitment may lead to preference change. We need to consider several things at this point. We may think that people have two preference orderings. For example, one ranking represents the goal of maximizing welfare, the other is a ranking of options according to a particular norm. Suppose that people are able to form an ATC-ordering, i.e. they are able to form a single preference ordering which includes the norm. Consequently, nothing prevents them to choose the best available option given their preference ordering. But Morris points out, if agents already have ‘all-things-considered’ rankings when they come to face, say, a prisoner's dilemma situation, they will converge on the Pareto-inferior outcome (p. 54). Hence, according to Morris, ATC-preferences alone (without a change in rationality) do not overcome the dilemma. Morris then goes on to consider what happens if preferences are ‘virtually’ all-things-considered, which would mean that the norm is not yet included into the preference ordering. In this case, the inclusion of the norm may well lead to a preference change (what Pettit called ‘goal-modifying commitment’). But for Morris, this is not a different form of rationality, because people would, in the end, still apply their maximizing rationality, but based on changed preferences. Morris wants to show that goal-displacing commitment is possible (which Pettit thinks is not) and proposes in particular the ‘resolute choice model’ (p. 55). This is a situation in which people's preferences do not change, but people decide to act on the basis of the norm rather than on the basis of their preferences. For Morris, this is a typical situation of ‘counterpreferential choice’ (p. 56), which is how Sen originally defined commitment. On this account, commitment, that is norms, are the reasons for actions and not preferences.

Notice that even though the norm gives rise to a different ranking of the available options, this ranking does not correspond to the person's preferences. Indeed, Sen has distinguished between a person's preferences as they are typically assumed to be in a prisoners’ dilemma game, and his or her ‘as if’ preferences that would be observed if the person decided to choose on the basis of a particular norm. For Sen, a prisoner's dilemma game would be transformed into an assurance game if these ‘as if’ preferences were taken as the person's ‘true’ preferences. The question then is, why, if people recognize the advantage of behaving according to the norm-ordering, they would not eventually form an ATC-ordering that pays tribute to this. When Morris discusses ATC- preferences and suggests that only the standard non-Pareto solution would be achieved, he seems to think that with ATC-preferences, the self-welfare goal would come above the self-goal choice, i.e. that maximization of one's own welfare dominates norm following. But it is not clear why and when this is indeed the case. But if this is the case (the norm-preference ordering dominates the self-welfare-goal-ordering and an ATC-ordering can be formed), it seems to suggest that commitment, after taking everything into account, leads to maximizing rationality after all. As mentioned before, in the case of the prisoner's dilemma, norm following increases welfare. But this may not always be the case. Most of us would think that it is rational to follow norms, despite the fact that this would not give the highest possible welfare. In fact, one way to incorporate norm following is as an ‘external reference’, i.e. as a goal or a motivation the person has or cares for. The ‘external reference’ literature suggests that norm following can be rationally explained by accepting the violation of some standard consistency axioms and by introducing others instead (see e.g. Baigent and Gaertner Reference Baigent and Gaertner1996). Hence, finally, it seems that the question may be less, as Morris seems to think, whether commitment is a different form of rationality or not, whether preferences change or not, or whether commitment is a plausible solution concept for ‘games’ played among people, but rather what axioms a reasonable and socially embedded person may satisfy. This has not been discussed. Hence, despite the very good chapters by Dhongde and Pattanaik as well as by Morris, the rationality issue requires further clarification.

Morris's book Amartya Sen contains many more interesting discussions of Sen's philosophy and economics, but I now turn to Measuring Justice (MJ from now on), a book edited by Harry Brighouse and Ingrid Robeyns and in which different authors compare primary goods and capabilities. As is well known, John Rawls's (Reference Rawls1971) A Theory of Justice says that for the evaluation of any state of affairs, one should look at and compare the individual holdings of social primary goods. These are all-purpose means or resources that people need to carry out their life-plans. The capability approach developed by Amartya Sen, and also Martha Nussbaum, made an important critique in relation to Rawls's informational basis of justice: according to Sen, the primary goods approach takes too little notice of the diversity of human beings. What matters is not so much what and how much people possess, but what they are able to do with it. The capability approach therefore focuses on people's functionings, what they do and are, and on their capabilities, what they can do and be. In short, what Sen thinks is important to look at is the substantive freedom people have in leading the life they can have reason to value. The conversion of primary goods into freedoms may be different for people even if they hold the same amount of primary goods. For example, a lactating woman will need more food intake than another woman ceteris paribus, a one-legged man will not be able to ride a bicycle, even if he got one. Sen thinks that omitting to consider these inter-individual differences may lead to unjustified inequality and unfairness. While both approaches need to assess what people have, the capability approach takes possession only as a means to freedom and not as an end as the ‘resourcist’ or primary goods approach would do.

This crucial distinction between the capability and the resourcist approach leads to another important demarcation, regarding what can be considered to be a justice claim. Rawls's view was, as Thomas Pogge and Erin Kelly make clear in their respective contribution to the MJ volume, that the standard or normal needs and endowments of human beings are the reference value for a theory of justice, whilst the capability approach considers people's specific or heterogeneous needs and endowments. Pogge's contribution is to show how many heterogeneous needs of people can actually be taken into account from a resourcist point of view and thus demonstrates that the resourcist approach can be taken beyond the original scope of Rawls. These include needs that stem not only from variations in social climate and environmental diversities, for example, but also from personal heterogeneities such as whether a woman is breastfeeding or not, age-specific differences and more generally all those diverse needs that are socially caused. Pogge shows that the resourcist approach can be, and actually should be, sensitive to these differences. The difference between Rawls's and Sen's approach finally concerns only those needs that arise from natural diversity which may be a combination of any of the following three factors: ordinary genetic variations, self-caused factors and differential luck (p. 32). This however means, as Elizabeth Anderson and Richard Arneson, the defenders of the capability approach, point out in their respective contributions, that Pogge, or the resourcist view, considers the socially caused disadvantages as more important than the naturally caused ones.

In the social justice literature, there is a debate about where the ‘responsibility cut’ lies, i.e. the line that separates disadvantages that deserve a form of collective assistance distributed over many people (e.g. in form of monetary payments that stem from taxpayers’ contributions), from those disadvantages whose consequences must be solely borne by the individual person (see e.g. Dworkin Reference Dworkin1981; Arneson Reference Arneson1989; Cohen Reference Cohen1989). Pogge's critique of the capability approach is that it sets the responsibility cut too low as so many differences among individuals seem to lead to justifiable compensation claims. This has also as a consequence that for capability theorists claims for compensation are justice claims, while for resourcist theorists, a substantial set of compensation claims, namely those stemming from ordinary genetic variations, self-caused factors and differential luck, cannot appeal to justice but only to other moral virtues of persons, such as duties of humanity or solidarity as Pogge would say, or duties of assistance as Kelly would call them.

How can one or the other position be justified? Resourcists like Pogge eagerly point at two considerations: first, a justice claim for compensation requires money, and who will bear the cost? Hence, justice claims need to be restricted. Second, the advantage of the resourcist approach over the capability approach is that it has a clear public criterion of social justice that indicates what can be compensated and what not. It is not clear which capabilities are valuable, how their achievement is measured, and what happens if the simultaneous achievement of several capabilities is impossible. This is not an uncommon critique of the capability approach. It is often followed by the demand for a more precise definition of valuable capabilities, a clear ordering of capabilities and how they can be compared inter-individually. It is well known that Nussbaum has presented a list of valuable capabilities while Sen refuses to draw such a list and only sees the need of partial orderings of capabilities, leaving it to public discussion to decide which priority to give to which capabilities. But this is not enough for Pogge, who goes as far as demanding a complete interval ranking of capabilities.

Erin Kelly adds another consideration that should justify the advantage of the resourcist view over the capability approach regarding the responsibility cut. She reminds us that Rawls's conception of justice is relational or ‘procedural’, that is, justice is ‘[. . .] conceived as a virtue of relations that persons stand in when they cooperate with one another’ (MJ: 64). Therefore, ‘[p]ersons who fall below the normal range of functioning are not cooperating members of society. Their claims therefore, fall outside the domain of principles of justice, which regulate institutional schemes of social cooperation’ (MJ: 63). She writes, referring to Rawls that ‘[t]he distribution of goods, then, depends, on what, how much, and by what means persons decide to produce’ (MJ: 65). Cooperation thus seems production related, or more generally, work related. This may be acceptable to those who have jobs that are fulfilling, pleasant, interesting and satisfying. But in a world in which many people, even in economically more developed countries, still have to work under unpleasant or severe conditions, have to do monotonous tasks, are commanded or maybe even bullied by superiors and other more powerful people, and who have as their sole source of income their human capital and thus are, strictly speaking, forced to work to survive, it is harsh to think that this is the price they have to pay in order to qualify for justice. In Kelly's world, those who cannot ‘engage productively with others in society’ (MJ: 64) cannot rely on justice, but only on other ethical obligations. This seems to be a very strict view indeed.

Kelly opposes the ‘procedural’ type of justice to the ‘allocative’ type of justice, which is ‘non-historical’ in the sense that it does not attach any importance ‘to the cooperative social scheme by which goods are produced’ (MJ: 65), but rather follows a particular distributive pattern or wants to attain a particular result. She qualifies the capability approach as belonging to this allocative type of justice, an approach that can be defended she says, but such defence cannot use values such as reciprocity or mutuality, because these are the values that Rawls uses to justify his resourcist account.

Anderson gives such a defence of the capability approach in her contribution to MJ and Kelly may have overlooked this. In earlier work, Anderson has offered a capability-based theory that she calls ‘democratic equality’. In her theory, demands of justice are tied to the concept of citizenship. ‘Citizens have a claim to a capability set sufficient to enable them to function as equals in society’ (MJ: 83). The demands of equal standing in society are of course much more inclusive than the demands of productively engaging people; and they are independent of the history of people. There are no requirements other than the fact that they are members of a particular society. For Anderson therefore, ‘[d]eprivation of equal standing is always a concern of justice, not merely of humanity’ (MJ: 93).

Kelly also thinks that allocative theories require a strong commitment by people wanting to promote the well-being of all. She doubts however that productively cooperating members of society will want to share the fruits of their efforts with others (MJ: 66). Yet if it is the case that people do not want to share the fruits of their efforts, one wonders why they would feel the ethical obligation of assisting those who are not participating at the collective production and who may need extra help. A number of experiments on this subject have shown however that people are willing to redistribute. Actually, people tend to redistribute more to those who have lost out due to brute luck or who suffer from adverse social circumstances, than to those who show less effort and who voluntarily engage in risky decisions. This means that there are non-production related variables that cause more support for redistribution than production related ones (Chanel et al. Reference Chanel, Luchini, Teschl and Trannoy2011). Kelly's point can therefore not be considered a critique of the capability approach, but rather shows that people are differently committed to redistribution. Where this difference is coming from is possibly a question of moral psychology, but it can certainly not be considered a weakness of the capability approach. In fact, it may well be that the resourcist approach and the capability approach serve as self-selection devices to people with varying degree of commitment to redistribution.

Another hotly debated topic is whether or not the capability approach follows a comprehensive doctrine or value system. At times, this debate seems to be artificially inflated, because whether or not this is true depends very much on how the argument is formulated. The original critique came from Rawls himself who made a distinction between a political conception of justice and a comprehensive moral doctrine and said the former should be independent of the latter. But the reference to valuable capabilities that people can have reason to want seems to refer exactly to such doctrines. However, as even Pogge acknowledges, Sen does not present any particular comprehensive doctrine. Pogge reminds us of what Sen himself said, namely that if the resourcist view did not refer to such general goods such as political liberties, income and wealth, but to more concrete goods such as being governed by pious Catholics or having access to Verdi performances, the same critique could be extended to the primary goods approach. On the other hand, the capability approach may be formulated very generally without reference to any particular doctrine and would thus escape such critique.

Only Kelly continues to argue that a capability approach proposes a comprehensive doctrine. Her own argumentation however appears equally controversial. Overall, she questions whether the achievement of equal capabilities is actually such a good thing. It might be that equal capabilities produce a conflict between political liberties and liberty of conscience and the question then is if and how the conflict should be solved. For example, writes Kelly, it could be that religious practices and traditions restrict the worth of political liberties to female practitioners by trying to discourage women to use those liberties (MJ: 68). For Kelly, this is to a certain degree acceptable, as long as women do not mind, because that is the price they are prepared to pay for their freedom of conscience. Interference to prevent such discrimination would threaten the liberty of conscience according to Kelly. This therefore means that in her view, a liberal theory of justice should not try to interfere with this conflict as long as its outcome, unequal capabilities, is acceptable to those who are worse off, i.e. women in this case. She uses this argument to charge especially Nussbaum with presenting a comprehensive doctrine. Nussbaum, in several of her contributions to the capability approach, has defended equal capabilities in all areas of life, including religious practices and has argued that any differential treatment should be eradicated. Hence, trying to solve such a conflict appeals to a particular moral system according to Kelly, and does not conform to political liberalism.

This is a typical liberal argument, but one may still think that it should be possible to be liberal and not to accept gender inequality. We know that language informs people's value systems, we know that social positions, including observing people in particular positions, informs people's values and this impacts on how they then interact with other people. Even Sen (Reference Sen1985) himself, in a different, but related context, has pointed out that in traditional societies, where women do not have the right or the possibility to gain access to the labour market and to earn their own living, they and others will perceive women to have a lower status than men who are the ‘breadwinners’. But gender inequality of course threatens the basis of self-respect, as Nussbaum also argues. It seems therefore, that if one is supporting a liberal theory, one cannot say anything against such inequalities without being charged of supporting particular sectarian values. If this is the case, it is certainly very odd for a liberal theory.

The crucial point in all this is the assumption that those who have less or are in a more unfavourable situation accept this. Yet in this context, one must raise again the question of adaptive preferences. It may well be that in certain situations psychological mechanisms come into play that make people accept their lot. Or it may be the case that there is only a restricted set of options and people learn to put up with that. Is it then acceptable, for example, that women engage less in civil society because of the internal rules of their faith? Kelly translates this question into asking whether it would be permissible to trade liberty of conscience, freedom of association or the free exercise of religion against an increase in political and economic power. She argues that one would have to consider this from the perspective of the original position, situated behind the veil of ignorance. From there, nobody would want to accept such an exchange. Kelly reminds us that Rawls thought that denying religious freedom would be oppressive.

But are we really talking about denying religious freedom? The discussion here is more about whether one should accept gender inequalities in religious practices that carry over to inequalities in political and economic powers. The question is if women with faith practice their religion despite gender inequalities because that is the only option they have, or if they believe that it is part of that religion that women have a lower status than men and thus cannot execute the same functions as men. It may be, for example, that many women in the Catholic Church would not mind having female priests and bishops, but given that a male-dominated church is against that, most women are simply putting up with the current gendered hierarchies. If that were the case, they may not see interference as a restriction of their liberty of conscience or free exercise of religion. They may regret that external interference needs to resolve this issue, and that the church itself cannot become more open to these ideas, but this is a different matter. Another question is whether it would be possible that behind a veil of ignorance, one would defend the principle of equality in a political sphere, but at the same time accept inequality in a different sphere, such as religious practice. In fact, one may rather think that because people behind the veil of ignorance are not aware of the social position they will be born in, and whether or not they will hold any particular comprehensive view of the good, they would not accept whatever sort of inequality. If equality is ‘good’ in terms of equal basic rights, why should it be less ‘good’ for particular comprehensive doctrines other than by assuming (or by adapting to believe?) that there are some intrinsic differences between men and women that do not allow them to engage in the same activities? This all means that the problem of adaptive preferences is not resolved by taking a stand behind the veil of ignorance such as Kelly supposes.

Martha Nussbaum has written an interesting contribution to the above debate in the third book Against Injustice, edited by Reiko Gotoh and Paul Dumouchel (AI from now on). In general, Gotoh and Dumouchel's book has the ambitious aim to present ‘The New Economics of Amartya Sen’ as the subtitle says. It follows a conference that was held in Kyoto a few years ago and presents a collection of papers that extend and discuss Sen's thought. Nussbaum's article singles out some of the issues that need to be reassessed in the theories of justice in the Western tradition of liberalism, to which Rawls's primary goods approach obviously belongs. She argues that these theories contain ‘gross structural inadequacies where women's issues are concerned’ (AI: 95). First, she stresses that the private–public distinction should be criticized. Liberal theories usually restrict their application of justice to the public realm, but not to the private one and thus overlook many injustices. Second, she criticizes an inadequate conception of equal treatment. ‘If I begin as a peasant and you begin as a king, and then we are treated exactly the same, I will probably continue to lag way behind you [. . .]’, she writes (p. 101). Third, she singles out the fact that ‘deformed preferences’ (p. 104) need to be recognized. And finally, theories of justice need to deal with the problem of care. Hence liberal theories need to be extended and revised to include these issues to satisfy notions of justice to which many people nowadays are committed. In fact, it seems that as much as available theories of justice influence the way people think, other, new ideas and values continue to inform and change people's perception of justice too, so that in the end, the existing theories of justice need to be updated. This is in stark contrast to Pogge and Kelly's view of the timeliness of Rawls's original contribution.

Finally, I turn to Marcel Hénaff's interesting chapter on ‘Gift, market, and social justice’ in AI. This chapter is in some sense complementary to Nussbaum's as Hénaff is discussing the issue of recognition and social justice in different spheres – public, common and private. The public sphere is the embodiment of the relationship that is instituted between citizens and sovereign power and contains the rights and obligations of each individual as citizen. Like Nussbaum, Hénaff thinks that humankind has made a lot of progress in terms of social justice in this sphere. However, critical issues of social justice need also to be resolved at the common and private level. The common sphere is that of the social interactions at a daily or frequent level, like everyday work and neighbourhood relationships, religious, athletic, cultural and other social activities. The private sphere concerns the space of personal life with all its personal relationships and interactions. Hénaff opposes those who argue that the struggle for social justice is now followed by the struggle for recognition, i.e. the instauration of equality among different cultural, religious, professional and gender identities. He rather sees recognition as profoundly involved in any kind of social justice.

Hénaff traces recognition back to ritual gift-exchanges and thus presents an illuminating counter-interpretation to Marcel Mauss's famous account in The Gift. Hénaff sees the reciprocity that is necessarily involved in gift-exchanges transformed at an institutional level into a right of equality. For him, gift exchange has evolved within a legal realm and not an economic one as Mauss would have thought. Hence equality (and thus social justice) cannot be seen simply at an economic level as equality of income, for example. It must be understood in terms of equality of capabilities, which includes many more dimensions. In particular those actions and states that are necessary for self-esteem. This is how Hénaff can reconcile social justice with recognition. ‘Self-esteem requires that respect be offered and esteem received’ (p. 135). Hénaff thus succeeds in bringing a new issue to the forefront of the social justice debate because he thinks that what is at stake is the dignity of the person and not only the equality of capabilities.

So here are three different edited volumes on the work of Amartya Sen, which contain a number of interesting and thought-provoking articles. They reflect the breadth of Sen's contribution to economics and philosophy. Do they present, as Gotoh and Dumouchel try to do with AI, a picture of a ‘new’ economics, namely a ‘Senian’ economics? I would not go that far. On the one hand, with respect to the capability approach, Sen succeeded in stimulating and influencing an enormous amount of work. On the other hand, despite Sen's celebrity and success as an economist, it is interesting to see that in comparison with the huge amount of behavioural literature that also deals with irrationalities, Sen's contributions have only been taken up to a limited extent by other researchers (see e.g. Baigent and Gaertner Reference Baigent and Gaertner1996; Bhattacharyya et al. Reference Bhattacharyya, Pattanaik and Xu2011). The reasons for this are unclear; maybe it is because ‘Sen's economics’ would really require an overhaul of long established ways of thinking about consistency conditions in economics and thus about rationality in general. Hence to fully grasp Sen's influence on the economic thought and to understand what Senian economics would look like, more would need to be said.