In general, contractualists derive moral obligations from the content of a contract to which we’d all agree under certain conditions.
Hobbesian contractualism [contractarian]: moral norms are those that agents would accept self-interestedly.
Kantian contractualism (e.g. Scanlon): contracting parties are motivated by a desire for mutual recognition and respect.
Scanlon: “an act is wrong if its performance under the circumstances would be disallowed by any set of principles for the general regulation of behavior that no one could reasonably reject as a basis for informed, unforced general agreement”
For Scanlon, the principle-rejection is prior to the wrongness. Whereas others might think: if an act X wrongs someone, that is the reason to reject any principle permitting X.
There are three criteria for what reasons can ground a reasonable rejection:
Generic reasons: i.e., those that any person could have by virtue of being in a particular situation, not those that are contingent on our own distinctive circumstances or preferences.
This is so the theory is impartial.
Individualist restriction: the grounds for rejection must focus on how the principle affects individuals. No aggregation allowed.
You can appeal to non-welfare values like friendship or autonomy, insofar as they’re about individuals.
Impersonalist restriction: claims about impersonal goodness or badness of outcomes cannot be grounds for reasonable rejection.
So there’s no axiology involved.
Greater Burden Principle: “it would be unreasonable […] to reject a principle because it imposed a burden on you when every other alternative would impose much greater burdens on others.”
One gloss: we ought to do the thing which minimises the greatest possible complaint.
So we could determine rightness by enumerating all possible decision-making principles, find the weightiest complaint against each one, select the principle with the smallest weightiest complaint, and then check whether the action is right according to it.
But Scanlon doesn’t actually want to do this; he’s trying to create a framework to organise moral thinking rather than a moral algorithm.
One objection, therefore, is that it’s a poor guide to action.
It follows that assessment of principles is comparative: there’s no reason that’s automatically sufficient for reasonable rejection.
The realm of morality
Contractualism might be able to answer the question of moral motivation (“why be moral?”) better than other theories; this is why Scanlon likes it.
According to contractualism, the principles of right action are governed by an ideal of mutual recognition, via justifiability to others. So the importance of being moral is grounded in the value of this way of living in relation to others.
This is less trivial than something like “be moral because it’s [analytically] the right thing to do/thing you have most reason to”, and more characteristically moral than “be moral because it’s in your self-interest”.
But it’s not clear this is unique to contractualism – adherents of other theories can still say that an action which impermissibly wrongs another is one that can’t be justified to others (Hieronymi 2011).
Conflation of morality and politics (see Sinclair slides #44)
Limitations and objections
It might seem like there’s redundancy in the reasonable rejection apparatus (Pettit 1999, Blackburn 1999).
Surely the wrong-making feature of an action X is just whatever property it has that means it’s reasonable to reject any principle permitting X, not the fact that such principles are reasonably rejected.
However, this objection is sidestepped (Stratton-Lake 2003), because Scanlon distinguishes between what wrongness consists in and what makes an act wrong.
His view is not that the wrongness is because of reasonable rejection – he accepts that the thing that makes the action wrong is indeed the rejection-enabling property.
But his account says prohibition by principles nobody could reasonably reject is what it is to be wrong.
The cost of this response is that wrongness is no longer itself reason-giving (that’s done by the actual harm, not the fact of rejection-enabling).
Alternatively, contractualism isn’t redundant because it explains why agent-relative reasons for rejection should imply agent-neutral reasons for everyone else.
[Some further stuff from Ridge here.]
Still, it does seem like we can do moral reasoning without needing to appeal to the hypothetical agreement. [is this just disputing that the agreement really does characterise what it is to be right?]
Scanlon’s theory is about what we owe to each other, which means it doesn’t characterise our obligations to non-humans (e.g., to animals, or natural wonders, etc).
A more precise characterisation of the contractualist principle might say: “an action impermissibly wrongs someone iff it is prohibited by principles nobody could reasonably reject”.
And Scanlon is happy to acknowledge there are obligations that fall outside the scope of his theory; he thinks this is where duties to non-human animals lie.
This accords with the idea that non-human animals do have intrinsic moral significance, but of a different, lower moral status to humans.
More difficult: what about humans who lack sophisticated cognitive abilities, and couldn’t enter into contracts?
Scanlon thinks they get the same moral status as other humans just by virtue of their species.
Unlike contractarians, contractualists can accommodate obligations to future generations, because they assume that agents are morally-motivated (rather than acting in line with some self-interest, and given that future people can’t benefit us in any way). But non-identity cases pose problems for them.
Summer or Winter Child (Parfit 1984): Mary could have a child in summer or winter. If born in winter, the child would suffer serious ailments (but have a life worth living); they’d be fine if born in summer. But it’s mildly inconvenient for Mary to have the child in summer, so she does it in winter.
This seems morally wrong. But the contractualist can’t explain it: who gets to reject to the principle permitting this behaviour? Not the Winter Child, as he wouldn’t exist otherwise; not the Summer Child because he simply doesn’t exist.
Maybe the problem is that Mary doesn’t show adequate respect for the (unnamed) individual of her future child. But it’s hard to translate this into a reasonable complaint without relaxing the individualist / impersonalist restrictions.
Contractualism seems to have less justifiable demands than utilitarianism (Ashford 2006). In particular, while both are very demanding at present because of the state of the world and global poverty, contractualism would be highly demanding in all practically realisable worlds (because it rejects all interpersonal tradeoffs) – and that seems wrong.
Scanlon accepts the rescue principle: if you can prevent something very bad at moderate cost, it’s wrong to fail to prevent it.
Ashford argues this extends to foreign aid in the face of global poverty, like how utilitarianism makes great demands of us.
This follows from contractualist assessments of principles being comparative: people in poverty have complaints of starvation/death, whereas our complaints are only e.g. reduced leisure. So we can’t reasonably reject very demanding aid principles, even if they’re highly costly.
The conscientious agent might complain it’s unfair they have to bear a huge burden because of others’ non-compliance. While this has force against the slackers, it doesn’t reduce her own obligations. (This is relevant to utilitarianism & demandingness, too!)
Kumar argues that the consequences of general acceptance of extremely demanding principles might be reasonable grounds for rejection, though.
E.g., loss of autonomy over one’s life, the undermining of relationships (including supererogatory assistance), etc.
Contractualists can talk about goods other than welfare!
But contractualism is extremely demanding in any practically realisable world, because of non-aggregation plus the stipulation that low probability doesn’t discount ex post complaints.
For example, air travel would be prohibited because of the 1-in-a-billion chance that one falls on & kills someone who’s never flown.
(See more in the section below on risk).
Aggregation and risk
Contractualism does not allow for any interpersonal aggregation, which means it accommodates some widely-held intuitions about the separateness of persons.
Transmitter Room: Jones is trapped under electrical equipment in the TV station transmitter room receiving electrical shocks. To rescue him we must interrupt the broadcast of a World Cup match.
If you add together the inconvenience for every watcher, in principle the total disvalue could outweigh badness of leaving Jones to suffer.
But many insist you should save Jones no matter how many watch.
This follows straightforwardly from Scanlon’s Greater Burden Principle.
However, you can construct cases where non-aggregation seems like it is extensionally incorrect.
Suppose you have one dose of medicine, which could save one person from death or millions from cluster headaches (excruciatingly painful but just meaningfully less bad than death). Surely we ought to save the million?
Or, consider taking 1 unit of welfare away from a million impoverished people, and giving it as 1 million to a very wealthy person.
[perhaps here Scanlon can appeal to non-welfare values though, so it doesn’t go through?]
And see below on the difficulty of explaining saving the greater number which also results from non-aggregation.
Non-aggregation means that contractualism seems blind to the moral impact of numbers.
Saving the greater number: suppose there are two sinking boats, one with 1 person and one with 5. I can save all the people on one boat. Contractualism seems to require a coinflip, rather a principle of saving the greater number.
Each person on each boat has the same reason (that she’ll die) to reject any principle permitting the rescue of the others.
You can’t aggregate the deaths of the five on boat II, nor appeal to impersonal goodness.
So the reasons look equally weighty in both directions; you need a coinflip.
Scanlon suggests a tie-breaker solution which he thinks can explain SGN.
In a 1-1 case, that presumably does require a coinflip. But if this was unchanged when we move from 1-1 to 1-2, then the extra person would’ve made no moral difference.
So the additional person could reject a principle requiring a coinflip in the 1-2 case for that reason.
However, Otsuka points out the extra person’s presence does make a difference: we rescue her and person B.
Maybe a way to rescue Kumar: pairwise-over-options reading (see Hayden comment on essay)
Moreover, Hirose shows this proves too much.
If we move from 1-5 case to 2-5 case, couldn’t the new person in boat I reasonably object to any principle telling us to save the 5?
Or if we move from 1-5 to 1-6, couldn’t the person in boat II complain that his presence made no difference? Surely he’s not got any complaint at all, though – he’s being saved!
Kumar (2001) suggests claim-balancing should be interpreted as elimination of paired equal claims from each person, until there’s no more elimination and you have a decisive reason to go to boat II. This way, there’s no aggregation being smuggled in.
But suppose it’s a 1-1 situation where you could either save A or save B. Then both their claims neutralise each other and you’re permitted to save nobody, which is surely wrong (Otsuka 2006).
The best-looking solution is Anonymous Pareto principle (Hirose 2001).
He introduces two non-aggregative principles: (i) impartiality [if alternatives differ only wrt identities of people, they’re equally good], (ii) Pareto.
There’s no moral difference between rescuing one person on boat I vs one person on boat II. And rescuing one person on II plus 4 others is Pareto-better. So we should do that.
Otsuka points out this would justify going to boat II even if all but one person on board could swim ashore with mild inconvenience, which then does start to look like aggregation.
Also, note that this only works for consolidating equal-strength claims; it doesn’t deliver a verdict in Transmitter Room.
There are other potential solutions too; e.g. a weighted lottery where we flip an n-sided coin to decide who to save, then save everybody else on their boat.
The advantage is that A alone in boat I can no longer complain that, when saving the greater number, you’re acting exactly as you would if she wasn’t there.
But this does foreseeably (on occasion) lead to intuitively worse outcomes than just always saving the greater number, if you end up e.g. saving 1 person over 1 million.
Alternatively, Scanlon could give up on the impersonalist or individualist restrictions.
Parfit favours this. (what does he mean about “only prima facie wrong”?)
Maybe we should reject the intuition (Taurek 1977)
Taurek suggests we really should just flip a coin: that gives each person a 50:50 chance of surviving, and that expresses our equal concern and respect for each.
Anscombe thinks what matters is that you’re acting for good reasons. Saving the smaller number because you’re acting for their good might be sufficient reason.
Contractualism also struggles with explaining when it’s permissible to impose risks for societal benefit.
Scanlon says that you could reasonably reject a principle that imposes severe harms at random in order to benefit the majority (e.g. randomly selected medical experiments), even though each person benefits ex ante.
i.e., complaints can be grounded in ex post burdens.
This seems to imply that any risk-imposing practice is impermissible, if the ex post burdens are sufficiently large (no matter how small the probability of realisation). For example, no air travel is permitted because a plane might crash on someone.
See Ashford
Scanlon disagrees, and says you just need to minimise the risk. [this seems like special pleading?]
Mass Vaccination (Frick 2015): 1 million children will by default die. You can administer Drug A to save them all but paralyse legs, or Drug B to save each person with 999/1000 probability (i.e. 1000 are expected to die).
Ex post contractualism requires Drug A since it has the weaker basis for rejection (paralysis, not death). But that ignores the fact that Drug B is a good bet from each individual’s point of view.
Ex ante contractualism would discount each complaint by the probability. So it allows us to impose risks randomly, but it would still let us reject killing 1000 preselected individuals to save everyone.
Ex ante contractualism has its own problems, though: it permits the loss of any number of statistical lives to save one identified life.
At the extreme, it suggests a government should choose to treat one person to avoid certain death, rather than fund a vaccination scheme that reduces a million people’s risk of fatal disease from 99% to 0%.
Frick accepts ex ante approach but thinks that contractualism is only an account of equity, not all-things-considered-interpersonal-wrongness, because we need to incorporate impersonal reasons for the latter.
e.g. reasons grounded in the fact that more deaths is impersonally worse than fewer deaths, regardless of whether any specific individual can complain.
Horton (2017) thinks we should just accept interpersonal complaint aggregation.
The debate between ex ante and ex post contractualism really is about whether you prefer ex ante Pareto or dominance principle.
Dominance says: if you’d judge P as morally better than Q in every possible state, you should choose P over Q overall.
In Mass Vaccination (Unknown Victims), the contractualist framework will, in every realised state with identifiable victims, say it’s better to administer Drug A (since the paralysis complaint is smaller than death complaint). So dominance pulls towards ex post.
Obviously ex ante Pareto pulls towards ex ante contractualism.
Some more stuff in Emma Curran slides about [omniscient guardians and Boat Race; I think too in-the-weeds]