Do the reactive attitudes justify public reason?

Gaus’s public justification principle

Gaus holds the following public justification principle:

A moral imperative ‘Ф!’ in context C, based on a rule L, is an authoritative requirement of social morality only if each normal moral agent has sufficient reasons to (a) internalize rule L, (b) hold that L requires Ф-type acts in circumstances C and (c) moral agents generally conform to L. (Gaus, 2011a: 263)

Here, Gaus draws a distinction between internalist and externalist accounts of having a reason. On the externalist view, an agent has a reason whenever there is a reason that applies to her. Gaus rejects the externalist picture because it ‘misconstrues the relation between having a reason and being a rational agent’ (Gaus, 2011a: 233). The externalist view implies, for instance, that Aristotle had a reason to accept particle physics. But, says Gaus, Aristotle had no such reason; indeed, only by ignoring his evidence and by being irrational could he have come to endorse such a reason. When we move from theoretical reasoning to practical reasoning, externalism might seem more plausible. For surely everyone has a reason to avoid genocide? But consider the case of a benighted tribal society that lacks the concept of an ethnic group. Of this case, Gaus (2011a: 235) says:

to attribute a reason to these peoples, to say that they possessed a reason not to eliminate other groups, while acknowledging that any justificatory force of this reason is inaccessible to them as reasoning beings is, I think, not only a misuse of language, but undermines the point of discourse about reasons and rationality.

Instead, Gaus adopts an internalist account of having a reason. This view, of course, need not say that the reasons we have are those that we would immediately affirm at any particular moment. Some moderate idealization of the reasoner is permitted. The point is just that there must be ‘some not-too-difficult bridge to cross connecting what reasons a person can be said to possess and what the exercise of her rational faculties can lead her to’ (Gaus, 2011a: 35). Thus:

Alf has a sufficient reason R if and only if a ‘respectable amount’ of good reasoning by Alf would conclude that R is an undefeated reason (to act or believe) (Gaus, 2011a: 250).

Slotting this into the earlier public justification principle, we see that Gaus’s view, abstractly stated, is that one should not make an authoritative demand on others unless those subject to the demand have sufficient reason to accept it, in the sense that they would, after a respectable amount of good reasoning, accept the demand. This is the principle that Gaus takes to have profound and far-reaching implications for social and political life and that he seeks to justify through the reactive attitudes argument.

Why the reactive attitudes presuppose public justification

The term ‘reactive attitudes’ was introduced by Strawson (2003) in his seminal essay ‘Freedom and Resentment’. For Strawson (2003: 79), the reactive attitudes are the ‘attitudes belonging to involvement or participation with others in inter-personal human relationships’. Paradigm examples of such reactive attitudes are resentment, indignation and gratitude. What makes these attitudes reactive, says Strawson (2003: 83), is that they are ‘reactions to the quality of others’ wills towards us’ – to their good will or ill will.

A crucial point made by both Strawson and Gaus is that these reactive attitudes have certain ‘appropriateness conditions’ (Gaus, 2011a: 206). For instance, while it makes perfect sense to be indignant if my brother bites me, it seems quite inappropriate to be indignant if my dog bites me. Perhaps I can in such moments appropriately experience non-reactive attitudes such as frustration and annoyance. But insofar as my dog is incapable of ill will towards me, it seems misplaced and inappropriate to direct at it the reactive attitudes of resentment, blame and indignation.

For Gaus, there are two appropriateness conditions for the reactive attitudes. The first is that an individual must be ‘capable of caring for a moral rule even when it does not promote her wants, ends, or goals’ (Gaus, 2011a: 211). Following Taylor (2018: 103), we can call this the Capacity Condition. This condition explains the common belief that young children, psychopaths, dogs and others who lack minimal capacities for moral motivation should not be the targets of blame and resentment.

I shall not contest the Capacity Condition. My worry is with the second condition for the appropriateness of the reactive attitudes, which, again following Taylor (2018: 103), we can term the Sufficient Reason Condition. To introduce this second condition, Gaus notes that sometimes the reactive attitudes are inappropriate even when the first condition is met. Consider the case of systematic indoctrination, such as Winston Smith at the end of 1984. Even though he is a fully-fledged moral agent, capable of putting aside his own wants and acting on the basis of moral reasons, it is still inappropriate to feel resentment and indignation towards Smith when he embraces Big Brother, because the relevant moral reasons ‘are now beyond his comprehension’ (Gaus, 2011a: 218). The Sufficient Reason Condition, then, says that the reactive attitudes are appropriately directed towards an individual who violates a moral rule only if that individual ‘has sufficient reasons to endorse the relevant rule’ (Gaus, 2011a: 222). To motivate this condition, Gaus imagines making a moral demand on an individual:

However, I know that she does not see that it is the moral thing to do, and suppose I think her lack of appreciation is quite genuine. I demand ‘Ф!’ and she does not see why she is obligated to Ф. She is puzzled that anyone would think Ф is obligatory. If I think this, then again I cannot reasonably feel resentment or indignation that she fails to Ф, any more than I can feel indignation at a four-year-old who is unable to detach himself from what he most wants to do and so steals some favourite candy. She just cannot see how ‘Ф!’ has any internal authority over her. (Gaus, 2011a: 219)

Why the reactive attitudes cannot be rationally renounced

The second key premise in Gaus’s argument is that the reactive attitudes cannot be rationally renounced or repudiated. In making this argument, Gaus follows Strawson, who similarly claims that our commitment to the reactive attitudes is ‘too thoroughgoing and deeply rooted for us to take seriously the thought’ that these attitudes should be repudiated (Strawson, 2003: 81). Such a thought, says Strawson (2003: 81), ‘is, for us as we are, practically inconceivable’. As Gaus puts this Strawsonian point:

We are embedded in certain sorts of practices, with certain beliefs and emotions. They form part of the reasons from which we must judge, criticize, and propose changes. A practice such as social morality is deeply embedded in our view of the world; it affects our understanding of interpersonal relations, including love and friendship, and so of what sort of life is worth living. If the presuppositions of our moral practices are so deep a part of the way we see the world, then to renounce the practice would be to renounce most of what we care for and value. But how could we have reason to do that? How can we survey all that matters to us and come to the conclusion that our reasons lead us to give it up, by renouncing the view of the world on which our reasons depend? Where would that reason come from? It is, I think, as difficult to argue a moral person out of her moral practices as it would be to argue the psychopath into them; given who they are, they do not have reasons to change their view of the world. (Gaus, 2011a: 192)

Some cases: genocide, animal cruelty, racism, sexism, littering

The most vivid type of counterexample to Gaus’s claim that liability to the reactive attitudes presupposes sufficient reason involves cases of monstrous yet rationalized evil. Consider an individual such as Goebbels, who performs monstrously evil actions yet who, after a respectable amount of deliberation, finds that he sees no sufficient reason to refrain from these actions. ² He is capable of caring about morality – in this sense he satisfies Gaus’s Capacity Condition – and sees that there is some moral reason against harming others in general and against genocide in particular. But he assigns this moral reason such little weight that it fails to defeat the more important reason (as he sees it) in favour of genocide. So he simply cannot see how the rule prohibiting the murder of some racial or ethnic group has any authority over him. Yet are we seriously to think that, therefore, Goebbels is off the hook – that we cannot appropriately feel resentment and indignation towards him? Assuming that Goebbels ‘is puzzled that anyone would think’ that refraining from genocide is obligatory, Gaus’s view has the implausible implication that we ‘cannot reasonably feel resentment or indignation that [Goebbels] fails to’ refrain from genocide. ³

It is not only at the heights of vice that the Sufficient Reason Condition misfires, however. Though I will primarily focus on the Goebbels case – because it elicits the relevant intuition most strongly – it is only the most extreme instance of a common kind of structure. Suppose that, at a dinner party, an animal rights activist discovers that she is sat opposite a producer of foie gras – the luxury food made by inserting a 12-inch steel tube into the throat of a goose in order to force-feed it corn mash (a procedure known as gavage). Aware of the controversy surrounding gavage, the foie gras producer proceeds to spend the evening justifying this horrible procedure by reference to the delicious taste of foie gras and its place in traditional French cuisine. In such a situation, must the animal rights activist suspend her indignation simply because the foie gras producer sees no sufficient reason, from within his system of values and beliefs, not to practice gavage? Consider, next, an individual who often tells racist jokes interspersed with terrible racial slurs. He has been called out on these jokes before, but he just cannot see what all the fuss is about. He considers his jokes to be side-splittingly funny and interprets others’ offense as a manifestation of ‘Generation Snowflake’ political correctness. Does the fact that he sees no sufficient reason not to tell racist jokes make it inappropriate for his interlocutors, or those who overhear, to feel outrage and indignation? Or suppose a woman discovers that her job application was rejected due to the employer’s sexist implicit bias against résumés with female names. Insofar as the bias was implicit, the employer would likely not have been able to root it out even after a respectable about of good reasoning. Would it therefore be inappropriate for the woman to feel resentment in this case? Consider, finally, the case of a group of youth who genuinely do not think it is a problem to throw litter on the ground. Again, is indignation really an unreasonable reaction in this case just because the youth would be puzzled by the objections of those who oppose littering?

Gaus may, of course, respond to the rhetorical question associated with each of these cases – Goebbels, the foie gras producer, the racist joke-teller, the sexist employer, the littering youth – by saying: ‘But they don’t get it. Your demand makes no sense to them. How can you be outraged at them for failing to conform to moral rules that they do not grasp?’ But this typically does not make the reactive attitudes seem any less appropriate. Indeed, a natural reply here is to say: ‘Yes, and that’s the whole point! Their failure to see (variously) genocide, gavage, racism, sexism and littering as something they shouldn’t do is precisely why I am so indignant!’ It is natural to feel indignation towards these individuals, in other words, not in spite of their inability to see the reasons that apply to them but precisely because of it.

The agents have sufficient reasons

In response, Gaus might want to argue that these are not, in fact, cases where the transgressors lack sufficient reasons not to transgress. Consider the case of the littering youth. Perhaps, at the moment in which they throw litter onto the ground, the youth see no reason not to litter. But Gaus’s notion of what it means to have a reason involves some degree of idealization. So perhaps Gaus could say that the youth do have sufficient reason not to litter in the sense that, were they to spend a respectable amount of time thinking about the ethics of littering, they would come to see that it is wrong. The same strategy can of course be used in relation to the other cases, including the central case of Goebbels.

In fact, this move is the response that Gaus has employed on another occasion when presented with the Goebbels-style objection:

Do we really think that Nazi politicians were models of good reasoners — that they reasoned well, albeit from bad commitments? … Students of ideology typically see fascism as held together more by emotion and myth than by reasoned coherence. An ideology that constituted a revolt against reason strikes me, at any rate, as a distinctly inappropriate model of a rationally coherent immoral view. (Gaus, 2013: 78)

So the problem with responding to one counterexample after another by pointing out how, in each case, the agent does not spend enough time reasoning is that, still, nothing in Gaus’s theory rules out the prospect of an individual who satisfies all that is required for Gaus’s procedural account of having a reason, yet who nonetheless arrives at the same morally abhorrent conclusions. And so we are still left with the question: What should we say about him? Would not the reactive attitudes be appropriate here? The short-term strategy, then, of quibbling with details of how cases are described fails as a long-term solution.

Consider, then, a rationally upgraded version of Goebbels: Goebbels+, say. Like Goebbels, Goebbels+ believes in the permissibility of killing entire racial and ethnic minorities. But, unlike Goebbels, Goebbels+ does not base these views on ‘emotion and myth’. Instead, he simply assigns enormous weight to his country’s racial purity. The weight he assigns to racial purity is so great that, even though he recognizes pro tanto reasons not to harm others, these reasons against harm are outweighed by reasons to promote racial purity. And so, after a respectable amount of time, the exercise of his rational faculties leads him to formulate and carry out genocide. Goebbels+ knows that some people consider him immoral, but he just cannot understand why they do so. In this case, Goebbels+ lacks a sufficient reason (in the Gausian internalist sense) to desist from killing racial and ethnic minorities. Yet he is surely still liable to the reactive attitudes for doing so.

The reactive attitudes really are inappropriate here

Even if Gaus can contest the details and rational coherence of the reasoning process of the actual, historical Goebbels, the interesting and difficult question is this: What should we say about Goebbels+? Again, this is only the most extreme case. We can imagine rationally upgraded versions of the other agents mentioned above. What should we say about foie gras producer+, racist joke-teller+, sexist employer+, littering youth+: agents who, by stipulation, have engaged in whatever amount of good reasoning Gaus wishes to see, yet who are still repellent in their thoughts and actions?

Here Gaus must, I think, bite the bullet and insist that, despite appearances to the contrary, the reactive attitudes of blame and indignation are inappropriate in these cases, assuming that the agents really do see no sufficient reason to comply with the moral rules that are contravened by their actions. Here, that is, Gaus must simply say: we cannot feel morally outraged at Goebbels+ for failing to conform to a moral rule against genocide that makes no sense to him after rational reflection on the matter and that he regards as a kind of indefensible taboo rather than a strong moral requirement applicable to him (see Gaus, 2011a: 221). But this is highly counterintuitive and thus rather unsatisfying. If the reactive attitudes of resentment and indignation are appropriate anywhere, they are surely appropriate here in the case of a racist genocidal killer. So, is there anything more that Gaus can say in support of the claim that feelings of indignation towards Goebbels+ are inappropriate? In this subsection, I consider and reject two ways in which Gaus might support this claim.

We begin with the error-theory strategy. Gaus could seek to provide an error theory that explains why, if the reactive attitudes are inappropriate in the case of Goebbels+, we had the contrary intuition. To do this, Gaus could point out that there is no reason to deny that Goebbels+ can elicit plenty of non-reactive attitudes, such as frustration and annoyance. Nor need he deny the appropriateness of certain kinds of reactive attitude in this case, such as a resentment that is directed at the racialized structure of society and at other related structural injustices that encourage beliefs about racial purity. Indeed, Gaus can even accept a certain kind of reactive attitude towards Goebbels+ himself: not for performing genocide (the moral rule against genocide, after all, is beyond his comprehension), but for allowing himself to become the kind of person who could fail to grasp the wrongness of genocide (Gaus, 2011a: 221). So, on this error-theoretic proposal, when forming intuitions about whether it is appropriate to feel reactive attitudes towards Goebbels+ for engaging in genocide, we are sensitive to other questions in the picture – for instance, whether it is appropriate to feel reactive attitudes towards Goebbels+ for allowing himself to be the way he is – which are similar to, and thus conflatable with, but strictly distinct from, the question of whether it is appropriate to feel reactive attitudes towards Goebbels+ for his acts of genocide. Once we realize this, so Gaus might argue, we become less confident that our intuitions are really tracking what they are meant to.

My worry with this strategy is that, like all error theories, it is too speculative to be convincing. For my own part, the relevant intuition seems suitably fine-grained: when contemplating the case of Goebbels+, the indignation I feel is directed not (or not merely) at the structure of society, nor even at Goebbels+ himself for letting himself become the kind of person he is, but rather, and quite specifically, at Goebbels+ for doing what he does. To see this, compare Goebbels+ with Distracted Goebbels+. Distracted Goebbels+ lives in the same structurally unjust society as Goebbels+, and Distracted Goebbels+ similarly sees nothing wrong with conducting acts of genocide. Yet, unlike Goebbels+, Distracted Goebbels+ is too occupied with reading novels to act on his beliefs. Now, while one might reasonably feel indignation towards both individuals for being the way they are – for being the kinds of people that would conduct genocide – is it not reasonable to feel some special additional indignation towards Goebbels+ for actually doing the terrible deeds? It seems highly plausible that Goebbels+, in carrying out acts of mass murder, disgraces himself and makes himself liable to the reactive attitude in a way that Distracted Goebbels+, in virtue of not perpetrating any such acts, is not. This suggests that our reactive attitudes are tracking what they are meant to – namely, Goebbels+’s violation of a moral rule against genocide – and not other factors, such as being the kind of person who sees nothing wrong with genocide, because these other intuition-confounding factors are controlled across Goebbels+ and Distracted Goebbels+.

A second strategy that Gaus could employ in order to support the claim that feelings of indignation are inappropriate in the case of Goebbels+ is to appeal to the Winston Smith analogy. ⁵ Gaus could, that is, argue that Goebbels+ is relevantly analogous to Winston Smith and so is no more liable to the reactive attitudes for conducting genocide than is Winston Smith for embracing Big Brother. Both, after all, while capable of caring about morality, lack sufficient reasons to endorse the moral rule contravened by their action. If Winston Smith genuinely isn’t to blame, then maybe Goebbels+ isn’t either?

However, while it is true that a systematically manipulated agent such as Winston Smith and a mass murderer such as Goebbels+ are analogous in the sense of lacking a sufficient reason to endorse some moral rule, they are also disanalogous in a number of morally important respects. The first and most obvious disanalogy is that Winston Smith is subject to intense psychological manipulation by another agent who intentionally and wrongfully causes him to have the views he has, whereas Goebbels+ is not. Perhaps, in other words, Winston Smith is not liable to the reactive attitudes, whereas Goebbels+ is, because manipulation is irreducibly exonerating. Second, Winston Smith’s embrace of Big Brother is completely out of character (until the final scene, 1984 sees Winston Smith constantly rebelling against Big Brother), whereas Goebbels+’s genocidal activities are not. More precisely, and as Sher would put this point, Winston Smith’s action is not, but Goebbels+’s is, traceable to ‘the interaction of some combination of his constitutive attitudes, dispositions, and traits’ (Sher, 2009: 88). Third, Winston Smith’s embrace of Big Brother at the first order is in conflict with his second-order desire to destroy Big Brother, whereas Goebbels+’s first-order desires for racial purity are plausibly aligned with his second-order desires. Fourth, given natural assumptions about the severity of Big Brother’s intensive manipulation of Winston Smith, he loses the capacity to be appropriately reasons-responsive, whereas Goebbels+ has such a capacity. Fifth, in arriving at his views, Goebbels+ plausibly exhibits ‘epistemic vices’ such as ‘overconfidence, arrogance, dismissiveness, laziness, dogmatism, incuriosity, self-indulgence, contempt’, whereas Winston Smith does not (FitzPatrick, 2008: 609). In light of these five morally significant differences, the analogy between Winston Smith and Goebbels+ is weak and should not dent our confidence that Goebbels+ is liable to the reactive attitudes.

It will not have gone unnoticed that, in sketching out the disanalogies between Winston Smith and Goebbels+, I have gestured towards certain leading theories of moral responsibility. ⁶ Each of these theories can be understood as an account of when the reactive attitudes of praise and blame are appropriate. Here it is important to stress, though, that I am not taking a stand on the question of which of these various disanalogies grounds the difference in liability to the reactive attitudes between Goebbels+ and Winston Smith. Another way of putting this is that I am not endorsing any particular alternative appropriateness condition for the reactive attitudes that would replace the Sufficient Reason Condition and that would explain why Goebbels+ is, and why Winston Smith is not, liable to the reactive attitudes. ⁷

But nor do I need to endorse an alternative condition, for my objection to Gaus’s Sufficient Reason Condition is case-based rather than principle-based. The objection is not that the Sufficient Reason Condition is false because some alternative condition (one that refers to manipulation, epistemic vices, the constitutive elements of the deep self, reasons-responsiveness, etc.) better captures and explains our general practices of praise and blame. Rather, the objection is that the Sufficient Reason Condition is false because it implies, highly counterintuitively, that it is not appropriate to resent genocidal mass murderers like Goebbels+ (nor other agents, such as foie gras producer+, racist joke-teller+, sexist employer+ and littering youth+). This counterintuitiveness does not subside once we consider Winston Smith, who is analogous to Goebbels+ in one respect yet disanalogous in many other respects. So, to reject the Sufficient Reason Condition, it is enough to show that it implies that Goebbels is off the reactive-attitudes hook; we need not in addition offer, in place of this condition, some alternative theory of liability to the reactive attitudes that explains where among a range of important disanalogies the line between Goebbels+ and Winston Smith should be drawn. ⁸

Pereboom’s alternative account

The reactive attitudes, recall, are ‘reactions to the quality of others’ wills towards us’ – to their good will or ill will (Strawson, 2003: 83). There are debates about which attitudes are reactive in this sense, but it is possible to categorize the reactive attitudes into two classes and to work from certain paradigmatic cases. On the one hand, there are negative reactive attitudes such as resentment, blame and indignation; on the other, there are positive reactive attitudes such as gratitude and reciprocal love. Drawing on the Pereboomian strategy of identifying non-reactive aspects of these reactive attitudes, as well as non-reactive alternative attitudes, I shall now argue against Gaus’s claim that renouncing the positive and negative reactive attitudes would so impoverish social life as to make it no longer worth living.

Let us begin with the negative reactive attitudes. As Pereboom argues, resentment and indignation are ‘suboptimal as modes of communication in relationships relative to alternative attitudes available to us’ (Pereboom, 2014: 179). When someone wrongs me or wrongs a third party, there are other available emotions beyond resentment and indignation: I might feel hurt or shock or disappointment or sorrow. These alternative non-reactive attitudes preserve what (if anything) is valuable about resentment and indignation. Disappointment, for instance, can quite effectively play the communicative role typically played by resentment and indignation. Indeed, Pereboom even argues that these non-reactive substitutes are often preferable because resentment and indignation can ‘damage or destroy relationships’, can ‘give rise to destructive resistance instead of reconciliation’ and can, at the social level, lead to excessively punitive and retributive policies (Pereboom, 2014: 180).

One might object that even if the reactive attitudes are often detrimental, there is also a significant positive side to emotions such as anger, resentment and indignation. First, these emotions seem to play an important role in motivating people to combat oppression and injustice. Without the reactive attitudes, would we still fight the good fight with all our might? Second, the reactive attitudes seem essential to our practice of punishment, which in turn seems essential to social cooperation. But how would we account for punishment without the reactive attitudes of judgement, resentment and indignation? And, in the absence of the threat of punishment, would people still cooperate and follow rules? So, even if Pereboom’s approach works for one-on-one interpersonal relations, would it really be able to sustain large-scale social cooperation within highly diverse societies? ⁹

These worries, both of which point towards benefits of the practice of the reactive attitudes that would supposedly be lost under Pereboom’s alternative practice, raise complex empirical and normative questions that I cannot address fully here. But let me make three brief points. First, regarding the fight against injustice: while anger and indignation do motivate, there are also plenty of powerful non-reactive sources of motivation for resistance and social change, such as compassion, sympathy, sorrow and a sense of justice. Mahatma Gandhi and Martin Luther King, for instance, were both able to successfully resist and protest injustice by appealing to moral principle and not primarily to resentment, indignation and other reactive attitudes that are based on the perceived ill will of others (Pereboom, 2014: 149).

Second, regarding cooperation: neither (a) the premise that cooperation requires punishment nor (b) the premise that punishment requires the reactive attitudes is entirely convincing. To take only (b): it is true that punishment is often driven by reactive attitudes such as resentment, blame and indignation; generally, when these emotions are absent, so too is the impulse to punish. But there are various alternative ways of thinking about and justifying punishment – alternatives that do not involve the reactive attitudes. For instance, on Pereboom’s ‘quarantine’ account of punishment, the punishment of criminals and cheaters is modelled on the containment of carriers of dangerous diseases: preventative measures are taken to protect society and to preserve commitment to its rules, but these are not accompanied by feelings of blame or resentment towards those who pose the harm (Pereboom, 2014: 169ff.).

The third and more general point is about the dialectical bar and about the modesty of my aims in this section. I need not show, and am not seeking to show, that Pereboom’s alternative account is perfect – nor even, in fact, that it is as good as our current practice of the reactive attitudes of resentment, indignation, blame and so on. Perhaps, for instance, the quarantine model is on balance worse than the reactive attitudes at driving punishment. Rather, I need merely to show that Pereboom’s account is liveable or good enough – that it leaves intact a solid amount of what we care about. Satisfying this relatively low dialectical bar would be sufficient to undermine Gaus’s Strawsonian claim according to which renouncing the reactive attitudes would be catastrophic for social life. For Gaus, recall, the practice of the reactive attitudes is ‘so deep a part of the way we see the world [that] to renounce the practice would be to renounce most of what we care for and value’ (Gaus, 2011a: 192). This language might be plausible if the only alternative to the reactive attitudes were some kind of bleak state of nature devoid of emotion and cooperation. But the availability of a liveable alternative practice such as Pereboom’s, even if it is a downgrade from our current practice, shows Gaus’s language here to be overstated. Even if there would be some costs to giving up the reactive attitudes, that is, Gaus cannot import the enormous Strawsonian costs to which he appeals.

Perhaps it will be thought that negative reactive attitudes are the easy cases. Abandoning resentment is no great loss. The hard cases are positive reactive attitudes, such as love and gratitude. How could we live without these? However, as Pereboom points out, while love and gratitude are reactive attitudes in some respects, they are also non-reactive attitudes in other respects. So, to reject the reactive attitudes one need not reject the attitudes of love and gratitude wholesale; one need only reject the reactive aspects of these attitudes. The question is: How much is left? Are enough aspects of love and gratitude in the clear for some recognizable and valuable form of that attitude to remain in place?

Consider gratitude. It might well be true that gratitude has certain reactive aspects – that part of what it means to be grateful is to be responsive to another person’s good will. But gratitude also has important non-reactive aspects. For instance, one can be thankful to a young child for some kind gift, despite the child being too young to be able to express good will. As Pereboom (2014: 190) says, ‘gratitude involves, first of all, being thankful toward someone who has acted beneficially’ – and insofar as this core sense of gratitude makes no essential reference to the quality of others’ wills, it could survive the rejection of the reactive attitudes.

The same goes for love. Again, it might seem undeniable that love is completely or at least primarily a reactive emotion – that it is a response to the quality of another’s will. But in fact love has many non-reactive aspects. A mother does not in general love her children because of their good will towards her – this being most obviously true in the case of very young children who are incapable of good will or ill will. Even when mature adults love each other, this love is nourished by intelligence, appearance, mannerisms and other factors that make no essential reference to perceived good will. As Pereboom argues, it is possible to see love in largely non-reactive terms: ‘love of another involves, most fundamentally, wishing well for the other, taking on aims and projects of the other as one’s own, and a desire to be together with the other’, none of which is threatened by the rejection of the reactive attitudes (Pereboom, 2014: 190).

All of this casts considerable doubt on Gaus’s (2011a: 192) claim that ‘to renounce the practice [of the reactive attitudes] would be to renounce most of what we care for and value’. Gaus worries, following Strawson, that to reject the reactive attitudes would be to adopt the dreaded ‘objective attitude’ towards other persons whereby we see them 'as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled or cured or trained'. (Strawson, 2003: 79) But, while Pereboom’s alternative perspective does involve a greater degree of objectivity in interpersonal relations, it does not justify assuming an objective attitude in this sense. Pereboom’s alternative account thus shows that it is possible to renounce the reactive attitudes whilst leaving intact much of what we care about in life. In fact, his account may even serve to improve our interpersonal relationships by making us calmer, and less angry, judgmental and confrontational. And, with determination, education and the right social and cultural environment, there is no reason to think that a gradual adoption of this alternative moral practice is impossible: ‘the past several centuries have witnessed very significant changes in attitudes toward criminals, the insane, and children, and thus it cannot plausibly be argued that significant emotional change over time is not possible for us’ (Pereboom, 2014: 185).

Of course, for Pereboom, the search for non-reactive analogues and aspects of the reactive attitudes is motivated by the question of how personal relationships and meaning in life could be rationally sustained in the face of hard determinism. But one does not have to subscribe to Pereboom’s views on free will to accept that his alternative picture of social life is an interesting and plausible contender and thus to accept that the reactive attitudes can be rationally renounced. So, until Gaus engages with this possibility and explains where Pereboom goes wrong, it is not clear that he has grounds for saying that to renounce the reactive attitudes would be to renounce most of what we care for and value.