I think what you’re picking up on here is that you need an account of appropriate equitable treatment in order to determine what discrimination is unjust. This is why, for example, people do not object to depriving people below the age of 18 of a bunch of legal rights but would object if the age in question was 50. In fact, most legal formulations of the disparate impact standard effectively state that there is an exception for when something is a good idea. Indeed, they generally amount to saying that disadvantage a group beyond what would be a good idea is discriminatory, which of course makes total sense on your view. Criticism of things like QALYs implicitly rely on the different and controversial and contestable conception of equality where it is considered equal treatment to give you the same medical resources, even if it will only extend your life by a few days, whereas it will extend somebody else’s life by much longer. Obviously, you can and should contest whether this is in fact, equitable on moral reflection.
To be clear, I am not suggesting that you cannot determine what is unjust discrimination without a full fledged theory of justice, as you can definitely rely on intuition to some extent, but often in the absence of additional reflection, you will notice that your intuitions on particular cases are either self-contradictory or sensitive to framing. So you need additional theorising, and of course you have to balance it against the intuitive appeal of certain general principles. So you cannot use particular judgements to shortcut the need for a broad investigation of what is just and also in many of the particular cases you highlight while I’m sure there are many people who genuinely found them counterintuitive. However, I also think many of the people objecting to such cases are just picking up on a family resemblance between these situations and situations that they find objectionable instead of actually having a direct intuition about the case.. ultimately like with the full of moral philosophy, there is no silver bullet that avoids all conflicting intuitions and you have to balance your intuitions and pick the bullets you’re willing to bite while acknowledging that your answers may depend on hard to obtain empirical information.
It occurs to me that in terms of triaging on the basis of social value, a flaw would be that this might encourage people to lie and deceive in ways that are socially inefficient (and would invalidate the utility of the practice). I can almost see it in my mind "We can only save one of these two patients, according to their charts this one has 5 Nobel prizes, is gifted neurosurgeon and runs a charity that feeds the hungry around the world, this one has 10 Nobel prizes and is a cardiac specialist at the Mayo clinic and has developed 10 highly deployed life saving vaccines for communicable diseases common in the developing world. Amazing how many Nobel laureates we see come through here."
If one limits judgements of social utility to cases were it would be easy to gather reliable evidence (hard to fake credentials etc.), I suspect the number of situations where it could matter would be reduced significantly, if not entirely.
In terms of the prioritizing treating health care workers, doesn't that more follow the logic of putting your own oxygen mask first in the plane emergency before helping your neighbour put on their mask? Healthcare workers who are sick (enough to benefit from treatment) are not providing care, usually social utility is indirect relative to the crisis in question, should you save the engineer now versus the ditch digger, but neither's utility is directed at an immediate current crisis but at a remove. Whereas treating healthcare workers is about treating the people who are currently under a crisis of manpower to deal with the very disaster that has laid them low. Also, there is the morale consideration, if you know the person you just worked with his waiting in the hall on a gurney because they are low priority it probably erodes motivation, leads to more burnout etc. Also, also assessing whether the hospital worker in front of you who you worked with his in fac a healthcare workers is an easier task of discrimination and harder to fake ("I only faked working with you for the last 20 years by coming in an actually doing the assigned work so I could secure this hospital bed you fool!"). A similar issue would be giving special treatment to the ill pilot landing the plane during a food poisoning incident at 40 000 feet (why did he have the fish?) this is not so much social utility as immediate necessary load baring utility.
As someone severely "disabled", in practical ways, by multiple sclerosis medical interventions I find that your argument isn't how the real world works. First off, I was extremely healthy and greatful for majority of life. I'm well educated, worked hard, contributed quite a bit to "society". Volunteered, donated money, etc. Live in a state that hates the disabled, provides no help or follows laws even. However loves medicaid recipients that get huge amounts to remodel homes. Get modified cars, free health care even ultra expensive interventions that don't help them. But provide big dollars and control to an entire slate of people. But we who pay our own way can't even have useable restrooms.
So it may be easy to put it in the term of people who've never had to deal with disability, but is that reality? Or a nice sounding doctrine that is impossible to honestly quantify because of so very many variables?
I'd likely agree that states should (in many cases) do more to accommodate the disabled (though I'd need to know more empirical details to have much confidence either way). If we listed all possible interventions in priority ordering, governments spend a LOT of money on things that I expect should be ranked lower than disability accommodations. So I think a broadly optimizing perspective can support a view in this vicinity.
The main practical verdict I'm defending in this post is just that it's reasonable (and not unjust discrimination) for a philanthropist to prioritize life-saving donations to help the global poor over donating to help Americans with disabilities. You can agree with this even if you think that states should generally be doing more for their disabled residents.
1) I agree that welfarist considerations can probably explain why it is wrong, collectively, for employers to engage in rational discrimination and why we ought to outlaw this [though I think the case for this is stronger if we hold that certain types of hedonic "benefits" to discriminating groups do not count as boosts to their welfare]. Unless we're doing some kind of rule-consequentialism, it's not as obvious to me that welfarist considerations can explain the injustice of individual employers engaging in rational discrimination on their own. It's not hard to imagine an employer, especially prior to the enactment of anti-discrimination laws, who faces a choice between e.g. (i) hiring a woman and losing a large amount--more than would be gained by the woman--due to angry customers and (ii) refusing to hire the woman to appease the customers and making her worse off, but keeping the business of the customers. I think many people would call this a case of unjust discrimination even if the economic or other admissible benefits of sex discrimination substantially outweigh the costs.
2) Just want to flag that the term "disparate impact" is also a legal term whose definition is much narrower than in this discussion [and is actually closer to your own view?]. At least in the U.S. a successful defense to a charge of disparate impact is for the defendant to show that it was justified in making the choices that it did. If a theme part operates on weekends and has a rule that all employees must be available to work Saturdays, there is no disparate impact liability on employees who are Jewish or Seventh-Day Adventists since the rule is reasonable. The fact that an agent disproportionately harms members of a protected class doesn't mean that there's a disparate impact in a legal sense. I know that you're not talking about the law--just found the use of the term a bit off-putting
re: 1, I think the force of calling it "unjust discrimination" is best understood as suggesting that there ought to be a universal law against it. At least, it would seem pretty unreasonable to demand that the only non-sexist business owner in the state must shoot themselves in the foot by making a hire that their social context guarantees will turn out to have negative value for their business! What's really needed there is a systemic solution (and welfarist considerations can explain this important fact).
I was mostly pointing out that to the extent your account is meant to capture what people think counts as unjust discrimination, I'm skeptical the everyday notion is necessarily captured just by welfarist considerations. There are definitely some philosophers who defend harm-based accounts of what makes discrimination wrong [e.g. Lippert-Rasmusson] but my sense is that it's a minority view, though the discrimination literature is pretty enormous these days.
It seems like “unjust” is supposed to be tied to some notion of desert.
For example, imagine Bob Billionaire and Joe Janitor and Joe Janitor’s wife Wendy. Bob Billionaire is an investor who was born with a keen eye for value but other than that is a terrible person who abuses his family and does not give to charity. Joe Janitor and wife Wendy are a lovely churchgoing couple who are kind and decent. If Bob Billionaire is alone swimming east and the janitor couple are swimming west, I suppose you have to save Bob Billionaire (the taxes alone on his billions are worth several VSLs per year). But can we at least call this is unjust? He doesn’t deserve it!
A reasonable response might be to say that saving the janitor couple would be unjustly discriminating against the unknown interests who benefit from Bob Billionaire’s investing skill.
But I can press you further and say that Bob Billionaire, as a selfish billionaire, lives a much better life than Joe Janitor. So even ignoring the interests of anyone else, if you had to choose between saving Bob Billionaire or Joe Janitor, even flipping a coin would be “unjustly discriminating against” Bob Billionaire and his greater interests. It seems like we are perverting the definition of justice at this stage.
I’m not necessarily claiming that one’s interests should be weighted by desert or other Rawlsian notions of fairness. I’m just not sure an equal weighting should be called “just”.
Yeah, in the instrumental case you're doing it for the sake of the downstream beneficiaries, not for Bob himself.
But I like your next challenge. It does intuitively seem very unfair to treat Bob's advantages in life as a reason to advantage him over Joe *even more* (by preferentially saving him). I wonder how best to systematize this intuition without getting absurd results like flipping a coin between saving children vs 90 year olds, or humans vs chickens. Maybe we bracket certain *kinds* of (socially contingent) advantages, while treating biological differences like age and species as legitimate to consider? Would it make a difference if Bob's wealth were truly deserved, so the idea is just to bracket *undeserved* advantages? I'm not really sure what to think about the case.
Worth flagging that I still don't think any concern along these lines would justify prioritizing disabled Americans over the global poor, at least. But it may be that a slightly different systematic account would be a better way to secure this verdict. Thanks for the comment!
I think, in your example, part of the intuition is that bad people who treat their family badly deserve some punishment, which could, for example, take the form of giving their welfare less importance. However, if Bob was a good person, your second challenge becomes a lot easier to resolve. After all, I think you would agree that reducing BOB to the status of a Janitor if he was a good person with no benefit to anyone else would hurt him, but if so being alive and billionaire is obviously better than being alive and a Janitor and so saving life for Janitor is actually doing him a smaller favour than saving a life of a billionaire in terms of how much you benefit them. It’s obvious from transitivity, since if being a billionaire is better than being a Janitor, you cannot treat them as equally valuable states of being when comparing them to the alternative of being dead. In fact, if you think about it, selfishly, a Janitor, who is not happy with his life might potentially be willing to take a risk of death with greater willingness compared to if that same janitor became a billionaire and was happier with their life because the cost of dying increases with your wealth.
Timely post for me, as I've just been to a talk about DCEA (distributional cost-effectiveness analysis) in health economics and I'm trying to work out what I think about it all.
I share your view that inequality does not ground a moral objection to optimisation. I also agree that in an ideal world we'd include externalities associated with targeting specific groups in our optimisations, but that this is very hard in practice.
Wondering what you think about proposals to incorporate population-elicited inequality aversion parameters into social welfare functions?
I'd need to hear more details, but my gut reaction is to be dubious of the idea, since I don't think inequality has intrinsic significance. (Could be fine to consider as a heuristic in many cases though.)
Yeah, I wonder if it could be a reasonable heuristic to capture intuitions about instrumental costs of inequality. These parameters are actually notoriously hard to elicit though (vary a lot between people and studies) so I'm not sure it's a great solution.
Looks like a discussion of what is more important, the action itself or its result. In other words, what ethics should guide our lives, deontological or teleological?
The question I'm curious to ask, considering having unjust discrimination, is it possible to have a just one?
I think what you’re picking up on here is that you need an account of appropriate equitable treatment in order to determine what discrimination is unjust. This is why, for example, people do not object to depriving people below the age of 18 of a bunch of legal rights but would object if the age in question was 50. In fact, most legal formulations of the disparate impact standard effectively state that there is an exception for when something is a good idea. Indeed, they generally amount to saying that disadvantage a group beyond what would be a good idea is discriminatory, which of course makes total sense on your view. Criticism of things like QALYs implicitly rely on the different and controversial and contestable conception of equality where it is considered equal treatment to give you the same medical resources, even if it will only extend your life by a few days, whereas it will extend somebody else’s life by much longer. Obviously, you can and should contest whether this is in fact, equitable on moral reflection.
To be clear, I am not suggesting that you cannot determine what is unjust discrimination without a full fledged theory of justice, as you can definitely rely on intuition to some extent, but often in the absence of additional reflection, you will notice that your intuitions on particular cases are either self-contradictory or sensitive to framing. So you need additional theorising, and of course you have to balance it against the intuitive appeal of certain general principles. So you cannot use particular judgements to shortcut the need for a broad investigation of what is just and also in many of the particular cases you highlight while I’m sure there are many people who genuinely found them counterintuitive. However, I also think many of the people objecting to such cases are just picking up on a family resemblance between these situations and situations that they find objectionable instead of actually having a direct intuition about the case.. ultimately like with the full of moral philosophy, there is no silver bullet that avoids all conflicting intuitions and you have to balance your intuitions and pick the bullets you’re willing to bite while acknowledging that your answers may depend on hard to obtain empirical information.
It occurs to me that in terms of triaging on the basis of social value, a flaw would be that this might encourage people to lie and deceive in ways that are socially inefficient (and would invalidate the utility of the practice). I can almost see it in my mind "We can only save one of these two patients, according to their charts this one has 5 Nobel prizes, is gifted neurosurgeon and runs a charity that feeds the hungry around the world, this one has 10 Nobel prizes and is a cardiac specialist at the Mayo clinic and has developed 10 highly deployed life saving vaccines for communicable diseases common in the developing world. Amazing how many Nobel laureates we see come through here."
If one limits judgements of social utility to cases were it would be easy to gather reliable evidence (hard to fake credentials etc.), I suspect the number of situations where it could matter would be reduced significantly, if not entirely.
In terms of the prioritizing treating health care workers, doesn't that more follow the logic of putting your own oxygen mask first in the plane emergency before helping your neighbour put on their mask? Healthcare workers who are sick (enough to benefit from treatment) are not providing care, usually social utility is indirect relative to the crisis in question, should you save the engineer now versus the ditch digger, but neither's utility is directed at an immediate current crisis but at a remove. Whereas treating healthcare workers is about treating the people who are currently under a crisis of manpower to deal with the very disaster that has laid them low. Also, there is the morale consideration, if you know the person you just worked with his waiting in the hall on a gurney because they are low priority it probably erodes motivation, leads to more burnout etc. Also, also assessing whether the hospital worker in front of you who you worked with his in fac a healthcare workers is an easier task of discrimination and harder to fake ("I only faked working with you for the last 20 years by coming in an actually doing the assigned work so I could secure this hospital bed you fool!"). A similar issue would be giving special treatment to the ill pilot landing the plane during a food poisoning incident at 40 000 feet (why did he have the fish?) this is not so much social utility as immediate necessary load baring utility.
As someone severely "disabled", in practical ways, by multiple sclerosis medical interventions I find that your argument isn't how the real world works. First off, I was extremely healthy and greatful for majority of life. I'm well educated, worked hard, contributed quite a bit to "society". Volunteered, donated money, etc. Live in a state that hates the disabled, provides no help or follows laws even. However loves medicaid recipients that get huge amounts to remodel homes. Get modified cars, free health care even ultra expensive interventions that don't help them. But provide big dollars and control to an entire slate of people. But we who pay our own way can't even have useable restrooms.
So it may be easy to put it in the term of people who've never had to deal with disability, but is that reality? Or a nice sounding doctrine that is impossible to honestly quantify because of so very many variables?
I'd likely agree that states should (in many cases) do more to accommodate the disabled (though I'd need to know more empirical details to have much confidence either way). If we listed all possible interventions in priority ordering, governments spend a LOT of money on things that I expect should be ranked lower than disability accommodations. So I think a broadly optimizing perspective can support a view in this vicinity.
The main practical verdict I'm defending in this post is just that it's reasonable (and not unjust discrimination) for a philanthropist to prioritize life-saving donations to help the global poor over donating to help Americans with disabilities. You can agree with this even if you think that states should generally be doing more for their disabled residents.
1) I agree that welfarist considerations can probably explain why it is wrong, collectively, for employers to engage in rational discrimination and why we ought to outlaw this [though I think the case for this is stronger if we hold that certain types of hedonic "benefits" to discriminating groups do not count as boosts to their welfare]. Unless we're doing some kind of rule-consequentialism, it's not as obvious to me that welfarist considerations can explain the injustice of individual employers engaging in rational discrimination on their own. It's not hard to imagine an employer, especially prior to the enactment of anti-discrimination laws, who faces a choice between e.g. (i) hiring a woman and losing a large amount--more than would be gained by the woman--due to angry customers and (ii) refusing to hire the woman to appease the customers and making her worse off, but keeping the business of the customers. I think many people would call this a case of unjust discrimination even if the economic or other admissible benefits of sex discrimination substantially outweigh the costs.
2) Just want to flag that the term "disparate impact" is also a legal term whose definition is much narrower than in this discussion [and is actually closer to your own view?]. At least in the U.S. a successful defense to a charge of disparate impact is for the defendant to show that it was justified in making the choices that it did. If a theme part operates on weekends and has a rule that all employees must be available to work Saturdays, there is no disparate impact liability on employees who are Jewish or Seventh-Day Adventists since the rule is reasonable. The fact that an agent disproportionately harms members of a protected class doesn't mean that there's a disparate impact in a legal sense. I know that you're not talking about the law--just found the use of the term a bit off-putting
Thanks for clarifying the legal understanding!
re: 1, I think the force of calling it "unjust discrimination" is best understood as suggesting that there ought to be a universal law against it. At least, it would seem pretty unreasonable to demand that the only non-sexist business owner in the state must shoot themselves in the foot by making a hire that their social context guarantees will turn out to have negative value for their business! What's really needed there is a systemic solution (and welfarist considerations can explain this important fact).
I was mostly pointing out that to the extent your account is meant to capture what people think counts as unjust discrimination, I'm skeptical the everyday notion is necessarily captured just by welfarist considerations. There are definitely some philosophers who defend harm-based accounts of what makes discrimination wrong [e.g. Lippert-Rasmusson] but my sense is that it's a minority view, though the discrimination literature is pretty enormous these days.
It seems like “unjust” is supposed to be tied to some notion of desert.
For example, imagine Bob Billionaire and Joe Janitor and Joe Janitor’s wife Wendy. Bob Billionaire is an investor who was born with a keen eye for value but other than that is a terrible person who abuses his family and does not give to charity. Joe Janitor and wife Wendy are a lovely churchgoing couple who are kind and decent. If Bob Billionaire is alone swimming east and the janitor couple are swimming west, I suppose you have to save Bob Billionaire (the taxes alone on his billions are worth several VSLs per year). But can we at least call this is unjust? He doesn’t deserve it!
A reasonable response might be to say that saving the janitor couple would be unjustly discriminating against the unknown interests who benefit from Bob Billionaire’s investing skill.
But I can press you further and say that Bob Billionaire, as a selfish billionaire, lives a much better life than Joe Janitor. So even ignoring the interests of anyone else, if you had to choose between saving Bob Billionaire or Joe Janitor, even flipping a coin would be “unjustly discriminating against” Bob Billionaire and his greater interests. It seems like we are perverting the definition of justice at this stage.
I’m not necessarily claiming that one’s interests should be weighted by desert or other Rawlsian notions of fairness. I’m just not sure an equal weighting should be called “just”.
Yeah, in the instrumental case you're doing it for the sake of the downstream beneficiaries, not for Bob himself.
But I like your next challenge. It does intuitively seem very unfair to treat Bob's advantages in life as a reason to advantage him over Joe *even more* (by preferentially saving him). I wonder how best to systematize this intuition without getting absurd results like flipping a coin between saving children vs 90 year olds, or humans vs chickens. Maybe we bracket certain *kinds* of (socially contingent) advantages, while treating biological differences like age and species as legitimate to consider? Would it make a difference if Bob's wealth were truly deserved, so the idea is just to bracket *undeserved* advantages? I'm not really sure what to think about the case.
Worth flagging that I still don't think any concern along these lines would justify prioritizing disabled Americans over the global poor, at least. But it may be that a slightly different systematic account would be a better way to secure this verdict. Thanks for the comment!
I think, in your example, part of the intuition is that bad people who treat their family badly deserve some punishment, which could, for example, take the form of giving their welfare less importance. However, if Bob was a good person, your second challenge becomes a lot easier to resolve. After all, I think you would agree that reducing BOB to the status of a Janitor if he was a good person with no benefit to anyone else would hurt him, but if so being alive and billionaire is obviously better than being alive and a Janitor and so saving life for Janitor is actually doing him a smaller favour than saving a life of a billionaire in terms of how much you benefit them. It’s obvious from transitivity, since if being a billionaire is better than being a Janitor, you cannot treat them as equally valuable states of being when comparing them to the alternative of being dead. In fact, if you think about it, selfishly, a Janitor, who is not happy with his life might potentially be willing to take a risk of death with greater willingness compared to if that same janitor became a billionaire and was happier with their life because the cost of dying increases with your wealth.
Timely post for me, as I've just been to a talk about DCEA (distributional cost-effectiveness analysis) in health economics and I'm trying to work out what I think about it all.
I share your view that inequality does not ground a moral objection to optimisation. I also agree that in an ideal world we'd include externalities associated with targeting specific groups in our optimisations, but that this is very hard in practice.
Wondering what you think about proposals to incorporate population-elicited inequality aversion parameters into social welfare functions?
I'd need to hear more details, but my gut reaction is to be dubious of the idea, since I don't think inequality has intrinsic significance. (Could be fine to consider as a heuristic in many cases though.)
Yeah, I wonder if it could be a reasonable heuristic to capture intuitions about instrumental costs of inequality. These parameters are actually notoriously hard to elicit though (vary a lot between people and studies) so I'm not sure it's a great solution.
See e.g. https://pmc.ncbi.nlm.nih.gov/articles/PMC6120144/
A lot of disabled people face such discrimination because they are seen as less than.
Looks like a discussion of what is more important, the action itself or its result. In other words, what ethics should guide our lives, deontological or teleological?
The question I'm curious to ask, considering having unjust discrimination, is it possible to have a just one?