In 'Offsetting and Risk Imposition,' we argued that when carbon emissions are accompanied by offsetting, this can result in a set of actions that worsens no one's prospect, and therefore imposes no risk. We argued that this is true for some forms of offsetting but not others. Here, we reply to the criticisms of that argument that Scott Hill makes in 'No Special Morality for Carbon Emitting,' and examine more closely the relationship between risk-imposition and the worsening of prospects.

States and other climate actors now commonly set ‘net zero’ targets – pledging that, by a certain date, they will put no more carbon into the atmosphere than they take out. However, there is controversy over what exactly should count as attaining such targets. The method of emissions accounting that states currently use – territorial emissions accounting – is often criticized as problematic, but a fully satisfactory explanation of the problem is needed. We argue that the key both to understanding the problem and to solving it is that there is a specific kind of cooperative contribution – a mereological contribution – that a national net zero commitment needs to make to global climate action in order to meet the standards that global citizens should hold states to. On this basis, we propose an alternative method – Contributory Value-Chain Accounting – showing how it can be qualified to reflect trade relationships between net-zero-aligned countries, and defending it against the objections that it involves double counting, is infeasible and too informationally demanding, and that it would send the wrong incentive signals.

Abstract Rossian pluralism can be defined as the conjunction of two claims: there is more than one derivationally foundational right-maker, and there is no further principle governing the contributions they make to determining the overall rightness of an act. So understood, Ross’s own moral theory is only one possible version of Rossian pluralism. Six distinctive features of his version are its self-containment, its subsumptive picture of how the surface content of morality derives from its foundations, its hybrid view of the relation between the right and the good, the kinds of consideration it treats as foundational, its balancing view of how competing moral considerations interact, and its treatment of those considerations as being captured by exceptionless contributory moral principles. Examining these features, I point out some of the problems they invite, and explore some of the further possibilities that are open to a Rossian pluralist in dealing with those problems.

This Perspective describes the various dimensions of Carbon Dioxide Removal (CDR) durability and interprets them in the context of current policy making. Durability - together with scalability and sustainability - is an essential condition of CDR. It depends on (i) the duration of CO2 storage and (ii) the risk of reversing such storage. The risk profile of durability varies widely across CDR methods. Because engineered, novel CDR methods involve more stable forms of CO2 storage than nature-based CDR, these methods are often promoted as a priority for CDR mitigation investments. However, shorter-term CDR plays an essential role in balancing sources and sinks of greenhouse gases in the second half of this century. Decision makers must also consider CDR policies in a larger context that takes into account readiness and feasibility, policy alignment and co-benefits of different CDR methods. They must also address durability in CDR policies and contracts, which tend to span much shorter timeframes than those contemplated by science when discussing durability. We argue that nature-based conventional CDR and novel engineered CDR that show complementary timing and risk profiles can be deployed in synergistic CDR portfoliosto balance the conditions of durability, feasibility and social and environmental sustainability.

In general, otherwise permissible actions do not become wrong when agents act on bad attitudes. But cases of discrimination can be exceptions to this generalization. It could "be morally permissible for someone to rent her house to any one of several prospective tenants but not morally permissible to refuse to rent it to one of those people because of his race" (Scanlon 2008: 71). These two claims are plausible and widely accepted, but they call for explanation. Why is it that in some cases of discrimination, agents' bad attitudes make otherwise permissible actions wrong? Why doesn't this imply that otherwise permissible actions generally become impermissible when agents act on bad attitudes?? Together, these questions produce a puzzle which existing accounts of wrongful discrimination have not solved. We argue that the solution comes from appreciating how the manifestation of bad attitudes in action is a central constituent in the unjust social condition to which discrimination subjects members of subordinated groups: the deprivation of positive moral standing.

According to Bernard Williams, the ethical questions we face, and the resources we have for answering them, are historically contingent and local. There is no ‘Archimedean point’ from which to seek universal answers to ethical questions about the living of a human life. This poses a challenge that is both philosophical and practical: how can a reflective awareness of the cultural contingency of ethical thought be reconciled with a commitment to the seriousness of ethical questions, and the objectivity that this seriousness requires us to attribute to them? This chapter raises some questions about the success of Williams’s own response to that challenge, and makes the case for a different but related response, drawing on themes from one of Williams’s intellectual mentors, Isaiah Berlin.

There is a difference between acting with a probability of making a difference to who is harmed, and worsening someone's prospect. This difference is relevant to debates about the ethics of offsetting, since it means that showing that emitting-and-offsetting has the first feature is not a way of showing that it has the second feature. In an earlier paper, we illustrate this difference with an example of a lottery in which you shake the bag from which a ball will be drawn to determine the identity of the person who will suffer a harm. Here, we reply to Stefansson and Willners' criticism of that example, explaining how their reply makes a mistake about the point of the example.

A familiar part of ordinary moral thought is this idea: when other people are doing something worthwhile together, there is a reason for you to join in on the same terms as them. Morality does not tell you that you must always do this; but it exerts some pressure on you to join in. Suppose we take this idea seriously: just how should it be developed and applied? More particularly, just which groups and which actions are the ones with respect to which you have participatory moral reasons? And just when is it wrong not to join in? I present answers to these questions, and discuss the implications—including for joint action on a very large scale, like global efforts to address climate change or fight a pandemic.

Glen Pettigrove's 'What Virtue Adds to Value' maintains that sometimes virtue is fundamental in the order of value, and that we should reject the general thesis that the value of our responses depends on their proportionality to the value of the objects toward which they are directed. He argues that this view is needed to account for the moral phenomena surrounding love, forgiveness and ambition. I object that his view is unable to explain the forms of discrimination that distinguish the good kinds of love, forgiveness and ambition from the bad ones; and that explaining this requires a version of the proportionality thesis he opposes.

AbstractAccording to the ‘fitting response’ tradition of thinking about value, good things are fit objects of favour-responses and bad things of disfavour-responses. What is the relationship between the relation of fit, so understood, and the normative reason relation? A widespread view treats normative reasons as primitive, and explains fit responses as those for which there is sufficient reason of the ‘right kind’. This chapter adds to recent efforts to develop a fitness-priority view with the opposite explanatory direction. Its proposal, in outline, is that a normative reason for an action, thought, or feeling is a consideration that is fit to determine what one will do, think, or feel. When this proposal is carefully developed, it provides appealing explanations of the different sources of our normative reasons, of how reasons have a directiveness that value-facts need not have, and of what our responsiveness to reasons consists in.

Suppose one of your actions imposes a risk of harm that, on its own, would be excessive, but a second reduces the risk of harm by a corresponding amount. By pairing the two actions together to form a set of actions that is risk neutral, can you thereby make your overall course of conduct permissible? We argue that the answer to this is sometimes yes and sometimes no. We propose a criterion, the Principle of Aggregate Risk Imposition, that distinguishes the two kinds of cases and apply it to the offsetting of personal greenhouse gas emissions.

Three things often recognized as central to morality are concern for others’ welfare, respect for their self-expression, and cooperation in worthwhile collective activity. When philosophers have proposed theories of the substance of morality, they have typically looked to one of these three sources to provide a single, fundamental principle of morality—or they have tried to formulate a master-principle for morality that combines these three ideas in some way. This book views them instead as three independently important foundations of morality. It sets out a plural-foundation moral theory with affinities to that of W. D. Ross. There are major differences: the account of the foundations of morality differs from Ross’s, and there is a more elaborate explanation of how the rest of morality derives from them. However, the overall aim is similar. This is to illuminate the structure of morality by showing how its complex content is generated from a relatively simple set of underlying elements—the complexity results from the various ways in which one part of morality can derive from another, and the various ways in which the derived parts of morality can interact. Plural-foundation moral theories are sometimes criticized for having nothing helpful to say about cases in which their fundamental norms conflict. Responding to this, the book concludes with three detailed applications of the theory: to the questions surrounding paternalism, the use of others as means, and our moral responsibilities as consumers.

We have previously argued that there are forms of greenhouse gas offsetting for which, when one emits and offsets, one imposes no risk. Orri Stefansson objects that our argument fails to distinguish properly between the people who stand to be harmed by one’s emissions and the people who stand to be benefited by one’s offsetting. We reply by emphasizing the difference between acting with a probability of making a difference to the distribution of harm and acting in a way that worsen’s someone’s prospect.

Since it tackles many topics, Concern, Respect, and Cooperation invites questions and challenges from many directions. Rather than attempting to cover all of the points raised in the commentaries, I will concentrate on what I see as the most prominent ones, grouping them thematically under five headings. The book sets itself three main desiderata for success in a substantive moral theory: truth, illumination, and usefulness (1-3). I talk of “illumination” rather than just explanatory power, in order to emphasize that when it comes to theorizing about morality, what we are looking for is a description of its structure that casts light on why it makes sense for us to care about it as much as, and in the way that, we do (2). However, explanatory power is essential for illumination, and this is increased when a theory is relatively parsimonious with respect to what it explains (168). Bykvist and Tenenbaum both press me on this. There are two main ways in which a moral theory could be more parsimonious than the one I present. It could be less pluralistic in its foundations, or it could restrict more tightly the types of derivation-relation by reference to which it seeks to explain how the surface content of morality derives from its foundations. Tenenbaum presses the first of these points; Bykvist the second. Is the theory unnecessarily elaborate? Speaking very broadly, my reply to Bykvist and Tenenbaum is the same as Moore's reply to the classical utilitarians, and Ross's (in turn) to Moore: “it is more important that our theory fit the facts than that it be simple.”1 The three desiderata are ranked: truth comes first.2 However, that way of putting the point is a bit flat-footed. The issue is not just that our theory must get the extension of morality right, by making the correct overall classifications of morally right actions, morally good dispositions, and so forth (although of course it must do that). The more crucial test of the adequacy of a moral theory is whether it convincingly explains (echoing Ross again) what makes right acts right—what I call the “determinative justification” of their rightness (13). Thus, I do not doubt that, for any credible set of overall moral verdicts, you could devise a theory that preserves those verdicts while being more parsimonious than the one I propose, either by having fewer foundations or by working with fewer principles of derivation. (I don't think you could do both: then you would get the extension of morality wrong, as classical utilitarianism does.) However, I think you would then face a choice between two unsatisfactory alternatives. Actions of type A are morally wrong Dispositions of type D are morally good Ross's appreciation of this problem, as I see it, is one of the main influences that draws him towards a theory that has at its foundations not verdictive principles but prima facie duties. His kind of theory drops the ambition of producing deductive derivations of overall moral verdicts from foundational principles, but looks to identify the most basic kinds of pro tanto moral reasons, seeking to analyse all the moral reasons there are into combinations of these basic types. But although this is a significant step forward, we need to go further. Ross's principles of prima facie duty invite counterexamples too (39). It is not true that there is always a pro tanto reason for gratitude, understood as ‘the returning of services, irrespective of motive.’3 And it is not true that there is always a pro tanto reason for beneficence, understood to include increasing others’ pleasure.4 This is the holist's objection to Ross. It is because I endorse this objection that I think we need to take the process he pioneered a step further and formulate the foundations of morality as exception-hedged principles. (And, for reasons we will come to in §2, I formulate them in terms of the fitness, not the normative reason, relation.) The second alternative (Tenenbaum's) is instead to simplify the theory's foundations, discarding my pluralism in favour of a monistic, single-foundation view. The most familiar form that this takes is a welfare-foundational theory on which there is ultimately one fundamental moral reason: the reason we have to promote others’ welfare (19-20). Tenenbaum is attracted to a respect-foundational alternative. I actually doubt whether that is the best way to develop a Kantian moral theory—treating respect for others’ self-expression as a single foundational norm from which all of the content of morality derives (16-17, 20). However, my objection to proposals of either kind is similar (19).5 Monistic theories of these kinds can be configured in a way that yields credible verdicts about right and wrong; but it is hard to accept their account of why some actions are wrong. A welfarist view, by making autonomy part of welfare, can say that paternalism is wrong; but in telling me to respect your autonomy because it is good for you, it fails to provide me with the right reason for non-interference. A respect-foundational view can find ways of deriving the wrongness of cruelty to non-human animals, infants and the severely mentally impaired from considerations of respect for rational agency, but that is too roundabout to be convincing: it misses the direct moral importance of their welfare. I do not claim that objections of these simple kinds provide quick and fully conclusive refutations of the monistic alternatives. The comparison of rival moral theories is a matter of assessing their overall relative plausibility; and that assessment goes beyond what I attempt in the book. However, my general concern is that a foundationally more parsimonious theory will fail to reflect the direct moral importance of the three different forms of unselfishness that I treat as fundamental, and thereby misrepresent what ultimately makes right actions right. Given how complex the content of morality is, and how resistant it is to completely exceptionless, non-trivial generalizations, I actually think the theory is likelier to be too simple than too elaborate. Tenenbaum also raises worries of this kind (and of course he is right that it is possible for a theory to be too elaborate in some respects and too simple in others). He discusses two main examples: the treatment of harms and benefits, which I come to in §3 below; and the treatment of friendship. He asks how my view could accommodate Stroud's treatment of friendship as requiring epistemic partiality. To this, my reply is that a plausible view about welfare sees goods of fellowship as one of its main constituents, with friendship being one of those goods, and if Stroud is right then epistemic partiality is a disposition you need to have in order both to possess that good yourself and to confer it on your friends—friendship being a good which, when created, is both given and received by both parties. So, while no doubt there are moral phenomena that my theory is under-equipped to explain, I don't think this is one of them. The theory of the constitution of welfare is a supplementary project, independent of my own attempt to spell out the concern-responses that morality enjoins towards it (45). I think the answer to Tenenbaum's question about friendship lies within that supplementary theory. Bykvist and Tenenbaum also both raise the question of the relationship between my view and particularism. This is not something I discuss explicitly in the book, preferring to avoid the definitional discussions that would require: the scope for misunderstanding here is demonstrated by the fact that Bykvist sees me as embracing a kind of particularism and Tenenbaum as opposing it. What they are both pointing to is the way in which the theory does not seek to ground morality in exception-free principles—its foundational principles are exception hedged—but it seeks to give regular explanations of the exceptions. It rejects Ross's “atomism” in favour of a holism about reasons (and also fitness) according to which the status of one fact as reason-providing can depend on the other surrounding facts. But despite this, it departs from the spirit of a view like Dancy's particularism in an essential respect. For Dancy, the relation of determination (that is, the relation Ross calls “making” when he asks “What makes right acts right?”) runs from the particular to the general: if there are any general moral truths, their truth is explained by the truth of the instances that constitute them.6 My project runs in the other, more orthodox direction, from general to particular. The foundational principles I work from admit of particular exceptions; but the exceptions are explained by the way in which applications of these exception-hedged principles interact with each other (110). This leads Bykvist and Tenenbaum to a shared worry, which they express differently. This is that the foundations and superstructure of my project are disconnected from each other. I start with a simple picture of morality as constituted by three basic forms of unselfishness; but on top of this I want to provide elaborate, contextually nuanced explanations of the deontic and aretaic content of morality which are properly responsive to all of the factors that can bear on them. The worry is that whatever plausibility the latter explanations may have is achieved by making them float free of the claimed foundations. Bykvist expresses this as a concern that the combination of system-building (in aspiring to derivations from simple foundations) and particularism (in aspiring to contextually nuanced explanations that eschew exceptionless principles) is unstable. Tenenbaum expresses a worry of the same kind as the concern that the foundations of the theory are arbitrary (since the contextual explanations could just as readily have been overlaid on a different set of foundations). I take this up in more detail in §3 (Foundations) and §4 (Derivations) below. However, my initial response is this. I think this is an excellent worry to raise—but that if a moral theory is to attempt all of the work we need it to do, it should raise this worry. The aim of illumination encourages the search for compression and simplicity in our account of the foundations of morality; but the aspiration to accuracy and truth points in the direction of complexity in an account of its full content. These aims are in tension, and attempting to meet them both is a big part of what makes moral theorizing so hard (22-3). I think there is less disagreement between myself and Arpaly than meets the eye. I use “fitness” (not “fittingness”) as a name for the relation between the praiseworthy and praise, the admirable and admiration, the desirable and desire (36). Unfortunately, ordinary English does not supply a name for that relation, so theorists need to make one up if we want to talk about it. One possibility would be to call it “-worthiness”; a popular choice amongst other theorists is “fittingness”; Arpaly calls it “warrant.” I think all of these terminological choices are to some degree unsatisfactory, since these terms all have uses in ordinary English that do not refer to this relation. Arpaly points out that, in ordinary speech, if someone says that praise is fitting, they might either be asserting that it is directed towards something praiseworthy, or that it is desirable. But the same is true of the remark that praise is warranted. The OED, illustrating its definition of “warrant” = “justify,” offers this from a medical writer: “I said that I did not think a surgeon warranted in tying the external iliac artery.” Here, what is being assessed is the instrumental justification for an action: is it likely to save the patient? So on this usage, you will be warranted (i.e., justified) in praising a torturer when it is likely to save his victims. I use “fit” and “called for” since (to my ear) they attract a more “object-given” reading. When there are merely instrumental reasons for praising the torturer, it sounds strange to say that the torturer is fit to be praised, or calls for praise—stranger than it does to say that praising him is fitting or warranted. However, no doubt my terminological choice is awkward in other ways. The point of substance is that I am treating “Karen's A-ing is praiseworthy iff praise is a fit response to Karen's A-ing” as true by stipulation—not as an attempted analysis of a concept that is expressed in ordinary use of the word “fitness.” This might provoke a suspicion about whether the claims I make about the structure of morality could possibly be right. How could the relation I call “fitness” be so fundamental to the structure of morality if we don't even have a name for it in ordinary moral discourse? However, I do not see that as a problem. This relation is one that is referred to all the time in ordinary moral discourse: we are talking about it every time we use evaluative suffixes like “-able” (desirable), “-some” (awesome); “-ful” (shameful) and “-worthy” (praiseworthy). We only need a noun to refer to this relation once we turn from moral practice itself to the activity of theorizing about it. Ordinary moral practice is not in the business of theorizing about itself, so ordinary English has no need to furnish us with this noun: we theorists therefore need to supply our own. Tenenbaum raises a worry that follows on from this. Charitably, he accepts that fitness can be introduced as a theoretical notion, with its value to be judged by its theoretical contribution—but he then worries about the explanatory credentials of the resulting view. The theoretical concept of fitness now needs to be interpreted in terms of the moral phenomena the theory relates it to: so the direction of explanation runs from those phenomena to fitness, rather than the other way around. But to this, I reply that fitness is not a theoretical notion. My use of the word “fitness,” it is true, is a theoretical use—but the notion I use it to express is not a theoretical notion. It is the ordinary, common and garden, evaluative notion that we deploy when we say that something is desirable or praiseworthy or awesome. The key theoretical claim I want to make about this relation is that it is that all of our normative reasons derive from it, either directly or indirectly. Reasons for praising Mandela derive directly from his praiseworthiness; reasons for praising the torturer derive indirectly from the prevention-worthiness of torture. So fitness-relationships are more fundamental than normative reasons (36). In saying this, I make two contentious assumptions. One is that there is a single fitness relation: the relation between the desirable and desire, the praiseworthy and praise, the loathsome and loathing, is the same relation. I operate with this assumption by default: if there is more than one fitness-relation, my theory will need to be revised to reflect this. The other contentious assumption is that this relation is one which relates objects not only to attitudinal responses such as admiration and loathing, but also to actions—and not just actions such as praise and choice, for which we have fitness-involving terms such as “praiseworthy” and “choiceworthy,” but actions such as promotion and protection, for which we do not. In the fitting-attitudes literature, one common claim about the fittingness relation is that it is the relation that an attitude stands in to its object when it correctly represents the object as possessing the corresponding evaluative property: for your admiration to be fitting is for it correctly to represent its object as admirable. If that account is correct, then actions cannot stand in such relations, since actions do not represent objects as having evaluative properties. This might make you reluctant to accept that the relation I refer to as “fitness” really does play the important role within the structure of morality that I attribute to it. However, there are ways of using the language of reasons to talk about the relation I call “fitness” while making substantially the same claims (37-8). For example, you could follow Parfit in saying that we have “object-given” reasons for praising the praiseworthy and “state-given” reasons for praising people like the torturer whom it is merely useful to praise.7 Speaking this way, my view could be reformulated by saying that the most fundamental kind of reason is the object-given kind: all other reasons derive from those. I don't follow that Parfitian line because I think it faces two objections (35). One is that normative reasons are capacity-dependent whereas fitness-relationships are not. If my mental powers become so impaired that I lose the capacity to praise Mandela, then I cease to have a reason to praise him; but he doesn't cease to be praiseworthy. The other is that the reasons a person has are subject to a condition of personal relevance that does not apply to fitness-relationships. The deftness and sensitivity that Julia displays in her job as a corporate lawyer in Hong Kong may make her praiseworthy, but that is not enough to give me a reason to praise her, in the absence of any relevant connection between her and me. I do not venture an account of just what is needed in order to satisfy this “personal relevance” condition, but examples like this one suggests that we must make room for this condition: it is another way in which an action's praiseworthiness—the relationship of fitness between it and praise—can fall short of generating a reason for someone else to praise it. A response that some will be happy to make to this is to distinguish between whether there is a reason and whether some particular person has that reason. If all of us lost the capacity to praise Mandela, then no one would have a reason to praise him-, but his fine qualities would still be reasons to praise him, nonetheless. I avoid this way of talking, since using “reasons” for these two different kinds of relation seems to me to invite confusion (38). But anyone who prefers that terminology could use it to describe the structure of my view without any change of substance: my talk of “fitness” can be replaced with talk of the object-given reasons there are. Phrased in those terms, my claim is: others’ welfare is something there is an object-given reason to promote, unless it is misdirected in a way that subjects it to content-undermining. To get from this to your having a reason to promote someone's welfare—say, in the cases Tenenbaum imagines—the conditions of capacity and personal relevance need to be met, and there must be no context-undermining; and to get beyond this to an overall conclusion about whether you ought to promote it, all of the other reasons bearing on your action must be taken into account. Tenenbaum asks whether the foundational structure of my theory is optional. However, as I use the term “foundations,” all actual substantive moral theories have them (44). The foundations of morality are those parts of the content of morality that do not derive from other parts of its content: stopping points for the derivation-relation. So the only way to avoid attributing foundations to morality would be to advocate the “interminist” view that, for every part of morality's content, the question “Why is this part of the content of morality?” can be answered by citing some other, deeper part of its content, without any stopping point. However, no actual moral theory is interminist. And even if you reject the enterprise of substantive moral theory altogether, you will still treat some parts of the content of morality as not deriving their importance from others. To say that the substantive content of morality has foundations does not commit you to a foundationalist rather that a coherentist moral epistemology (I am not a Rossian about that—23-6). And to say that a part of the content of morality is foundational is not to say that the question “Why is this part of the content of morality?” has no answer: it is to say that it does not have an answer that derives it from some other part of morality (162). This still, of course, leaves open that the particular foundational structure I impute to morality is optional. Above, I explained why I resist the idea of simplifying its foundations by seeking to derive concern for welfare from respect for self-expression or vice versa. Theories of either of those monistic kinds get their credibility from their claim to have identified something of underived moral importance. But the credibility of both claims to fundamentality speaks against either form of monism. More unusual is the further claim I make: that morality's foundations include non-derivative norms of cooperation too. This is the subject of Steele's commentary. As she rightly says, a compelling case for treating this as morally foundational needs to be supported by convincing examples in which reasons to cooperate are not derivable from considerations of concern and respect. She discusses the case of Many Hikers, and raises some good questions about that. However, Many Hikers is not the sort of case I would want to cite as a compelling example of reasons of cooperation. I agree with her that—once we specify that your joining in will neither ease the burden on the other co-operators nor improve the assistance to the trapped person—piling on is pointless. You should surely express a willingness to help and stand by to do so if needed; but you needn't pointlessly push the rock just so you can say you are joining in. Instead, I think the most compelling examples of reasons of cooperation that are not derivable from the other two moral sources come from cases of free riding (55). When others are cooperating to produce a good, and I parasitically help myself to it without contributing to the cooperative effort, this can be unfair (223). That can be true when the good in question is non-rival, so that my consumption of it does not diminish the benefits available to anyone else (I sneak into the theatre and enjoy the show without paying, but without disturbing anyone else). Here, it is unconvincing to explain the reason against behaving in this way by deriving it from my reason to promote and protect others’ welfare, or my reason not to interfere with others’ self-expression. Instead, a fundamental form of moral decency is a willingness to join in on the same terms as other similarly placed people to achieve something worthwhile together (52-3). However, as I see it, the same basic point applies much more broadly than just to cases of free riding. As Steele points out, my talk of “worthwhile collective action” casts a wide net: it includes all those collective actions (themselves broadly conceived—53) that there are sufficient reasons for us, collectively, to perform (54). So it covers cases in which the reasons that make our collective action worthwhile are reasons of collective prudence; but it also covers cases in which our collective action is one that there are sufficient moral reasons for us to perform—for example, reasons of concern or respect (55). It does not require that the collective action be optimal: only that it is worth doing. The claim is that when other similarly situated people are cooperating to achieve something worthwhile, I can have a reason (provided I meet the capacity and relevance conditions) to join in. And it is a claim about joining in existing collective actions: I do not maintain that there are reasons of cooperation for me to do what would make sense as a contribution to something we ought to be doing, when we are not (223). A collective action's being worthwhile, then, depends on whether there are further reasons that make it worthwhile, and those reasons can be moral ones. If they were always moral ones, then (as Tenenbaum points out) there would be a responsive derivation of norms of cooperation from more fundamental norms of concern and respect. However, the reasons that make a collective action worthwhile need not have a moral source (79). The question “Why are there moral reasons against free riding?” is not answered by deriving these reasons from reasons of concern or respect. So (C+) is needed as an independent foundation for morality. Steele's main focus is on what it is for individuals to be “similarly placed” with respect to participation in a collective action. I approach this by asking: first, what is the group cooperating to do; next, what are the reasons making its action worthwhile; and then, what is the property possessed by members of the group in virtue of which they constitute a group whose collective action is supported by those reasons? (55) When a group constituted by individuals of kind K has sufficient reasons to be performing a collective action, and I belong to kind K, the action calls for my participation. In the theatre example, where other viewers of the performance are cooperating to make it happen by paying for it, K is being a viewer of the performance, and my having this property gives me a reason of cooperation to join in by paying. If the action is demonstrating against our country's involvement in an unjust war, then K is being a citizen with the capacity to participate without serious personal cost, and my having this property gives me a reason of cooperation to join in by marching. I agree with Steele that this would be implausible if it implied that I have a reason to pile on in a case like Many Hikers. Nonetheless, I do also want to recognize reasons of cooperation to join in large-scale collective action despite the fact that your own individual action will not make a significant difference to what is achieved by the group. When a large group is achieving something important, kind K can include being someone whose individual contribution will make a negligible difference to what is achieved or expressed by the collective action (222). This applies to joining the demonstration. When the demonstration is worthwhile (and a protest against injustice need not be likely to prevent it to be worthwhile), it can be achieving what it does because others whose individual contribution makes a negligible difference are making the effort to participate. So the question still arises: am I willing to participate in this worthwhile collective action on the same terms as everyone else? How, then, does Many Hikers differ from joining the demonstration? We can approach this by considering a more extreme case, which I discuss in more detail elsewhere.8 One form that actions of group beneficence can take is administering a limited capacity pool for beneficence. Jonathan Glover's example (taken up by Parfit) of a thousand people each adding a pint to a tank that will be taken to succour a thousand severely thirsty men is of this type.9 After a thousand pints have been added, the tank is full. The collective action is worthwhile; but surely there is nothing to be said for my wasting a further pint of water by pouring it into the already-filled tank and causing it to overflow. It is tempting to explain why there is no reason to do this by saying that my action of adding a pint would make no one any better off; but I think that is the wrong explanation. (That is equally true of my action of joining the demonstration.) Instead, the important point is this: in the Drops of Water case, there is no reason for us collectively to add any water to the tank once it is full. The worthwhile collective action is filling the tank to help the thirsty men; making the tank overflow is not something it is worthwhile for us to do. So the feature possessed by members of the group performing a worthwhile action—the relevant kind K—is being able to contribute to filling the tank. If the tank is not full yet, then morality gives those with the ability to contribute a reason to do so; once it is full, that is no longer true. However, in order for morality to give a potential contributor a reason to join in, it is not necessary that she herself will make a significant difference to the beneficiaries of the collective action: if the group is very large, none of the contributors does that. (So that—each person's not making an individual difference—is also a relevant similarity between the contributors.) A similar set of points applies to Steele's Many Hikers case. The group's current action is worthwhile, but there is no reason for it (the group of helpers) to increase its size by adding further members: its labour capacity is now full. If it added further members, it would be wasting their labour, in the same way that making the tank overflow would be wasting water: that collective action would not be worthwhile. The respect in which contributors to the worthwhile collective action are relevantly similar is their having the ability to contribute to the labour force needed to free the trapped person. After that labour force is full—full in the sense the group has no reason to increase it—someone who piles on is not contributing to a worthwhile collective action and therefore lacks a reason of cooperation for doing so. The case of joining the demonstration is different. This is not a limited-capacity collective action: as the size of the group increases, so does the strength of its protest. Here, there is no point at which the protest is full and beyond which the group has no reason to increase its size further. So (assuming the aim of the protest is itself worthwhile), every additional person who joins in is contributing to a worthwhile collective action. Given this treatment of Steele's challenge, one might wonder about its mirror-image: not piling on once a limited-capacity pool is full, but instead hanging back while others fill it. If I am one of the Many Hikers who comes across the trapped person, and I hang back until enough people have stepped forward to act, then this could be morally criticizable too. Here, we should say this: while I was hanging back, I ha

What constraints should be imposed on individual liberty for the sake of protecting our collective security? A helpful approach to answering this question is offered by a theory that grounds political obligation and authority in a moral requirement of fair contribution to mutually beneficial cooperative schemes. This approach encourages us to split the opening question into two—a question of correctness and a question of legitimacy—and generates a detailed set of answers to both subsidiary questions, with a nuanced and plausible set of implications. The plausibility of its treatment of the issues surrounding liberty and security, I argue, helps to confer credibility on the fairness-based theory that carries these implications.

For many of the moral beliefs we hold, we know that other people hold moral beliefs that contradict them. If you think that moral beliefs can be correct or incorrect, what difference should your awareness of others' disagreement make to your conviction that you, and not those who think otherwise, have the correct belief? Are there circumstances in which an awareness of others' disagreement should lead you to suspend a moral belief? If so, what are they, and why? This paper argues that three principles, taken together, give us a good answer to these questions; that they license a form of provisional moral self-trust; and that they reveal an interestingly distinctive form of pragmatic encroachment in relation to the epistemic standards governing moral belief.