The Robust Demands of the Good is based on the Uehiro Lectures in Ethics that Pettit gave in Oxford in June 2011. Building on his significant work on the Republican conception of freedom, Pettit presented a compelling new version of consequentialism founded on a deeply innovative yet intuitively attractive account of interpersonal goods such as love, respect and honesty. Instead of relegating these goods to peripheral significance, as many familiar variants of consequentialism do, Pettit affirmed these goods as amongst the most valuable. In doing so, he fashioned a version of consequentialism that unlike many predecessors is appropriately responsive to the valuable role the norms of attachment and virtue play in human life. In the intervening years since those lectures, Pettit has significantly refined and extended his views and arguments. The result is an impressive and novel contribution to value theory, the philosophy of action and normative ethics. Indeed, the book is at its best when it defends a new view, rather than objecting to existing views. Even so, consequentialists and non-consequentialists alike will benefit from engaging with this important work.
In chapters 1–3, Pettit argues compellingly that many of the most important goods we provide to others, such as love, depend not only on how we act in the actual circumstances but also on how we would act if circumstances were different. For example, for Alice to love Betty, it is necessary but not sufficient that Alice recognize and respond to the stimulus of Betty's needs or wishes with care when she has sufficient reason to do so in the actual circumstances. Alice must also be disposed to provide such care in all and only alternative possibilities where analogous stimulus and reason are present and Alice's disposition to care for Betty is not impaired. These necessary conditions are jointly sufficient for this love. Thus this love requires both actual and counterfactual care. It would not be love if, for example, Alice would not provide care just in case caring would be inconvenient, or just in case Betty adored dogs instead of cats.
Furthermore these counterfactuals are not only relevant to the necessary and sufficient conditions of love, but are also relevant to how valuable love is for its beneficiary. When one provides love, and not merely actual care, to another, then the beloved receives what Pettit calls a modally demanding good. It is modally demanding because it requires the provision of modally non-demanding goods in each of the aforementioned actual and counterfactual circumstances. This modally demanding good is more valuable than just the actual care that is necessary for that modally demanding good. For Pettit, all other things being equal, the more counterfactual possibilities in which Alice cares for Betty, the better it is for Betty.
According to Pettit, other important interpersonal goods, including the attachments of friendship and solidarity; the other-regarding virtues of honesty and tolerance; and respect for others’ liberties have this same modal structure. For example, it is not enough to be honest to tell the truth when it is actually expedient to do so, but the agent must also have a disposition to tell the truth in all and only circumstances where they have sufficient reason and stimulus to do so and there is no impairment of this disposition.
In chapter 4, Pettit argues that not only do these goods share the same modal structure, but they share the same justification. Being treated in this way by suitably disposed agents in different circumstances is desirable to us, not because it maximizes our expected good, but instead because it reduces our vulnerability to other free agents. For example, if others are disposed to tell us the truth even when it is inconvenient to them, we will be less vulnerable to being lied to.
In Chapter 5, Pettit distinguishes between two terms of art that he refers to as acts and actions, in order to argue against the view that agents’ behaviour should be morally evaluated solely or primarily on its disposition-independent properties. Acts refer only to an agent's behaviour in the actual circumstances, independent from how that agent would have acted under different circumstances and thus independent from how they are disposed to act. For example, an actual act of truth-telling is still an act of truth-telling whether or not it comes from a truly honest person who is disposed to tell the truth or from an opportunistic truth-teller who would instead lie whenever it is amusing to do so. An action, on the other hand, refers to an agent's behaviour in both actual and counterfactual circumstances and thus its disposition-dependent properties are essential to its identity. Pettit argues that it is important to morally evaluate both acts and actions. This entails that it is important to evaluate an agent's behaviour in terms of its disposition-dependent properties as one does when one evaluates action. This runs counter to a prominent view that dispositions are irrelevant to the moral evaluation of behaviour and instead are only relevant to the moral evaluation of the agent themselves.
While Pettit's account of doing good by being loving and honest is compelling, his account of the weaker, asymmetrical counterfactual requirements of doing evil in chapter 6 is not. He argues that we do not and should not hold doing evil to the same modal requirements as doing good. For Pettit, while doing good in a virtuous way requires that benefactor to provide benefits in certain scenarios when the provision of those benefits goes against the benefactor's inclinations or interests, doing evil does not require malefactors to impose costs on others in scenarios where doing so would go against the malefactor's inclinations or interests. It is sufficient to do evil, according to Pettit, if one actively fails to pursue the modally robust demands of the good when one imposes costs on others. While Pettit's theory of doing good is supported by a common, intuitive conception of doing good, his asymmetry lacks this support. Under a common conception, doing evil is a morally serious ascription that involves more than the above kind of failure. For example, if Celine intentionally returns her brother's library book one day late only because it is slightly inconvenient for her to return it on time and therefore her brother suffers a monetary fine, we are not inclined to characterize Celine as doing evil. Celine may be doing something that is bad and perhaps blameworthy, but the ascription of evil carries a moral severity that makes it inappropriate in this case. It would be both more theoretically parsimonious and intuitively supported if doing evil had the same counterfactual requirements that apply to doing modally demanding good.
Furthermore, Pettit's attempt to use this value asymmetry to explain the Knobe Effect also fails. Famously, Knobe showed that more people would ascribe intentional action to someone in a case involving harming than in an otherwise identical case involving helping (Intention, Intentional Action, and Moral Considerations, 2003). In both cases, a company chairman is asked whether to pursue a profit-increasing programme. In the first variant, the chairman is told that the programme will harm the environment. In the second variant, the chairman is told that the programme will help the environment. In both vignettes, the chairman agrees to pursue the programme and says that he just cares about profits and that he does not care at all about the environment. The so-called Knobe Effect is that most first variant respondents claim that the chairman intentionally harms but most second variant respondents claim that the chairman does not intentionally help.
According to Pettit, the difference in response is explained by a difference between the counterfactual requirements of helping and harming and not by intentional side-effects. Pettit claims that the asymmetry supports the suggestion that helping requires doing modally demanding good, whereas harming only requires doing actual evil. Given what we know about the chairman's profit-focus, in the second vignette he does not do the modally demanding good of environmental help by improving the environment in the relevant range of alternative circumstances. Instead, the chairman only improves the environment when it increases profits. On this view, the chairman does not intentionally help the environment because he does not help it at all. On the other hand, in the first vignette the chairman does harm the environment because he does what for Pettit counts as doing evil: actual damage to the environment. Furthermore, because the chairman is aware of this damage, he counts as intentionally harming the environment. Pettit argues that his explanation can be generalized to analogues of the original Knobe Effect case including non-moral cases as well as cases that do not involve intention. The difference in action ascription is purportedly explained by differing counterfactual requirements that must be satisfied in order for the action in each respective variant of the case to count as an action of the relevant kind.
This explanation fails because there have been replications of the Knobe Effect that do not involve differing counterfactual requirements. For example, Machery ran an experiment with two vignettes involving someone requesting the largest smoothie available (The Folk Concept of Intentional Action, 2008). In the first vignette, the customer is told that the largest smoothie is $1 more than it used to cost, and in the second vignette the customer is told that the largest smoothie is provided in a commemorative cup. In both vignettes, the customer responded that he did not care about paying $1 more/a commemorative cup respectively, but he just wanted the largest smoothie available and subsequently bought that smoothie. Most respondents said that in the first vignette, the customer intentionally paid one dollar more but most respondents said that in the second vignette, the customer did not intentionally obtain the commemorative cup. Pettit's purported explanation of this difference would have to claim that paying one dollar more has lower counterfactual requirements than obtaining the commemorative cup. On this view, it is not enough to count as obtaining the commemorative cup for the customer to get the cup in the actual circumstances but instead they must get the cup in a range of actual and counterfactual circumstances. However, it is enough to spend the extra dollar in the actual circumstances for it to count as paying one dollar more. These differing counterfactual requirements are implausible and this casts significant doubt on Pettit's explanation of the Knobe Effect.
Even if my objections to Pettit's views about doing evil and the Knobe Effect are sound, Pettit's subsequent arguments in chapter 7 that demonstrate that his view about modally demanding goods can be adopted by consequentialists and non-consequentialists alike are unaffected. That said, Pettit favours a variant of consequentialism that does not explicitly guide agents’ behaviour except in exceptional circumstances. In normal circumstances agents should instead be guided by the aforementioned dispositions involved in the provision of modally demanding goods. This version of consequentialism is resilient to many of the standard objections because amongst other reasons it allows us to privilege those near and dear to us and it allows us to be motivated by the dispositions that are responsive to their needs, wishes and liberty. For this reason, while normative theorists of all stripes will benefit from this book, it is an especially important contribution to the consequentialist normative tradition.
