Paul Hurley's book is a multi-pronged critique of consequentialism ‘from within’, as it were. Roughly the first half of the book identifies tensions within consequentialism. Specifically, Hurley identifies tensions between consequentialism as a moral theory (that the right action is what produces the impersonally best state of affairs) and various accounts of practical reason that consequentialists have tended to accept that undermine the idea that impersonal reasoning about what will produce the best state of affairs is fundamental to morality. The second half of the book aims to block support for consequentialism more directly by analysing some intuitively plausible ideas about the nature of moral thought that have been taken to provide theoretical support for consequentialism. Here Hurley argues that these intuitive ideas support consequentialism only when misinterpreted, and that when properly understood they lead to a moral framework that is an alternative to consequentialism. In that sense, Hurley's aim is to argue that intuitive lines of thought that are accepted by consequentialists or have been taken to support consequentialism in fact lead ‘beyond consequentialism’. Hurley's alternative conception of moral evaluation is ‘interpersonal’ (what he calls an ‘interpersonal conception of impartiality’). Actions are to be assessed by taking into account the point of view of each person and according proper moral weight to the reasons that individuals have. Such a framework, he argues, leads straightforwardly to agent-centred permissions not to bring about the impersonally best state of affairs as well as standard deontological restrictions on choice. Most of Hurley's arguments target direct act-consequentialism, but the concluding chapter outlines how his main arguments extend to other forms of consequentialism, such as indirect consequentialism or non-foundational consequentialism (moral conceptions that base consequentialist normative principles on non-consequentialist foundations such as the idea of fair treatment or respect for persons as equals).

Hurley's book is thoroughly and densely argued and presents a number of powerful objections to consequentialism. I may not be the ideal reviewer since I am squarely in Hurley's camp and find his arguments to be generally persuasive. Surely defenders of consequentialism will have responses and it is important to hear from them what they are. In the balance of this review, I shall limit myself to outlining some of Hurley's most important challenges to consequentialism.

Hurley initiates his critique by identifying three claims to which consequentialists are drawn that are in tension with each other. The first is that morality provides agents with decisive or overriding reasons for action (‘RAMS’ for the rational authority of moral standards). The second claim is that ‘some practical reasons are fundamentally non-impersonal’ (p. 13). That is, agents sometimes have sufficient reasons for action that are not based in and not reducible to what will bring about the impersonally best state of affairs. For example personal interests and projects, relationships and loyalties, or individual commitments and values, and so on, can provide agents with sufficient practical reasons regardless of whether acting on these reasons contributes to bringing about what is impersonally best (‘NIR’ for the non-impersonality of practical reason). (‘Non-impersonal reasons’ can of course include standard deontological restrictions on choice, but Hurley does not build this point into NIR at the outset.) The third claim is a consequentialist theory of moral standards (CMS) according to which the standard of right action is what produces the best overall state of affairs. The main point of tension is between NIR and CMS, since NIR holds that agents can have sufficient practical reason not to do what is morally required by CMS. Thus NIR and RAMS create pressure to reject CMS, here allowing that NIR and RAMS shape the content of morality by setting limits to what morality can reasonably demand of agents. Or if one accepts NIR and CMS, it will be difficult to accept RAMS: NIR opens the door to situations in which an agent can have sufficient reason not to promote what is impersonally best – that is, not to do what consequentialist moral standards require – and that is to deny that moral reasons are overriding. Hurley suggests that many consequentialist theories must accept this result. It should be noted that his argument, here and throughout the book, requires a strong form of NIR. Genuine tension with CMS requires more than the weaker thesis that agents’ personal projects, relationships and loyalties, etc., normally generate good reasons. Rather, for the argument to work, NIR must hold that non-impersonal reasons are sometimes sufficient or decisive practical reasons (even when they conflict with impersonal reasons to bring about what is best overall). Otherwise a consequentialist could resolve the tension through a weaker form of NIR, holding that personal projects and relationships, etc., provide weighty reasons that can be overridden by impersonal reasons. One might think that a consequentialist should accept only the weaker version of NIR, but Hurley shows that prominent consequentialist theorists have in fact wanted to accept it in its strong form.

Hurley's third and fourth chapters draw on NIR to develop this tension in interesting ways. A standard worry about consequentialism is that it is unreasonably demanding, requiring agents to set aside individual interests to pursue impersonal good. In chapter 3, Hurley stands this objection on its head by claiming that in fact consequentialist moral theories make no demands on individuals – at least not as generally articulated. A central concern of many leading non-consequentialist moral theories (such as Kant's moral theory, Hobbesianism or Scanlon's contractualism) has been to give an account of the authority of moral norms. Hurley observes, correctly I believe, that this has not been a focus of consequentialist theories. They have been concerned to develop an account of the content of morality, what Hurley calls a ‘theory of moral standards’, though not an account of the rational authority of such standards (a ‘theory of moral reasons’). But unless individuals have decisive reasons to comply with consequentialist principles – that is, absent an account of the authority of consequentialist moral standards – they make no demands on individuals and are reasonably ignored when they conflict with weighty personal reasons. Moreover, accepting NIR (in its strong form) appears to be an obstacle to establishing the authority of consequentialist moral standards. But isn't it just obvious that we should always do what is right, and can't a consequentialist theory simply apply this insight to its theory of moral standards? Hurley argues that this intuitive idea is tied to a non-consequentialist account of the content of morality, and in later chapters argues that a related intuitive idea, that it is always right to do what is best, does not lead to consequentialism when properly understood.

One might think that consequentialists should jettison NIR, at least in its strong form. But Hurley's fourth chapter, drawing on features of Bernard Williams's critique of utilitarianism, argues that consequentialists ‘presuppose in their very account of impersonal evaluation of states of affairs that rational agents have good non-impersonal reasons at the foundational level’ (p. 66) This is an ambitious and clever argument if it can be made to work, because it would show that consequentialism presupposes the strong version of NIR, in a way that precludes it from establishing the rational authority of consequentialist moral standards. The argument is complex and has several components. First, Hurley argues that the impersonally best state of affairs is one in which the interests, concerns and projects that are the source of individuals’ non-impersonal reasons (or that individuals have non-impersonal reason to care about) are maximally satisfied. Thus, in order to specify the impersonally best state of affairs, consequentialism presupposes NIR (though, it should be noted, so far only in its weaker form). Second, how could agents have reason to promote what is impersonally best? One way, Hurley claims, is through a higher-order project of impartial beneficence – the project of promoting the satisfaction of individual interests generally (pp. 71ff.). But impartial beneficence will be just one of an individual's projects among other (personal) projects with no special priority – or at least no rationally necessary priority when the demands of beneficence conflict with an agent's other reasons. The argument here is that consequentialism relies on NIR to give content to the notion of the best state of affairs in a way that creates a barrier to establishing the authority of CMS. Now one might respond that general beneficence is not rationally optional, but a demand of reason. So Hurley argues, third, that consequentialists have tacitly assumed that the impersonal standpoint is the standpoint of rational agency in order to secure RAMS. But one needs an argument for that claim, and it will be harder to supply once one has acknowledged that agents have good non-impersonal reasons (as consequentialism must by the first argument). (Hurley develops a version of this argument to show that the application of the impersonal standpoint to non-impersonal moral values, such as fairness or respect for rights, will likewise undermine the claim of impersonal reasoning to carry special authority or to be fundamental to morality – but space does not permit discussion here.)

This chapter contains several important arguments whose success depends on showing that consequentialism is committed to NIR in its strong form. I am not convinced that consequentialists cannot get by with the weaker version of NIR. That aside, the stronger version is prima facie quite plausible and leads to problems for consequentialism, and as I've indicated, Hurley shows that many consequentialists have been drawn to the stronger thesis.

Later chapters (6 and 7) target arguments that aim to provide theoretical support for consequentialism from intuitively plausible assumptions about the nature of morality. We might agree, for example, that it is always right to do what is best, and once ‘what is best’ is interpreted as ‘what produces the best overall outcome’, some form of consequentialism follows. But Hurley argues that the consequentialist reading ‘misappropriates’ this plausible general intuition. Since the aim of deliberation is to settle on what there is good or best reason to do, ‘what is best’ is most naturally understood as ‘the action supported by the best reasons’ or ‘best action’. And given plausible accounts of practical reason (e.g. NIR) there is no reason to identify the ‘best action’ with ‘action that produces the best outcome’. Variants of this argument are applied to other intuitive forms of support for consequentialism, such as the idea that overall consequences are the only considerations relevant to the moral assessment of actions (suggesting that it can never be wrong to do what produces the best outcome) and Philip Pettit's claim that the rational response to value is to promote it.

Likewise we might agree that morality is based on impartial reasoning that recognizes the equal moral significance of all persons (everybody to count for one, nobody for more than one). The consequentialist takes impartiality to involve assessing actions through the impersonal goodness of their outcomes. Here Hurley argues for an alternative conception of impartiality that applies directly to actions – roughly that impartial reasoning about right action involves recognizing the independent moral significance of each person's point of view and giving proper moral weight to the reasons of each individual, a form of reasoning that he argues supports standard deontological principles. This is where Hurley develops his ‘interpersonal’ conception of impartiality. This conception is not fully worked out, but that is not Hurley's aim in the book and one can see how this form of reasoning can be taken in a contractualist or neo-Kantian direction. This chapter develops an interesting argument against Samuel Scheffler to the effect that what provides a rationale for agent-centred permissions not to promote the impersonally best state of affairs – that individual projects and interests have independent moral significance – equally supports or makes intelligible standard deontological restrictions.

Beyond Consequentialism is an important book which contains many innovative arguments. It deserves serious attention from defenders of consequentialism.