To set the stage for our discussion, we start with this assumption: unless a society has completely eradicated the COVID-19 virus in its borders, any decision to re-open that society from the stay-at-home orders that were common, at least across the U.S., in April, will accelerate the spread of the virus. This will likely have deadly effects. A protracted shut down may also take a large toll on human life, and the justification for reopening could be framed in terms of maximum lives saved. But in the near-term, the justification for opening up must appeal to the importance of avoiding costs other than loss of life, such as unemployment, poverty and the psychological stress of isolation. Can opening up in the near term be justified in that way? If so, how?

*              *              *

ALEC:  I want to start by thanking you two, Sophia and Christian, for agreeing to discuss the ethics of reopening society with me.

I take the position that it is, in principle, justifiable to reopen society for the sake of people’s economic and psychological welfare. This follows from my views on risk and aggregation.

There are two ways to think about the moral significance of risk: the ex ante model discounts claims with regard to an outcome (a harm or benefit) by the probability that the claimant will experience that outcome; the ex post model weighs the claims of those who experience some outcome in the end, whoever they turn out to be, discounting only insofar as it is uncertain that anyone will experience a particular outcome. Both models can be combined with different views on the permissibility of aggregating claims.

Many contemporary ethicists embrace the ex ante model. Because people can sort into groups with different COVID-19 risk profiles, people have conflicting claims on public policy. Low-risk individuals often reasonably prefer risking infection for the sake of being able to carry on a normal life, while those at high risk often reasonably prefer that society remain on lockdown.

Many who embrace the ex ante model also oppose aggregation. They think that the only morally relevant question is who has the strongest complaint if their claims are not honored. The complaints of those who prefer not to die—at least if they are not already very old or very sick—would seem to dominate.

In an article of mine, “Risks and Weak Aggregation: Why Different Models of Risk Suit Different Types of Cases” (forthcoming in Ethics), I argue that when people can sort themselves into different groups with competing claims on public policy, the ex post model should be used. I also argue for a variation on what is often called “limited aggregation.”  

According to limited aggregation, if interests are far enough apart in moral significance, no number of less significant interests can outweigh more significant ones. For example, no number of people’s desires not to suffer an annoying itch for one minute could justify leaving another person to die years earlier than she would otherwise.

In the end, I think we can plausibly justify opening up, if sufficient precautions are taken, because people’s claims based on economic and psychological welfare are not like claims not to suffer an annoying itch for a minute. Taken in the aggregate, they plausibly outweigh the claims not to be allowed to die early. [Post script: I now realize that this formulation, framing the issue as though opening up merely allows people to die, is flawed—see Christian’s comment below and a recent article by Helen Frowe.]


SOPHIA:  Thanks for sharing your thoughts, Alec.  I’m going to raise a few questions about your characterization of our dilemma, and then some questions about the methodology you’re using to address it.

First, I don’t think our dilemma maps neatly onto the scenarios that philosophers standardly discuss, when thinking about whether to aggregate and how to measure people’s complaints.  Our dilemma in deciding when and how to open up our societies is not how to weigh the smaller complaints of the many under lockdown against the larger complaints of those who will die when we open up.  Many lives are at risk under lockdowns, too.  The UN has reported a surge in serious abuse of women and children; women are being denied access to life-saving reproductive healthcare; people with life-threatening medical conditions have had their treatment put on hold.  There are many with large complaints about all alternatives.

Second, your methodology.  COVID-19 is not just a combination of a health crisis and an economic crisis.  Our responses to it risk perpetuating serious social injustices, such as racial injustices, gender-based injustices, and injustices against those with disabilities.  And I worry that we, as philosophers, risk leaving these injustices invisible when we simply weigh the different complaints of “people” taken in the abstract.  It matters, for instance, that our essential workers are disproportionately members of racial minorities, whose current vulnerability is the result of prior social injustices.  It also matters that the recession is concentrated in the service sector, where most employees are women.  Somehow, we need to find a way to factor this into our assessments of different policies.  One way to do so — though I’m not sure it’s the most satisfactory way — is to suggest that the size of a person’s complaint about a policy depends partly on whether they are already disadvantaged, and also on whether this disadvantage stems from a prior social injustice.  

Lastly, I think we need to ask about the risks imposed by particular policies, rather than about the risks of “lockdowns” or “opening up” in general.  Different policies raise different moral dilemmas.  For instance, digital immunity passports raise questions about whether economic or health benefits to the many can ever justify a government legislating two classes of people into existence.  Proposals to impose mandatory lockdowns on the elderly for long periods of time raise questions about parity of treatment and what it takes to justify paternalism.  And there are questions that nobody has been asking about what society owes to those with disabilities who normally live in the kind of isolation that many of us have now experienced for the first time.


CHRISTIAN: I’d like to follow up on Sophia’s insightful comments on how Alec’s interesting model might apply to policy responses to COVID-19 by drawing attention to two further respects that the tradeoffs we face in deciding whether to ease lockdown restrictions differ from tradeoffs involved in simple cases of choosing which of two groups of people to protect from harm. One of these makes it easier to justify continuing lockdown restrictions, while the other makes it harder.

First, one principal justification of these restrictions is that they are necessary to prevent people from doing or enabling harm to others. (In the simple cases, by contrast, there is no question of the one group we might save doing or enabling harm to the other group we might save—these groups are related only in the minimal sense that our saving one group will mean that we do not save the other.) Generally speaking, it seems permissible to impose more cost on a person that might do or enable harm others than simply as a means of conferring benefits on others. Now just how much more cost is debatable, especially when those that risk doing or enabling harm do not do so intentionally. How great the harms and how large the risks that their conduct will result in them surely matters, but so too will whether those put at risk can plausibly be seen as having validly consented to the imposition of these risks, or if those on whom the risks are imposed are, reciprocally, imposing risks on others. Still, this feature of the situation weakens the complaints of those who would be made worse off by the continuation of lockdown, if these restrictions are necessary to reduce the risks that they will do or enable harm to others.

Second, if people do suffer harm as a result of lockdown restrictions, those who are responsible for imposing these restrictions will not just have allowed these harms to occur.  At the very least they will have contributed to them by enabling them to occur. If we save one group at the expense of the other in the simple cases, by contrast, we will have allowed a group to suffer harm but will not have enabled harm to them. Generally speaking, it seems easier to justify allowing some to be harmed when this is necessary to protect others from suffering than it is to justify doing or enabling harm on some for this reason. This consideration strengthens the complaints of those who would be made worse off by the continuation of lockdown, if these restrictions will indeed do or enable harm.

*              *              *

Join the conversation in the comments.

45 Replies to “A Philosophical Conversation on COVID-19: with Alec Walen, Sophia Moreau, and Christian Barry

  1. Just another, perhaps minor, complication to throw in the mix. Much of the harm that results from easing a lockdown will be due to voluntary acts of negligent or even intentional risk imposition. But those acts have not yet happened and the estimation of expected harm requires predicting their frequency and magnitude. If the decision to continue the lockdown is made on that basis, people for whom it would have been far better to ease up will suffer harm on the basis of statistical generalizations. They might feel aggrieved about the assumption, however justified by statistical evidence, that a large number of those who seek re-opening would act recklessly. If they have reason to believe that they themselves would not have done so, could they have a complaint about the harmful use of statistical evidence? (They couldn’t place the blame on those who would have flouted safety measures, unless they have already done so.) What if the prediction wasn’t strictly scientific, but involved generalizations about groups to which the complainant belonged, like Red State residents, or Trump supporters? Maybe, since the generalization was not applied to them specifically, they have no complaint.

  2. I’m going to respond first to Sophia, then Christian, then David.

    To Sophia: you write, “Many lives are at risk under lockdowns, too. The UN has reported a surge in serious abuse of women and children; women are being denied access to life-saving reproductive healthcare; people with life-threatening medical conditions have had their treatment put on hold. There are many with large complaints about all alternatives.” I agree. My point is that I think it’s reasonable to assume that *more* lives will be lost if we open up now than if we keep more restrictions on social activity. If aggregation matters, and lives dominate the balance of claims, then we should not open up yet. I’m not sure why, then, you frame your point by saying “I don’t think our dilemma maps neatly onto the scenarios that philosophers standardly discuss, when thinking about whether to aggregate and how to measure people’s complaints.”

    That said, I am very sympathetic with your point that social disadvantage and being the victim of injustice is relevant. And I agree, it’s unclear how best to account for that, though it should count for something. I worry, however, about being too quick to assume that the right response is “open up” rather than “do more for those who will be most harmed by a prolonged shutdown.”

    To Christian: I really appreciate the focus on causing or enabling harm vs allowing harm. But I’m skeptical about the claim that it is significantly more difficult to justify locking society down and causing or enabling harm than merely allowing harm by not locking down. The reason is that I think the most important issue is the claims of those affected. If those who go about their business when the lockdown is lifted are imposing unjustifiable risks on others, and if there’s no way to prevent them from doing so except by restricting their liberty (perhaps not with a “lockdown”, but some other lesser restrictions), AND those they expose to the risk of infection have a right not to be so exposed, then those who face lockdown do NOT have a right not to be locked down. They should voluntarily submit to the same treatment for the sake of those they would otherwise expose to risk. That is an enforceable duty, and it’s not hard to justify enforcing it.

    I don’t think this is the condition we’re in because I think the claims to be free to engage in social and economic activity are–if proper precautions are followed–stronger than the claims to be maximally free from risk of infection. But that’s where the action is, in my view.

    To David: Interesting point about being restricted on the basis of statistical generalizations. Why should *I* have to be limited to being in my house if I’ll be careful when I go out, just because some other yahoos won’t be careful? But this is a quite general problem about over-inclusive laws, restricting gun ownership (hey, I’ll be careful with my guns), drunk driving (actually, I’ve got great reflexes when my blood alcohol level is over .1), sex between adults and 15 year olds (actually, she is very mature and fully capable of consenting), etc. The question is: should one recognize that there are reasons why the law has to be somewhat crude, and accept, for the sake of the greater good, the over-broad restrictions. I think the answer is generally yes, but not always.

  3. Many thanks to Alec, Christian, and Sophia for sharing their thoughtful views on reopening.

    First, I want to follow up on Sophia’s call for more context. Alec proposes that a cautious reopening might be justifiable because many people’s claims “based on economic and psychological welfare” can “plausibly outweigh the claims [of other people] not to be allowed to die early.” As Sophia urges, reopening isn’t an all-or-nothing matter. Different policies for reopening implicate the varied moral interests we have (in health, in privacy, in employment, and so forth) in similarly varied ways. And I’m unsure how Alec’s proposal guides us in sorting through people’s competing claims—to economic and psychological welfare, and to continued life—in particular policy settings.

    For example, should more visitors be allowed to nursing homes so that some people can reconnect with their elderly parents and family members? Should we offer in-person classes to support the psychological welfare of students (and the reputation and hence economic welfare of the university and its members), even if doing so means putting students, staff, and faculty who are elderly or have certain medical conditions at serious risk? I’m not sure. In each case, some people have strong claims of economic and psychological welfare to reopening, yet I am uncertain what moral principle, on Alec’s view, would explain why those claims have priority over the lives of the individuals at risk. And if they do not have priority, why not?

    It may be that Alec’s proposal is not supposed to apply at this more granular level of public policy. If that’s so, at what level of policy making is the proposal supposed to apply, and can the proposal still provide guidance in these more low-level cases that currently confront us?

  4. I also want to follow up on another theme in Sophia’s comments: that how to reopen is a social justice question. Sophia explains that analyzing reopening policies in terms of the “complaints of ‘people’ taken in the abstract” risks obscuring how social injustice has left the members of our societies differentially situated with respect to various reopening policies. If Sophia is right, such abstraction would be regrettable not just because we’d risk getting the answer to the question of how to reopen wrong. We’d also risk further entrenching the “colorblind” approach to lawmaking that, as critical race theorists have long argued, treats the experiences of white men as our public moral default.

    I certainly do not think Alec means to advocate for such a colorblind approach (and abstraction can instead aim at inclusivity). But I think that responsiveness to social injustice is something that has to be built into our theories—something that we think about while we are crafting our moral principles for reopening—because we have historically tended to treat social injustice as an afterthought. And so I look forward to hearing more from Alec about how his proposal would treat the differential interests people have based on their race, gender, class, and so forth.

  5. Thanks for the important and interesting discussion, everyone.

    I agree with Sophia that it pays to think about specific policies, as well as (rather than?) “opening up” in general. For example, I think that it can be much more difficult to justify *requiring* individuals to return to normality (e.g. to work) than it is to justify *permitting* individuals to return to normality. To take an example close to my heart, my reason not to go to the pub may be weaker if the bar staff are there voluntarily than if the bar staff are there only because they’d be fired if they didn’t turn up. This is because, when it comes to the easing of lockdown, we’ve weighty (though not always decisive) reasons to ensure that individuals have adequate opportunities to shield themselves from infection.

  6. One thing that the discussion so far reveals for me is the limits to which philosophers are well placed to make a contribution by coming to conclusions about different policies, whether large or small scale. COVID throws up a myriad of morally tricky issues. So I think that rather than thinking about whether we should open up more or less, or even about more specific policies, where the empirical and moral questions multiply quickly, we should think about the significance of certain kinds of consideration in deciding matters of policy. To what extent should the widespread smaller frustrations of lockdown matter in policy decisions when lives are at stake insofar as they do not lead to any larger harm? To what extent does it matter, when assessing the significance of harm, that the person who suffered the harm would have been able to avoid it? To what extent should we assess policies to respond to COVID by considering not only the number of people suffering more or less serious harms, but how those harms are distributed in a way that exacerbates and compounds pre-existing inequalities and injustices on, for example, racial, disability, gender, socioeconomic, and also global lines?

    I also think that whist the framing and the question is interesting, a much more important distributive question arises. What can political and legal philosophers contribute to arguing for the development of more impressive global institutions, laws and regulations that can help the human race respond more effectively, legitimately and fairly to challenges like this, that will almost certainly continue to arise – not only to preventing and limiting the effects of pandemics, but to ensuring that the dire economic, social and health burdens that will almost certainly be caused by the pandemic beyond death and illness caused directly by COVID are shared fairly and responded to appropriately across the globe. I’m surprised how little this last idea has figured in academic discourse, given the catastrophic response of some of the champions of nationalism and localism, and the fact that national and local responses to the crises that the pandemic is causing and will continue to cause seem to be all that is on the table in popular public discourse. The WTO is responding to the crisis in a range of ways, for example, but with almost no scrutiny at all either in public discourse or those concerned with global distributive justice.

  7. Hi Alec,

    Let me check I understand the form of your response to Christian. As Christian says, it’s more difficult, generally, to justify enabling persons to inflict harm on others than it is to justify allowing harm to befall others. I.e., it’s hard enough to justify not curing Mary’s infection; it’s even harder to justify licensing John to engage in behaviour that will lead to his infecting Mary (and *then* not curing Mary’s infection).

    But — you’re claiming — some cases that meet the causal conditions of ‘enabling harm’ lack the harder-to-justify status of typical such cases. In the John-and-Mary example, if John has a prior moral entitlement to do X, and his Xing infects Mary, then my licensing John to do X *enables* harm to Mary in the causal sense, only. My licensing him to do what he has a prior entitlement to do has the *moral* status of allowing harm to Mary.

    Is that the idea? I.e., if John’s entitled to a degree of freedom of movement within society, my licensing him to move about in public spaces doesn’t *enable* him to infect Mary (in the morally significant sense), even if his moving about does infect her.

    I couldn’t tell from your remarks whether you think enabling harm will always be *either* merely causal *or* all-things-considered unjustified. I.e., John is a lot less likely to infect Mary if he wears a mask. Plausibly, there’s no prior entitlement go out among people without a mask (at least not one comparable to retaining freedom of movement). Accordingly, licensing John to go mask-less would enable him to harm Mary (or to impose a greater risk of harm on her), not merely causally, but in the morally significant sense. This makes it harder to justify licensing John to go mask-less — regardless of whether, all things considered, so licensing him turns out to be justified.

    Do you disagree with that last claim?

  8. Thanks to all of you for so thoughtfully helping to complicate the picture even further.

    Alec, you express puzzlement over why I said that our dilemma doesn’t map neatly onto the kind of philosophical scenario in which we have to decide whether the smaller complaints of the many outweigh the larger complaints of the few. There are several reasons why I think our situation now is much more complicated. Some of these reasons have been pointed out by others in the Comments, and pertain to the many different kinds and magnitudes of complaints that are at issue in our current situation, beyond complaints about social and economic welfare, on the one hand, and lives lost, on the other. I’ll turn to these in a moment. But the reason that I meant initially to call attention to is this: I don’t think it’s evident that there are more serious complaints (at the level of lost lives) about opening up than there are about lockdown policies. One reason for this is the one I mentioned in my initial comments: home is not a place of safety for everyone. But, equally importantly, our current lockdowns have not protected everyone: front line health care workers, paramedics, grocery store clerks, truck drivers, cleaners, meat-packers, and support workers for seniors all still need to go to work, and it has become clear to us that a disproportionately large number of fatalities come from these groups, their families, and those who live near them. If we opened up, but did so in ways that actually enabled us to produce and distribute enough proper PPE to these groups that they could be adequately protected, then it is not evident to me that opening up would threaten more lives than do our current lockdowns.

    But furthermore, as Sabine mentions in her Comment, there are so many different kinds of complaints about different ways of opening up and locking down, and, as I mentioned in my initial remarks, many of these complaints raise issues of social justice and so cast some doubt on whether the right approach is to simply to assess the magnitude of different people’s complaints and then aggregate. What about the right of children to a continuing education, or the right of vulnerable children from impoverished backgrounds not to be set even further behind their more privileged peers, who have internet access and an online education? What about the position of women who aren’t at risk of abuse, but are nevertheless having to shoulder a hugely disproportionately burden of the childcare and eldercare during lockdown –what weight do we give to their complaints? Does it matter, when we think about the magnitude of a complaint, whether this person has a right to the good in question, or whether it is just an important interest? And what about the racial injustices that are being perpetuated by various policies –both under certain lockdown policies and under certain ways of opening up? Are these individual complaints – or are they (as I suspect, but didn’t have the space to elaborate upon in my initial thoughts) complaints made by groups, where we lose sight of the real nature of the injustice if we try to paint it as a complaint made by a single individual? How does this model allow us to think about complaints of injustice against a group of people? The point I made about justice was not supposed to be independent of the point I made about the difficulty of discussing our current situation using the simplified models we use when discussing aggregation. If we don’t know how to conceptualize these more complicated complaints, or how to factor in considerations of justice, then it’s not clear whether the model is really a helpful model of our current situation.

    I would also like to reply to some of the other Comments, including your interesting comment, Tom, and yours, Victor. So I will be back momentarily to do that. (Like most of you, I am multi-tasking during lockdown, and this afternoon’s task involves assisting with a Grade 2 science project: building a roller coaster out of recyclables that stretches all the way down your staircase. Just what my home needs during lockdown).

  9. Tom, I agree very much with what I take to be the spirit of your comment –that it is morally significant how great a cost particular individuals face, when they choose an option such as going back to work, and that governments, employers, and clients all have a responsibility to ensure that the costs imposed on particular workers when they return to work are fair (I’ll come back to what this might mean in a minute). But I think we need to be careful of relying on the “permitted / required” distinction here, and also careful of the term “voluntary.” In your example of the pub staff, those who would be fired if they didn’t show up for work are nevertheless at work voluntarily, in the sense that they have chosen to show up for work rather than to be fired. So the problem isn’t that they didn’t have a choice. It’s that there were severe –and, I would want to say, unfair– consequences attached to one of their two options. If you don’t think this counts as “voluntary,” then we can tweak the example to make the same point. Imagine that the staff are not legally required to return to work. Instead, their employers are required to hold their jobs for them even if they choose to stay home. But there is no government scheme in place to pay their wages during the time when they are not at work; so if they stay home, they will have no income. They therefore choose to go to work. In this version of the example, the employees are certainly voluntarily at work. But because one of their options has such severe consequences attached to it, it doesn’t seem like they had a fair choice. What seems most important here is not whether they are at work voluntarily or under compulsion, or whether the government is requiring them to be at work or permitting them to be at work, but what the costs of going back to work are for them.

    And here’s where I think it gets even more complicated, in two ways. You note that we have weighty reasons to ensure that individuals “have adequate opportunities to shield themselves from infection” –and I’m assuming that you are thinking of the option of staying home from work while being paid as one such opportunity. While I agree, I think a lot of work needs to go into thinking further about what counts as an “adequate” opportunity to avoid infection. Is a person’s opportunity to avoid infection adequate when the government provides protective equipment that’s appropriate for, and commensurate with, the risks they face –N95 masks for front-line health care workers, but nothing more than instructions on how to sew your own cotton mask for ordinary citizens lining up at grocery stores, given the different risks these groups face? Is a person’s opportunity to avoid infection “adequate” only if they have no worse a chance than anyone else in their society, of avoiding infection? What about the very real concern that, given the enormous background inequalities in our societies in quality of housing, quality of sanitation, neighbourhood layouts, etc, the rich have many more opportunities to avoid infection than the poor? Does this impose further obligations on the government, or on private employers, to do more for members of vulnerable social groups than they do for more privileged groups –eg. more mobile testing units in impoverished areas, more social benefits, more PPE and medical supplies diverted to these neighbourhood hospitals than to others? The second complication here is that it matters who the “we” is that we are imagining has the reasons to ensure others have adequate opportunities to shield themselves from infection. Tom, your example is nice because it draws out the many kinds of agents whose obligations we need to think about here. We need to think about the government’s obligations: what obligations does it have, to put the background conditions in place that would enable some people to choose not to work. We need to think about the obligations of employers and those in similar positions of power. And we need to think about the obligations of ordinary people like the clients of the pub, whose choices over matters that used to seem trivial can now significantly increase the risks that others face. (We also need to think about how individuals’ and employers’ obligations are affected by the failure of governments to meet their obligations. Perhaps, as Barbara Herman once emphasized to me, we have more stringent such obligations when the government is not doing what it needs to do, unfair though this sounds).

  10. Victor, I think it’s a really excellent point that philosophers ought to be thinking more globally, about what kinds of international institutions and laws are necessary in order to achieve social justice during and after the pandemic. (Though I don’t think such questions are more important than the question of what to do, right now, given the institutions we have, and given that local authorities are the ones currently making decisions about how to open up. I think they are equally important questions).

    You also suggest that, rather than thinking about the significance of certain considerations in relation either to specific or general policies, we ought first to decide on the moral relevance of these considerations in the abstract. I can see why you and others take this approach. But for some of us, it seems a mistake to think that we can settle questions about, eg. the relevance of past injustices, in the abstract, without thinking about what the other morally relevant variables are at play in a given situation, and in particular, without thinking about the distinct institutional context in which a given policy is being proposed (eg. which type of agent is making the policy, what that agent’s pre-existing obligations and responsibilities are, and what kinds of relationships obtain between the different people to whom that agent has responsibilities). I don’t see why we should suppose that the answers to these questions about the moral significance about particular factors would be the same, regardless of the institutional context and regardless of other morally relevant factors. That’s why I think it’s important to talk about specific policies, and to assess the moral relevance of different considerations within the relevant institutional context.

  11. I’m back and want to thank Sabine, Tom, Victor, Kerah, and Sophia for their extensive engagement. Stepping back from the content I think we see the emergence of all sorts of relevant details, a richer and more nuanced picture. That’s what a discussion should be about. What I want to resist is making it into an *argument* when we are really meaning to say, “yes, and let me add…” And, of course, Sophia, Christian and I had *very* few words with which to try to spell out something to kick the conversation into gear. So… principle of charity here.

    In that spirit, I want to try to engage with everyone briefly in what follows. I’ll do it in a string of comments, with hopefully clear headers, that will allow the reader to choose which ones to read and which to skip.

  12. ROLE OF MORAL PHILOSOPHY: One of the main themes in the comments has been on the proper role of moral philosophy. Victor, in particular, was pushing on this when he wrote: “rather than thinking about whether we should open up more or less, or even about more specific policies, where the empirical and moral questions multiply quickly, we should think about the significance of certain kinds of consideration in deciding matters of policy.”

    I want to say, yes, the kinds of considerations that matter is one important thing we can do with discussions like this. And yes, given the empirical uncertainties, ultimate prescriptions is not what we should be in the business of providing. But I want to add that another function of engaging with questions like: “how should we weigh the competing claims in the context of Covid-19” is to use the moral urgency and ultimately the moral clarity regarding what is an implausibly extreme thing to demand as a test case for moral theories.

    I wrote my initial comment thinking about my forthcoming paper on risk and aggregation, and it still seems useful to me to ask how my model and different models would try to handle the policy questions around Covid-19. For example, compare my model with traditional Scanlonian contractualism. If one does not believe in aggregation, and is concerned with ex post claims, how does one weigh out the options? It seems that if one thinks that claims not to be killed are stronger than claims to be saved, and stronger than any other claims that might speak in favor of opening up the economy, then the clear prescription for Scanlonian contractualists should be: stay locked down (or at least under a highly restrictive regime) until doing so is likely to produce more deaths than an alternative. Indeed, if we are concerned with the strongest complaint against a policy, we would have to reach the point at which those who are dying from the lockdown can say to policy-makers: you are killing me. Then their claims not to be killed would be as strong as (or stronger than) the claims of those who are killed by an infection whose claim is the arguably slightly weaker claim against enabling others to kill them.

    My point is not to explore the details of this here and now. It is to say that without aggregation, it seems an impoverished way of thinking. Then one has to ask: are we best off accepting the model and complementing it with some sort of consequentialist reasoning as suggested by Johann Frick in his paper on Contractualism and Social Risk? (Note: his model of contractualism was ex ante, not ex post; that’s beside my point). I think not. I think aggregation of lesser claims has to be part of the picture from the get go.

    In sum, my point is that we can and should use real world problems to test theories. That will, I presume, also pay off when it comes to the task of using moral philosophy to guide real world decision-making. But we don’t have to go straight for applied moral philosophy. The discourse should move back and forth in search of reflective equilibrium.

  13. David’s intervention raises some very interesting issues. One is whether, when considering what restrictions can be instituted to prevent individuals in some population from imposing risks of harm, it matters how much of the harm that might thereby expect to be prevented would have been caused culpably (I’ll use the phrase ‘caused culpably’ to capture harms that result either from negligence, recklessness, or intent to harm).

    I tend to think it does matter: an outcome in which some quantum of harm has been culpably caused seems worse, all else being equal, than one in which such harm has not been caused culpably, so we have more reason to prevent it. But I’m not sure how much it matters. To figure out how much it matters we’d need to consider cases where we face the choice between two policies, one of which would have a greater expectation of harm reduction, while the other would have a greater expectation of culpably caused harm reduction. I’m having a difficult time thinking of a good examples off the top of my head. In the abstract, however, I do feel the pull of prioritizing the policy that results in the greatest overall harm reduction, at least where the differences between the two policies are significant on this dimension.

    But let’s suppose that some restriction is partly justified on the ground that it is necessary to prevent harm that would be caused culpably. Would individuals that are restricted on this basis have complaints if they have reason to believe that they themselves would not have culpably caused harm in its absence? [It’s interesting to think if this case is any different from cases where a policy is justified simply on harm reduction grounds without taking the factor of culpability into account. Does a person who regards themselves as quite competent at driving safely at high speeds have a complaint that imposing speed limits on them involves the harmful use of statistical evidence?]

    I’m inclined to adopt the position that David suggests in closing: that in the case where some restriction is applied to all within some jurisdiction, these individuals would lack complaints because the statistical generalization that is being used to justify the restriction is not being applied to them specifically but as a basis for a restriction on all. This seems in contrast with those cases where the use of statistical generalisations seems most questionable, such as sentencing a particular person for a criminal offense on the basis of some reference class to which they belong, or restricting only the freedom of movement of republicans on the ground that they are more likely to forgo the use of protective masks than democrats.

    I also wanted to take up further Sabine’s point concerning the importance of remaining responsive to social justice when thinking about responses to episodic crises like COVID-19. In my opening contribution, for example, I drew a contrast between harms that result from what a government does or enables from those that it allows to occur. But it is pretty clear that when thinking about lockdown it becomes quite difficult to apply cleanly any such distinction. For example, it isn’t right to say that many of those who suffer harm as a result of the easing of restrictions will merely have been allowed by their government to suffer harm. The reason is that many of the people that will be harmed by the spread of COVID-19 may have been made particularly vulnerable to such harm as a result of other things their government has done to them or enabled others to do to them. In that case we are no longer weighing the significance of harms that the government might do or enable against those it might allow, but those it might now do or enable against those it will otherwise have done or enabled as result of things it has done in the past.

    I will reply to Alec separately after I’ve read his reply to Kerah.

  14. GRANULARITY OF OPTIONS AND ANALYSIS. Sabine asked whether my framework is meant to apply to more fine grained options and contexts. She offers this helpful example: “should more visitors be allowed to nursing homes so that some people can reconnect with their elderly parents and family members?”

    Following Victor’s warning about empirical uncertainty, I should refrain from saying anything too concrete. But I do think that my views should help us to see what’s at stake here. In my recent book on rights theory, The Mechanics of Claims and Permissible Killing in War, I argue that we should think about rights as, first, a matter of the balance of patient-claims on an agent (including artificial agents, like the state or a corporation), then weighing the agent’s claim not to have to act on the balance of patient claims (a claim that only real human agents have), and then taking into account special claims on whatever liberty the agent may have.

    So, what can we say about the nursing home context? There are many different possible agents we could focus on: the state regulatory body, a particular nursing home, a doctor taking care of a patient, a patient, a family member who wants to visit a patient. Just to keep it “short,” let’s focus on the state.

    The state has to consider the claims of people at the nursing home who want to get visits from their families and the claim of family members who want to visit their relatives–those are the primary claims pulling in favor of allowing visitation. Then there are the claims of patients who (presumably) want to avoid infection and death (disclosure, my step-father’s brother recently died of Covid-19 in a nursing home, so this topic hits close to home). Some of these people may be too demented to express their own desires, but we don’t need to recur to paternalism to judge what to do; their family members can presumably press their claims for them. In addition, there are staff who have reason to fear infection and have claims to avoid being exposed.

    Now there come questions about whether having responsibility for putting oneself in a difficult situation might weaken one’s claims. For example, staff might have weaker claims because they choose to work in these settings, whereas many older people may have not much choice to leave. But then, realistically, many staff have no good economic alternatives to quit, so their claims may not be weakened. And, of course, staff will go home and their families have little choice about that, and they have claims not to be on the receiving end of infections.

    Then we come to the range of options. If the question is simply: allow visitors or don’t allow visitors, the interest in avoiding killing the vulnerable outweighs the interests in maintaining personal connection. But, of course, the options are not so black and white. There are a range of options to consider. Perhaps the facilities could develop wings where those who want visitors can get visitors and those who want to remain safe can remain safe. Or hygienic policies can be put in place, with masks and shields. Such things cost money, of course. So the state should ask: given the cost to taxpayers, should we socialize the costs of making it relatively safe for people to visit their relatives in nursing homes? If so, how much cost can be passed on to the taxpayer? Should we expect rich families to fund the bill and help poor families?

    And if we can mitigate but not eliminate risk with measures like masks and shields, what should we say about those who still object to having visitors come and *raise* the risk of infection and death? Should they still have dominant claims, or when the risks get low enough, do they lose to the claims of those who want to have more connection with their loved ones?

    All of these questions naturally unfold, but I think they can be handled in the Mechanics of Claims framework.

  15. CLAIMS IN THE ABSTRACT VS CLAIMS TAKING ACCOUNT VARIOUS FORMS OF INJUSTICE. I have less to say here. Yes, being a victim of injustice surely matters. We should, at the very least, be very mindful of how, for example, poor people tend to hold down the essential jobs that expose them to more risk of infection. And we should think about how, for example, being black in a racist country induces a level of stress that could very well explain why blacks in the U.S. are dying at a higher rate than whites.

    But how do these claims matter? Do they imply that we should use the response to Covid-19 to rectify past injustices? Or do they mean only that we should be very mindful of the pressures that the poor face and the vulnerability that oppressed minorities may have in thinking about their claims to be accommodated? I suspect the latter.

  16. REPLY TO KERAH RE MY COMMENT IN REPLY TO CHRISTIAN. Kerah, I really like the way you put it. If R is the regulator, and R sees that John has a right, all things considered, to move about in society (let’s say, with a mask), then acknowledging that right does *not* count as enabling him to harm Mary in a way that is meaningful in itself. For regulators, ensuring that people have the freedom they should have and do not have the freedom to wrong others is simply what they should do. In a word, the focus on enabling harm is just misplaced once you know the balance of claims between Mary and John, and the claims each has on R.

    So what are the claims on R? Being somewhat oversimple, Mary’s claim is: save me from being exposed to risk by limiting John’s freedom of movement (or, maybe, by providing me with other resources to protect myself from contact with John). John’s claim is: allow me to move about freely to pursue my ends; do not prevent me from doing important things like earning a living, visiting friends and family, going to political protests, what have you. (This makes it all seem rather ex ante, but that’s just for simple exposition. We could also reframe this in terms of ex post claims and aggregation.)

    But, crucially, R is not really the important agent in this scenario: Fundamentally, John is. Fundamentally, Mary’s claim is on John: don’t expose me to risk for the sake of the goods you want to pursue. And John’s claim back to Mary is: I should be free to expose you to risk for the sake of the goods I want to pursue.

    R’s role comes into view to try to set rules of sufficient clarity, and with sufficient room for case-specific exceptions, so that the Johns of the world have all the freedom they should have, and no more. The rules will always be over-and under-inclusive. But they must be laid down to try to get the cases, generally, right, watching out to avoid rare but extreme injustices.

    Interestingly, the specific nature of the claims on R may actually be the *reverse* of what was suggested in Christian’s comment. Mary’s claim is: don’t allow me to be harmed by John. That’s a positive claim for help. John’s claim is: don’t harm me by restricting me unjustly. That’s a negative claim not to be harmed. All else equal, *John’s* claim is the stronger. But, again, this is a secondary consideration. The primary balancing is done at the stage of determining what John may do to Mary, and in *that* context, Mary’s claim is the stronger, negative claim.

    I am not denying that there’s some role for the notion of enabling. But it presupposes a baseline. Suppose John would not normally have the ability to move about in public with N95 masks, because they are not available, but R can order a large number of N95 masks to be made and get them distributed. Then R can enable John to move about with an N95 mask. If this creates a new option that exposes Mary to more risk than she would face if John were simply locked down in his house, then Mary might have a claim: do not enable John to get an N95 mask and then move about with it, exposing me to risk. But I think this is a fairly marginal sort of issue.

    In sum, we have to pay attention to where the action is, and it shifts around depending on who is the agent. We need a two-stage analysis to get it right. Simply saying that R enables John to expose Mary to risk if R permits him to move about oversimplifies that moral landscape.

  17. Thanks to Alec, Sophia, Christian, and Kerah for this wonderful discussion. I don’t have much to add, but I thought I’d follow up on Sophia’s idea that lockdown policy should be sensitive to social injustice.

    Sophia writes:

    “COVID-19 is not just a combination of a health crisis and an economic crisis. Our responses to it risk perpetuating serious social injustices…It matters, for instance, that our essential workers are disproportionately members of racial minorities, whose current vulnerability is the result of prior social injustices…Somehow, we need to find a way to factor this into our assessments of different policies. One way to do so — though I’m not sure it’s the most satisfactory way — is to suggest that the size of a person’s complaint about a policy depends partly on whether they are already disadvantaged, and also on whether this disadvantage stems from a prior social injustice.”

    This sounds undeniable in the abstract—if injustice is happening, why ignore it?—but I think Sophia is right to be cautious. There is something paradoxical about this way of doing policy. (And now I’ll rip off Schelling’s “Economic Reasoning and the Ethics of Policy.”)

    We have two problems here. The “lockdown” problem of whom to keep inside, and the “injustice” problem of how to help the oppressed. Sophia’s basic idea is that we need to be mindful of both. (No doubt.) The more specific, and tentative, proposal was to prioritize justice when deciding whom to lock down—in effect, trying to solve both problems at once. (Or at least, to solve the lockdown problem in way that helps with the injustice problem.) For example, this might mean hiring people who aren’t minorities to do more of the risky essential work.

    But there’s something inefficient about this. The people who are currently doing the essential work are the ones most willing to do it. To hire someone more privileged, we’d have to pay them more. That makes inequality even worse, which means we’d have to offset *that* by giving still *more* cash to the essential workers stuck at home. A simpler solution is to let the essential workers do their work, then correct for any resulting or lingering injustice directly, via your favorite means of redistribution. Solve the lockdown problem on its own terms, then account for the effects when solving the injustice problem on its own terms. Sometimes, one stone isn’t enough!

    Caveat 1: this assumes that we have the *option* of redistributing wealth by our favorite direct means. If that’s politically infeasible, it’s harder to object to lockdown policies as a method of redistribution. (Mutatis mutandis for other roundabout redistribution policies, like free parking, which help the poor but in wasteful ways.)

    Caveat 2: I’m assuming that the lockdown policy only “perpetuates injustice” in the sense that it doesn’t help. If the policy is plain unjust—e.g. because it discriminates against women—that’s a death blow.

  18. Thanks for the response, Sophia — I agree with almost everything that you say here. I was envisaging the case of a vulnerable member of staff who’s told that she must turn up for work or she’ll lose her job. It’s not exactly “your money or your life?”, but it could be close to it — and that’s why I thought her attendance might be involuntary. But, to be clear, I intend this only as an example: involuntariness of this kind might be sufficient to trigger alarm bells, but not necessary. I agree with you that concerns of can fairness arise in cases short of this, and that what seems most important is what the costs of going back to work are for different individuals.

  19. Alec and Kerah, your exchange is really helpful (and clarifying.)

    What motivated my initial comment was the general precept that we have especially stringent reasons to ensure that the state, acting on our behalf, avoids causing harms directly through its agency, even when this means that greater harm will result, overall. This precept seems implicit in restrictions in criminal procedure, constraints on punishment, and many other public policies (or at least so some of us have argued). I was assuming that if the state doesn’t prevent John from moving freely about, then if he ends up infecting and harming Mary, he does harm to her, but the state will have allowed harm to occur. I wasn’t in this sort of case understanding the state to have enabled him to harm her by ‘licensing’ him to move about. [I realise this raises some challenging questions about how to interpret distinctions between what it means to do/enable/allow when it comes to the conduct of institutional actors.]

    The case that you introduce of John and Mary doesn’t really contest this precept. It suggests, rather, that in this instance the restriction on John’s freedom does not infringe his rights or expose him to harm he was not liable to bear if we assume that Mary has a claim that he not impose excessive risk on her by freely moving freely about.

    However, I think that there are other cases that have a different structure that may be relevant to the current context. One is where some of those who stand to be harmed as the result of restrictions imposed by the state would not be likely to impose excessive risk of harm on others in the absence of such restrictions. [Some groups of people, as Sophia points out, may be particularly vulnerable to harm under lockdown policies but unlikely to do much to increase the risk to others in the absence of such policies.] Unlike John in the simple 2 person case, these people are not liable to the harms that result from the restrictions imposed by the state. From their perspective, the state is enabling harm to them as a consequence of acting to prevent others from causing harm. Here the fact that the agency of the state is involved does seem important, and we need to consider whether our reasons to ensure that the state prevent some from being harmed are sufficient to justify its enabling harm to others that would pose little risk were the state to refrain from doing so.

    I also wanted to follow up, Sophia, on something you raised in your interesting comments on the thread. You wrote that you think it is a mistake to think that we can settle the questions about things like the relevance of thing such as past injustices “in the abstract”. I wasn’t sure what you meant by this, or what approach you take yourself to challenging. No one, I think, regards it as possible to settle questions about particular injustices and the claims to which they give rise in the abstract. I take it, rather, that some of us think we can identify general considerations that are relevant in determining the strength of claims on the basis of features of the world such as past injustices. That is, we think of ourselves as in the business of finding out what factors are morally relevant in determining the strength of different claims, acknowledging that figuring out which of these factors are present in any particular situation requires a careful examination of institutional context, historical details, and so on. This is not to deny that looking carefully at particular contexts can help us to think about what general variables are relevant in assessing different claims.] Indeed, I take some of your comments to Alec as challenging his understanding of what the context is (e.g. whether it best seen as involving tradeoffs between lives and lesser goods, among other things) rather than on the factors that he deems morally relevant to evaluating different policy options. So I’m wondering if there is really the deep methodological divide between you and Alec that your comments suggest.

  20. Sophia, you are absolutely right to say that sometimes the moral significance of some fact to what a person should do depends on features of that person. And that might be true in the case of injustice – those who perpetrate injustice, have stronger reasons to ensure that they mitigate, and do not exacerbate, the injustice than those who are not, we might think. But I’m still for abstract thinking to help us resolve the issues, even if we should be aware of potential contextual interaction of different moral factors. The issues are really tricky, even when we simplify cases. Here are some cases that pick up on a bit of Christian and David’s discussion (though not the interesting statistical question).

    Two groups are at risk of being infected with a disease in different countries Country A and Country B: {A} and {B}

    Case 1: {A} are at risk of infection as a result of injustice because they were wrongly left vulnerable to the infection by their government. {B} were not at risk as a result of injustice. Other things are equal between {A} and {B}: their governments are otherwise equally unjust. A vaccine becomes available in a third country, Country C, and there are sufficient doses to give it to {A} or {B}, or both.

    Should the vaccine be given to {A}? Should it even if {B} is larger so that more lives will be saved? I am not sure that unjust infection is even a tie-breaker in this case.

    Now suppose that in Case 2, the vaccine is developed by Country A. Should A’s government give the vaccine to {A} over {B} (this is a real world issue: as I understand it, the UK government has secured a large share of any potential vaccine developed by the Oxford group to be rolled out by AstraZeneca; though, of course, it has committed plenty of international as well as national injustices!!! They prioritise their citizens on a nationalist basis; but might they be justified because their unjust policies led UK citizens to be vulnerable to infection)?

    This case is tricky, and I’m not sure what to think. It might seem clear to some that the government of A should mitigate the injustice it will otherwise cause to {A}. But in reply, it might be noted that {B} will then be worse off as a result of the injustice that the government of A has perpetrated against {A}. I am at least initially inclined to think that the nature of the unjust policies might make a difference here. I’m at least tempted by the view that there is only a relatively weak reason for the government of A to prioritise {A} on the basis of the injustice it perpetrated against its citizens if they simply acted with neglect (and in the UK there is plenty of that), and this is an ‘allowing’ case rather than a ‘doing’ case (how to draw that contrast, of course, is fraught).

    Things are different, I think, if {A} were intentionally left vulnerable, or were left vulnerable as a result of racist government policies (also plenty of that in the UK). The government has stronger reasons to protect the direct victims of its racist policies than preventing indirect victims, even if it is an allowing case.

    Things may also be different where the unjust actions themselves can be prevented, and that is what Christian’s original post envisaged (that case was focused on costs to potential perpetrators, where the issue is clearer, but there is a distinct issue about costs to third parties) Consider:

    Case 3: Country C can prevent one of two groups being harmed. Either it can prevent the government of A unjustly failing to prevent {A} from being harmed by making A behave more justly, or it can prevent {B} being harmed, but not as a result of injustice.

    In this case, I am inclined to give quite a bit more weight to the injustice. Other things equal, I am at least initially inclined to think that C should prevent {A} from acting unjustly. I think that this is true for the sake of those who will otherwise act unjustly, but also for the sake of the victims of injustice. If that is right, we have weaker reasons to prevent the results of injustice than we have to prevent injustice occurring with similar results. And that is a further reason prioritise developing institutions that will lead governments not to act unjustly rather than mitigating injustices they cause.

  21. Really helpful cases Victor! I think I am with you on the first case (and probably the second-see below), but was a bit surprised by your take on the third. If I can either (1) save 100 innocent people from being unjustly harmed by their government or (2) 100 innocent people from being killed by some unfortunate non-human cause like an earthquake, I don’t see that I have substantially greater reason to opt for 1 than for 2. It seems permissible to me to save either group, and if the number of people in scenario 2 is greater, then insofar as I think numbers should count anywhere they should count here. Moreover, I don’t think I can be required to bear more cost for saving in (1) than in (2) (or, consequently that more cost could be imposed on my by a third party to compel me to save those in (1) than to save those in (2).) So our reasons in these cases seem to me to be symmetrical to those when we are faced with two groups of people that have already suffered injustice or misfortune (I have just as much reason to save 100 from the wreckage of an earthquake as to save 100 from the fallout of an unjust attack, so long as I am in no relevant way connected to the attack.)

    About the second case I’m not completely sure what stance you are suggesting. If the idea that an agents’ reasons to prioritise the mitigation of harms that result from injustices for which they are responsible become continuously more stringent the more serious the injustice and the more the direct is their agential contribution to it, that seems plausible. But your description made me think that you might perhaps see things as much more discontinuous, such that there are very flimsy/easily overridden reasons to prioritise the mitigation of harms resulting even from serious injustices so long as they were not intended, but quite stringent reasons to do so when they were intended, which seemed to me less plausible.

    Ok, it’s late here in Australia so that’s it from me for awhile.

  22. Hi All.

    This is a fascinating discussion, and some of the things I would have picked up on have been mentioned already (in particular the question of how to understand when and whether legislation does harm/enables harm/allows harm).

    Something else that interests me is Alec’s ex post approach. In thinking about the ex post approach, we are inclined to think first and foremost of those who will die if restrictions are relaxed. Alec says we should think of those people as having claims against death/being killed, rather than claims against a risk of death/being killed.

    However, the ex post approach requires us, I think to (broadly speaking) divide people into six categories. First we split people into low risk versus high risk from the disease (this is obviously simplistic, but it’ll do for now, I think). Those who are at high risk must be split into those who will die/be seriously harmed versus those who won’t. Those who are at low risk must be split into four groups across two axes – those who will die/be seriously harmed and those who won’t, and those who will seriously harm others, and those who won’t.

    I’m going to raise some questions about some of these groups and how we should understand their claims according to the ex post view (and Alec’s version of it in particular). In doing so I’m going to simplify things, and fail to take account a lot of the important things that have been mentioned. This isn’t because I don’t think these things are relevant, but just as way of getting to *some* of the issues, in the sort of way Victor mentioned.

    Group 1. High risk who will die/be seriously harmed if lockdown is eased
    As I said above, this group clearly have ex post claims against death/serious injury.

    Group 2. High risk who won’t die/be seriously harmed if lockdown is eased.
    Obviously, under the ex post view these people do not have claims against death. But since they are also under lockdown, should we regard them as having claims against the lockdown? If they did have such claims, the case for lockdown would be weaker than it otherwise would be.

    From a quick re-read of his paper, I think what Alec would say here is that these people would waive their ex post claim to free movement/economic benefits etc, since the policy was ex ante in their interests. So a lot of people who are, ex post, harmed by lockdown do not have those harms/claims counted. Is that right Alec? If so, this makes the case for lockdown stronger than it otherwise would be, since there are a lot of people who would ex post benefit from relaxed restrictions, but whose interests in ending lockdown are not counted. But given the appeal to ex ante interests, this makes the position less of a pure ‘ex post’ position that it otherwise would be.

    Group 3. Low Risk and would be harmed if lockdown is eased.
    A pure ex post view would surely count these deaths as significant in deciding what to do. But on Alec’s version of the ex post view, these deaths may not count – provided lockdown is not ex ante in these people’s interests, they would (or rationally should) waive their ex post claims against death. This would weaken the case for lockdown, since thousands of deaths will be irrelevant to deciding whether to ease lockdown. This is surprising, I think, for an ‘ex post’ view.

    Group 4. Low risk and would harm others if lockdown is eased.
    David Wasserman has pointed out that some will infect others through negligence and has raised questions about how this affects how we should treat them/their claims. But whether or not conduct is negligent is a matter of evidence-relative morality. If we were being purely ex post about this, shouldn’t we simply divide those who pose a risk to others into two groups: those who will actually harm others and those who won’t (because they won’t contract the virus, or won’t pass it on, or won’t seriously harm those they pass it on to)?

    The group who will harm people arguably have ex post interests in being locked down. Insofar as we have interests in avoiding harming others, or perhaps interests in wrongly harming others, these people have interests in being told to stay put. Certainly, I would be grateful to know that someone had withdrawn my liberty for a bit if it meant I wasn’t causally central to the death of another person (imagine someone took control of my car to stop me hitting a pedestrian). This would make the case for lockdown stronger than it would otherwise be. (It should be also be remembered here that we’re much more likely to pass the virus on to those we spend time with indoors – colleagues, friends, family).

    Group 5. Low risk and won’t harm others if lockdown is eased.
    This group has interests in being allowed out and about, and these interests are not extinguished by or to be balanced against interests in not seriously harming others. But these people have evidence-relative/ex ante reasons to avoid harming others (certainly other-regarding, and potentially self-regarding reasons, as explored above). Given that those who are at high risk but won’t actually be harmed are taken to have waived their claims to being out and about, by appeal to the ex ante situation, what should we say of the low risk but won’t harm group? Wouldn’t/shouldn’t they also waive their ex post claims to be out and about? Alec’s view allows ex ante interests to cause us to waive ex post rights for self-regarding reasons, but what about for other-regarding reasons? If so, this would make the case for lockdown stronger than it otherwise would be.

    I hope the above made some sense. It has been written over a couple of hours whilst also engaging in all sorts of other tasks that lockdown has imposed – like homeschooling and playing Avengers.

  23. A quick reply to Christian, and not expecting a response!

    1) On the first point, one thing to consider is how strong our obligations are to others to act in a way that results in their not being serious wrongdoers. As I think that it’s pretty awful to be a serious wrongdoer, I think that these reasons are strong. But I don’t think that we have very strong reasons, owed to wrongdoers, to prevent their attempts from being successful. Suppose that I wrongly threaten an innocent person, but the harm they will suffer is overdetermined. Do others have strong reasons, grounded in my interests in not being a completer rather than an attempter, to prevent the threat that I have caused from being realised, so that the person is harmed by a naturally occurring threat? I don’t feel the force of that, or at least not much.

    A second question is the extent to which we owe it to people to ensure that others don’t have and act on wrongful attitudes towards them. I think that it’s pretty bad that others have and act on these reasons, independently of the harm that I will suffer. Suppose that you are at the end of your life, and you are unsure whether you lived in a society of racists who have tried to kill you, but who failed, or whether you lived in a society of well-intentioned people who have not tried to kill you. If I were you, I’d quite strongly prefer it that the latter is true. But when it comes to the difference between the way in which outcomes are caused, my preferences are quite a bit less strong. Suppose that I am about to be crushed by a boulder. I know that a person who hates me rolled a boulder at me, but I don’t know whether that boulder missed me and another naturally rolling boulder is about to crush me, or whether the attempter will succeed. I don’t have very strong preferences between these things. So I am least tempted by the view that we have pretty strong reasons to prevent people from having and acting on bad attitudes – both perpetrator and victim centred reasons – but not very strong reasons to ensure that they don’t succeed, when we hold harm equal. Christian seems to disagree with the first of these points. I’m interested in others’ views.

    2) Suppose that I am negligent, and for that reason I fail to notice that a lethal threat will arise. As a result X faces that threat. But I now find that Y is subject to a similar threat that I could not have prevented. I can now save either X or Y.

    I am not completely sure about how to analyse this case, but I think that I don’t have a strong reason to prioritise X. I think that I owe a more stringent obligation to save a person than a bystander would. But if I save X, my injustice will still result in my having failed to save Y, and so I then it is less obvious that X has a significantly stronger claim on me than Y. Christian, I think, thinks otherwise. Again, I’d be interested to know what others think.

  24. A reply to Daniel: thanks for your helpful thoughts. I agree with you that hiring the more privileged to do the essential work would create further problems: that isn’t a solution. But I disagree with the assumption –which I think also underlies Alec’s comments—that there are two separate kinds of problems here, a lockdown problem and a problem of injustice. My suggestion was not that we should use our Covid-19 policies to ameliorate a set of separate background problems involving racial and gender injustices, thereby solving two problems at once, the lockdown problem and the injustice problem. My suggestion –and I take this to be Sabine’s point as well– was that these problems are not separate. Given the background social inequalities that persist in our societies, and that have resulted from past injustices, most of the policy options that are available to us now risk, to a greater or a lesser extent, (i) worsening existing social inequalities, or (ii) constituting ongoing discrimination of a systemic kind. (I don’t want to stray far from the theme of this blog, so I will simply say here that if you think that none of these policies could be discriminatory because discrimination requires animus, then you need to read my book, Faces of Inequality, which focuses on indirect discrimination and has just come out with OUP). So there is a question we need to answer, about how we factor (i) and (ii) into our assessment of various exit strategies –particularly if we are using an individual complaint-based model to determine which strategies are morally acceptable.

  25. Hi Victor,

    Here’s having a go at Case 2 (in your last):

    “Suppose that I am negligent, and for that reason I fail to notice that a lethal threat will arise. As a result X faces that threat. But I now find that Y is subject to a similar threat that I could not have prevented. I can now save either X or Y.”

    On your (tentative) analysis, the fact that my negligence contributed to X being lethally threatened gives me a more stringent obligation to save *someone,* but that someone needn’t be X.

    This doesn’t seem right to me — it seems like I have a duty to X either to prevent my negligence eventuating in her harm, or to compensate her (this over and above any general obligation I have not to allow harm to her.) If I owe such a duty to X, how can I claim to have fulfilled it by saving Y instead?

    I wasn’t sure from the description of the case whether, absent the threat to X, I would have been required to save Y, or saving Y would be supererogatory. If the former, I can see that I might be left owing something to Y (or his estate), on account of how a downstream effect of my negligence was Y’s death. (Had I not been negligent, I would be have required to save Y and saved Y). But Y’s (estate’s) claim here seems weaker than X’s would have been had I saved Y, probably a lot weaker. (Among other reasons, the implications of my negligence for Y was not something I ought to have anticipated.)

    Victor, I’d love to know where you think this sort of analysis goes awry. (It wouldn’t be the first time you convince me of something I initially thought HAD to be wrong.)

  26. Thanks for your response, Christian. It looks like we’re pretty much on the same page. But I need to add some refinements that occurred to me when I woke up this morning. I woke up thinking that I’d missed something important in my own discussion of R, the regulator, and the claims on R. I was thinking of R operating in a context in which you can independently make sense of whether those who want to move about are imposing unjustifiable risks of harm on others who wish they would not move about. The model I was implicitly relying on is that of conscientious drivers and pedestrians (the last part of my paper on Risks and Weak Aggregation). But of course much government regulation has to address effects that arise because of human behavior on a large scale, a scale where the moral effects cannot be analyzed down into individual actions as some libertarians might wish to do.

    A better model, then, might be pollution, where one person being an emitter does no harm at all, because it is simply diluted to levels where it falls into the background noise of life, but if many people emit, then the pollution starts to accumulate and people start to be effected. In that kind of case, R has reason to act and is an agent in a stronger sense. Now, R’s regulations will restrict people who individually could free ride on the good behavior of others with no harm, and R’s choices about how to regulate will determine how many suffer harm, and, on the other side, how many live with what restrictions on their liberties. And, of course, issues of distribution of impact on restrictions and permissions really arise here too.

    I’ll come back in a moment to say something more about whether the behavior of people with a transmissible disease is more like driving or polluting. But first I want to say something about the significance of regulations ethically. I think it’s important to recognize that well-designed regulations will be over and under-inclusive in various ways, but nonetheless impose moral duties on people not to free ride on the cooperation of others. That’s not to say that they have to be mindlessly obeyed. The state doesn’t have the resources to grant approval to all reasonable exceptions. People should think through whether they really are in exceptional circumstances–the kind that the state *should* recognize as exceptional if it only had the resources to consider most pleas for exception. But in general they should take reasonable regulations as relevant to setting their baseline liberties and duties and expectations of others. In other words, in the modern world, the old model of the law tracking a prior set of rights and wrongs has to give way to a model that includes mala prohibita, where the wrong is wrong because the state reasonably chose to make it so.

    How do we think about what the state is doing, then, when it permits certain forms of risk-imposing behavior? Is it merely allowing those who suffer to die? I think we need to be careful of an equivocation between the moral and the causal sense of allow. In permitting (allowing) a certain level of activity, it is responsible, morally and causally, for bringing about (causing) a certain number of deaths and other harms. In restricting other activities, it is preventing a certain number of deaths and other harms that would have resulted from that activity, but also *causing* a certain number of deaths and harms among those who are restricted. The common asymmetry between causing and allowing slips away. I think in these cases the state is more or less in the position of someone who wants to turn the trolley away from a large group and has to choose whom to turn it on: it will be a killing any way it acts. Simply by taking the wheel, it assumed that responsibility. Of course, the state can fail to take the wheel at all. It can simply not step up to a problem and merely allow bad things to happen. Unlike when a private person fails to act, however, this is not a morally better/safer position to be in (nor is it always better for private individuals… but sometimes it is); it is an even worse failure because the raison d’être of the state is to address the needs of the people under its jurisdiction.

    So much for the relevant deontological category. Now, back to whether Covid-19 is more like driving or polluting. I think it has to be more like the latter. It is not completely that way, as one sneeze from an infected person onto an uninfected person is like a driver hitting a pedestrian. But much of the pattern of transmission and the ultimate risks to people seems to turn on group behavior: are crowds gathering? Also, if other people are being safe, one can free ride on their good behavior to act in a more or less risk free way, as with pollution. So, the ethical frame for thinking about regulation should be more like that of enforcing regulation that all have to obey to create a decent balance of avoiding death by infection and avoiding harm by restriction.

    Last two thoughts of the morning (I see there’s a message from Patrick, but I’ll come back to that). First, when thinking in these group based ways–rather than about what John might do to Mary–it becomes clear, I think, that we should be using an ex post framework with aggregation, not an an ex ante one of individual risk imposition or risk exposure.

    Second, it’s also important think about the right sized units for regulation. The politics of Covid-19, at least in the U.S., is now thoroughly antagonistic In Part because the regulations did not pay sufficient attention to right-sizing. New York City needed to be on lockdown. Upstate NY did not. What it needed was to quarantine people from NYC, admitting people in only if they were willing to spend two weeks in isolation. Unfortunately, at least in the U.S., the local, state, and national level responses were not coordinated and correctly designed, and now we not only have a simmering ongoing infection problem, but have made the response to regulations itself into a politically divisive matter 🙁

  27. Sophia wrote: “Proposals to impose mandatory lockdowns on the elderly for long periods of time raise questions about parity of treatment and what it takes to justify paternalism. And there are questions that nobody has been asking about what society owes to those with disabilities who normally live in the kind of isolation that many of us have now experienced for the first time.”

    This comment is addressed to Sophia in particular. I think your remarks in the passage I excerpted are vital. I think it is important to avoid naturalizing the vulnerability of seniors and elders in discussions about COVID-19. (see this: Approximately 80% of cases in our province (Ontario) have been in nursing homes. I have not seen any data that indicate that seniors and elders *in the community* here (Ontario) are testing positive at a disproportionate rate. Community spread here has been low, with the exception of Toronto and surrounding areas. In other words, the relevant factor here is the institutional setting. In the U.S., the rate of infection in nursing homes is more than 60% (community spread has been higher across the U.S., presumably because of the ineffectual COVID 19 policies there), suggesting that in the U.S., too, the institutional setting is the relevant factor, not the age of the population. In short, I think the question of whether nursing homes should be reopened should not be conflated with the question of whether seniors and elders (in general) should be allowed to do what everyone else is allowed to do. Nursing homes are the problem; they must be abolished.

    Societies and governments must develop new living arrangements in the community. The seniors, elders, and younger disabled people who live in nursing homes and similar congregate settings at present are isolated (as you note in regard to the latter). Indeed, many, if not most, disabled people live relatively isolated lives due to widespread inaccessibility, poverty, ableism, and social injustice. I am glad that you drew attention to our isolation.

    Although philosophers in general have largely ignored issues about the ableism, oppression, and injustice that disabled people confront, it is nevertheless not true that “nobody” has been asking questions about what we are owed. A growing number of disabled philosophers work in philosophy and theory of disability, asking these sort of questions, among others. For too long, philosophers have medicalized questions about social justice for disabled people, sequestering them in the subfield of bioethics. Indeed, the very existence of bioethics depends upon our medicalization and pathologization. I would like to invite you to visit BIOPOLITICAL PHILOSOPHY, the philosophy blog that I coordinate. I am a disabled philosopher of disability and thus the question of social responsibility to disabled people and disability and ableism more generally are central to the writing that I do on the blog. The link that I provided above takes you to the blog.

    I think that philosophers should ask themselves how they contribute to the continued social isolation of disabled people (social isolation that preexists COVID-19) and the continued refusal to address issues pertinent to us. How many philosophers reading this post have disabled colleagues in their departments? How many philosophers reading this post think that questions of social justice for disabled people are questions that bioethicists are best suited to address? It is a mainstay of mainstream ethicists to naturalize the social disadvantages that disabled people confront. It is a mainstream of analytic philosophy to circumscribe such questions about power and oppression, to miscontrue power as fundamentally negative rather than primarily productive.

    If you, Sophia, or anyone else reading this comment is interested in reading a more comprehensive treatment of these arguments, please take a look at my book Foucault and Feminist Philosophy of Disability or some of my other work.

  28. Hi Kerah,

    Great response.

    As I said earlier, I’m not at all sure about this case. Let’s start with a variation where I don’t know the identities of anyone affected, but I do know what my rescuing situation will be generically.

    KNOWLEDGE. I know that a person will be drowning in a pond tomorrow, on Thursday, and I will be able to save her at little cost to myself. And I know that I will only be able to rescue one person. I could put up a barrier today, on Wednesday, to prevent another person from falling in at no cost, but I fail to do that, so tomorrow the second person falls in.

    Do you now agree that there is not much to favour my rescuing the second person? Let’s suppose that there isn’t much.

    Now consider whether my ignorance about my ability to rescue more than one person makes a difference. You suggest, I think, that it does because my failing to put up a barrier in KNOWLEDGE is negligent with respect to the first person, as well as the second person, in that it creates the conditions under which the first person, whoever that is, has a greater risk of not being saved.

    I’m not sure. I can see that this might have some force, but I tend to think this doesn’t have much force. The question is why the first person should be deprived of a chance of being saved because I wasn’t negligent about the possibility that I would face a conflict about whom to save. After all, had I had the choice of putting up a barrier to save the second person on Wednesday, or rescuing the first person on Thursday, there would have been nothing to choose between the two. So why does my violation of the duty to put up the barrier make a difference, so that I must now pick the person who I failed to protect on Wednesday? Why shouldn’t I just treat this situation as similar to the situation where I can only rescue one of the two, and my negligence is not involved in creating a dilemma?

    Does that depend on whether I would have been required to save the first person, or whether doing so would have been supererogatory, without the presence of the second? I don’t think so. For if I am required to save the second person, the second person’s presence makes my rescuing the first person wrong. In other contexts, if I can either save one person at no cost, or save another at high cost to myself, I am permitted to save the second, even though saving the second alone would be supererogatory. So the supererogatory case is, I think, no different from the duty case.

    A final thought. It’s worth noting that the COVID situation seems more like KNOWLEDGE, in that the UK government, for example, knows that there are plenty of people in need of help, and then the UK government would have no reason to prioritise people who become vulnerable through its own negligence.

  29. To add to Sophia’s response re: the “injustice problem” versus the “lockdown problem” –

    By characterizing the question of how to reopen as a question of justice, I (and I think Sophia) mean to say that how to reopen is a public policy question that implicates our democratic public values—of antidiscrimination, of equal membership, and so forth—and that those values should guide how we reopen, and not merely operate ex post to correct inequalities produced by the disparate impact reopening policies may have on people of color, women, and other less privileged members of society. Whatever our reopening policies should be, they should be justifiable to those least advantaged members and compatible with treating them as equal members of society.

    The risk of discriminatory disparate impact that Sophia mentions is well illustrated by our current lockdown policies. We’ve been told to shelter in place unless we’re performing essential work. Women and people of color are overrepresented in essential fields like delivery and grocery work, nursing, and the like. So while many of us more privileged members of society are able to safely work from home, it is the working poor, as well as people of color and women, who go out and expose themselves every day so that we can stay safe. (I suspect that this fact of disparate impact has played a major role in generating the outrage that many people currently feel today.)

    That’s not to say that it was wrong to issue a shelter-in-place order. But rather, that even though the order was on its face neutral, it has operated such that many of the least advantaged members of society have born the heavy costs and physical risks of slowing the spread of COVID-19. And given our histories (and the current state of affairs) of racial and gender subordination, we should be really disturbed by this picture and how it reifies status inequality. This is a state of affairs that really requires a moral justification that references our equal status.

    And, going forward, it’s going to be hard to justify such disparate impacts if we do not take steps to avoid those impacts as we plan for reopening. Now that we know that the potential for disparate impact with COVID-19 policies is real and foreseeable, it would be negligent of us not to be guided by that risk in developing reopening policies. This is really not something we should wait correct for once we’ve already created yet another manifestation of social inequality. The “injustice problem” is thus not morally separable from the “lockdown problem.”

    So, like Sophia, I’d like to hear more about how our individual economic, psychological, and health interests, considered in abstraction from our membership in various protected groups, can be deployed at the outset to decide who stays home and who goes back to work in a way that either avoids such disparate impacts or can justify them.

  30. Hi Victor,

    Thanks for this challenging response. Now let me try to get my head around it (at least part of it).

    If I’m envisioning the case right, I live by a pond. I know that if I do nothing today, two people will fall in tomorrow. I will be able to rescue one cheaply, but unable to rescue the other. Consequently, the second person will drown. But if I erect a barrier today, also cheaply, then only one person will fall in tomorrow. I’ll still be able, and obligated, to rescue the one victim tomorrow, so on-one will drown.

    This case is different than the one I had in mind earlier, since today, you’re symmetrically situated with respect to both potential victims. The each have claims on you (i) to erect the barrier, and (ii) to rescue them tomorrow should they be in need of rescue. Since their *ex ante* claims on your are the symmetrical (shout out to Alec!), I find it plausible that, if they’re both drowning tomorrow, their claims on you tomorrow will be symmetrical, too. But I do think the lack of *ex ante* knowledge of which victim would have been saved by the barrier (on the description of the case, it might even be indeterminate) makes a big difference — a bit more tentatively, all the difference.

  31. Hi Sabine and Sophia (and Tom and Daniel),

    I want to chime in on how to justify the elevated risks of harms faced by essential workers (many whom are already the victims of substantial social injustices) to those workers. Part of the answer (per Sophia’s response to Tom) must be the terms under which they’re subjected to those risks. I’d add to what been said already that these terms can include not only the protections they’re granted (adequate PPE, etc.), but also forms of compensation. Examples of the latter might be guarantees of free health care for any COVID-related medical issues (including long-term health-effects), hazard pay, free childcare, fast-tracking to citizenship, and etc.

  32. Thanks, Sophia and Sabine, for these illuminating replies, though I fear we might be talking past each other a bit. (Perhaps my comment wasn’t clear enough.)

    I never meant to say that the problems of lockdown and justice are “separate.” No doubt they are linked in just the ways you have laid out. Nor did I have in mind the rather cruel idea, criticized by Sabine, that we should “wait to correct” the problem of injustice, leaving the poor in the lurch. (I’m no angel, but I’m nicer than that!)

    I was just hoping to bring out a point that is in tension with one of Sophia’s suggestions, though I expect you’d ultimately agree with it. The point was that policy choice is holistic, and sometimes we can’t evaluate individual policies – even ones with clearly disparate impacts – in isolation. Congestion taxes on busy streets disparately hurt the poor (who may be mostly non-white). But which city would you rather live in: one with no tax and ghastly traffic, or one with less traffic and a modest tax made up for by higher welfare payments? Which city would a poor black person rather live in?

    Analogously: would you prefer a city with 10k infections and a fair lockdown, or a city with 5k infections, a lockdown that burdens the poor, and a welfare system that corrects for the burdens? If you prefer the second city, that is in tension with the idea that, in choosing lockdown policy *specifically*, we should be giving extra weight to the claims of the poor.

    (Now, you might object that it’s unrealistic to suppose that we can simply “make up for” the burdens of lockdown with a change to the welfare system. Fair point – welfare systems are hard to change. This is what I was trying to acknowledge with my “Caveat 1.”)

    Sincere thanks again, Sophia, for the thought-provoking post, and for the pointer towards your book. Looking forward to reading your take on systemic discrimination.

    (Bonus kudos to Alec for poignant use of emoji.)

  33. Shelley, I am very grateful to you for commenting, and I want to thank you for helping to make a number of important issues visible on this blog.

    I agree that it is crucial for us not to assume that seniors or people with disabilities of any age are naturally vulnerable. Instead, we need to acknowledge and try to change current living arrangements and social institutions (and, I might add, our public transportation systems –or the absence of them) in order not to isolate those with disabilities and not to obscure the social causes of their situations.

    I also think you are right to suggest that we, as philosophers, have a responsibility to think about the political effects of our discussions. It is deeply troubling, as you note, that we so often treat the problems faced by people with disabilities as special problems that are to be addressed only either in a medical context or by specialized philosophers working in disability studies –rather than treating them as issues that all political philosophers have an obligation to grapple with, when they ask about what social policies are morally acceptable in a given situation. I think I was making a similar point when I suggested, in my last Comment, that social injustices are not separate, specialized problems, which we can put to one side. None of the injustices faced by marginalized groups in our society is a problem stemming from unique needs of that group, which should be addressed only by specialists in that field. They are problems caused by our social arrangements and our social institutions (I write about this at length in my book). We need to acknowledge that they are *our* problems, and work collectively to understand them and to try to change them.

    I am more optimistic, though, about the prospects of including these injustices in our current discussions about which policies to adopt as we exit our lockdowns. When I raised such questions initially, I didn’t mean just to raise them and leave it at that. I was hoping we could actually have a discussion about some of the injustices and about some of the ways of recognizing their moral salience in our evaluations of different policies. (The kind of discussion that Daniel has helpfully tried to initiate, which I will follow up on in a little). Perhaps, as a way of helping us along here, Shelley, you could raise some further questions for us which you think we should consider when evaluating various COVID-19 exit strategies. Which interests, in particular, should we be careful to bear in mind, in your view? Which sorts of policies may have a particularly unfair impact, and on which particular sub-groups of people with disabilities?

  34. Kerah and Daniel – Thanks for those helpful examples and clarifications.

    One small clarification: Daniel, in your earlier post you suggested that we should “solve the lockdown problem on its own terms, then account for the effects when solving the injustice problem on its own terms.” I agree that it would indeed be cruel to wait until the aftermath to worry about the injustice problem! And I understand that you are suggesting we plan to mitigate deleterious effects in advance. But even if we did not wait to turn to social justice until the aftermath of a given reopening policy, the point I understood Sophia to be making (and with which I agree) is that we still go wrong in our conceptualization of reopening if we think reopening is not in itself a matter of social justice. In particular, as a methodological matter, we shouldn’t start by asking what lockdown (or reopening) policy would be best under conditions in which we do not account for social injustice, and then, once we know the answer to that, turn to how the idealized lockdown policy would operate on the ground, only then designing a response for disparate effects. I worry that such a methodology would systematically privilege the already privileged. But your driving example suggests that perhaps we are all in agreement on this point!

    Daniel and Kerah, your thoughtful examples raise the really hard and interesting question of how much financial compensation can help either prevent or remedy potentially disparate impacts. Daniel, in your example, the alternative of the tax + redistributive policy seems attractive in part because the disparate impact is a financial one (it’s harder for the poor to pay the tax). But what should do when the burden is not largely economic, but tied to status, health, or emotional wellbeing? Kerah, you helpfully suggest hazard pay and a variety of other goods might ameliorate what might otherwise be unfair burdens on essential workers. This is attractive, I think, not just because what’s being provided is somehow an equivalent for what’s lost, but also because it communicates a genuine concern about the risks of the job and a willingness to share the costs somehow. (I suspect that how effective these responses are will also turn on what the alternatives were—whether a fairer way to distribute the burdens was available to us ex ante, for instance.)

  35. Daniel, thanks for your follow-up. I think we are actually in agreement. When I said that we needed to look at specific policies, I didn’t mean that we should or could look at them in isolation from other policies or without considering the background social regulations and systems of social support that the state has or does not have in place. I meant rather that we need to look not at “lockdowns” and “opening up” in general, but at particular *ways* of opening up, and the social conditions that form the backdrop to them.

    For instance, in my own province, businesses, marinas, golf courses, and massage parlours are currently permitted to be open, but schools and public playgrounds are all required to be closed. Some have called this “the opening up strategy for middle-aged, middle-class white men” –precisely because, given the background social conditions in which women are disproportionately tasked with child care, and racial minorities and younger people are disproportionately unable to afford yachts and golf equipment, this way of opening up benefits privileged white men at a certain stage in their lives, at the expense of others. My point was that different policies are going to benefit different groups and exacerbate different injustices, and so we need to look at the details of the different social policies, along with the background social context, in order to make moral judgments about their acceptability.

  36. Sabine and Sophia, it does sound like we’re in agreement! I’m going to think more about the question of incompensable harms, which is surely relevant here — thanks for this insight, Sabine.

    And thank you Sophia for that example. Now I see what kinds of disparate impact you’re concerned with; your point seems eminently reasonable.

    But I must say, it sounds like your province is allowing…one yacht too many.

  37. RESPONSE TO PATRICK ON EX POST REASONING WITH WAIVER. Patrick, this is a typically wonderful breakdown of a case. I’m grateful for the question as a way of testing my own ideas. I’ve got one main response, and then a few small ones.

    By way of reminder for the 2 people (at most) who will read this other than you and me Patrick, you’re challenge was primarily that since my view holds that ex post claims are involuntarily “waived” if the ex ante interest in enjoying it is small enough, and the interests of others in an alternative policy are strong enough, my view is going to have all sorts of people taking unexpected positions for an ex post view. For example, high risk people who won’t die/be seriously harmed if lockdown is eased might be expected to waive their claim to freedom from lockdown because of their ex ante expectation that it will not be a net benefit to them. As you wrote Patrick, “given the appeal to ex ante interests, this makes the position less of a pure ‘ex post’ position that it otherwise would be.” Conversely, those who are low risk but would be harmed if lockdown is eased “would (or rationally should) waive their ex post claims against death. This would weaken the case for lockdown, since thousands of deaths will be irrelevant to deciding whether to ease lockdown. This is surprising.”

    To this I say YES. And here’s what it shows, the reason the difference between ex ante and ex post model matters is that they have different implications when it comes to aggregation. Here’s how I put it in my paper (discussing a case I call Random Killing): “Angela has an option in which she exposes each of 100 people to a dependent 1% chance that they will be killed, such that, at the end of the day, one of them will die. The ex ante model says that each person has a claim not to be killed that is discounted to reflect his facing a 1% risk of suffering that fate. The ex post model says that one of those 100 people has an undiscounted claim not to be killed. These could come to the same thing, but weak aggregation suggests that they do not; it suggests that the sum of the 100 ex ante claims has, or at least might have, less weight than the single ex post claim. This difference can be applied to help sort out which is the right model to use.”

    So, let’s move that over to the Covid-19 case. What your categories show is that the ex post view is going to look a lot like the ex ante view. We’ll divide society into those who ex ante expect to benefit from a lockdown and those who expect to suffer from it. But by using the ex post model we can avoid the problem of limited or weak aggregation, where we discount people’s claims not to suffer a possible bad outcome by the odds and then aggregate. The danger in the ex ante model is that if we discount by both the odds, And keep that there’s an expected benefit many would enjoy (those in the high risk group also suffer under lockdown), then there’s a fear that we get into weak aggregation territory. Those with claims not to suffer death from ending lockdown may look like they lose too easily against those who want to regain their liberty (the benefits of which are more certain).

    That’s my main comment. I’m going to pass on the culpability point; I don’t see any problem in discounting ex post claims if they are based on culpable behavior. But I do want to reproduce your last two paragraphs, which my comments on them.

    First, you say: The group who will harm people arguably have ex post interests in being locked down. Insofar as we have interests in avoiding harming others, or perhaps interests in wrongly harming others, these people have interests in being told to stay put. Certainly, I would be grateful to know that someone had withdrawn my liberty for a bit if it meant I wasn’t causally central to the death of another person (imagine someone took control of my car to stop me hitting a pedestrian). This would make the case for lockdown stronger than it would otherwise be.”

    I respond: Yes, I suppose that’s the reasonable thing to say—they should prefer to be locked down. If there were enough people who would play this role in a sine qua non sense, that would surely favor continued lockdown.

    Then you say: “Group 5. Low risk and won’t harm others if lockdown is eased.
    This group has interests in being allowed out and about, and these interests are not extinguished by or to be balanced against interests in not seriously harming others. But these people have evidence-relative/ex ante reasons to avoid harming others (certainly other-regarding, and potentially self-regarding reasons, as explored above). Given that those who are at high risk but won’t actually be harmed are taken to have waived their claims to being out and about, by appeal to the ex ante situation, what should we say of the low risk but won’t harm group? Wouldn’t/shouldn’t they also waive their ex post claims to be out and about? Alec’s view allows ex ante interests to cause us to waive ex post rights for self-regarding reasons, but what about for other-regarding reasons? If so, this would make the case for lockdown stronger than it otherwise would be.”

    I say: If the sacrifice for them were low enough, they should waive their claim to freedom. But if the sacrifice for them is high, I think I should say that they need not waive it for the sake of those who will die. They can trust the overall balance to take into account the interests of those who will die; otherwise, we’ll get a double counting that would skew the decision.

    I hope these comments make sense Patrick (or anyone else who has read this far). If not, let me know.

  38. TO VICTOR, RE YOUR RESPONSE TO CHRISTIAN, PART I: You write: “I think that it’s pretty awful to be a serious wrongdoer, I think that these reasons are strong. But I don’t think that we have very strong reasons, owed to wrongdoers, to prevent their attempts from being successful. Suppose that I wrongly threaten an innocent person, but the harm they will suffer is overdetermined. Do others have strong reasons, grounded in my interests in not being a completer rather than an attempter, to prevent the threat that I have caused from being realised, so that the person is harmed by a naturally occurring threat? I don’t feel the force of that, or at least not much.”

    I agree, but I wonder what’s doing the work here. Here’s a background scenario: I know you have a horrible jealous streak, and you would form the intention to kill your wife if you find her with another man, and as it turns out, she has been with another man. As it turns out, he is a bad man who will soon kill her, and there’s nothing I can do to prevent that. But I could interfere with you doing it just a bit sooner in the following 3 ways: (1) I can hide the evidence of this fact so that you never form the intention to kill her. (2) You have found out and have formed the intention to kill her, but I can drug you and keep you from attempting to kill her until she is dead of other means. (3) I could let you get right up to the verge of killing her and then intercede, at which point you will have made a failed attempt to kill her. I take it that you think there’s not much reason to do (3) rather than just let you kill her, given that she would soon die anyway at the hands of her lover (put aside her special horror at seeing you come for her; let’s suppose you’d kill her in her sleep, and he lover would do the same an hour or so later if I didn’t stop you). I take it that you think there’s not much reason to interfere at stage 3 because you’re already as culpable as you’re going to get, the effect doesn’t matter. But what about stages (1) and (2) where your culpability has yet to be fully developed. Do I have strong reason to stop you there in those ways. How does that weigh against anti-paternalism?

  39. Thanks for the really interesting reply Alec.

    I think this question you ask is really important, “How do we think about what the state is doing, then, when it permits certain forms of risk-imposing behavior? Is it merely allowing those who suffer to die? I think we need to be careful of an equivocation between the moral and the causal sense of allow.”

    I agree that, in permitting a certain level of activity, the state becomes responsible, morally and causally (at least on any plausible account of causation) for a certain number of deaths.

    But I disagree that this obviates the importance of deontological distinctions such as that between doing and allowing harm when it comes to evaluating the conduct of the state when it considers measures like lockdown restrictions. The reason is that there are different ways in which the agency of the state can be involved in the causation of harms, and these differences may be morally quite important. There is a difference, for example, between the agency of the state being involved in the wrongful killing of an innocent person, and a state’s having failed to adopt policies that prevent the wrongful killing of an innocent people by other ordinary citizens. The state should, to be sure, try its best to ensure that people are protected from wrongful harm, but when the agents of the state do harm to others, this constitutes a particularly serious injustice. That this is widely held to be so is clearly in evidence in the justifies rage over police killings and torture, and it seems also to underlie the principle of the so-called ‘golden thread’ of the legal system (which you and Patrick have both written interestingly about) that tolerates a higher incidence exonerating those guilty of serious crimes to avoid wrongful convictions of innocents. So I do think there is at least a kind of analogue with the individual case—the “safer” position for the state is to ensure that its agents refrain from doing wrongful harm to others, even when this may involve sacrifice in its ability to protect people overall from suffering wrongful harm. This does not mean that the state can “simply step away from the problem and allow bad things to happen”, but it does mean that there are limits to the risks that it can take that its agency will be directly involved in doing wrongful harm that it should respect in trying to prevent other bad things from happening.

    So far I’ve been using paradigmatic examples of the state’s being actively involved in doing harm to illustrate the general point. But clearly the state’s agency is not going to be like that in upholding ‘lockdown’ restrictions. This leads to the interesting question of the nature and significance of its involvement. In my view, if the state coercively prevents people from engaging in activities that are essential to their well-being or escape conditions that will lead to their abuse, that is quite a serious form of agential involvement, different from failing to prevent some from imposing risks on others. That is not to say that its duties to prevent some from imposing risks on others are not very stringent, or that they cannot override the states’ reasons to refrain from various forms of harmful coercion, just as saying that we have especially stringent duties to refrain from doing harm does not mean that we cannot sometimes do harm when it is necessary to prevent a great deal of other harm from occurring.

  40. Thanks Kerah and Victor, your exchange is really making me think!

    On your negligence case, Victor, you write:
    “I am not completely sure about how to analyse this case, but I think that I don’t have a strong reason to prioritise X. I think that I owe a more stringent obligation to save a person than a bystander would. But if I save X, my injustice will still result in my having failed to save Y, and so I then it is less obvious that X has a significantly stronger claim on me than Y.”

    This kind of case is fascinating, but I don’t think that we should infer too quickly from the fact that you don’t have a strong reason to prioritise X that X lacks a stronger claim on you than Y. To see why consider what seem to me some differences between your duties to X and Y. The first is that I think you are required to take on more cost to assist X than to assist Y. If X’s plight is due to your negligence, you can be required to take on a good bit of cost, relative to what is at stake for X, to assist her. I don’t think this is true of your reasons to assist Y. At the very least, there will be a point at which the cost that you must bear to assist X will still be required but will make it be supererogatory to assist Y. This also gets reflected in enforceability: we can impose more cost on You as a necessary side effect of saving X than we can to save Y. The third difference concerns a moral remainder. If you save X in this case, then after that occurs Y will not have any unresolved claim against you. But that is not true if you assist Y instead. In that case, X can still rightly claim compensation from you, since your negligence cause a harm they suffered.

    I confess it is puzzling that these different normative characteristics of our duties to assist seem to come apart in these cases, but I think it’s genuinely a puzzle rather than an indication that our duties to those we have negligently put at risk are no different from those that we can protect from risks that have been caused without our agency.

  41. Christian,

    I agree with everything that you say in substance – great points. But I don’t think that they are best understood as establishing a greater claim that X has that I rescue her on Thursday. Here’s how I’m inclined to understand what is going on. Suppose that it would be more costly for me to rescue X than Y on Thursday. Normally, this would be sufficient reason for me to rescue Y, and I would not wrong X if I did so on that basis. But because the extra costs arose due to my negligence on Wednesday, it is harder for me to justify appealing to them on Thursday as reasons to be permitted to rescue Y. So I must flip a coin. That is not because X has a greater claim on me to be rescued – the claim is just to be saved from drowning, which is identical to Y’s claim – but rather that certain considerations that would normally legitimately count against my rescuing X have been disabled from playing this role by my negligence.

    Similarly, with respect to compensation. Suppose that I flip a coin on Thursday and rescue Y. And suppose that this is not unfair, as you seem to accept. You are right that X (or X’s family) is now entitled to compensation, and that would not have been true of Y. But what is X entitled to compensation for? Not for my infringing her right to be rescued on Thursday, but for my negligently failing to rescue her by putting up the railing on Wednesday. X’s claim, here, is no different I think to the claim that she would have were it impossible for me to rescue her on Thursday.

    So I think that X has no greater claim than Y to be rescued on Thursday, although you are right that X has various other claims that Y does not have due to my negligence on Wednesday.

  42. TO CHRISTIAN ON THE STATE’S AGENCY: I think we’re fundamentally in agreement. I agree with you that “There is a difference … between the agency of the state being involved in the wrongful killing of an innocent person, and a state’s having failed to adopt policies that prevent the wrongful killing of an innocent people by other ordinary citizens.” I meant to capture that when I said, “Of course, the state can fail to take the wheel at all. It can simply not step up to a problem and merely allow bad things to happen.” My main point was that in case like this, the better view is that once having taken the wheel, which it should have done, the state will be responsible for killings/harming whatever it does. By continuing a lockdown (or lesser limitations on freedom), it harms those whose liberties are restricted. But by RELEASING a lockdown, it unleashes a threat to others. That’s not a mere allowing or enabling, it’s a causing. Compare blowing up a dam and releasing flood waters on a town.

    Of course, removing an obstacle can count as either allowing or causing. But I think that’s a moralized judgment. If the little Dutch boy removes his finger from the dyke, he doesn’t kill the folks in the town below, he allows them to die. It was his finger and they had no right to demand he leave it in place. But if someone else comes and removes a key rock from the dyke and it fails, they *cause* the flooding of the town below. I think the state has no claim to simply stop a lockdown in this situation; it’s not like it would lift the lockdown because it has run out of resources (the analog of the Dutch boy withdrawing his finger); it lifts the lockdown for the sake of those who are harmed by the lockdown. That’s why I suggest that the right analog is someone who has decided to turn a trolley away from many and now has to decide what to do with it. Either way, it’s a killing.

    OK, that’s my best defense of what I said above. But I also confess that there’s something appealing about your point, Christian, that ending the lockdown feels like an allowing. It’s allowing others to do things that constitute killing/harming. If that’s right, the interesting Q is: how does the back and forth of claims work?

    On the one hand, the claims of potential victims on those who would harm them is the strong negative claim not to be killed (Helen’s point). On the other hand, their claims on the state could be seen as claims not to be *allowed* to suffer killing at the hands of others. Are those claims stronger than claims not to be allowed to die of natural causes? Maybe, but not by much. If I had to choose whether to rescue A from murder or B from a falling rock, I think I’d choose A. But if I had to choose whether to try to rescue A from murder with a 90% chance of success, or to try to rescue B from death by falling rock with a 95% chance of success, and the cost to me is the same either way, I think I should choose to try to rescue B.

    One might also try to rescue the idea that it is more important to prevent deaths at the hands of others than to prevent deaths by natural causes because the former involve the claims of potential wrongdoers not to put in a position that they at least *should* regret. But there are two problems with this line of thought. First, it’s not clear that many of them would be acting wrongly, even if they cause death. If they have the right to move about as they do the deaths that result may be regrettable side-effects, but not a sign of wrongful action. Of course, some will regret what they’ve done if they come to learn that they’ve caused the death of others. But it would be unjustifiably paternalistic to prevent them from exercising their rightful liberty because they might come to regret the results. The idea that they have the right to exercise that liberty presumably implies that the overall, ex ante judgment about the balance of effects is that it’s better to be free and run the risk of harming others than to be constrained. Second, even if some of them would behave wrongly, and would not have done so if kept under lockdown, I think Victor is correct: that’s not a reason to choose a policy that makes others worse off. It’s their problem.

    In conclusion, insofar as you are right that the right deontic framework is that the state has to choose between harming some by locking them down and *allowing* harm to others by lifting or not-imposing a lockdown, then the thing that is harder to justify is the harm that results from the locking down.

    One last point on this: there may be an important asymmetry between not locking down in the first place and lifting a lockdown. A lockdown, once in place, may form a new baseline from which lifting it would count as doing harm, whereas if it’s never in place, then failing to impose it is *merely* allowing harm.

    Now, I want to add a note to two to other things you said: First, you wrote: “The state should, to be sure, try its best to ensure that people are protected from wrongful harm, but when the agents of the state do harm to others, this constitutes a particularly serious injustice. That this is widely held to be so is clearly in evidence in the justifies rage over police killings and torture.” ABSOLUTELY RIGHT, and so important. But I think it also matters not just that it is the state ACTING, but also that it is the STATE acting. That is, I think it’s not *just* the doing-allowing/+ vs – claims distinction that explains what’s so awful about the state acting wrongly. It’s that the state is meant to be the servant of the people (to quote Zelensky). It’s not just a bad guy when it behaves badly. It’s abusing its power, power that it has a right to only for limited purposes. That abuse is so dangerous and offensive that I think we ALSO seriously resent state INACTION if it’s based on corruption: e.g., standing by and watching private linch-mobs do their thing.

    You also wrote: “it seems also to underlie the principle of the so-called ‘golden thread’ of the legal system (which you and Patrick have both written interestingly about) that tolerates a higher incidence exonerating those guilty of serious crimes to avoid wrongful convictions of innocents.”

    Actually, my position on that is that it CANNOT be explained (to any substantial degree) by the difference between + and – claims. See my “Proof Beyond a Reasonable Doubt,” pp. 390-395.

    FINALLY, Thanks Again Christian… this has been a really thought provoking exchange, and I hope to go back to it in my writing soon. I hope it’s also of value to you and anyone else with the patience to have read this far (ha ha).

  43. I have just one brief comment for now. I am puzzled by Sophia’s insistence that race and gender are important considerations. It seems to me sufficient to note that many of our so-called “essential workers” take the brunt of what opening up entails, especially since most of these workers cannot “work at home” and work under conditions where close proximity to others is a relative necessity. As such, they are more susceptible to transmission, as recent outbreaks in meat processing facilities has demonstrated. It doesn’t matter what their gender or race is. They are still more vulnerable and that should somehow be taken into account when considering “opening up” the economy. What this demonstrates is that opening up is not a homogeneous affair: it means different things to different people and employments. What needs too to be discussed is what “opening up” entails. Can it be done piecemeal, in sectors, or must it be attempted more or less in toto? When we speak of “essential workers” it sounds like these workers are not part of the “opening up.” They, being essential, are already at work in some way. “Opening up” for them might mean working more often, or with greater density of workers. Another aspect of this issue that I don’t think was raised was that of how we ought to fund the stalled economy and the startup. How we decide to do so is a normative issue. I suggest two broad ways and encourage comments. One is by borrowing and the other is by taxation. Borrowing lays a burden on future generations. Taxation can be considered a means of distributing limited and reduced wealth among the population as a ,means of sharing in our impoverished conditions.

  44. Bill, I am puzzled as to how you could think that race and gender are not relevant. You say that what matters, in the case of vulnerable workers, is simply their vulnerability. But the reasons why visible minorities and women disproportionately belong to the most vulnerable parts of our essential work force have to do with systematic racism and gender-based discrimination in American society, which have deprived many of them of the chance to be employed at anything else. As I suggested initially, we might think that when some communities are already disadvantaged as a result of prior injustices, they now have a special claim on society –perhaps a claim for a greater share of its resources during an emergency, or perhaps a claim for certain special sorts of consideration or prioritizing of their interests over those of others.

    But the reason it is crucial for us to consider the impact of different possible exit strategies on racial minorities and on women isn’t only because these groups are disproportionately represented among essential workers. It is also because, as I explained in one of my Comments, and as Sabine also discussed, the background racial and gender inequalities in American society mean that many of the apparently neutral policies that we adopt either during our lockdowns or during opening up risk perpetuating these longstanding injustices against these groups, in ways that exacerbate –and arguably also constitute– ongoing discrimination. Because many rules look neutral, it’s easy to miss these racial or gendered consequences. For instance, a rule that closes schools and daycares for months in order to limit the spread of Covid-19 seems perfectly neutral on its face. But when children and adults are spending their time at home without access to the other individuals –teachers, counsellors, extended family members– who normally share the work of raising a family and supporting its members emotionally, this has a greater impact on women as a group than on men, because women are already burdened with a disproportionate share of childcare and what is sometimes called the “emotional load” of raising a family. To take another example, when rapid testing for Covid-19 is not available because governments have not prioritized it, those who live in crowded, impoverished neighbourhoods and cannot self-isolate suffer most –and such neighbourhoods are disproportionately occupied by racial minorities. And when different hospitals have an equally limited set of resources, it is the hospitals in the most impoverished neighbourhoods which tend to fare worst, unless resources are redistributed to them. I am sure you have also read that African-American and Latino communities have higher rates of the kinds of underlying medical conditions that increase one’s risk of dying from Covid-19 –such as asthma, diabetes and heart disease. This isn’t genetic. It is because of the conditions in which they live and the stresses that they live with, many of which are caused by racial inequalities and by racism.

    So decisions about what policies we enact in light of Covid-19 –which resources we choose to make available to which social groups and which neighbourhoods, which institutions we choose to open up and which we allow to stay closed—have the potential to contribute to structural and systemic injustices in our society. That is why Gretchen Whitmer, Governor of Michigan –where African-Americans have suffered 40% of all deaths from Covid-19 yet make up 14% of the population — wrote in the New York Times a few days ago that “The coronavirus is a civil rights battle too.”

    I don’t mean to suggest that there isn’t room here for a lot of philosophical disagreement. Of course there is. There are many philosophical questions we can and indeed need to ask about what exactly the moral significance is of the fact that these groups stand to face further discrimination if we adopt certain policies rather than others. I don’t have an easy answer, for instance, to the question of how we weigh one person’s interest in not being discriminated against in relation to another person’s interest in prolonging their life. Nor am I sure whether the claims of members of these groups are best thought of as individual claims or as claims brought by particular communities, on the basis of group wrongs. But these are the questions we need to be asking –not questions about whether racial inequalities and gender inequalities are relevant, but about how they are relevant.

Comments are closed.