Abstract
Several important questions in applied ethics – like whether to switch to a plant-based diet, reduce greenhouse gas emissions, or vote in elections – seem to share the following structure: if enough people ‘cooperate’ and become vegan for example, we bring about a better outcome; but what you do as an individual seems to make no difference whatsoever. Such collective action problems are often thought to pose a serious challenge to consequentialism. In response, I defend the Reactive Attitude Approach: rather than focus only on the deontic status of (non-)cooperative acts – as existing consequentialist responses mostly do – consequentialists should also focus on when to praise or blame individuals. For example, even if your veganism does not change the overall outcome, consequentialists might still recommend praising you for your cooperative behavior. I make a consequentialist case for including praise and blame in our ethical practices around collective action and argue that doing so helps us capture important intuitions. I develop and defend these arguments by drawing on recent instrumentalist theories of blame and responsibility.
1 Introduction
Meat: If enough people stopped eating meat, a much better outcome would be brought about, because there would be less suffering and death in factory farms. But any one individual buying plant-based products instead of meat likely makes no difference at all. Worse, some individuals might even enjoy eating meat so much that switching to a plant-based diet would bring about a worse outcome.
Climate Change: If enough people reduced their greenhouse gas emissions, a much better outcome would be brought about. But any one individual reducing their emissions likely makes no difference at all. Worse still, if an individual foregoes activities that emit greenhouse gases but which she would enjoy greatly – like going on a nice holiday – she would bring about a worse outcome overall.
Again, consequentialism seems to say that any one individual in Climate Change need not reduce emissions. Worse, an individual even ought to emit emissions, when doing so has best consequences.
Intuitively, it seems that we still want to say something positive about people who cooperate – by for example reducing greenhouse gas emissions – even if individual acts seem to make no difference. Call this the No Difference Challenge. Some consequentialists take up the challenge and argue that, contrary to appearance, we individually do make a difference, because cooperative acts make a difference to the expected value of outcomes. Others instead argue that we should reject the challenge: individuals should not perform cooperative acts in collective action problems when doing so is futile and costly. This is a plausible feature of the view, not a bug.
In this article, I argue that consequentialists should move beyond the deontic status of (non-)cooperative acts and include a consequentialist ethics of praise and blame for collective action cases. I present two lines of argument for what I call the Reactive Attitude Approach.
First, including praise and blame makes consequentialism more plausible as an ethical theory for collective action, because it captures some intuitions. For example, even if eating meat sometimes has better consequences, there still seems something commendable about someone sticking to a plant-based diet. A consequentialist ethics of praise would typically praise such cooperation, even when it does not do the most good. Conversely, it might sometimes recommend blaming non-cooperative behaviour, such as causing unnecessary emissions, even if such non-cooperation has good consequences.
Second, I also make a ‘practical’ consequentialist case. Real-world ethical practices around collective action will be more effective when they include practices of praise and blame, most importantly because those help resolve collective action problems. For example, they help establish, spread, or entrench collective normative expectations around cooperation, increase the expected impartial value of cooperative acts – particularly through ‘indirect effects’ – and foreground the social importance of cooperative dispositions.
To develop the Reactive Attitude Approach, I draw on recent writing on instrumentalism about responsibility and blame. Traditional instrumentalist accounts, such as J.J.C. Smart’s, are often seen as too simplistic. Recent theorizing, however, provides a richer story about the function and value of reactive attitudes in social life. This story, I argue, provides the right framework for a consequentialist ethics of collective action.
I proceed as follows. In section 2, I briefly survey how consequentialists have responded to the No Difference Challenge. In section 3, I describe my instrumentalist framework for praise and blame. In section 4, I use this instrumentalist framework to defend and spell out the Reactive Attitude Approach. I conclude in section 5.
2 Consequentialism and Collective Action
I now discuss two ways in which consequentialists have tried to approach collective action problems by focusing on the deontic status of acts. This will set the stage for my own account which argues that we should move beyond deontic status and also consider praise and blame.
2.1 Existing Responses
The first response to the No Difference Challenge is what I call the Expected Value Response: instead of actual outcomes, we should focus on the expected value of our actions. Often, cooperative acts – eating vegetarian, reducing emissions, etc. – have higher expected value than non-cooperation.3
The expected value of an act is the sum of the value of all possible outcomes multiplied by their probabilities. Even if you often make no actual difference in collective action problems, ex ante you have a small probability of making a big difference.4 For example, if you buy a chicken in a supermarket, there is a large chance that you make no difference whatsoever but a small chance your purchase will push demand above a threshold. Say you buy chicken number 10,001. If demand goes above 10,000, the supermarket will next month order 11,000 rather than the usual 10,000. You could thus make a big difference, which compensates for the low probability. Accordingly, buying a chicken has lower expected value than buying the plant-based option.
Mark Budolfson argues that the Expected Value Response moves too quickly (Budolfson 2019). It typically assumes that the expected impact of one person’s cooperative action equals the average expected impact of all (relevant) people cooperating. In real life, this assumption is often false.5 For example, several empirical factors suggest that the chance that one chicken purchase will make a difference is negligible, as suppliers might order meat with buffers and carry more inventory than there is demand on a normal day. Moreover, your chance of affecting the market price is tiny and meat producers’ investment decisions are not that price-sensitive to begin with. In response, McCullen and Halteman (2019) contest Budolfson’s empirical claims. And Brian Hedden responds that Budolfson has not shown that a lower probability would not be offset by a bigger potential impact (Hedden 2020).6
Instead of the above ‘direct effects,’ a second argument for the Expected Value Response foregrounds indirect effects of cooperation.
First, my cooperative acts might affect other people. When things go well, my cooperation can inspire others to cooperate also. Conversely, my non-cooperation might sometimes signal to others that it is okay not to work together to solve collective action problems or that, given my non-cooperation, it might be pointless for them to try.
Second, my cooperation might help change choice environments. For example, if people switch to a vegetarian diet, more and better vegetarian options will become available, which in turn makes it easier for others to reduce meat. Moreover, cooperating – and convincing others to cooperate – can produce economies of scale: if enough people become vegetarians, vegetarian products become cheaper as they are produced on a larger scale (McMullen & Halteman, 2019, 107).
Of course, it is unclear whether indirect effects are weighty enough to counter the No Difference Challenge.
First, will my expected indirect effects ever be big enough to tip the balance? For example, can my expected indirect effects ever be large enough to prevent or even slow down climate change?
Second, for indirect effects, it matters that I am seen to be cooperating. If I enjoy eating meat when no one is watching, my buying meat might still have greater expected value than my buying vegetarian food.
Finally, some indirect effects might themselves encounter the No Difference Challenge. For example, economies of scale and changing choice environments only happen if enough people act cooperatively. My single act might make no difference whatsoever.
The Reactive Attitude Approach works whether the Expected Value Response works or not. However, I think it is reasonable to assume the following: first, there is still significant uncertainty around how far individual behavior might have an impact both in terms of direct and indirect effects; second, the debate around the Expected Value Response is not yet settled. Accordingly, it is rational to assign a significant credence to the view that in at least some central collective action problems our expected impact is too low to make a difference.
The second response that some consequentialists proffer is what I call the Bullet-Biting Response: rather than contest that consequentialism faces the No Difference Challenge, consequentialists might instead accept its implication for the deontic status of cooperative acts.7 Consider two arguments.
First, when lone cooperative acts are costly and futile, it seems that we have good reason not to do them. Act consequentialism would judge that we ought not do such acts. Given their costliness, is this judgment not intuitive?
Second, collective problems require collective solutions. For example, obsess less over the little things that we can do as individuals to reduce emissions and instead tackle the systematic changes – legislation, energy transition, carbon taxes, and so on – that are our only hope of meeting climate goals (Budolfson 2020a; Hindriks 2019; Kingston & Sinnott-Armstrong 2018). In fact, focusing too much on individual consumers can have negative indirect effects and distract us from tackling more important collective solutions.
2.2 Praising and Blaming
As said above, I defend the Reactive Attitude Approach on both practical and theoretical grounds: I argue, first, that there are consequentialist reasons to attend to praise and blame in an ethical practice around collective action and, second, that it helps us capture intuitions.
Justine: Two candidates are left in a presidential election decided by proportional voting. One candidate is a fascist, the other one is a decent person and politician. The fascist candidate has a comfortable margin in the polls and the polls are very reliable. Justine is enjoying a funny tv show, so much so that if she stays at home watching tv, she will bring about a better overall outcome than if she went out to vote for a candidate who will lose anyway.
Assume that Justine knows that her vote will not make a difference. Motivated to bring about the best outcome, she decides to stay home. However, even though Justine did bring about a better outcome (both in expectation and objectively), intuitively, we might not find her decision praiseworthy. Should Justine not feel more bound by collective norms of citizenship and a collective sense of responsibility instead of opting out so easily?
Henk: Henk loves the taste of meat. But he is convinced by the moral arguments against the farming of animals for food, particularly the unnecessary suffering and killing of sentient beings. Even though he would derive significant benefits from eating meat, he does not do it.
Assume, arguendo, that Budolfson is right for the decisions that Henk faces: he knows he will not make a difference, neither through direct nor indirect effects. What should we say about Henk’s behavior seeing that he did not do the most good? Intuitively, it seems commendable that Henk eats a plant-based diet even if he does not do the most good. Moreover, some of us might have the intuition that not eating meat – and foregoing one’s own pleasure – seems an appropriate response to the industrial torture and slaughter of billions of animals every year. However, neither the Expected Value Response nor the Bullet-Biting Response seems to capture such intuitions.
More generally, I – and certainly many of my students – have the intuition that there is something desirable and commendable about people being ‘the change they want to see in the world’ and about people, like Henk, foregoing their own self-interest to do so. For Justine, I also have the intuition that political action to avoid fascism is inherently collective and that there is something commendable when citizens do not opt out of collective action too easily.
The Reactive Attitude Approach has a simple answer. We should distinguish an act’s deontic status from whether one should blame or praise someone. The responses considered so far all focus on the deontic status of individual acts. But deontic status should not exhaust a consequentialist ethics of collective action. Instead, praising and blaming play an important, and sometimes separate, role too.
For consequentialists, whether and how we should blame is itself a consequentialist question: we ought to praise or blame the person, if doing so has good (or best) consequences. By decoupling praise and blame from an act’s deontic status, consequentialists have the resources to capture intuitions that we have about collective action problems.
First, in Henk, consequentialists should recommend that we typically praise people like Henk for adopting a plant-based diet. And they should likely do so, even if Henk does less good by not buying meat (both in expectation and objectively). While his act might not have been right, it likely is appropriate to praise Henk.
Second, in Justine, consequentialists should likely be inclined to hold that, even if Justine’s act were right according to act consequentialism, she should typically not be praised for her decision to stay home – if anything, it might be appropriate to blame her. Or, conversely, it seems appropriate to praise citizens who display a robust disposition to go and do their part in preventing fascism, for example if they are disposed to go and vote even when there is little chance to be pivotal in an election.
More generally, consequentialists can allow that it is sometimes appropriate to not praise someone, or even to blame them, even though their act did more good than cooperation (act consequentialists would then allow for ‘praiseless rightdoing’ and ‘blameworthy rightdoing’). And sometimes it is appropriate to not blame someone, or even to praise them, even though their act did not do the most good (act consequentialists would allow for ‘blameless wrongdoing’ and ‘praiseworthy wrongdoing’) (Parfit 1984, 25). So, moving beyond the deontic status of acts gives consequentialists a richer structure to capture more intuitions about collective action problems.8
Before presenting my central arguments, I first outline a more systematic account of how consequentialists should think about the ethics of praise and blame.
3 Instrumentalism about Blame
Praise itself comes to have some of the social functions of medal giving: we come to like praise for its own sake, and are thus influenced by the possibility of being given it. Praising a person is thus an important act in itself – it has significant effects. A utilitarian must therefore learn to control his acts of praise and dispraise …
smart 1973, 49–50
The idea is simple: we like praise and dislike blame. The justification of praise and blame lies in their capacity to incentivise better moral behavior (Smart 1961, 1973). Let us call the view associated with Smart, Simple Instrumentalism.9
Instrumentalism, more generically, is the view that the justification of holding other people responsible lies exclusively in the instrumental value brought about by doing so. For a long time, instrumentalism was very unpopular. However, with the emergence of more sophisticated theories, instrumentalism has recently seen a revival.
I here outline central moves within what I call New-Wave Instrumentalism.10 I will not defend those moves – even though I find them plausible – nor will I defend any one person’s theory in particular. Instead, I wish to show that those moves provide a plausible structure for a consequentialist ethics of collective action. I spell out those moves by highlighting two ways in which New-Wave differs from Simple Instrumentalism.
3.1 Practices
Simple Instrumentalism focuses on acts of praising and blaming. New-Wave Instrumentalism, in contrast, focuses on blaming practices.11 By switching to practices, New-Wave Instrumentalism can avoid two problems that beset Simple Instrumentalism.
First, by focusing directly on acts, Simple Instrumentalism provides a purely forward-looking perspective: whether to blame Henk is entirely determined by future effects. However, one might worry that purely forward-looking justifications are normatively inappropriate: we typically think that backward-looking considerations should figure centrally in an ethics of blame. Whether blaming Henk is appropriate seems to depend primarily not on whether blaming him is expedient but on what Henk has done in the past and on whether he fulfills responsibility conditions. As one might put it, purely forward-looking considerations are the wrong type of reasons.
This is where the move to justifying blaming practices helps, because blaming practices will only be effective if they include backward-looking appraisal. Here are two reasons.
… as one drops off the defining features of punishment one ends up with an institution whose utilitarian justification is highly doubtful. One reason for this is that punishment works like a kind of price system by altering the prices one has to pay for the performance of actions it supplies a motive for avoiding some actions and doing others. The defining features are essential if punishment is to work in this way; so that an institution which lacks these features, e.g., an institution which is set up to “punish” the innocent, is likely to have about as much point as a price system (if one may call it that) where the prices of things change at random from day to day and one learns the price of something after one has agreed to buy it.
rawls 1955, 12
Similarly, entirely forward-looking blaming practices will be ineffective, as they are not action-guiding: they would not be effective at letting me know how I can adjust my behavior so as to avoid blame or garner praise.12 A practice of blame with backward-looking considerations will be far more effective.
Second, without backward-looking appraisal, blaming practices will not be effective, because blame is about not just acts but also people’s dispositions. Intuitively, people distinguish between an act’s deontic status and the question of whether someone should be praised or blamed. For the latter question, we care about an agent’s disposition. We typically blame people for manifesting certain behavioral dispositions, such as vices or undesirable behavioral patterns and habits.13 Now, if dispositions – in a broad sense – are the typical target of reactive attitudes, then blaming practices will naturally include backward-looking considerations, as someone’s dispositions reach into the past (and often the future).
So, switching to practices instead of blaming acts helps New-Wave Instrumentalism include backward-looking considerations.
A second problem with Simple Instrumentalism is that by focusing only on acts, it does not give the right answer in an ethics of blame: the ethics of blame is primarily about whether blaming or praising Henk is appropriate or not (or whether Henk is blameworthy or praiseworthy). It is not simply about whether blaming Henk has good consequences or not. For it could easily happen that it might have best consequences to blame Henk, even though Henk does not strike us as blameworthy.
Again, this is where moving towards practices helps. New-Wave Instrumentalism will endorse effective blaming practices that will have norms and rules about when and how to blame – which in turn will feature backward-looking considerations. With those, it will have practice-internal normative standards for when it is appropriate to praise or blame someone.14 And those appropriateness standards are no longer determined just by contingent effects of individual acts of blaming and praising. Consider punishment as an analogy. A practice of punishment that just randomly punished people or that had absolutely no concern for whether someone had indeed done certain acts or not would likely be far less effective at guiding behavior than a practice that had such standards. An effective institution for punishment will have practice-internal appropriateness standards. Similarly, blaming practices that completely neglected who had done what and eschewed any responsibility standards would be unlikely to guide behavior very well. Accordingly, effective blaming practices will feature appropriateness standards.15
3.2 The Function of Blame
Smart compared praising to handing out medals: their function is to incentivize better moral behavior. New-Wave Instrumentalists, in contrast, tell a richer story about the function of blame. Before I tell this story, a quick clarification is in order. New-Wave Instrumentalists often lean towards a broadly Strawsonian picture of responsibility, praise, and blame, which I also assume for now. When I talk about ‘praise’ and ‘blame,’ or sometimes just ‘blame’ as a shorthand, it includes a richer set of reactive attitudes, including gratitude, resentment, disappointment, reproach, and so on. Note also that New-Wave Instrumentalism could be developed, mutatis mutandis, with different accounts of blame.16 For example, along with Scanlon, it could put greater emphasis on cognitive rather than emotive attitudes or move from ‘blame’ strictly conceived towards ‘less angry’ reactive attitudes (Brandenburg 2018; Fricker 2016; Scanlon 2008).
I now describe three functions: blame has a constitutive, motivational, and communicative function.
… adopt the objective attitude to another human being is to see him, perhaps, as an object of social policy; as a subject for what … might be called treatment … it cannot include the range of reactive feelings and attitudes which belong to involvement or participation with others in interpersonal human relationships;
strawson 1962, 126–127
New-Wave Instrumentalists argue that, rather than a problem for instrumentalism, Strawson’s point can buttress it. If relationships require reactive attitudes and if such relationships are valuable – as they often will be – then instrumentalists can hold that this makes reactive attitudes valuable.17 The point becomes even stronger when we focus on practices of praise and blame, and the dispositions underlying them. Such blaming dispositions are entwined in the social practices and relationships that Strawson holds are valuable. As such, they become valuable too.
Building on this, we can now appreciate the constitutive value of dispositions to blame. The idea is that social practices and relationships involve norms and normative expectations. And those normative expectations in turn require dispositions to respond with reactive attitudes. Imagine that several members of an underrepresented group form an informal group in which they vow to support each other in their professional lives. There is an explicit commitment to solidarity and friendship. Such a commitment only seems genuine if participants are also somewhat disposed to respond with some reactive attitudes should someone egregiously violate their solidaristic aim. Imagine, for example, that one member took confidential information about another group member and used it as leverage to gain advantage in their career. Commitment to the solidaristic goal implies that other members of the group should respond with some negative reactive attitude upon hearing of this flagrant violation of trust and solidarity. If they were disposed to respond with utter indifference, their relationship would not be structured around genuine normative expectations.
The more general point is that genuine commitment to norms and normative expectations in social practices and relationships often implies that participants are disposed to respond with some reactive attitude when others violate such normative expectations. And if such relationships and practices are valuable, and if dispositions to blame partly constitute such relationships and practices, blame can be constitutively valuable.18
The second function of blame is motivational. Simple Instrumentalism of course already argued that blame is justified by its motivational or ‘influencing’ function. For New-Wave Instrumentalism, even this motivational function works in a more complex manner than Simple Instrumentalism suggests.
First, the motivating function does not boil down to the motivational function of individual acts of blaming but to that of the practice. As an analogy, most people are deterred from breaking the law without the law ever having to punish them. The existence of the practice is often already enough to achieve that effect.
Second, the motivational function does not boil down to simply incentivizing better acts. Instead, several instrumentalists argue that a central aim of reactive attitudes is to cultivate better moral agency.19 Praise and blame are different from simply communicating that an act was right or wrong. Typically, they have an agent’s dispositions or the behavioral pattern as their object. The aim of blame is then typically also for people to adopt better dispositions and, in a sense, become better moral agents.
The third function of blaming practices is communicative.20 In social practices, our wider moral community, and in smaller-scale human relationships, we often need to publicly uphold and coordinate shared normative expectations. Blaming practices play a role in doing so. Through them we communicate what we expect others to do but also what we expect of us as joint agents to do together. Moreover, through blame – and discourses around when blame is appropriate – we can also coordinate and negotiate such expectations. Finally, in addressing those who do not live up to such expectations, blame connects the communicative with an ‘educational’ function: we let others know what we expect of them but also that we view them as moral agents whom we expect to live up to normative expectations.21
To sum up, New-Wave Instrumentalism justifies practices of blame through their instrumental value. Unlike Simple Instrumentalism, however, it typically focuses on blaming practices. This allows for backward-looking blaming practices with internal appropriateness standards. Moreover, it tells a richer story about the value of blame: practices of blame can have constitutive value by intimately connecting to the normative expectations underlying our practices and relationships, can have a complex motivational function that primarily targets people’s dispositions, and a communicative function in upholding, negotiating, and coordinating socially shared normative expectations.22
4 The Reactive Attitude Approach
A. There is a good consequentialist case to endorse ethical practices around collective action that go beyond the deontic status of individual (non-)cooperative acts and also include effective practices of praise and blame.
B. Such praising and blaming practices would at least sometimes praise co-operation and sometimes blame non-cooperation in a way that accounts for some of our intuitions.
By ‘ethical practices’ I mean our ethical discourse, norms, and normative expectations around collective action as they happen practically in society, rather than just in academic philosophical discourse.
Of course, A and B are (partly) empirical hypotheses. But I now outline various structural features to motivate A and give reasons for why (act) consequentialists should adopt this ‘richer’ ethics of collective action. In the process, I also give armchair reasons for B. To defend both A and B, I draw on New-Wave Instrumentalism to fill in some details of the Reactive Attitude Approach. I do so by going through its different features, starting with blame’s different functions.
4.1 The Function of Blame
I above distinguished blame’s motivational, constitutive, and communicative function. Start with the first.
Blaming practices have a motivational or influencing effect, which helps with collective action problems. One way to solve collective action problems is through social norms and normative expectations. Blaming practices here tie into those expectations and incentivize cooperation. People do not like blame but like praise – and the argument mostly extends, mutatis mutandis, to the more complex reactive attitudes that will feature in social life. For example, even if an individual decision not to eat meat were to do less good than eating meat, consequentialists should favor practices and social norms that move enough individuals to switch to a plant-based diet. Blaming those who do not switch to a plant-based diet or praising those who do might motivate people to switch to a plant-based diet.
Of course, one might wonder whether the argument would still work if everyone were a perfect consequentialist. Perfect consequentialists, one could object, would not care so much about whether they are praised or blamed. After all, they care about everyone’s wellbeing. However, I mean for my argument primarily to apply to the non-ideal real world in which even the more impartial among us do care about the reactive attitudes that we encounter (also see section 4.3). Moreover, few people are committed consequentialists. But even in a world populated with consequentialists who otherwise psychologically resemble real-life agents, the argument should still go through.
Second, dispositions to blame can have constitutive value. Commitment to certain normative expectations in social practices and relationships often comes with a commitment to respond with reactive attitudes. This feature comes in handy for the Reactive Attitude Approach: stable cooperation often requires not just a statistical expectation that others will act a certain way. Instead, robust commitment to collective action often requires normative expectations towards cooperation (Bicchieri 2017, Chapters 1; 3). Being disposed to blame when some diverge strongly from the normative expectation to cooperate is thus often a necessary part of a normative commitment to cooperation. Without dispositions to respond with reactive attitudes, an effective and sincere commitment to cooperative social norms is hard to imagine.
For example, consider Justine again. The more general claim is that a commitment to democratic citizenship and its concomitant duties implies normative expectations. Among such expectations might be that one should show up and vote when there is a choice between a fascist and a decent democrat. The constitutive claim simply holds that a genuine commitment to a normative ideal of citizenship implies that we are somewhat disposed to disapprove when groups and individuals fail to meet central expectations around citizenship. Having such norms around citizenship – and a normative commitment to democracy more generally – might be valuable in itself. Or it might just be instrumentally valuable in helping us mobilize enough citizens to overcome collective action problems, such as those often present in general elections.
Finally, New-Wave Instrumentalism holds that reactive attitudes have a communicative function. By expressing reactive attitudes, we sometimes communicate shared normative expectations but also facilitate discourses through which we can negotiate and coordinate normative expectations.
For example, if we focused only on Justine’s action and her individual impact, we would miss other aspects that often play a role in moral discourse around collective action problems. If instead we also think about whether Justine’s behavior is worthy of certain reactive attitudes, we shift the discourse away from Justine’s individual action towards the normative expectations that we have towards citizens like Justine. Expressing reactive attitudes in cases like Justine foregrounds the normative expectation which we apply to groups of agents to which Justine belongs. It is not just about the person being blamed in isolation but also about how we should act together to solve a collective action problem and about how individual participants are expected to play their part.
Or consider Henk. I argued that it can make sense to praise Henk for his decision to switch to a plant-based diet, even though he would be happier eating meat. Our praising then is not primarily about tracking his acts’ deontic status. Instead, praise’s communicative function here provides information and signals to others. We communicate that it is valuable if enough of us switch to a plant-based diet and we approve of Henk playing his part. Expressing reactive attitudes can signal norms or norm changes. Moreover, it can express a willingness or ‘open invitation’ to others to also switch to a plant-based diet and thereby help solve a collective action problem.
Of course, discourses around blame and reactive attitudes need not always take all normative expectations as settled. Sometimes discussing whether individuals should be blamed for something – for example, should Justine be blamed for staying home? – can be ways for groups or whole societies to discuss, negotiate, or coordinate around normative expectations or to make more precise what a general normative expectation implies in specific cases. Discourses around when reactive attitudes are appropriate can thus help us coordinate how we solve problems together.
So, reactive attitudes have a motivational, constitutive, and communicative function that make it more likely that real-world ethical practices can solve (some) collective action problems. Accordingly, there is a good consequentialist case for them.
4.2 Dispositions
New-Wave Instrumentalism holds that blame typically does not just appraise a single act but has a person’s dispositions (or motives or behavioral patterns) as its object. I now argue, first, that dispositions are important in our ethical practices around collective action and, second, that the Reactive Attitude Approach captures their importance.
Individual impact: if I adopt a particular disposition, what difference would I make?
Collective impact: if I and some number n (>0) of other actors adopt a particular disposition, what difference would we make?
Start with individual impact. Dispositions and their individual impact matter in regard to how we can practically approach collective action problems. More generally, consequentialists of all stripes, including hardened act consequentialists, distinguish between consequentialism as a criterion of rightness and consequentialism as a decision-making procedure.23 How we should make decisions is in itself a consequentialist question, and its answer is not always ‘actively try to do the most good in every decision.’ Consequentialists hold that dispositions where we do not try to maximize the good in every decision are extremely valuable. They are necessary, as the alternative is simply not available in the real world, given cognitive, epistemic, and time constraints. Sometimes, such non-maximizing dispositions even outperform case-based maximization (Todd & Gigerenzer 2012).
More to the point, dispositions are important in collective action problems.
First, many salient examples of collective action problems, such as climate change and meat consumption, are either recurring or stretch across time. For a consequentialist agent, it will typically be more effective to adopt a somewhat stable disposition, such as not eating meat, rather than calculate expected consequences in every decision. For example, it would not be effective to calculate afresh one’s expected consequences for any food-related or transport-related decision.24
Second, remember the Expected Value Response and the idea of thresholds. I earlier left it open whether the threshold argument works or not. But, still, it seems reasonable to believe that across different decision domains, there can be significant uncertainty about whether or where thresholds might lie. If I go to an ‘eat local’ restaurant, I might have to do a different calculation about thresholds than in a big supermarket chain (McMullen & Halteman 2019). I cannot just extrapolate across food environments. Such uncertainty will increase further when we consider collective actions across domains. Calculations about meat purchasing decisions might not easily extrapolate to voting or climate change decisions. And, again, for any emissions decisions, the calculations should vary between decision domains.
If there is enough uncertainty about whether one might affect any thresholds, it might become rational to adopt cooperative dispositions. When an agent is faced with severe constraints on how much ‘action-specific’ and ‘domain-specific’ reasoning they want to do, they might adopt more general cooperative dispositions. For example, it might be more effective to adopt a stable disposition across ‘food domains’ than to reason afresh about the deontic status of individual dietary choices whenever you go to a different restaurant or shop. And adopting such dispositions might then capture that there is enough uncertainty about potential thresholds, but insufficient time or information to reduce it, such that adopting cooperative dispositions can be a rational moral shortcut.25
Additionally, we also face significant uncertainty around indirect effects. It seems hard to estimate how much one’s behavior in collective actions might affect others. We might underestimate this effect, particularly when it is unintended. For example, after I became vegetarian, several people engaged me in conversation about it. I was then surprised to see that a handful of people switched to either a vegetarian or a meat-reduced diet as a result. The real number is probably higher, as some of those who switch might in turn inspire more people to switch. At the same time, we can sometimes overestimate our effect: some of those who switched might have done so anyway at some point, so my impact was only to make them switch earlier. So, estimating one’s indirect effects involves a lot of uncertainty. It seems even harder to estimate one’s indirect effects for individual actions. Given this uncertainty, it can make much sense to adopt more stable dispositions and habits.26
Finally, indirect effects support the consequentialist case for cooperative dispositions in another way. You are more likely to inspire others and instill cooperative norms if your cooperative behavior expresses a stable disposition instead of a one-off fluke. The person who has the vegetarian dish just once because she fancies some rigatoni is unlikely to inspire others to switch to a plant-based diet. But if someone switches consistently to a plant-based diet, she might have some chance to inspire others and bring about changes in choice environments.
Overall, in collective action problems, we should pay attention to the individual impact of dispositions rather than focus only on the deontic status of individual acts.
Collective impact: if I and some number n (>0) of other actors adopt some disposition, what difference would we make?
Justine*: Like in Justine, the fascist candidate A is expected to have a clear majority over the decent candidate B. There are two differences. First, a big portion of democratically minded voters are not inclined to go. If they went, they would all vote for B, and B would win. Call them the ‘decent non-voters.’ Second, the election happens during a global pandemic. So, there is a tiny but existent risk that if one goes to vote, one might contribute to the spread of the virus. However, the risk is far smaller than, for example, going to the supermarket. Therefore, despite the small infection risk, the outcome would be better, if enough of the decent non-voters went and voted for B rather than A.
Justine knows that the ‘decent non-voters’ are not going to vote and that A will win. So, given the small but existent risk of adding to the pandemic, it is better if Justine does not go.
Disposition 1 (D1): ‘I only go to vote, if I get good signals that enough other pro-democracy voters are going such that there is a decent chance that we will make a difference.’
Justine does not receive strong signals that others are going and does not go. Imagine also that enough others in the group of ‘decent non-voters’ also have disposition D1. As a result, they do not go either, and the fascist wins.
Individually, adopting D1 helped them do the right act according to consequentialism (let us assume): they avoided engaging in costly and futile acts. However, the collective impact of adopting D1 was bad, as it brought about an outcome in which a fascist was elected. If consequentialists evaluated D1 only by its individual impact – and we can ignore potential indirect effects here – D1 might not be a bad disposition. But when judged by its collective impact, it is bad.
D2: ‘In elections where there is a chance that a fascist might win, I go and vote for the democratic over the fascist candidate to do my bit, unless the consequences of my going out to vote would be much worse than my not going.’
D2 might lead Justine to do acts that are wrong according to consequentialism, namely when going out to vote has expected consequences that are only a little bit worse than not voting. Moreover, in many plausible scenarios, D1 might outperform D2 in terms of individual impact. However, D2 has likely better expected collective impact. If enough of the decent voters in Justine* adopted D2, they would together prevent a fascist from rising to power.
So, from a consequentialist perspective, dispositions matter in collective action problems both through their individual and their collective impact. The Reactive Attitude Approach helps us capture the importance of dispositions and helps anchor them in our ethical practices of collective action.
First, the Reactive Attitude Approach provides an evaluative framework that naturally focuses on dispositions rather than (just) individual acts. As mentioned earlier, praise and blame are rarely about individual acts and their deontic status alone. Instead, they tend to carry a message about and sometimes to the agent. So, for consequentialists, the Reactive Attitude Approach provides a good framework to capture how we (should) evaluate dispositions in collective action problems.
Moreover, blaming practices can function to motivate people to adopt good dispositions – rather than just doing what is right in individual instances – including cooperative dispositions. This, as we have seen above, is important in collective action problems, where dispositions can have an individual and collective impact.
Second, the Reactive Attitude Approach captures particularly well the collective impact of dispositions. Some argue that to capture this collective dimension, consequentialists need to abandon simple act consequentialism and switch towards either rule or collective consequentialism (Parfit 1984, Chapter 1; Regan 1980). Others instead argue that this brings out the inherent limitations of consequentialism: non-consequentialist theories like Kantianism or Contractualism will much better capture such ‘collective reasoning.’
With the Reactive Attitude Approach, consequentialists can capture this collective dimension. Moreover, they can do so without rejecting act consequentialism (although the view is compatible with other forms of consequentialism, like rule consequentialism, too). Imagine that in Justine*, our protagonist Justine adopts disposition D2. D2 is a type of disposition that, at least in the conditions considered here, is likely to have a comparatively good collective impact. Practices of blame would thus do well to help create collective normative expectations towards adopting D2 (or dispositions like it). Praising Justine can help in that. So, an effective praising practice would likely consider it appropriate to recommend praising Justine. Moreover, praise and blame have a communicative function: praise and blame help communicate normative expectations towards other citizens. When praising Justine for adopting D2, we communicate not only an evaluation about Justine. Instead, we often also communicate expectations to her and other people about the dispositions we expect them to adopt. Blame’s communicative function thus naturally connects with a form of reasoning sometimes thought to escape consequentialism: what would happen if not just I but also others adopted this disposition? By considering dispositions’ collective impact, the Reactive Attitude Approach captures such reasoning.
Overall, the Reactive Attitude Approach captures well that in our ethical practices around collective action we should care not only about actions and their deontic status but also dispositions and their individual and collective impact. The latter is no small gain: it shows that one need not give up act consequentialism – let alone consequentialism more generally – to capture the collective dimension of moral reasoning.
4.3 The Reactive Attitude Approach and Expected Value
I now present three ways in which effective practices of praise and blame – and the normative expectations they help sustain – can increase the expected value of cooperation. This further shows how effective practices of praise and blame help resolve collective action problems.
First, in earlier cases, non-cooperation was all-things-considered better because of its prudential value, for example because someone might enjoy flying somewhere nice. Effective praise and blame practices will often increase the comparative prudential value of cooperation. But interestingly, the expected impartial value of cooperative acts can go up too. For consequentialist agents, their own good matters just as much as anyone else’s. So, for agents who really dislike blame or like praise, that might sometimes be enough to tip the impartial balance.
Second, norms around collective action problems can change the social meaning of their constituent acts. For example, when I was young, taking the train rather than a plane to faraway places was likely interpreted as a sign of not being rich enough to buy a plane ticket. Nowadays when someone goes on a long business trip and takes the train rather than the plane, I tend to interpret this as a commitment to cutting emissions. Once act types are seen as part of a social practice meant to tackle a collective action problem, their social meaning can change. Along with the new meaning, their value can change too. For example, such acts can then signal to others one’s willingness to cooperate (Lawford-Smith 2016); this, in turn, can increase the chances of affecting others, which in turn increases the expected value of cooperative acts.27 But note that for acts to acquire this social meaning, certain norms – for example against flying or eating meat – must typically be sufficiently well-known and widespread. And, as argued above, that will often first require practices of praise and blame that play an important role in establishing, communicating, and upholding such norms.
Finally, cooperative acts can be necessary for a person to have standing to praise and blame. If you express reactive attitudes, you might nudge more people towards cooperation. Moreover, you can communicate with your peers and might even convince people outside your social circle (through social media, for example). So, your opportunities to affect positive change in collective action problems are not exhausted by your cooperative acts. However, to be effective in this way, you typically require appropriate standing. To have standing, you should not be (too much of) a hypocrite. For example, to have standing to blame someone for a certain disposition, one should not display the same disposition oneself: if I blame you for eating meat whilst chomping on a steak, you will typically deny me standing to blame. Standing to praise has been less analyzed, but existing work suggests that hypocrisy threatens (or at least lowers) one’s standing to praise too. So, being a co-operator affords you standing to praise and blame and thus the credibility to influence others. This, in turn, drives up the expected value of cooperative acts, if such acts are followed up by attempts to change other people’s behavior.28
We thus get the intuitive result that effective praise and blame practices can drive up the expected value of cooperation and thus add consequentialist pro tanto reasons to adopt cooperative dispositions.29
4.4 Is Non-Cooperation Always Blameworthy?
As I earlier wrote, the hypothesis is that effective praising and blaming practices would at least sometimes praise cooperation and sometimes blame non-cooperation in a way that accounts for some of our intuitions. I have not argued that cooperative acts are always praiseworthy and non-cooperative acts always blameworthy. Still, readers have wondered about the scope of this relationship: when is cooperation praiseworthy and non-cooperation blameworthy? The Reactive Attitude Approach should have enough empirical ‘flexibility’ for two reasons: (i) such flexibility is practically important: for example, a practice that always blamed non-cooperators would likely not be the most effective; (ii) it is theoretically important: in some cases, it seems inappropriate to praise cooperators or blame non-cooperators.30
Entrenchment: In situation S1, nearly everyone eats meat and next to no one considers this behaviour wrong. In S2, most people are either vegetarians or akratic meat eaters.
Assume that in S2, dispositions to blame are likely instrumentally justified but not in S1 where there is too little uptake for blaming to do any good. How far behavior and norms are already entrenched seems to affect whether blaming does any good. Can the Reactive Attitude Approach capture that?
I think so. For this, we should ask more general questions: for any situation S, should the Reactive Attitude Approach recommend any dispositions to respond with reactive attitudes (scope)? And, if yes, what type of reactive attitude (praise, blame, or both) would it recommend in S (valence)?
For the consequentialist, scope and valence will be determined by which practices of praise and blame are most effective in bringing about good outcomes. How far norms are already entrenched will impact when and which reactive attitudes will bring about good outcomes. For example, in S1, blaming meat eaters is unlikely to turn them vegetarian or to establish a cooperative norm. It might even cause what psychologists call ‘reactance’ where people react to a ‘behavior change intervention’ by doubling down on their previous behavior. However, S1 might still fall under the Reactive Attitude Approach’s scope: praising meat eaters who make gradual changes might work quite well and also help with gradual norm change. In S2, in contrast, the vegetarian norm is already entrenched such that praise as well as occasional negative reactive attitudes towards norm violation might help sustain and further entrench it.
Single Mother: Judy, a single mother of two, lives in a place with poor public transport and refuses to give up her car to reduce her greenhouse gas emissions.
If the Reactive Attitude Approach always blamed non-cooperation, it would recommend blaming Judy. To many, such blame would seem inappropriate.31
However, concerns around effectiveness likely prevent this counterintuitive result. For praising and blaming to be effective, one must pick effective exemplars: the behavior, disposition or person blamed must be well-chosen to affect behavior and norms. For several reasons, Judy’s non-cooperation is not an effective exemplar.
First, effective normative expectations around cooperation would make room for ‘exemption clauses,’ particularly when the moral opportunity costs are high. There are important goods beyond carbon emission. And for Judy, ditching her car would come with high moral opportunity costs, impacting both her own and her children’s wellbeing. Moreover, an effective exemplar should not make it seem that reaching ‘blameless cooperative behavior’ is an excessively demanding standard; otherwise, people might be discouraged to even try. Again, this makes Judy an ineffective exemplar.
Second, people dislike being the person (the ‘sucker’) who makes sacrifices to achieve a collective goal only for others to free-ride or, worse, take advantage. Similarly, when addressing collective action problems, people care about how the benefits and burdens of cooperation are distributed. Compared with Judy, there likely are people and collective agents (like very rich individuals or polluting companies) who engage in more polluting behavior and for whom cooperation would be less burdensome.32
Overall, effectiveness will guide when and how to use reactive attitudes in a way that likely accounts for our intuitions. Of course, given limited space and ‘data,’ I have only motivated this claim by responding to two examples that were put to me. A thorough account of the scope and valence conditions would require more space (and empirical input).
4.5 Is this Still Compatible with Act Consequentialism?
I have defended the Reactive Attitude Approach and argued that it is available even for act consequentialism. Which version of consequentialism to adopt is orthogonal to the Reactive Attitude Approach. I have implicitly assumed act consequentialism, because it most obviously and directly encounters the No Difference Challenge. Moreover, that my account works even when combined with act consequentialism makes it quite robust, as it can be combined with other forms of consequentialism that face the No Difference Challenge less strongly. But two readers have wondered whether, ultimately, the Reactive Attitude Approach moves too far away from act consequentialism and requires something like rule consequentialism. Two reasons were given for this suggestion.
First, New-Wave Instrumentalism focuses on practices of blame rather than acts. Moreover, it draws on the norms internal to good practices of blame to reconstruct norms around when to praise and blame. Does this not imply a rule or practice-based consequentialist framework where the rightness of a blaming act is determined by the best practices?
However, this impression is erroneous. Remember that act consequentialists recognize that social practices can be immensely valuable and that we should often adopt decision-making procedures (which I grouped under ‘dispositions’) other than the act consequentialist criterion of rightness. Putting these two points together, act consequentialists recognize that it can often be very valuable when agents are disposed to act in line with valuable social practices instead of always and consciously trying to do the most good with each act. Now, importantly, this point applies to practices of praise and blame too. As argued in section 3, for praise and blame to better fulfill their function, they should be practice-based. Accordingly, if such a practice is valuable, we often ought to adopt dispositions to praise and blame in line with such a practice. So, agents often ought to be disposed to act as part of the blaming practice rather than trying to maximize the good with each individual act of praising or blaming. So, classical (act) consequentialist arguments for adopting decision-making procedures other than the act-consequentialist criterion of rightness apply to dispositions to praise and blame themselves. Accordingly, act consequentialism and New-Wave Instrumentalism are perfectly compatible.33
The second worry expressed to me was the following: combining the Reactive Attitude Approach with act consequentialism, we get combinations like ‘blameworthy rightdoing,’ ‘blameless wrongdoing,’ and even ‘praiseworthy wrongdoing.’ Some simply object to notions like ‘blameworthy rightdoing,’ because they take it as axiomatic that blameworthiness implies wrongness (call this biw).34
However, taking biw as an axiomatic truth does not make much sense from within consequentialism. As explained earlier, consequentialists should view the norms around when blame is appropriate as internal to effective and valuable blaming practices. But once we think such norms should themselves have a consequentialist justification, we lack any ‘consequentialist’ reason to restrict them axiomatically, that is, for non-empirical and non-consequentialist reasons. Accordingly, there seems little reason to view biw as axiomatic.
Of course, a reader might respond: well, so much the worse for act consequentialism, any plausible theory should view biw as axiomatic! However, there are also good independent reasons to think that biw is not a self-evident axiom. While the details are beyond my current scope, several authors provide good arguments against biw (Capes 2012; Haji 1998, 148–51; Vranas 2007; Zimmerman 1997).
Of course, some readers might still feel uneasy about combinations like ‘blameworthy rightdoing.’ Even though I do not share this feeling, let me give such readers further options. I here operated with act consequentialism, but we can combine the Reactive Attitude Approach with other versions of consequentialism like practice-based, rule, motive, or global consequentialism. One combination I find promising is to adopt a scalar instead of a deontic version of act consequentialism. Instead of giving us a criterion of rightness (a deontic criterion), scalar consequentialism is about which acts are better than others (Norcross 1997, 2008). Scalar consequentialism disposes of ‘right’ and ‘wrong’ and just ranks acts as better and worse instead. Now, once we dispose of deontic categories like ‘right’ and ‘wrong,’ the Reactive Attitude Approach would trivially avoid combinations like ‘blameworthy rightdoing’ and ‘blameless wrongdoing,’ because right and wrong would not feature at the fundamental level of morality.35 Accordingly, instead of blameworthy rightdoing and praiseworthy wrongdoing, we would say that it can be appropriate to blame someone even if they did a comparatively good act and appropriate to praise someone even though they did not do the best available act. So, for act consequentialists bent on avoiding combinations like ‘blameworthy rightdoing,’ scalar consequentialism might be an attractive option.
5 Conclusions
I have argued that consequentialists should take a Reactive Attitude Approach to collective action. Consequentialists should distinguish between the deontic status of acts and whether cooperation or non-cooperation should be praised or blamed. They should endorse practices of praise and blame in our ethical practices around collective action, because doing so leads to better outcomes (primarily by solving collective action problems). Moreover, effective blaming and praising practices will sometimes praise those who cooperate – or blame those who do not – even when an individual act of non-cooperation might do more good. This, I argued, helps us capture some intuitions around collective action.
Recent work on instrumentalism about praise and blame here supplies plausible details. Dispositions to respond with reactive attitudes play an important role in constituting the normative expectations that we have in collective action problems, help us communicate and negotiate our shared normative expectations, and can motivate people to cooperate. They thus play an important role in providing collective solutions to collective problems. Moreover, rather than focusing mostly on one-off behavior, reactive attitudes foreground the individual and collective importance of dispositions in collective action. Finally, including practices of praise and blame in our ethical practices around collective action can drive up the expected value of cooperation, inter alia through positive indirect effects.
Overall, the Reactive Attitude Approach provides a rich and intuitive framework and should add much plausibility to a consequentialist ethics of collective action.
Acknowledgements
I would like to thank two anonymous reviewers for their very helpful written suggestions, Jacob Barrett for insightful written comments and discussions, and my colleagues in Groningen for their many helpful comments at the Grundlegung session and beyond (Daphne Brandenburg and Frank Hindriks in particular).
Bibliographical note
Andreas T. Schmidt is Professor of Moral and Political Philosophy at the Faculty of Philosophy in Groningen. He works in political theory, normative and applied ethics, and the philosophy of public policy. He is particularly interested in longtermist political philosophy, freedom, consequentialism, inequality, behavioural policies, public health, and biomedical ethics.
References
Arneson, R. (2007). The Smart Theory of Moral Responsibility and Desert. In S. Olsaretti (Ed.), Desert and Justice. Clarendon Press.
Arntzenius, F., & McCarthy, D. (1997). Self Torture and Group Beneficence. Erkenntnis, 47(1), 129–144. https://doi.org/10.1023/A:1005376607563.
Bales, R. E. (1971). Act-Utilitarianism: Account of Right-Making Characteristics or Decision-Making Procedure? American Philosophical Quarterly, 8(3), 257–265.
Barnett, Z. (2018). No free lunch: The significance of tiny contributions. Analysis, 78(1), 3–13. https://doi.org/10.1093/analys/anx112.
Barnett, Z. (2020). Why You Should Vote to Change the Outcome. Philosophy & Public Affairs, 48(4), 422–446. https://doi.org/10.1111/papa.12177.
Barrett, J. (2020). Optimism about Responsibility. Philosophers’ Imprint, 20(33), 1–17.
Bicchieri, C. (2017). Norms in the Wild: How to Diagnose, Measure, and Change Social Norms. Oxford University Press.
Brandenburg, D. (2018). The Nurturing Stance: Making Sense of Responsibility without Blame. Pacific Philosophical Quarterly, 99(S1), 5–22. https://doi.org/10.1111/papq.12210.
Brandt, R. B. (1969). A Utilitarian Theory of Excuses. Philosophical Review, 78(3), 337–361.
Brighton, H., & Gigerenzer, G. (2012). How Heuristics Handle Uncertainty. In P. M. Todd, G. Gigerenzer, & A. R. Group (Eds.), Ecological Rationality: Intelligence in the World (pp. 33–60). Oxford University Press.
Broome, J. (2013). Climate Matters: Ethics in a Warming World. W.W. Norton.
Budolfson, M. (2019). The Inefficacy Objection to Consequentialism and the Problem with the Expected Consequences Response. Philosophical Studies, 176(7), 1711–1724. https://doi.org/10.1007/s11098-018-1087-6.
Budolfson, M. (2020a). Collective Action, Climate Change, and the Ethical Significance of Futility [Unpublished]. http://www.budolfson.com/papers.
Budolfson, M. (2020b). The Inefficacy Objection to Deontology: What it is, Why it is Important, and How to Respond to it [Unpublished]. http://www.budolfson.com/papers.
Capes, J. A. (2012). Blameworthiness Without Wrongdoing. Pacific Philosophical Quarterly, 93(3), 417–437.
Fricker, M. (2016). What’s the Point of Blame? A Paradigm Based Explanation. Noûs, 50(1), 165–183. https://doi.org/10.1111/nous.12067.
Fritz, K. G., & Miller, D. (2018). Hypocrisy and the Standing to Blame. Pacific Philosophical Quarterly, 99(1), 118–139. https://doi.org/10.1111/papq.12104.
Haji, I. (1998). Moral Appraisability: Puzzles, Proposals and Perplexities. Oxford University Press.
Hare, R. M. (1981). Moral Thinking: Its Levels, Method, and Point. Oxford University Press.
Hedden, B. (2020). Consequentialism and Collective Action. Ethics, 130(4), 530–554. https://doi.org/10.1086/708535.
Hindriks, F. (2019). The Duty to Join Forces: When Individuals Lack Control. The Monist, 102(2), 204–220. https://doi.org/10.1093/monist/onz006.
Holroyd, J. (2007). A Communicative Conception of Moral Appraisal. Ethical Theory and Moral Practice, 10(3), 267–278. https://doi.org/10.1007/s10677-007-9067-5.
Jamieson, D. (2007). When Utilitarians Should Be Virtue Theorists. Utilitas, 19(2), 160–183. https://doi.org/10.1017/S0953820807002452.
Jefferson, A. (2019). Instrumentalism about Moral Responsibility Revisited. The Philosophical Quarterly, 69(276), 555–573. https://doi.org/10.1093/pq/pqy062.
Kagan, S. (2011). Do I Make a Difference? Philosophy & Public Affairs, 39(2), 105–141. https://doi.org/10.1111/j.1088-4963.2011.01203.x.
Kingston, E., & Sinnott-Armstrong, W. (2018). What’s Wrong with Joyguzzling? Ethical Theory and Moral Practice, 21(1), 169–186. https://doi.org/10.1007/s10677-017-9859-1.
Lawford-Smith, H. (2016). Difference-Making and Individuals’ Climate-Related Obligations. In C. Heyward & D. Roser (Eds.), Climate Justice in a Non-Ideal World. Oxford University Press. http://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780198744047.001.0001/acprof-9780198744047-chapter-4.
Malle, B. F., Guglielmo, S., & Monroe, A. E. (2014). A Theory of Blame. Psychological Inquiry, 25(2), 147–186. https://doi.org/10.1080/1047840X.2014.877340.
Mason, E. (2020). Consequentialism, Blame, and Moral Responsibility. In D. W. Portmore (Ed.), The Oxford Handbook of Consequentialism (pp. 162–179). Oxford University Press.
Matheny, G. (2002). Expected Utility, Contributory Causation, and Vegetarianism. Journal of Applied Philosophy, 19(3), 293–297. https://doi.org/10.1111/1468-5930.00223.
McGeer, V. (2012). Civilizing Blame. In D. J. Coates & N. A. Tognazzini (Eds.), Blame: Its Nature and Norms (pp. 162–188). Oxford University Press.
McGeer, V. (2014). P.F. Strawson’s Consequentialism. In D. Shoemaker & N. A. Tognazzini (Eds.), Oxford Studies in Agency and Responsibility (Vol. 2). Oxford University Press.
McGeer, V. (2015). Building a better theory of responsibility. Philosophical Studies, 172(10), 2635–2649. https://doi.org/10.1007/s11098-015-0478-1.
McGeer, V., & Pettit, P. (2013). The Hard Problem of Responsibility. In D. Shoemaker (Ed.), Oxford Studies in Agency and Responsibility, Vol.1. Oxford University Press.
McKenna, M. (2012). Conversation & Responsibility. Oxford University Press.
McMullen, S., & Halteman, M. C. (2019). Against Inefficacy Objections: The Real Economic Impact of Individual Consumer Choices on Animal Agriculture. Food Ethics, 2(2), 93–110. https://doi.org/10.1007/s41055-018-00030-4.
Nefsky, J. (2011). Consequentialism and the Problem of Collective Harm: A Reply to Kagan. Philosophy & Public Affairs, 39(4), 364–395. https://doi.org/10.1111/j.1088-4963.2012.01209.x.
Nefsky, J. (2019). Collective harm and the inefficacy problem. Philosophy Compass, 14(4), e12587. https://doi.org/10.1111/phc3.12587.
Norcross, A. (1997). Good and Bad Actions. Philosophical Review, 106(1), 1–34.
Norcross, A. (2004). Puppies, Pigs, and People: Eating Meat and Marginal Cases. Philosophical Perspectives, 18(1), 229–245. https://doi.org/10.1111/j.1520-8583.2004.00027.x.
Norcross, A. (2008). The Scalar Approach to Utilitarianism. In H. West (Ed.), The Blackwell Guide to Mill’s Utilitarianism (pp. 217–232). Blackwell Publishing.
Parfit, D. (1984). Reasons and Persons. Oxford University Press.
Pettit, P., & Brennan, G. (1986). Restrictive consequentialism. Australasian Journal of Philosophy, 64(4), 438–455. https://doi.org/10.1080/00048408612342631.
Pinkert, F. (2015). What If I Cannot Make a Difference (and Know It). Ethics, 125(4), 971–998. https://doi.org/10.1086/680909.
Railton, P. (1984). Alienation, Consequentialism, and the Demands of Morality. Philosophy & Public Affairs, 13(2), 134–171.
Rawls, J. (1955). Two Concepts of Rules. The Philosophical Review, 64(1), 3–32. https://doi.org/10.2307/2182230.
Regan, D. (1980). Utilitarianism and Co-operation. Clarendon Press.
Scanlon, T. M. (2008). Moral Dimensions. Harvard University Press.
Schmidt, A. T. (2019). Consequentialism and the Ethics of Blame [Unpublished manuscript]. https://www.academia.edu/5476524/Consequentialism_and_the_ethics_of_blame_Draft_version.
Schmidt, A. T. (2023). Consequentialism and the Role of Practices in Political Philosophy. Res Publica. https://doi.org/10.1007/s11158-023-09631-1.
Sidgwick, H. (1907). The Methods of Ethics, 7th Edition (1982nd ed.). Hackett Publishing Company.
Singer, P. (1980). Utilitarianism and Vegetarianism. Philosophy & Public Affairs, 9(4), 325–337.
Sinhababu, N. (2018). Scalar consequentialism the right way. Philosophical Studies, 175(12), 3131–3144. https://doi.org/10.1007/s11098-017-0998-y.
Sinnott-Armstrong, W. (2005). It’s Not My Fault: Global Warming and Individual Moral Obligations. In W. Sinnott-Armstrong & R. Howarth (Eds.), Perspectives on Climate Change (pp. 221–253). Emerald Group Publishing.
Slote, M., & Pettit, P. (1984). Satisficing Consequentialism. Aristotelian Society Supplementary Volume, 58(1), 139–176. https://doi.org/10.1093/aristoteliansupp/58.1.139.
Smart, J. J. C. (1961). Free-Will, Praise and Blame. Mind, 70(279), 291–306.
Smart, J. J. C. (1973). An outline of a system of utilitarian ethics. In J. J. C. Smart & B. Williams, Utilitarianism: For and Against. Cambridge University Press.
Soon, V. (2021). An Intrapersonal, intertemporal solution to an interpersonal dilemma. Philosophical Studies, 178, 3353–3370. https://doi.org/10.1007/s11098-021-01604-5.
Stout, N. (2020). On the Significance of Praise. American Philosophical Quarterly, 57(3), 215–226. https://doi.org/10.2307/48574434
Strawson, P. F. (1962). Freedom and Resentment. Proceedings of the British Academy, 48, 1–25.
Todd, P. M., & Gigerenzer, G. (2012). What Is Ecological Rationality? In A. R. Group, P. M. Todd, & G. Gigerenzer (Eds.), Ecological Rationality: Intelligence in the World (pp. 3–30). Oxford University Press.
Vargas, M. (2013). Building Better Beings: A Theory of Moral Responsibility. Oxford University Press.
Vargas, M. (2015). Desert, responsibility, and justification: A reply to Doris, McGeer, and Robinson. Philosophical Studies, 172(10), 2659–2678. https://doi.org/10.1007/s11098-015-0480-7.
Vranas, P. (2007). I Ought, Therefore I Can. Philosophical Studies, 136. https://doi.org/10.1007/s11098-007-9071-6.
Wiblin, R. (2020). Why is voting important if you care about social impact? 80,000 Hours. https://80000hours.org/articles/is-voting-important/.
Zimmerman, M. (1997). A Plea for Accuses. American Philosophical Quarterly, 34.
While my discussion is only about consequentialism, it also matters for non-consequentialist views that accord pro tanto weight to consequences when determining an act’s deontic status. Moreover, while instrumentalism about responsibility and blame has a ‘consequentialist flavour,’ non-consequentialists can espouse it too (see, for example, Jefferson (2019)). Finally, non-consequentialists also face separate challenges in collective action problems (Budolfson, 2020b).
My broader definition is meant to include act consequentialists who do not subscribe to a maximizing duty (Slote & Pettit, 1984) or are scalar consequentialists who either reject rightness (Norcross, 2008) or think rightness is scalar (Sinhababu, 2018) (see section 4.4).
Subjective consequentialists define deontic status through expected value (although I use ‘expected value,’ some such arguments also work with decision-theories other than expected value maximization, but I ignore this complication here). But even objective consequentialists can hold that focusing on expected value in one’s decision-procedure can sometimes help bring about what is objectively best.
See Kagan (2011), Matheny (2002), Norcross (2004), and Singer (1980).
Some argue that cases seemingly without thresholds have thresholds too (Arntzenius & McCarthy 1997; Barnett 2018; Hedden 2020; Kagan 2011); for a response, see Nefsky (2011). However, even so, the Expected Value Response still needs to show that individuals are likely enough to hit those thresholds, which means addressing objections like Budolfson’s.
I have only discussed meat consumption, but the Expected Value Response is also discussed for climate (see Broome 2013, 72; Lawford-Smith 2016); see Budolfson (2020a) and Kingston & Sinnott-Armstrong (2018) for objections and for voting, see Barnett (2020) and Wiblin (2020).
See, for example, Budolfson (2020a), Kingston & Sinnott-Armstrong (2018), and Sinnott-Armstrong (2005). Those responses could either accept that one should not cooperate or could supplement difference-making with something else, such as virtue or a duty not to be complicit (see Nefsky (2019) for an overview). Alternatively, some suggested solutions are consequentialist in spirit but reject simple act consequentialism (Jamieson 2007; Pinkert 2015; Regan 1980). I here do not reject act consequentialism and thus do not discuss those alternatives.
But do we ever have relevant intuitions adequately captured by praiseworthiness/blameworthiness or are they not all about deontic status? I think we do. First, people do distinguish between the deontic status of actions and whether such acts are praiseworthy or blameworthy (Malle et al. 2014, 150–151). Our pre-theoretic intuitions about the latter are likely not reducible to the former. Second, many of us likely also have such intuitions in hot-button collective action cases, such as vegetarianism, the climate crisis, and what to do about rising threats to liberal democracy. Third, and more generally, many of us intuitively care not just about the deontic status of our acts but also about being good persons. Finally, I deliberately leave the intuitions to be captured somewhat vague and do not theorize them too much (like ‘there is something commendable about Henk’ rather than ‘we have the intuition that an adequate theory should judge Henk to be praiseworthy). Otherwise, they would stop being ‘pre-theoretic’ intuitions.
This association might be somewhat unfair, as some parts of his writings could be reinterpreted along New-Wave Instrumentalism lines.
See, for example, Arneson (2007), Barrett (2020), Jefferson (2019), McGeer (2014, 2015), McGeer & Pettit (2013), and Vargas (2013). My summary follows my own take on instrumentalism (Schmidt, 2019).
See (Schmidt, 2019) and Barrett (2020), Mason (2020), and Vargas (2015).
See (Schmidt, 2019) and Barrett (2020, 13–6).
See Malle et al. (2014, 150–1) for an empirical and Brandt (1969, 357–8), McGeer (2012, 173–4), and Sidgwick (1907, Chapter iii.ii.) for a philosophical take on this. Watson (1996) also distinguishes between appraising someone’s character and appraising someone’s actions as manifesting certain traits or dispositions. I here primarily focus on the latter.
See my (Schmidt, 2023) for the role of practices and practice-internal normativity in consequentialism.
Although showing that those track intuitively plausible standards would require further arguments far beyond this article. See Brandt (1969) and Vargas (2013) for some such arguments.
New-Wave Instrumentalists also somewhat disagree among themselves about the proper function and value of reactive attitudes. I here present a hotchpotch summary of different functions – which I also find attractive – without unifying them in a grand monistic theory. Moreover, I gloss over many details.
See, for example, Barrett (2020), McGeer (2014), McGeer & Pettit (2013), and Vargas (2013).
I do not assume that relationships are intrinsically valuable (although they might be). Even if only instrumentally valuable, consequentialists agree that many relationships and norm-governed social practices are robustly instrumentally valuable and very important for bringing about good outcomes.
Jefferson (2019), McGeer (2012), and Vargas (2013) are examples of the agency cultivation account.
See Fricker (2016), Holroyd (2007), and McKenna (2012) on communicative and conversational theories of blame.
Of course, ‘bad blaming practices’ can also be ostracizing and exclusive. But good praising and blaming practices can be educational and come with a ‘positive’ message that the addressee of the blame is considered a participant in a moral community.
Consequentialists have good reasons to adopt New-Wave Instrumentalism in general; see my (Schmidt, 2019) for more arguments. I do not repeat the case here, as my case is more specific: consequentialists have good reason to adopt blaming practices along New-Wave Instrumentalist lines to solve collective action problems and capture some relevant intuitions.
See, for example, Bales (1971), Hare (1981), Pettit & Brennan (1986), and Railton (1984).
See Soon (2021) for an exploration of the diachronic dimension of collective action.
More generally, for decisions under uncertainty, rather than risk, heuristics can often outperform more complex expected utility models (Brighton & Gigerenzer 2012).
I do not argue that adopting cooperative dispositions is always superior to adopting more flexible dispositions that allow for domain-specific and case-specific maximization. Second, I am not fully committed to there being enough uncertainty about thresholds for the Expected Value Response to go through. But if there is uncertainty about thresholds and indirect effects – which seems plausible to me but which I leave open – then this can give you (defeasible) reason to adopt cooperative dispositions, given real-world time and information constraints to reduce such uncertainty.
It is not necessary that everyone or even many people accept such norms. For example, if enough people know that vegetarian norms and convictions exist in some subsets of society, the moral meaning of acts can be understood widely.
I claim that, empirically, hypocrisy will render communicating reactive attitudes – and other attempts to elicit behavioral change – less effective. My argument does not assume that we objectively ought to deny hypocrites standing to praise/blame on non-instrumental grounds (for discussions on this stronger claim, see Fritz & Miller (2018) and Stout (2020)).
One might now object that this shows too much: with the expected value of cooperation going up, would the Expected Value Response not be sufficient and the Reactive Attitude Approach superfluous? No, for several reasons. First, in the above first and second argument, actions get the additional expected value only if certain norms and accompanying dispositions to blame are in place. And the third effect only holds instrumentally through the value of expressing reactive attitudes. Second, the above pro tanto arguments for cooperative acts might still not be enough to make them right. For many people, the prudential value of a nice holiday might still outweigh the expected downsides of non-cooperation. And, as mentioned before, the indirect effects argument can itself encounter a No Difference Challenge. Third, the indirect effects argument fails to capture our intuitions around secret acts of (non-)cooperation that have no indirect effects. The Reactive Attitude Approach, in contrast, can still judge agents praiseworthy (blameworthy) for secret acts of (non-)cooperation. Finally, and importantly, practices of praise and blame involve separate acts and norms and thus have effects different from (non-)cooperative acts. So, the consequentialist value of the former cannot be reduced to the value of the latter. My praising others for not eating meat are acts separate from my eating plant-based meals. So, even if the Expected Value Response works, there is a separate consequentialist case to include blaming practices in our ethical practices around collective action.
Thanks to two reviewers who put the below cases to me.
Of course, often it will be inappropriate to blame someone who did not cooperate, because they do not meet general responsibility conditions (a young child eating sausages, for example). I do not discuss conditions for responsibility here, but instrumentalists have written much about the subject (see footnote 15).
Also, when sustainability norms are not yet entrenched, it can be effective to first prioritize praising those who make changes, rather than blaming people like Judy, and prioritize changes with higher (average) impact or lower moral opportunity costs (cutting meat, for example).
See (Schmidt, 2019) for a more in-depth argument. Also see McGeer (2014).
They might also think it is axiomatic that praiseworthiness implies rightness. I here focus on biw, as it is more widely held and discussed.
Norcross also argues that even if we reject deontic criteria as part of the fundamental moral level, they can still figure in our ordinary moral discourse. As such, they have a contextual semantic. We could then also imagine ‘wrongness’ being linked to blameworthiness contextually. Alternatively, Sinhababu (2018) defends a version of scalar consequentialism that turns rightness into a scalar property. Both options are compatible with my account.