No, not every strategy for avoiding the demandingness of morality results in an implausibly undemanding theory – they do all result in significant costs of some sort or another. This gives us reason to think that so-called demandingness objections to normative theories should be taken lightly, since the alternatives are often equally or more unattractive. In this essay, I briefly clarify what it means for an ethical theory to be demanding or undemanding, and outline two conditions that must be met by a successful avoidance of demandingness. Then, I walk through two potential strategies which do have the implication of implausible undemandingness, before arguing that a strategy appealing to the intrinsic good of the supererogatory can escape this, but at the cost of other unintuitive verdicts. Finally, I conclude that on the basis of these problems, we ought not to take it as a substantial tell against a particular normative theory that it delivers what seem to be extremely demanding obligations.
Speaking informally, one might describe a set of rules as demanding if they are costly to comply with, relative to what one would otherwise do. Similarly, a set of rules are undemanding if one can be fully compliant with them without significant sacrifice. Here, I mean costs and sacrifice in the sense of reducing one’s own wellbeing, or promoting one’s prudential interests to a lesser extent than a noncompliant agent would. This can be generalised to normative theories: let’s say a theory T is demanding of an agent S if it has a large compliance cost – i.e., if her interests are substantially less well-promoted when she complies with the obligations that T generates, than if she were to act with only prudential considerations in mind. (I will leave the threshold for “substantially” here indeterminate, as it could reasonably be set at one of many sufficiently high levels, to capture the sense of “extremely” demanding.)
We haven’t yet defined when a theory T is extremely demanding simpliciter, rather than for a specific agent. This is a more difficult task, because it requires coming up with some sort of mechanism to aggregate the demandingness of T for each agent across the population – which, as David Sobel demonstrates, turns on prior moral convictions (for example, an analogue to the doing/allowing distinction about whether misfortunes a theory permits to be imposed on patients count as demands). For the remainder of this essay, I will adopt the position usually (implicitly) taken by those who raise demandingness objections against theories: T is extremely demanding if it has extremely high compliance costs for some agent S.
It is important to note that if one objects to a theory T on the basis of demandingness, the complaint is not that this theory generates incorrect obligations. For example, my primary objection to some theory telling me I was obligated to stand on my head for 16 hours a day would not be that this is extremely demanding, but rather that those are clearly absurd and extensionally incorrect verdicts. Demandingness objections concede that the theory is (at least in some respects) correct about what is valuable, but makes unreasonable demands of agents in requiring them to promote that good. So, if the goal is to avoid the conclusion that morality is extremely demanding, we need to somehow accommodate the position that there might be actions open to us which create very large amounts of value, but are nonetheless not required – i.e., that some actions are supererogatory. Following Kagan, we can split this into two distinct conditions: that supererogatory acts are (i) good and permissible, yet (ii) not obligatory.
I will now consider two attempts at avoiding the conclusion morality is extremely demanding which each imply implausible undemandingness: Scheffler’s agent-centred prerogatives, and arguments based on incommensurability of reasons.
Scheffler’s account is the most interesting of the four, claiming that when selecting actions, we are permitted to upweight the importance of considerations involving ourself, relative to those involving others. More precisely, if there’s some costly action X open to me, where the net prudential costs are C and net external benefits B, X is supererogatory just in case that C < B but M * C > B, where M is some pre-specified multiplier larger than 1 capturing the extent to which one can upweight prudential considerations. Under these conditions, even though X is both good and permissible, it is not obligatory.
One might hope that this can avoid the conclusion that morality is extremely demanding for the following reason: in instances where a theory T requires me to make a large uncompensated sacrifice for the sake of others, I can appeal to agent-centred prerogatives to show that I am permitted to fail to make that sacrifice, by upweighting the costs to myself in the moral calculus. But as previewed, this also leads to cases where theories are implausibly undemanding, and indeed permit agents to take actions which seem grossly impermissible. After Kagan, consider the following two cases:
Donate to Save: Your elderly distant cousin is suffering from a rare illness which could be cured if you put $10,000 towards their treatment, and otherwise they will die. This is a significant sum for you to part with, but it would not put you into penury.
Drown to Inherit: Your elderly distant cousin has fallen into his bathtub, and is about to drown. You’re just outside the room, and could rush in to rescue him – but if you do so, then you will forgo a $10,000 inheritance (for he plans to change his will tomorrow).
Intuitively, many would hold that in Donate to Save it is permissible to keep your own money, while in Drown to Inherit it is certainly not permissible to stand by. But on Scheffler’s account, we would treat the two equally, since in either case, the costs to oneself are $10,000, and the external benefits are your cousin’s life. So, it does indeed imply that morality is implausibly undemanding, in the sense that it fails to generate obligations that we think it ought to.
Other theories fall victim to similar problems. Some might appeal to the value of autonomy claiming that it is intrinsically good and moreover incommensurable with goods like wellbeing. On this account, morality is not extremely demanding because the costs to your autonomy of making uncompensated sacrifices simply cannot be compared with the benefits to others, and therefore the theory generates no obligations when they conflict. But this also seems to lead to an implausibly undemanding morality: it implies that you can never be required to make sacrifices for the good of others, even in situations like Singer’s Muddy Pond, where the cost to you is as trivial as damp shoes and the benefit to another their life.
There are some strategies which one could deploy to avoid demandingness which do not have implausible undemandingness as a corollary, but they bring other problems. For example, one could claim that the supererogatory is intrinsically valuable (perhaps because it allows us to demonstrate our commitment and altruism towards each other), and therefore that morality cannot be so demanding that there is no space for such actions. But this seems to confuse the truth-value of a moral theory with whether it is the one that allows us realise the most moral value by its lights. Even if supererogatory actions were intrinsically valuable, surely it might be that the moral landscape is such a way that there are no such actions – the view that objects of Platonic beauty would be extremely valuable does not entail that they in fact exist!
Other strategies, like Nagel’s privileging of agent-relative over agent-neutral reasons and Heyd on the injustice of extreme demandingness, likewise come with substantial costs. Of course, enumerating these theories alone does not demonstrate that every strategy fails in one way or another, but it does highlight that demandingness can be avoided only with other costs – and therefore, that the demandingness of a theory is not in itself a particularly strong objection.
In this essay, I explained what it means for a theory to be demanding of agents and demanding simpliciter, and outlined the dual challenge for those trying to avoid extreme demandingness. Then, I presented Scheffler’s view of agent-centred prerogatives, and argued that it leads to an implausibly undemanding theory. Finally, I reviewed other potential strategies, and concluded that this means we should take so-called demandingness objections more lightly.
Post-Claude & examiner report notes
- [My own reflections first] I think that I didn’t do a good job of explaining why the intrinsic value of supererogatory approach doesn’t end up being implausibly undemanding
- Also, I had forgotten what the main specific strategies for defusing demandingness are, besides Scheffler’s
- Should’ve noted explicitly that Scheffler wants to reject the doing/allowing distinction and this is important for symmetry of the two cases.
- For Nagel, the idea is that pain generates agent-neutral reasons, but adopted, voluntary desires (e.g. to be a pianist) generate only agent-relative reasons. The problem is that then you’re not even permitted to make the sacrifice.
- For Heyd, his is the grab-bag approach, but one of them is that it’s unjust to require someone else to work ceaselessly for another’s welfare. The problem is that even if it’s wrong to force S to do X, that doesn’t mean they can’t be morally required to do X.
- Also, I had forgotten what the main specific strategies for defusing demandingness are, besides Scheffler’s
- One important thing: the question asks about every known way. Obviously it’s not possible to enumerate them all, but it would’ve been better to take a more general approach and spend more time on this aspect, probably by using Singer’s shallow pond.
- In particular, if you take:
- Singer’s beneficence principle (if by doing X you can prevent something bad from happening without sacrificing anything of comparable moral significance, you ought to do X), or
- Scanlon’s rescue principle (if we can prevent something very bad from happening to someone by making a slight or even moderate sacrifice, it would be wrong not to do so); extended by Ashford’s stringent principle (which extends to great sacrifice, given sufficiently weighty stakes).
- Then there’s a dilemma – either you accept the principle or you do not.
- If you accept the principle, then we can ladder up into really very large demands.
- If you reject it, then that looks implausibly undemanding.
- This matters because some strategies to avoiding demandingness are just turning to a non-consequentialist theory! So we need to show those are affected too.
- And e.g. if you take the Scheffler approach but add a constraint against doing harm, then it might get the results we want – except probably it’s implausibly undemanding if there aren’t any Kantian-style imperfect duties, and from there you can escalate to drowning children etc.
- In particular, if you take:
- There are other approaches to defusing demandingness to bear in mind, too.
- Satisficing consequentialism; the gratuitous-harm objection.
- Murphy on Collective Principle of Beneficence / everyone does their fair share
- Objection: what about two children drowning & two strangers, but one walks away; other still required to save both!
- (Very costly: reject moral rationalism and say that we don’t always have most reason to do what morality demands)
- The intrinsic-value-of-supererogation move might indeed not avoid implausible undemandingness.
- We’d need an independent principle telling us where the line of supererogatory falls. Otherwise, maybe e.g. saving the drowning child counts as supererogatory, but then this seems too undemanding.