Monday, November 19, 2012

General and Particular Moral Explanations

In 'Recalcitrant Pluralism', Philip Stratton-Lake draws on Korsgaard's Symmetry Thesis -- that the reason why a good-willed person does an action, and the reason why the action is right, are the same -- to argue against Consequentialism.  He's making the kind of character-based objection that I think is worth taking seriously, but I think that PSL goes importantly astray in how he understands the fitting (good-willed) consequentialist agent.

By way of background: PSL considers a case where a villager can save either his father's life or that of a (relevantly similar) stranger.  He then considers a form of consequentialism which tries to accommodate the intuition that the villager ought to save his father by assigning value not just to welfare but also to states of affairs in which sons help their fathers.  PSL then raises the motive objection (p.380):

Why would he help his father in preference to a stranger? Given the symmetry thesis this will be because of the reason why he ought to help his father. So if the consequentialist account of why he ought to help his father is correct, then the reason why he would help him in preference to a stranger would be because he will be producing a better state of affairs in the world than if he did any other act. [...]
The deciding factor in whether this state of affairs is best is the fact that his act will have as a consequence a state in which a son helps his father. But this should not lead us to think that the relation that figures in this state will figure in the villager’s motivation after all. The fact that his act will have this consequence is the reason why the act will have the best outcome, since it breaks the tie between the value of the benefit conferred on either his father, or a stranger. But it is the fact that his act has produced the best outcome, not the fact that it has produced a state in which a son helps a father, that makes it the right act to do. Otherwise this really would look like a mere semantic victory for the consequentialist. What makes it stand out from the intuitionist account is that it is the production of impartial value that grounds this duty. This is not merely a rewording of the intuitionists view.

While I wouldn't defend this particular form of consequentialism, I think PSL is mistaken about what fitting motives follow from the view, and also mistaken to think that the view would otherwise collapse into Rossian intuitionism.

For any moral theory, we need to distinguish between (i) the particular right-making features it posits, or what makes some particular action right, and (ii) the general property shared by all right actions, or what makes actions in general right (when they are right).  Applied to Consequentialism: what makes actions in general right is that they produce the best outcome -- this is what all right actions have in common.  But each particular action will be right in a different way.  The good-making features of the outcome (the values it realizes) will differ from case to case, and hence so do the right-making features of the actions that bring about these good outcomes.  PSL confuses these when he gives the general answer to the particular question of what makes the villager's act of helping his father right.  The correct answer, according to the form of consequentialism he's considering, will appeal to the particular values at stake in the decision, including the value of sons helping their fathers.

Note that this same distinction applies to PSL's Rossianism.  Imagine if I argued against his view as follows:
The deciding factor in whether the balance of prima facie duties favours the son's helping his father over the stranger is the fact that he has a prima facie duty of fidelity (or whatever) to his father.  But this should not lead us to think that the relation that figures in this state will figure in the villager’s motivation after all. The fact that his act will satisfy this prima facie duty is the reason why the act will satisfy the balance of prima facie duties, since it breaks the tie between the prima facie duties of beneficence that apply equally to his father and the stranger. But it is the fact that his act satisfies the balance of prima facie duties, not the fact that it satisfies his prima facie duty of fidelty to his father, that makes it the right act to do.

This would be a bad objection to Rossianism.  Sure, what all right actions have in common, on this view, is that they best satisfy the balance of prima facie duties.  So one might, speaking loosely, say that it is possession of this general property that explains why an act is right.  But it would also be misleading to say this, for it may lead one to overlook the fact that in any given case there will be a more particular explanation, invoking particular right-making features, and it is these particular right-making features that feature in the motivations of the virtuous agent.

And so it is in the case of consequentialism!  I'm not sure why more people haven't previously realized this, but if my work only ever leads to the wider appreciation of one insight, I hope it is this: that the virtuous (or fitting) consequentialist agent desires each particular good, and not just the promotion of abstract value as such.

Of course, the form of non-welfarist consequentialism that PSL discusses here can still be exposed as perverse (and distinct from Rossian Pluralism).  For it implies that it's fitting to desire that sons help their fathers (generally), and hence he should let his father die if it would thereby enable two other sons to aid their (respective) fathers.  Nobody should hold that view.  Either one should endorse full-blown agent-relative value, such that the son should value his father's welfare over that of the stranger, or else simply stick with good old-fashioned utilitarianism, according to which the son should care deeply about the welfare of his father and the stranger, equally.

Similar remarks apply to PSL's resentment objection (p.382):
[T]he fact that the villager has failed to bring about the best state of affairs leaves the fact that his father has been wronged completely unexplained. If anyone is wronged here, it would be the world which has not been made as good as it could have been. But that, of course, makes no sense.

It's not clear whether PSL sees this form of objection as applying just to the specific view he discusses, or to consequentialism more generally.  So let me just flag that it's not generally true that consequentialism can't account for personal wrongs (and hence warranted resentment).  Even on utilitarianism, if someone chooses to prevent a lesser harm to another instead of a greater harm to Bob, then Bob -- and not just the world -- has been wronged by them.  They have failed to take Bob's interests sufficiently into account, which is a kind of disrespect to Bob -- a failure to appreciate his distinctive value as a person, or to respond to it in the manner that is due.

To think otherwise is, again, to let the general facts about value blind one to the more particular value facts -- such as the distinctive value of each individual person, which we must (morally) take into account, for the individual's own sake.


  1. Hi Richard,
    If I follow you, what makes it the case that a given outcome is best is explained by its particular right-making features in a given situation. So the general property of rightness possessed by an action is the property of being made right by whatever right-making feature is relevant in the situation. Is that right?

    But then the general property of being right (i.e. the best outcome) will not explain anything, otherwise I take it this would be a case of overdetermination in the explanation. Or is it the general property that makes it the case that a given action is right according to a given theory (e.g. consequentialism), while the particular properties cannot do that job?

    My worry is, I don't see how the general property can be efficiently explanatory if the rightness of each action is in fact determined by particular features.

    Yet I don't think this affects your general account of fitting consequentialism, except that it seems to render your moral theory more like an account of "consequentialist fittingness" than "fitting consequentialism" since whatever counts as best for an outcome is determined by particular circumstances, and this will be that which one should maximize in each case.

    (I may be completely misunderstanding your point, so take these as clarificatory questions)


    1. Hi Nicolas, I don't think there's generally any problem with having multiple ("overdetermining") explanations. We can often explain things in more or less abstract ways, depending on what "level" of explanation we're interested in. See, for example, the distinction between robust-process and actual-sequence explanations.

      Having said that, I'm not really committed to seeing the general property as "explanatory" in any deep sense. We can think of the general property as, in effect, the disjunction of all possible particular right-making properties. It may be convenient to use it to identify which acts are right, and to explain what all right acts have in common (e.g., they maximize value). But my main concern here is just to note that this isn't what's relevant to motivation. It's the particular right-making properties that provide the "normative explanations" relevant to the Symmetry Thesis, and hence provide the contents that should figure in the virtuous agent's motivations.

    2. I agree with your first point, I was pressing a possible reductivist objection. And I think your second characterization (as a disjunction used to identify which acts are right) avoids it more straightforwardly. If the general property of right acts is that they maximize value (the good), and if value is relative to token acts (as suggested by Alex below), then this much seems right on a consequentialist view.

      I also agree your distinction is indeed better suited to motivating (fitting) agents, but then do you mean that a fitting consequentialist agent will (should) not be motivated by the general property of choosing the best outcomes?

      If so, is it because it follows from your distinction that that property is superfluous given the role of particular right-making features (the point is related to my initial objection), so we could dispense with the very idea of choosing the best outcomes?

      Or is it that whenever one picks out the right-making features in a given token act, one is at the same time (and knowingly) choosing the best outcome in the given circumstances?

    3. I was thinking the latter. The good-willed agent, if sufficiently reflective, will recognize that they are choosing the best outcome -- and that if the outcome weren't best, they wouldn't be choosing it. But this is because "being best" is a kind of higher-order property that tracks the more concrete values that they are concerned with, e.g., each individual's welfare.

      As I put it in an old post: "we should think of the deep feature ("being utility-maximizing") as really a kind of higher-order property: it is merely alerting us to the fact that the act in question has other (surface) features in virtue of which it maximizes utility. The claim that φ-ing maximizes utility doesn't add anything further to the claims that φ-ing helps Bob, harms no-one, and has no significant opportunity costs. If surface features have explanatory priority in this way, that could explain why it is that proportionate concern for them, rather than the deep feature as such, is what's constitutive of good will (virtue)."

      (It's a bit like the old "moral fetishism" objections to Kantianism and Virtue Ethics. The Kantian should not be motivated by de dicto thoughts of "duty" as such, but rather by the actual moral laws that it is their duty to follow. Similarly, the virtue ethicist should not be motivated by the thought of her own "virtue" as such, but rather by the more particular motivations that constitute being virtuous.)

  2. Hi Richard, I'm kind of sympathetic to all this, but speaking to Philip earlier made me doubt that I really have a good grip on exactly what you're saying.

    So hopefully you'll indulge me with answers to some simple questions:
    Take some token act A. Imagine that A has natural features N. What are the relevant explanatory relations between N, the rightness of A, and the goodness of A? Is your view:
    (1) That N makes A right iff N makes A best.
    (2) That if N makes A best, A's bestness in turn makes it right.
    (3) That if N makes A right, and it must be independently (and coincidentally?) true that A is best.

    I'm a little worried that (1) isn't really consequentialism, since on this view, the good isn't prior to the right, that (2) is precisely the view it sounded like you wanted to deny, and that (3) is objectionable. I think you endorse (1), but I'm not entirely sure.

    I think part of the problem is that you draw the crucial distinction in the post using the phrase "what makes actions in general right", and I'm struggling to parse this phrase. I would have thought that actions in general have no normative properties: only token acts do (when we talk of the wrongness of a type of act, we really mean that every token of this type is wrong, not that the type itself is).

    (I should repeat that my remarks here are influenced by some things Philip said to me earlier, though of course he should bear no responsibility for what I've said.)

    1. Hi Alex, that's a good question. I'm inclined to want to combine (1) and (2) into a sort of chained explanation: A is right (proximally) because it's best, and hence (ultimately) because of N.

      I do think that the good is prior to the right, in the sense that our reasons for action are explained by our reasons for desiring various outcomes (value = desirability, so this is a sense in which value has explanatory priority). But it seems to me that this must be compatible with holding that what features in the good will is the particular content that's desirable, rather than the higher-order fact that the outcome is desirable. The content of what's desirable is obviously what figures in the fitting agent's desires, after all, and if we accept any sort of belief-desire psychology then it seems like it's going to have to be this same content that figures in their motivating thoughts (or "will").

      So, it's because N makes A best that A is right. A would still be right, but in a different way, if it had some other natural features N* that still made it best. So, that's the sense in which maximizing value is "what makes actions in general right" -- it's criterial for right action, and so we can see a sense in which right acts are right "because" they maximize value. But I want to stress that this high-level explanation is not the relevant one for purposes of the Symmetry Thesis. It's the more particular explanation that provides the content that figures in the good will.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)