Sunday, November 30, 2008

Evaluative Non-Integration

There are many possible targets for consequentialist evaluation: acts, rules, motives, etc. This might lead one to ask which should be privileged: should we be act, rule, or motive consequentialists? But the question is ill-formed. The only principled answer is offered by global consequentialism: none is to be privileged; each of these targets may be assessed in terms of their consequences.

One may object: "Don't these perspectives offer conflicting normative advice?" An example (from Parfit, I think): suppose that Dad's love for Son is overall for the best. But on particular occasions, it leads him to make sub-optimal choices: e.g. benefiting Son when he could have netted a greater benefit for some strangers. One is tempted to ask, "How are we to assess Dad? Is he good because he has the best character of those available to him, or did he act badly in failing to perform the best action available to him?" Call this 'the integration problem'. But note that - again - the question is ill-formed. Those two assessments are perfectly compatible, so the answer is 'both'. He has the best character, and this led him to fail to perform the best action. What's the problem?

It's tempting to think that agential evaluation must be perfectly integrated, such that the best character (or motivations) will inevitably lead one to perform the best actions. Now, it's true that there is some connection here: acting on good motivations can be expected to lead to good consequences more often than acting from some alternative motivational set -- that's precisely what makes the former motivations 'better'. But this integration is imperfect, as the above example illustrates. The various targets for moral evaluation can come apart, and when they do it would be arbitrary for a consequentialist to assess all others on the basis of how well they integrate with a privileged one. When the best motivations pulls us away from performing the best acts, we are neither wholly good nor wholly bad. Accurate evaluation requires differentiating the quality of the motives from the quality of the act.

My best attempt at achieving a wholly integrated perspective was to consider the agent's life as a whole (thus including their every act, motive, and all the rest). This seems to obviously be a privileged viewpoint, and - moreover - one that fully integrates all other perspectives. After all: any divergence from the aggregation of acts and motives specified in my best (most utility-promoting) possible life will necessarily lead to a worse outcome.

But while that might work for perfect agents, extending the theory to non-ideal agents complicates matters. After all, if my future self cannot necessarily be trusted to do the ideal thing, this could radically alter what current decision would be for the best. Suppose my currently φ-ing could lead to (i) the best possible outcome if I were to follow this up with a series of acts S which I could, but actually won't, perform; and (ii) the worst possible outcome otherwise. In those circumstances, it is not best for me to φ -- doing so would have very bad consequences, due to my subsequent failure to S -- even though it's part of the best possible life. The divergence arises because at any given time I can only choose how to act then; I cannot perform a lifetime's aggregation of acts with a single decision. And so we find that even our evaluations of acts and of act-aggregations fail to fully integrate.

The only way to preserve integration is to embrace implausibly fatalistic theses. For example, if my decision to φ renders it impossible for me to later S, then there is no ideal 'φ+S' life available to me. More generally: if each act entails the full lifetime's aggregation of acts, then since the two cannot come apart, neither can the evaluations. Similarly, if it turns out that one's motivational set strictly determines each of one's actions, then it's no longer true that Dad acts sub-optimally in favouring Son. The only way he could possibly do otherwise, on this account, is if he had a different motivation set, which ex hypothesi would lead to worse outcomes overall. Forcing the coincidence of acts and motivations in this way could dissolve the integration problem, but again this narrow form of determinism (whereby future outcomes are fixed by a mere part of present reality) is not remotely plausible.

So we should accept evaluative non-integration. The best motivations may pull us away from the best actions, which in turn may steer us away from the best life. And the best decision-procedure might tell us to ignore all of this! So it goes.


  1. I find this a little confusing. Motivations and decision-procedures don't have consequences, directly. They have consequences only insofar as they're instantiated in actions. So shouldn't actions have priority?

    That is, suppose I make choices pursuant to decision procedure Q, which produces the best action (an action with the best consequences) in 99 of the 100 decision points at which it's relevant. At the 100th, decision procedure P would have produced a better action, even though decision procedure P would have produced worse actions at the other 99 times.

    Surely, in such a situation, I ought to follow decision procedure Q 99 times and P once, rather than Q 100 times, and I ought to do so just because it produces the best actions to follow that set of decision procedures?

  2. Hi Paul, since you are simply asking what you ought to do, then yes there is a simple answer: you should do what would be best.

    But one might ask other questions, such as 'what dispositions would it be best for me to have?' And the answer to this question is independent of our answer to the previous question. In particular, it is not necessarily the case that you ought to be disposed to follow Q 99 times and P once. There are two reasons for this.

    Firstly, it is false that dispositions "have consequences only insofar as they're instantiated in actions". Cf. Newcomb's problem, Kavka's toxin puzzle, etc. The world may be such that we are rewarded or punished for being in a certain state no matter our later actions.

    Secondly, motivations do not strictly determine actions, so it is possible that this motivation set would in practice lead you to perform other, worse actions. (Maybe you would end up performing P-ish actions too often.)

    The Dad and Son case seems a pretty clear example of all this. There it is true that:
    (i) It's for the best that Dad loves Son. This is precisely the motivational set that Dad ought to have.
    (ii) It's not for the best that Dad acts to mildly benefit Son instead of greatly benefiting Stranger. He ought to act differently.
    (iii) The optimal motivations mentioned in (i) make it likely that Dad will fail to perform the optimal act mentioned in (ii).

    In other words: being the way you ought to be can cause you to fail to do what you ought to do. In this way, evaluations of 'being' (character) and 'doing' (actions) cannot be fully integrated.

  3. hmm...

    Let's set aside the question of direct consequences of dispositions -- I don't think there are any real cases like that.

    There are two questions here: 1) what disposition is it best to have in general, and 2) what disposition is it best to have at time t? So it might be best in general to have disposition Q, but isn't it best to have disposition P at time t, when disposition P leads to the best action at time P.

    Making this more concrete: suppose that we're evaluating the actions of a mad neuroscientist who can control dad's brain from moment to moment. Wouldn't it be best if that mad neuroscientist made dad love his son at all times EXCEPT when loving his son made him do suboptimal actions?

    If that's true, then it seems like the world is better when dad's dispositions can change. Now it might be that in a world without the mad neuroscientist, dad can't bring it about that his dispositions are sometimes loving, sometimes not. So dad isn't *blamable* for not having perfectly calibrated dispositions. But that's a different question from whether dad would be better, in a consequentialist sense, if he had dispositions that were perfectly calibrated to the best actions.

  4. Your #2 seems to have been conflated with a third question: (3) What disposition-at-time-t is part of the best lifetime aggregation of dispositions?

    (2) and (3) are distinct, in exactly the same way as the 'act' vs. 'act-aggregation' distinction I discuss in the main post. Having disposition P at time t may be straightforwardly bad if it turns out that having this disposition at t would also lead one to have it at later, less opportune, times.

    So it looks like you've independently hit upon my "best attempt at achieving a wholly integrated perspective", viz. adopt the "whole lifetime" perspective. But my post explains why even this approach is ultimately doomed to fail.

  5. It's doomed to fail because it extends poorly to non-ideal agents? But that seems like a feature rather than a bug, in that it gives us some content for what an ideal agent is: an ideal agent is one such that disposition and action can be integrated in this way. This gives us the resources we need to say -- which I take to be obviously true -- that dad would be better if he were able to perfectly calibrate his dispositions to the best actions.

  6. We don't need any special "resources" to say that Dad would be better if he were able to perfectly calibrate his dispositions to the best actions. That's a simple answer to a simple question, as per my first comment. The point is that there are other questions we can ask, and it turns out these don't all integrate.

  7. I think this argument makes a lot of sense; if a consequentialist is to be a global consequentialist (and, as you say, it seems to be the way for a consequentialist not to go about arbitrarily privileging some evaluand over another), then while it's reasonable to think that these will more or less integrate most or at least much of the time, this is a far cry from their doing so perfectly all of the time. And it would have the advantage of giving consequentialists a wider range of evaluative standards -- which, setting aside the inconvenience of non-integration of standards, would seem to give consequentialists a more powerful and flexible set of tools than they would otherwise have.

    It looks like such a view would be a form of value incommensurability. (Which would be interesting, since people often treat value incommensurability as something consequentialism can't handle properly.) Was that what you were aiming for, or is there some key difference?

  8. Hi Brandon, I'm not sure about that. Incommensurable values are usually thought to be 'competing' in some sense. (They all offer reasons to act, say, it's just that their relative weights cannot be compared or balanced against each other.) But I wanted to highlight that all the global consequentialist's answers are entirely compatible with each other, because they are answers to distinct questions. We are tempted to assume that the answer to one has implications for all the others -- and if that were so it would lead to a kind of Sidgwickian incommensurability -- but on more careful reflection we should reject this assumption.

  9. Just to clarify: the kind of 'incommensurability' rejected in my last sentence would arise if each possible privileged starting point gave rise to complete and self-contained evaluative frameworks: e.g. an "act consequentialism" that evaluated acts directly and motives indirectly (as good insofar as they aim at impartial benevolence) vs. "motive consequentialism" that evaluates motives directly and acts indirectly (as good insofar as they are the actions a person with the best motives would perform), etc.

    It would be a real problem if we couldn't decide between these competing integrated evaluative frameworks. But global consequentialism pre-empts any such possibility by rejecting the demand for integration, and offering principled, direct answers to each question (which is the best act? which are the best motives?) individually. As such, it conclusively settles all questions, leaving none open to interminable competition between competing perspectives. So in that sense it avoids any incommensurability.

  10. But apparently we do need special resources: On the integrated approach, we're making a mistake if we say dad has a good character because he loves his son at all times. He may have the best character that is available to a non-ideal agent like him, but he'd have a better character if he loved his son at all times except when loving his son leads to objectionable nepotism. On your approach, saying the dad has a good character is not a mistake -- but if it isn't a mistake, then isn't it the case that we're unable to give the right answer to the simple question?

    I think I'm misunderstanding something...

  11. Richard, I'm glad to see you are a proponent of global consequentialism. I've thought about the subject quite a lot and, indeed it forms the basis for both my BPhil thesis and my DPhil thesis. I presented a paper on it at the recent International Society for Utilitarian Studies conference at Berkley. My paper was called 'How to be a consequentialist about everything' and is available here:


    Over the last few decades, there has been an increasing interest in global consequentialism. Where act-consequentialism assesses acts in terms of their consequences, global consequentialism goes much further, assessing acts, rules, motives — and everything else — in terms of the relevant consequences. Compared to act-consequentialism it offers a number of advantages: it is more expressive, it is a simpler theory, and it captures some of the benefits of rule-consequentialism without the corresponding drawbacks. In this paper, I explore the four different approaches to global consequentialism made by Parfit, Pettit and Smith, Kagan, and Feldman. I break these up into their constituent components, demonstrating the space of possible global consequentialist theories, and I present two new theories within this space.

  12. I should add that Feldman's theory is a possibilist theory of global consequentialism in which a proposition P is obligatory if and only if it is true in the best world open to you. You might be interested in this as it is quite like your lifetime version.

    Roger Crisp and Elinor Mason both have distinctly lifetime versions (lifetimes are the only objects of normative assessment), but in conversation with them I have found that they both now prefer global consequentialism. Indeed I think that most consequentialists do (it was very popular at the ISUS conference). It is also fairly clear that Bentham, Mill and Sidgwick supported something like this, though they do not make any precise formulation. I think the focus on acts arose mostly out of being defined against rule-consequentialism.

  13. I don't think incommensurability requires competition, but I take your meaning: if there's any sense in which we have incommensurability, it couldn't be one where there is competition, because that would involve a sort of category mistake. It's not the case that your analysis of what's the best motive would change the analysis of what's the best action; the two analyses are simply different. The non-integration is not a competition; it's just that, since motives and actions are in fact connected in certain ways, due to facts about the world, what contributes to the best motive might not be conducive to the best action, and vice versa. I'd call that a sort of incommensurability; but the word is arguably used in so many ways when talking about values that that might not be helpful.

    I think one could argue that the sort of view you're suggesting is one that makes a great deal of sense of at least certain kinds of Greek tragedy and comedy (this is part of what suggested the connection to incommensurability to me). To take just one example, Euripides' Bacchae: Pentheus is trying to do what is most conducive to the best city (as the Greeks would have thought of it), since he is keeping the peace, preserving order, and working to eliminate obstacles to the virtue of its citizens; but he is caught in a circumstance where this involves very bad actions (again, as the Greeks would have thought it) because in doing so he is opposing both a god (the wine god Dionysus) and human nature. Of course, what makes it a tragedy is that this has horrible results; in a comedy we might have much the same, but (luckily) find that all it meant in the end was that we couldn't reach the ideal of perfection all over. If so, this strikes me as a considerable potential strength, although this might be an idiosyncrasy of mine: I always test out ethical and meta-ethical views by whether they'd make any sense of the sort of very human issues found in literature.

  14. Hi Toby - thanks, your paper has been on my reading list for a couple of weeks, actually! I just got around to it today.

    You write, "To my knowledge all global consequentialists accept that if something is right, it is also best."

    Have you considered Railton's valoric consequentialism? I think it's fundamentally a kind of scalar global consequentialism, but with an indirect account of moral "rightness" tacked on. To forestall the obvious objection, note that it does license the direct evaluation of acts (as more or less "morally fortunate"), and this may even have normative significance in terms of determining our reasons for acting (Railton himself is neutral on this, but that's how I prefer to develop the theory) -- it merely suggests that an additional framework of (externalist, not necessarily reason-giving) 'moral rightness' could be offered too.

    (On this account, 'moral rightness' is best understood not as specifying what acts we really ought to perform, but rather, what is recommended by the best moral code. So the theoretical importance of the category is somewhat deflated, even though in practice it may be the best "public face", so to speak, of moral theory more generally.)

  15. Richard,

    That's a good point (with the provisos you mentioned). I'll add a short discussion of Railton's view before I send the paper for publication.


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.)

Note: only a member of this blog may post a comment.