There can be no doubt that Brian Barry has made an enormous contribution to the clarification of the ideas of justice current in contemporary political thought. In Barry's Justice as Impartiality he explicitly distinguishes and sets in competition three models of justice: justice as mutual advantage; justice as reciprocity; and justice as impartiality (the ‘rational’, ‘fair’, and ‘reasonable’ of my title), and he argues that we should prefer the last of these. What I want to do here is to consider four questions. First, what is this competition a competition about? Second, has Barry adequately characterized the contenders? Third, can the competition be won on the grounds Barry suggests? Fourth, is it a competition that we should want to be won by a single theory? By contrast I want to argue that there are advantages in retaining a pluralist perspective in which all three approaches remain in play.

It is a commonplace that in many societies people adhere to profoundly different conceptions of the good. Given this we need to know what political principles are appropriate. How can we treat people who are committed to different accounts of the good with fairness? One recent answer to this pressing question is given by Brian Barry in his important work Justice as Impartiality. This book, of course, contains much more than this. It includes a powerful and incisive discussion of several accounts of distributive justice (‘justice as mutual advantage’ and ‘justice as reciprocity’), a critique of other attempts to defend liberal neutrality and a rebuttal of those who are critical of the ideal of impartiality. In this paper I wish, however, to focus on Barry's defence of liberal neutrality. The paper falls into three parts. Section I outlines the thesis that Barry wants to defend and gives a brief sketch of the argument he employs to defend it. Barry's argument makes two claims – what I have termed the Sceptical Thesis and the Agreement Thesis. Section II therefore critically assesses Barry's defence of the sceptical thesis and Section III examines the agreement thesis.

Brian Barry's Justice as Impartiality is a highly enjoyable and rewarding book. It throws new light on some familiar theories of justice, and shows how the idea that principles of justice are those principles which no one could reasonably reject can yield prescriptions for constitutional design. But I shall argue that Barry's defence of his theory is less robust than he thinks, and more generally that there is reason to suppose that principles of justice are as contestable as conceptions of the good.

In Justice as Impartiality Brian Barry seeks to present ‘a universally valid case in favour of liberal egalitarian principles’ (p. 3). It is an ambitious enterprise undertaken with originality, vigour, and wit; and containing a wealth of interesting argumentation. If, ultimately, Barry fails in the task he sets himself, as I shall argue he does, the attempt is none the less highly instructive; not only because of the many local successes in his arguments with proponents of alternative theories and his often illuminating discussions of particular issues, but also because of the lessons to be learnt from his failure. For if even Barry's formidable defence of a contractualist, neutralist liberalism does not succeed then there may be good reason to think that none will. While this larger claim will not be fully justified here, and the case against justice as impartiality does not depend upon it, one of my deeper motivating concerns is to try to show that the most important failures in Barry's argument are inextricably bound up with the nature of his undertaking. We are concerned, I believe, not just with some inevitable imperfections of execution but with a flawed project.

Brian Barry's Justice as Impartiality is an important book. One of its contributions to the discipline is a characteristically clear presentation of what follows if one accepts a commitment to equality, and the reasonableness of continuing and profound disagreements about the nature of the good life (the reasonableness of pluralism). I take the argument of Justice as Impartiality to be an important next step in the attempt to give an account of the content of justice which is impartial, fair, or neutral between conceptions of the good, and engaging with it has the great advantage that many of the criticisms that can be made of Barry apply to other liberal contractualist theories of social justice. It is faintly ironic that it is one of Barry's great virtues, his clarity, that makes it easier to see the problems inherent in the attempt to complete the impartialist project. I have not attempted below to offer a systematic summary and critique of Barry's book or any particular section of it. Instead I have opted to try to engage with the ideas that drive it at a more fundamental level.

With a book as wide ranging and insightful as Barry's Justice as Impartiality, it is perhaps a little churlish to criticize it for paying insufficient attention to one's own particular interests. That said, in what follows I am going to do just that and claim that in an important sense Barry does not take utilitarianism seriously. Utilitarianism does receive some discussion in Barry's book, and in an important section which I will discuss he even appears to concede that utilitarianism provides a rival though ultimately inadequate theory of justice. Nevertheless, utilitarianism is not considered a rival to ‘justice as impartiality’ in the way that ‘justice as mutual advantage’ and ‘justice as reciprocity’ are. One response, and perhaps the only adequate response, would be to construct a rival utilitarian theory. I cannot provide such a theory in this paper, and I certainly would be very cautious about claiming that I could provide such a theory elsewhere. What I want to suggest is that utilitarianism is a genuine third theory to contrast with ‘justice as mutual advantage’ and ‘justice as impartiality’ – ‘justice as reciprocity’ being merely a hybrid of ‘justice as mutual advantage’, at least as Barry presents it (pp. 46–51). I also want to argue that it poses a more significant challenge to a contractualist theory such as Barry's than his discussion of utilitarianism reveals.

As the author of Justice as Impartiality, I am not ashamed to admit that I was delighted by the liveliness of the discussion generated by it at the meeting on which this symposium is based. I am likewise grateful to the six authors for finding the book worthy of the careful attention that they have bestowed on it. Between them, the symposiasts take up many more points than I can cover in this response. I shall therefore focus on some themes that cluster round the contractual device that I associate with the notion of justice as impartiality. Is it necessary? If it is not necessary is it nevertheless useful? Within an overall contractual framework is the form of contract that I propose uniquely justifiable? And does the form of contract that I defend generate the implications that I claim for it?