Philosophy, et cetera

Are QALYs Discriminatory?

2009-11-02T20:51:00.003-05:00

In 'QALYfing the value of life' (J. Med. Ethics, 1987), John Harris claims that it is unjust "discrimination" to allocate scarce medical resources to the patients that would benefit most from them (in terms of "Quality-Adjusted Life Years", or QALYs). Instead, he says, we should try to save (or, rather, postpone death for) as many people as we can, without regard for how much different individuals stand to gain from continued life. Since each life "counts for one", Harris argues, postponing death for two 90-year olds (by a month) is more important than postponing a teenager's death by scores of years.

There's something strange about insisting that each person's life has "equal moral value", without bothering to assess how much each person stands to gain from continued life. One does not show equal concern towards two people by being indifferent to whether one receives a papercut or the other is beheaded. We should, of course, be indifferent to which of two people receives an equal benefit (else we would be treating the favoured person's interests as more important than the other person's). But it is no kind of favouritism to prefer that a greater benefit be bestowed upon whoever is able to receive it. Egalitarians are simply confused to suggest otherwise.

Harris claims that it is "ageist" to allocate medical resources efficiently, since it means we would sooner give a dose of life-saving antibiotics to a 20 year old (who can expect to gain 60+ years) than to a 90-year old with the same disease (who only expects to gain a couple of months from being cured of this particular ailment). I do not think that this should be considered "ageist". Again, we are not treating the elderly as less worthy of receiving an equal benefit. Rather, we are saying that, given a choice between offering a slight benefit to one person (who may be elderly), or else offering a much greater benefit to another person (who may be younger), the latter option is obviously preferable, on perfectly impartial grounds.

(Similar remarks apply to Harris' objections to the 'quality-adjusted' component of QALYs as being "discriminatory" towards those with a lower quality of life.)

Harris' central confusion is revealed in the following claim (p.121):

If for example some people were given life-saving treatment in preference to others because they had a better quality of life than those others... this would amount to regarding such people as more valuable than others on that account.

This simply isn't true. Again, we do not ordinarily think that equal moral concern for persons entails being indifferent between a smaller benefit for one or a larger benefit for another. Rather, we may regard people with equal concern (i.e. treat them as being equally "valuable" in themselves) by treating their welfare interests as mattering equally. And treating their interests equally should lead us to prefer that treatment go to whoever would benefit most from it.

Simply put: to say that someone would gain more from continued life is not to say that the person is more valuable than another. It is just to say that what the person gains is more valuable to that person than the alternative outcome is to the other person.

P.S. In principle, we need to assess possible harms and benefits in context: it's not as though a healthy year of life has constant value to all persons at all times. As I argue in 'Gambling Life for Immortality', the next forty years of healthy life are worth much more to me than a subsequent forty (even if they're equally healthy, etc.), in light of what I want to achieve during my lifetime. So the real problem with QALYs is that they are insufficiently discriminating! But of course practical policies cannot be so perfectly fine-tuned, so QALYs may be the best practicable guide to efficient resource allocation. They're certainly a vast improvement over Harris' proposed alternative.

Dimensions of Normative Inquiry

2009-10-25T18:10:00.004-04:00

There are a number of different 'dimensions' of normativity that call out for investigation. (Below I'll describe three: the criterial, epistemic, and structural.) These dimensions are largely independent of each other, so to make progress on one dimension we might as well make simplifying assumptions about the others, just for sake of exposition. (One could always restate one's point without making any such assumptions, but the result might be more clunky and long-winded.) It's important for readers to understand these differences, so as to avoid raising "objections" that aren't really relevant to the discussion at hand.

(1) The criterial dimension (as I'll call it) concerns the fundamental goal or objective. So, for example, a criterial 'subjectivist' (or 'instrumentalist') might think that normative facts must be ultimately grounded in the agent's own goals (whatever those might be), whereas objectivists of various sorts think that certain things are good, or worth aiming at, whether we personally happen to like the idea or not.

(2) The epistemic dimension instead concerns how 'ought' facts depend on the available evidence. (Moral philosophers invoke this dimension when they distinguish what one "objectively ought to do", given the actual facts, versus what one rationally or "subjectively ought to do", given one's beliefs or available evidence.)

(3) Finally, questions regarding the structure of normativity arise when considering act vs. rule consequentialism, indirect 'sophisticated' psychologies, etc.

Bad things happen when one confuses these dimensions. For example, one commenter mistakenly dismissed a puzzle about the epistemic dimension on the grounds that he's an instrumentalist. But, as I pointed out in response, that's a wholly independent issue: the same issues arise on the epistemic dimension regardless of whether we're instrumentalists or objectivists on the criterial dimension.

More recently, I've deleted several off-topic comments that sought to respond to my discussions of normative structure by instead raising criterial and/or epistemic objections. Again, it's important to see that such objections aren't really relevant to the point at hand -- it's like objecting to an acronym or abbreviation that you don't like. In this particular case, the author was sufficiently rude and trollish that I doubt he much cares that he was missing the point. But for any non-philosophers out there with a good-faith desire to contribute to a productive discussion, it's important to (i) work out which dimension of inquiry the discussion at hand is centrally concerned with, and then (ii) ensure that one's response doesn't change the topic, e.g. by mistaking expository assumptions for substantive ones.

(More generally, it's good practice to think about whether one's objection really speaks to the heart of the matter, or whether it's a more tangential point that could easily be accommodated without much affecting the rest of the argument.)

Ord on Global Consequentialism

2009-10-13T16:47:00.001-04:00

Toby Ord kindly sent a copy of his dissertation on global consequentialism ('Beyond Action'), with permission to quote. Let me begin with a striking passage that exemplifies the kind of rhetoric I think needs deflating:

The potential advantages of global consequentialism are many. The addition of new focal points increases the expressivity of consequentialism. It allows consequentialists to directly answer questions about the best motives, the best system of government, or the best way to decide what to do. This also allows consequentialists to bring consequentialism into the strongholds of deontologists and virtue ethicists: it provides an explanation of the importance of character and of rules of conduct, going so far as to show how many acclaimed virtues and rules have systematically led to good consequences, and even daring to suggest ways in which they might be improved.

The striking thing about this passage, to my eye, is that none of the benefits mentioned are distinctive to global consequentialism (GC). They're all benefits of having an axiology, which every consequentialist has. (It's not as though anybody else lacks the theoretical resources to "show how many acclaimed virtues and rules have systematically led to good consequences". Given an axiology, such evaluation comes along for free.)

So we may ask whether GC adds anything beyond the kinds of evaluations that are already open to any consequentialist. Compare: the distinctive addition of Act Consequentialism (AC) is to further claim that the best act is also right, i.e. what we have most reason to do. Some global consequentialists might think that we can similarly make substantive new claims by applying these deontic or "normative" terms to everything, not just actions. Ord calls this view "normative global consequentialism". Others (e.g. Kagan, Parfit) only use the term 'right' in relation to acts, and restrict themselves to evaluative terms (like 'best') when assessing other kinds of things, like climates. Ord calls this view 'semi-normative global consequentialism'.

As far as I can tell, 'semi-normative global consequentialism' is just another name for act consequentialism. It does not even appear to make any different claims. Normative GC appears to make different claims, but this appearance is (I argue) deceiving. My argument for this was given in my post, 'Reasons Deflate Global Consequentialism'. There I suggested that, in calling a disposition (or climate, or other arbitrary evaluand) 'right', there is nothing else that the global consequentialist could plausibly mean, except that it is the best (of those available). So they are not really making a new claim. They are making an old claim, simply using new words.

The case of actions is importantly disanalogous. There are reasons for action, in addition to reasons for desiring that some action be performed. Evaluations concern what is good or desirable, i.e. what we have reason to desire. But beyond evaluating whether we have reason to want that an action be performed, there is a further question to address: whether the agent has reason to act so. The question whether an act is 'right' may thus be taken to address this second substantive question.

There is no such second substantive question when considering arbitrary other evaluands, like the climate. We have reasons to desire that the climate be temperate, say. But the climate does not have reasons to be temperate. Climates are not subject to reasons in this way. That is why there is no substantive question whether a climate is 'right', over and above the evaluative question whether it is best (of those available).

(Motives are, like acts, susceptible to two substantive questions. Aside from the question whether a desire that P is desirable, we can also ask the first-order question whether P is desirable, and hence whether we have reason to desire that P. Similarly with beliefs, we can ask whether we have reason to believe that P, independently of the question whether such a belief-state is desirable. But global consequentialism does not address these substantive questions -- or, if it does then it gives incorrect answers! Not every fortunate belief or desire is reasonable.)

Let's close by considering Ord's argument that [normative] GC is a distinct view from AC:

[I]t is logically consistent to claim that Michael should follow decision procedure Y even though his following decision procedure X would lead to a better outcome. This is closely analogous to the way that deontologists and rule-consequentialists can consistently claim that Michael should do act Y even though his doing act X would lead to a better outcome.[*] Act-consequentialists must be consequentialists about acts, but nothing in the letter of the view stops them being nonconsequentialists about decision procedures.

As my above remarks suggest, this analogy won't work unless one thinks that decision-procedures are, like actions, subject to an autonomous class of reasons. (And it certainly won't carry across to things like climates.) There is room for substantive dispute about whether our reasons for action derive from considerations of desirability (i.e. evaluative facts). This is how we can make sense of being a (non)consequentialist about acts. It's far less clear how to make sense of (non)consequentialism about climates or decision procedures. There does not appear to be logical space for such a dispute. There is no open question here to invite such competing answers.

Ord claims that there is an open question - 'should Michael follow decision procedure Y?' - that is not analytically settled by the evaluative fact that it would be better that Michael follow decision procedure X. I have no idea what he has in mind.

For example, one might be asking whether it would be rational (fitting) for Michael to use decision procedure Y. But then the global consequentialist gives an incorrect answer to the question: the most fortunate decision procedure is not necessarily the most rational. So that cannot be the sense of 'should' that Ord has in mind.

More generally, I think, when we ask the global consequentialist, "Exactly what are you claiming here?", they will either answer with a merely evaluative fact (hence already contained within any consequentialist's axiology), or else they will say something false. If they think they can say something both new and true, I have yet to hear what it is.

Comment Box Woes?

2009-10-13T16:19:00.002-04:00

Alex notes:

perhaps this is just me - though I doubt it - but this comment box is weird. I can't move the text cursor around except by mouse, and nor can I cut, copy or paste in it

Anyone else having problems? (Please email if you somehow can't leave comments at all!)

N.B. I sometimes get an error the first time I click 'post comment' (if I'm not already logged in), but a second click always does the trick. I haven't noticed any other problems myself (using Firefox and Chrome). But if enough others are having persistent problems, I guess I'll have to shift back (yet again!) to the awful 'pop up' comments page.

Local Evaluation

2009-10-13T11:03:00.002-04:00

Any form of consequentialism begins with an axiology, which tells us the values of things, and it derives from this an account of what we ought to do. Act consequentialism (AC), for example, tells us to perform whatever action would be best. Global consequentialists want to make further claims like, "The best character to have, of those available, is the one such that your having it brings about the best outcome." This sounds to me like a tautology, hence my skepticism about whether Global Consequentialism really adds anything to AC. In this post, I want to explore the (seemingly dim) prospects for making global consequentialism (GC) a substantive view, with coherent "local" AC competitors.

Local - or, we might say, 'hegemonic' - AC is the view that non-act evaluands (such as rules and dispositions) are to be evaluated in terms of their propensity to produce goodness through actions. Harms and benefits effected by the disposition through other means (e.g. evil demons reacting directly to your dispositional state) are not relevant.

The problem is that our shared axiology already settled questions about the values of things (what things are best, etc.), or what we have reason to desire. It only left open the question of what we have reason to do. Since AC and GC agree on this latter question, it's hard to see what there is left for them to disagree about.

Hegemonic AC sounds like it is simply contradicting the presupposed axiology: claiming that a rule is 'best' when in fact it isn't (it's only best in respect of the acts produced; but it isn't best all things considered). If that's so, then we don't have a coherent competitor to GC here.

But perhaps I am building too much into the shared axiology. We might think that consequentialists merely begin with an axiology that provides them with a preference ordering over whole possible worlds. The "global" values are thus given, but there is room for dispute over more "localized" evaluations. In particular, it may seem coherent to desire ice-cream (say) without desiring that the nearest possible world in which you eat ice-cream obtains. Similarly, the hegemonic Act Consequentialist might think that the most desirable dispositions are those that conduce to desirable actions, compatibly with the shared axiological fact that some other disposition would bring about a better outcome.

This would be a bizarre and unmotivated view, of course, but our question is whether it is coherent. (It's okay if GC turns out to be obviously superior to its competitors. The worry is that it might be trivial, and so not have any coherent competitors at all.)

In fact, I do not think this view is even coherent. The ice-cream example is misleading, because it begins with a merely 'pro tanto', some-things-considered desire. Ice-cream may be appealing in respect of its taste, for example, but all things considered you would prefer not to eat it because of the consequences for your health. Hegemonic AC is not making such innocent claims as these. After all, any consequentialist will agree that a disposition to produce good actions is good in that respect. (That is surely entailed by our shared axiology.) The dispute is over whether it is good, or desirable, all things considered (and not just in respect of its intrinsic properties, or in any other limited respect). And it does not seem coherent to say both that something is desirable all things considered, and yet that it would be bad (undesirable) were it to obtain. Just as to believe is to 'believe true', so to desire (all things considered) is to 'desire that it be the case'.

So I conclude that there is not any coherent 'local' or hegemonic version of AC that is a competitor to GC. So GC adds nothing substantive to AC.

Consequentialist Agents: Fittingness and Fortune

2009-10-12T18:08:00.012-04:00

[This is something of a manifesto for my current research project...]

Critics of consequentialism often object to how a consequentialist agent would (allegedly) think. They claim that the consequentialist agent is, in some sense, a bad character. Defenders of consequentialism typically dismiss such objections by citing the distinction between 'criteria of rightness' and 'decision procedures'. (Utility provides the criterion that determines the moral status of an act; it's a further question whether agents ought to attempt to calculate utilities themselves.) This is not entirely satisfactory. There remains a real objection here that needs to be addressed, not just dismissed. As I will explain, consequentialists still need to say something about what a 'rational' or fitting moral (consequentialist) agent would look like -- and when they do, this leaves room for others to object that the agent thus pictured is not in any sense morally 'rational' or non-instrumentally ideal.

To begin, we must distinguish two very different kinds of normative evaluation: the 'fortunate', and the 'fitting'. On the one hand, we can ask whether an agent's psychology is recommended by the normative theory as something to aim at -- roughly, whether it is desirable, or ought to be pursued, or such like. This is to ask whether it is a good or fortunate psychology to have. On the other hand, we can ask whether the agent's psychology embodies or "fits with" the normative theory -- roughly, whether the agent is responsive to the reasons posited by the theory: whether he desires what the theory says is desirable, etc. This is to ask whether the agent is, in a sense, rational or (as I will say) fit.

This distinction is illustrated in cases of 'rational irrationality', where the best disposition to have is one that embodies irrationality. Parfit's threat ignorer (for example) is irrational, but his psychology is rationally recommended or fortunate, since by being intrinsically defective in this way the agent is more likely to attain rational goods (he will no longer be vulnerable to threats or blackmail).

Similarly, a hedonist might say that only desires for pleasure are fitting, but it's fortunate -- better achieves the goal of pleasure -- to have other desires besides. Hedonists will think there's some sense in which an agent with other desires is rationally defective: they're desiring things which don't really warrant desire, after all. But, they'll say, it's fortunate to be defective in this way.

Finally, in case of ethics, we should likewise distinguish 'morally fortunate' from 'morally fitting' character. The fortunate character is that which serves to promote the good. The fitting character is that which embodies an orientation towards the good. This is the sense in which someone might have "good intentions", even if the intention has bad consequences, and so is unfortunate. Talk of "virtuous" character also plausibly concerns the 'fitting' mode of evaluation.

To aid your intuitive grip of the distinction, we can identify two families of evaluative terms. In the first family, we find terms like 'desirable', 'fortunate', 'good [on net]', and their opposites. These mark a kind of evaluation that is at least partly instrumental. In the second family, we find terms like 'rational', competent, virtuous/vicious, fitting/perverse, well-meaning, 'well-functioning'/'defective', and perhaps 'intrinsically good'. I should emphasize that, while these terms mark a kind of intrinsic evaluation (whether an agent is rational, virtuous, etc., does not depend on the outside world), one needn't think that there is any value to being in this fitting state. That's a substantive axiological question.

We're now in position to distinguish two anti-consequentialist objections. One claims that the fitting consequentialist psychology is unfortunate or 'self-defeating'. This is a very poor objection, as I explain in my old post, 'What's wrong with self-effacing moral theories?'

But when deontologists complain about the bad character of a committed consequentialist agent, there is something else that they might mean. They might mean that the fitting consequentialist psychology is (contrary to the consequentialist's claims) not actually morally fitting. For example, they argue that the "ideally rational/virtuous" consequentialist agent is incapable of friendship or commitment to projects -- but, they add, this seems like an intrinsic defect: surely genuine virtue and rationality are not incompatible with these important goods. So, they conclude, the consequentialist's conception of rationality (virtue, fittingness) must be in error.

This objection is the real challenge. Consequentialists have typically neglected it, because they have focused exclusively on evaluations of fortunateness. They haven't appreciated that their theory also commits them to a conception of the morally fitting agent. To take up this challenge, we must either (i) bite the bullet and insist that what the deontologist identifies as moral 'defects' are not really so, or else (ii) argue that, properly understood, the fitting consequentialist agent would not in fact possess the identified defect. (See, for example, my response to Stocker: 'Satisficing and Salience'.)

Of course, the first step towards a solution is recognizing that you have a problem.

[One might understand the developers of 'indirect' or 'sophisticated' consequentialism as working in this vein. But they have not always been clear about whether their theory commends the 'indirect' decision procedure as fitting or merely fortunate. Hence my previous post exploring the relation between sophisticated consequentialism and 'rational irrationality'.]

Making Suicide an Option

2009-10-09T14:10:00.005-04:00

David Velleman offers a fascinating argument in 'Against the right to die'. Sometimes options can reduce our autonomy, e.g. by reshaping the 'meaning' or social significance of our choice. You might most prefer the option of staying home by default, but once someone invites you to dinner, this option is replaced by the very different (and less desirable) option of staying home by refusing their invitation.

Similarly, Velleman argues, establishing a 'right to die' (as per physician-assisted suicide or voluntary euthanasia) replaces the option of staying alive by default with the significantly different option of explicitly choosing to stay alive. Velleman continues:

[I]f others regard you as choosing a state of affairs, they will hold you responsible for it; and if they hold you responsible for a state of affairs, they can ask you to justify it. Hence if people ever come to regard you as existing by choice, they may expect you to justify you continued existence. If your daily arrival in the office is interpreted as meaning that you have once again declined to kill yourself, you may feel obliged to arrive with an answer to the question, "Why not?"

There's something intuitively distressing about this, though it's surprisingly difficult to pin down. After all, isn't it a good thing to conform one's actions to what's justified? I see two potential worries here. Firstly, demands for justification can be experienced as burdensome in a way that threatens to suck the joy and value out of life if they become too ubiquitous. But perhaps situations of terminal illness and the like are sufficiently rare and high-stakes that moral mindfulness here would be justified on rule utilitarian grounds.

The second concern, and Velleman's main worry, is that others might respond unreasonably in ways that change what the terminally ill patient has most reason to do. Suppose that the patient could still find value in his continued existence, so long as he could continue to interact with his friends and family in a happy and dignified manner. "Unfortunately," Velleman says, "our culture is extremely hostile to any attempt at justifying an existence of passivity and dependence."

[The patient] can therefore assume that his refusal to take the option of euthanasia will threaten his standing as a rational person in the eyes of friends and family, thereby threatening the very things that make his life worthwhile. This patient may rationally judge that he's better off taking the option of euthanasia, even though he would have been best off not having the option at all.

A curious thing about this argument is that it seemingly depends upon most people rejecting it. If everyone shared Velleman's positive attitude towards care-dependent life, then there would be no such problem of misguided social pressure to worry about. On the other hand, those who consider the terminal patient unjustified in extending his life might not be so bothered by the outcome of his feeling pressured towards doing what is (by their lights) the right thing.

Or would they? Matters are complicated by the fact that the (dis)value of the outcome is massively dependent upon the patient's attitudes. So, even if it would be best if the patient were happy to die now -- leaving more resources for his grandkids' college funds, or for other citizens with curable health problems -- our preference ordering might change upon learning that the patient isn't yet ready to die. For we may plausibly think that it's so egregiously bad for a person to feel pressured into a premature death (whether from social pressure or an alienating sense of 'duty') that this disvalue outweighs the costs of continued terminal care.

This seems pretty plausible to me. No loving family could possibly want Grandma to die before she was ready, just to save a few lousy bucks. On the other hand, if she has gotten all that she can out of life, and cares deeply about providing her descendants with the opportunity for a better life, then she might quite reasonably prefer to die now rather than deplete her grandchildren's college funds -- even if her experience is still intrinsically pleasant enough (such that she would have happily continued it a bit longer if there had been no such cost to doing so).

Such an attitude seems entirely admirable, and even perhaps 'ideal'. But if we acknowledge this, it leaves us in the awkward position of seeming critical of the terminal patient who clings to life, and hence of potentially depriving them of the respect that makes their life still worth living at all. Tricky. (Though this problem doesn't seem altogether inevitable, given that an ideal may be supererogatory, not 'required' or expected.)

Then again, the assumption that continued existence is a "given" seems so deeply ingrained in our ordinary thoughts that Velleman's worries may well be moot. Even if our political institutions introduced a 'right to die', would that really make it such a live question that we would come to reconceptualize of living as an explicit 'choice' that needed to be justified? The default expectation may persist even as alternative options are provided for those who want them.

P.S. It is said that old Eskimos, when they felt their time had come, would go for a "long walk" out into the snow. Does this alternative cultural expectation seem worse? Would it likely be experienced by many as a burdensome, alien 'duty' as they approached old age? Or, if internalized to seem an entirely appropriate and even welcome way to conclude one's life, might it even be more dignified than the drawn-out hospitalized endings more common in our culture?

When considering this topic, it does seem essential to ask both:
(1) What is the best option now, holding fixed local attitudes towards death?
and
(2) What would be the best attitudes towards death for a society to inculcate?

Answers welcome...

Are Sophisticated Consequentialists Irrational?

2009-10-07T18:15:00.005-04:00

The Railtonian 'sophisticated consequentialist' (SC) acquires, maintains, and acts upon whatever first-order desire set would be (expectably) best. So, if a bias towards one's own family is a fortunate disposition to have, SC will possess this and hence be biased towards his own family, rather than impartial like a direct utilitarian. My question: is SC rational?

The mere fact that a disposition is expectably fortunate (hence rational to acquire and maintain) does not entail that the disposition is rational (in the sense of manifesting 'sensitivity to reasons' or 'an orientation towards the good') or that it disposes you towards rational actions. As seen in Parfit's case of transparent "threat-fulfillers" and "threat-ignorers", it can be rational to make ourselves irrational. So, we may wonder, is SC like Parfit's threat-ignorer: someone who is now rationally impaired - aiming at the 'wrong' things, or unresponsive to certain reasons - though their past self did the right thing in bringing about this impairment (given its fortunate consequences)?

Perhaps so. The main reason to conclude that SC is irrational is that his dispositions are desirable for reasons other than the actions they produce. (Otherwise there would be no need for such 'sophistication' -- he could just be a straightforward act consequentialist with 'rationality-enhancing' dispositions, as described here.) And the possession of dispositions whose value depends on their external (non-act) effects is, we may think, the sign of fortunate irrationality. In this respect, SC is like the pure threat-ignorer.

But there are two notable differences:

(1) 'Rational irrationality' typically requires self-manipulation (e.g. by pill or brainwashing), whereas the 'sophisticated' psychology is meant to be acquirable by "natural" psychological processes. Railton gives the example of a competitive tennis player who learns that he would play better if he were less focused on winning, and more focused on enjoying the game for its own sake. It seems that his aim of winning can guide the development of this new intrinsic (but contingent) desire. This may in part be due to the fact that the content of the new desire (viz. to play and appreciate the game) is consonant with the overarching aim of winning the game. If the player had to acquire a desire to lose in order to better win, that does not seem possible without external manipulation.

So, there are at least important constraints on the Sophisticated Consequentialist psychology. SC cannot acquire just any old fortunate dispositions: even if a malicious desire would somehow prove fortunate, the cognitive dissonance would be too great to maintain this in virtue of its fortunate character. (He could take a pill to make himself just plain malicious, but that is to make himself no longer a consequentialist, just like the threat-ignorer.) But so long as SC only acquires the sorts of motivations that can be naturally maintained consistently with his overarching consequentialist aim -- so, perhaps, special motivations to attend particularly to certain goods, such as the welfare of his friends and family -- then we may think that he is at least free of any gross irrationality. It's not like he ever aims at the bad. At worst, he fails to aim at every good to the precise degree that is warranted. But this is a fairly modest rational failing, and one that needn't require any kind of self-deception (again, unlike the threat-ignorer).

(2) The second, and much more straightforward difference, is that SC continues to regulate his dispositions in a way that the threat-ignorer (TI) does not. In a world where dispositions are transparent, threat-ignoring initially has high expected utility, since the threat-fulfillers should realize there's no point in even trying to threaten the ignorers. TI can expect to remain safe. But now consider what happens when a threat-fulfiller slips up and stupidly threatens to blow up the world unless TI shines his shoes. At this point, the expected value of TI's disposition is extremely low. Since he maintains his disposition unconditionally, he ignores the threat, and -- predictably enough -- the world blows up. Disaster.

The Sophisticated Consequentialist would never bring about disaster in this way. He acts on the dispositions that are (expectably) best to have. But if circumstances change, then so do his dispositions. In this sense, SC as an agent remains rational (responsive to reasons), even if his particular actions aren't so much.

(Of course, his being rational in this way leaves him vulnerable to extreme threats in a way that the grossly irrational agent is not. But that's fine: I'm not claiming that it's always best to be a sophisticated consequentialist -- it may not be. My aim is just to clarify in what respects such an agent exemplifies rationality.)

Callous vs. Caring Consequentialists

2009-10-04T21:42:00.003-04:00

My recent post on 'Desirable vs. Rationality-Enhancing Dispositions' contains an important mistake. I began by distinguishing 'rationality-enhancing' dispositions (desirable in virtue of conducing to good actions) from dispositions that have other good effects. I then suggested, mistakenly, that rational status is always transmitted from rationality-enhancing dispositions to the particular actions they dispose one to perform. This transmission principle holds only when the 'rationality-enhancing' disposition serves as a generally reliable 'rule of thumb', like the rule against killing people even when murder might (prima facie) seem to promote utility. But in non-ideal agents, a disposition might also conduce to good actions for a very different kind of reason, as I will illustrate below.

Meet Cam, a callous consequentialist. Cam is one of those utilitarians who likes humanity but not people so much. Due to his lack of regard for those around him, he tends to act insensitively, and makes other people (not least his poor family) miserable. Upon reflection, Cam recognizes this to be unfortunate. So he takes a pill that makes him a much more caring and loving person. He is now disposed to attend to the welfare of those that are close to him. This causes him to act in much better ways: in particular, he finds it easier to refrain from making the kinds of insensitive remarks that previously caused so much harm. The one downside is that he is now much less inclined than before to donate to charities like Oxfam that promote the impartial good. He would rather spend that money on his family. This is bad, but (let's suppose) not nearly as bad as Cam's callous actions had been.

The structure of the case is that Cam was previously weak-willed in a very bad way. He then acquired a disposition that helped him overcome this weak will, and so perform better actions -- though at the cost of acquiring a new (less bad) weakness. Because it is clear (by stipulation) which of his newly-disposed actions are better than before and which are worse, the epistemic argument for following a generally beneficial disposition no longer applies (in the 'clearly worse' case). So, it's right for Cam to acquire this disposition in virtue of the actions it conduces to in general, but some particular actions it conduces to may still be considered wrong (or less than perfectly rational/beneficent). We can thus have cases of what Parfit would call 'blameless wrongdoing' without having to appeal to dispositions that have good effects besides action. (Of course, the disposition may also be good for other reasons -- the point is just that my case doesn't rely on this.)

Hallucination, Virtual Reality, and Reality

2009-09-28T20:23:00.007-04:00

Some consider virtual reality to be akin to mere hallucination, whereas others see it as potentially on a par with physical reality. Both seem possible to me (depending on the precise nature of the VR we're imagining), so let's try to clarify matters by drawing some distinctions. First, we can consider the epistemic dimension of whether you remain aware (on some level) that you're in a virtual world, or if the VR is all-consuming. More importantly, I think we can identify variation along a metaphysical dimension of sorts, as per the following three alternatives:

(1) Passive Experience Machine: You are passively fed an externally determined 'phenomenal soup' (including the phenomenal experience as of making certain decisions), effectively living out someone else's story from the inside.

(2) Active Solipsism: A genuinely interactive fiction, where you make decisions that affect how things turn out. But you are the only sentient being immersed in the world -- any others you "see" are mere simulacra (or 'Non-Player Characters', in geekspeak).

(3) Active Shared Virtual Worlds: In this final category, the virtual world serves as a medium for causally interacting with other people, as in The Matrix.

When it comes to assessing a life spent immersed in such VR, I think these differences are of immense normative significance. The Passive Experience Machine is indeed akin to an extended hallucination, and a life so bereft of agency may strike many of us as no life at all. The case of Active Solipsism is at least some improvement on this, though still abhorrent insofar as we are social animals who value genuine relationships, and see them as grounding much of the meaning in our lives.

What about the final option: Active Shared VR? In principle (i.e. if the VR faithfully reproduced all the multi-modal sensory richness and fine-grained environmental responsiveness that physical reality has to offer), I think that such a world must be acknowledged as no less 'real' - in any sense that matters - than our own. It seems a piece of metaphysical chauvinism to claim otherwise: to think that when lovers intentionally cause mutual sensations of kissing this only counts as really kissing when there aren't any bits or bytes (but only atoms serving in a non-computational capacity) in the causal chain. (Compare the absurdity of claiming that people on the phone aren't really talking to each other, but merely "simulating" talking.) What's so special about material reality? It only matters insofar as it provides a common causal medium for the interaction of minds against a stable backdrop; but any other equally-responsive and stable medium could fill this role just as well -- its intrinsic nature, as material or computational, cannot plausibly be thought to matter.

'Reality' is just the stable causal backdrop for interacting minds. As such, it is multiply-realizable. Material stuff can do the job of reality, as can computational bits and bytes (at least in principle), or we could even have a kind of Berkeleyan Idealism according to which our existence is fundamentally based in the mind of God. So long as the requisite stability and mutual (counterfactual-supporting) causal influence obtains, the fundamental grounds don't matter. Our everyday concepts, and hence the contents of our desires, typically concern the surface structure of reality (that which the materialist and idealist worlds have in common), not its fundamental nature. So this is not a difference that would make a difference, so far as most of us are concerned.

(N.B. It would make a difference if the causal structure mediating our experiences were to be excised or impaired. I'm no subjectivist: a world where people have no causal impact on each other's experiences is a world that's sorely lacking.)

So, I agree with Alexandre Erler that "there is an important difference between my actually going to see the pyramids of Egypt, or having sex, and my sitting alone in my house in an armchair with electrodes plugged into my brain... believing I am doing these things." But then, there's also an important difference between merely believing that you're doing these things, and actually doing them in a different medium. This distinction is occluded if we fail to carefully distinguish the different forms that a 'Virtual Reality' could take.

Marginally Beneficial Rule-breaking

2009-09-27T19:27:00.008-04:00

Most everyone agrees that you should break the rules if that's the only way to avoid disaster. But it seems intuitively objectionable that Act Consequentialism tells us to (say) break a promise whenever doing so would be even the slightest bit better than keeping it. Well, maybe. I agree that there's something troubling about the agent who breaks a promise the moment it seems like there's something (marginally) better he could do. But such an agent is not, I will argue, what a competent act consequentialist would look like.

As I argued in 'Defective Deliberateness', competent agents can't be constantly deliberating. In addition, we must recall that overt calculation often goes awry. So the competent Act Consequentialist largely relies on rationality-enhancing dispositions and rules of thumb in his everyday life, only pausing to reflect when his well-calibrated sub-personal mechanisms alert him to the need (say due to complex novel circumstances, that his "auto-pilot" wasn't designed to deal with). Everyday promise-keeping is not exactly novel, so for the competent agent the question whether to keep the promise shouldn't even arise. It's a no-brainer.

(This is not necessarily because it's always clear that keeping promises is objectively for the best. But it typically is for the best, and on the odd occasion where it isn't, this almost certainly won't be clear. In that case, the possible benefit from breaking the rule is so marginal that it generally won't be worth the cognitive costs of attempting to assess the precise balance of reasons.)

But suppose the agent comes to consider the question anyhow. What should he conclude? We can stipulate that in fact the outcome would be marginally better if he broke his promise, but does the agent himself have any way of knowing this? Not easily, at least. (Among other things, he'd need to first consider the possibility of self-serving bias corrupting his judgment, and weigh the apparent benefits of rule-breaking against the long-run value of retaining a reputation for trustworthiness.) Maybe if he heard the booming voice of God reassuring him of this fact, then he could rationally go ahead and break his promise without further worry. But in ordinary circumstances -- as we're supposed to be concerned with here -- it's simply never going to be clear when rule-breaking is marginally beneficial. So the agent is faced with an immediate choice: he can (i) break the rule even though it's unclear to him whether any good would come of this; (ii) sink further cognitive resources into investigating a question that he probably shouldn't have bothered to ask in the first place; or (iii) simply keep his promise and turn his attention to more important matters. It seems pretty clear that, in this sort of case, option (iii) is the way to go.

In sum: breaking a rule will only be clearly worthwhile in cases where it is also of significant benefit (in which case we all approve of rule-breaking anyway). If it's only of marginal benefit, this fact typically won't be clear enough for a rationally self-doubting agent to confidently act on it. And the low potential payoff means that it isn't really worth inquiring further: better just to stick with the generally-reliable rule of thumb. So a rational act consequentialist generally won't be found engaging in marginally beneficial rule-breaking after all. They'd even share our intuition that there's something awfully dubious about any agent who would act that way.

This seems to me to defang the original objection. What do you think?

Reasons and Rule Consequentialism

2009-09-24T13:35:00.004-04:00

It can be useful to formulate moral theories in terms of their implications for normative reasons, since this brings into view their substantive commitments. For example, I recently argued that this methodology allows us to deflate global consequentialism into mere act consequentialism. I now want to see what it can tell us about rule consequentialism.

As with any consequentialism, we begin with some theory of the good (equivalently - if we accept a 'fitting attitudes' analysis of value - what's desirable). The distinctive thesis of rule consequentialism (RC) is then to derive our reasons for action from the recommendations of the best rules, rather than just straightforwardly recommending the best action. RC thus implies that the act we ought to perform may not be the most desirable act. It may be that we should keep a promise, while hoping that we won't (and immediately regretting that we did). This seems odd, and perhaps even incoherent.

Alternatively, the Rule Consequentialist could hold that the fact that the best rules recommend φ-ing makes the action not just actionable ("right"), but also desirable. At this point we may question whether the view is still recognizably consequentialist in nature. It seems to be treating reasons for action ('the right') as prior to reasons for desire ('the good'), which one might reasonably take to be diagnostic of deontology. What the rule consequentialist now treats as fundamental is not value generally, but merely a component of value: welfare, say. Their more general value theory also accords intrinsic value to acting according to the welfare-maximizing code. This also seems odd.

The Rule Consequentialist might respond by restricting the fitting attitudes analysis of value. Much as some theorists restrict 'value' to objects that provide us with agent neutral reasons for desire (even though they think we may have agent-relative reasons to desire other things in addition), the rule consequentialist might analyze 'value' in terms of fundamental reasons for desire, i.e. reasons that are not derived from reasons for action. They could then maintain that only welfare is valuable, and hence that their code is selected on the basis of value generally, compatibly with maintaining that it's desirable to perform disvaluable acts that are in accordance with the valuable code. Perhaps this is a better-sounding way for them to word it. But whatever words we use to gloss the view, it is the fundamental claims about normative reasons that we need to evaluate for their plausibility.

Which option is more plausible? Should rule consequentialists take themselves to be advising undesirable actions? Or would they have us revise our preference ordering over possible worlds, in such a way that their 'ideal code' is selected not on the basis of all-things-considered desirability, but only on the basis of a restricted set of our reasons for preferring some outcomes to others? Neither option seems especially appealing...

Rule Consequentialism and Changing Circumstances

2009-09-24T12:57:00.002-04:00

Suppose an evil demon rules over a world for the first of two epochs. In this first epoch, he tortures anyone who fails to internalize a stringent rule requiring them (unconditionally) to greet others by punching them in the guts. (To be clear, the demon doesn't mind if by some psychological fluke this disposition fails to manifest in action. He just wants to make sure that everyone has the wicked disposition.) In the second epoch, the demon leaves the people to their own devices. Then, at the end of the second epoch, the world will end. These facts are common knowledge. What should the people do?

The best outcome is presumably for everyone to (i) initially internalize the gut-punching rule, (ii) hopefully fail to act on it (whenever possible without undermining the disposition), and (iii) change their internalized rules once the demon goes away.

Rule-consequentialism clearly agrees with (i). It also implies, contra (ii), that agents ought to act according to this rule, i.e. they ought to punch each other in the guts, even though no good comes from this. But what does it imply about changing circumstances? Once in the second epoch, does the society get to rewrite the rules, so that what was previously obligatory is now forbidden? Does it make any difference whether the demon had also incentivized our internalizing a rule according to which we should never question or revise his rules? (Though initially beneficial, the changing circumstances would mean this rule, too, no longer has high expected value in the second epoch.)

Suppose that the people can predict that they will end up changing the rules in these (forbidden but) beneficial ways. Can their current moral code still condemn those future actions as "impermissible"? Or do contemporaneous codes trump? (Compare my old 'Reflecting on Relativism'.)

Desirable vs. Rationality-Enhancing Dispositions

2009-09-21T13:09:00.004-04:00

We previously noted that dispositions (e.g. internalized rules) can have other consequences besides producing downstream acts in the agent herself. (In particular, other agents might harm or reward you directly on the basis of whether you possess some disposition.) This suggests that we can distinguish (i) dispositions that have high expected value, all things considered, and (ii) dispositions that have high expected value in respect of the downstream actions they'll tend to produce. We can call the former class of dispositions 'desirable', and the latter 'rationality-enhancing' (on the assumption that rational actions are those that maximize expected value).

These two classes will tend to overlap, but not perfectly, which leaves room for the possibility of "rational irrationality". We can acknowledge that the mere fact that a disposition would be rational to acquire (because desirable all things considered), does not in itself guarantee that acting on the disposition is rational. For there are situations in which it would be (antecedently) rational to bring it about that we are (subsequently) irrational.

While acknowledging this possibility, we may still think that there must be some "transmission" principles according to which the rational status of a general rule can be inherited by the particular acts it prescribes. And, indeed, the distinction I've highlighted suggests an obvious candidate principle: we just need to restrict the rules in question to those that are 'rationality enhancing', i.e. desirable for their (expected) impact on your downstream actions, rather than for extrinsic reasons. [Update: this isn't quite sufficient, as explained here.] Consider:

If S rightly adopts a rule R as (maximally) 'rationality-enhancing' (in my stipulated sense), and R prescribes φ-ing in circumstance C, then when S is in circumstance C, S rationally ought to φ.

This is supported by considerations of meta-coherence. Ex hypothesi, the rule R offers the most reliable guidance available to S -- in particular, it is more reliable than attempting to autonomously determine what the best result would be in each case. (And it is also more reliable than any identifiable alternative, e.g. "following R except in circumstances with the subjectively distinguishing feature F.") So, any given departure from R can be expected to have worse results than would be obtained by following R. So in any given case, the agent should follow R's advice.

My main cause for hesitation here comes from cases where you adopt a rule as a hedge against (as it happens, misleading) evidence that you might be biased in your subsequent judgments. So the rule might tell you to disregard certain first-order evidence, because you can't be trusted to evaluate it rationally. But if you actually are capable of evaluating it rationally, then we may think that there's an important sense in which you rationally ought to be guided by the (first order) evidence. Or, even if the higher-order evidence makes some contribution, we may still doubt the radical claim that it completely swamps the first-order evidence in determining what you rationally ought to do. Anyway, this raises a tricky epistemological issue, which I set out in more detail in my old post, 'Personal Bias and Peer Disagreement'.

Foot on Courageous Wrongdoers

2009-09-18T20:59:00.002-04:00

We might think of words such as ‘courage’ as naming characteristics of human beings in respect of a certain power, as words such as ‘poison’ and ‘solvent’ and ‘corrosive’ so name the properties of physical things. The power to which virtue-words are so related is the power of producing good action, and good desires. But just as poisons, solvents and corrosives do not always operate characteristically, so it could be with virtues. If P (say arsenic) is a poison it does not follow that P acts as a poison wherever it is found. It is quite natural to say on occasion ‘P does not act as a poison here’ though P is a poison and it is P that is acting here. Similarly courage is not operating as a virtue when the murderer turns his courage, which is a virtue, to bad ends. Not surprisingly the resistance that some of us registered was not to the expression ‘the courage of the murderer’ or to the assertion that what he did ‘took courage’ but rather to the description of that action as an act of courage or a courageous act. It is not that the action could not be so described, but that the fact that courage does not here have its characteristic operation is a reason for finding the description strange.

-- Philippa Foot, 'Virtues and Vices', p.16.

Defective Deliberateness

2009-09-17T14:22:00.005-04:00

Consequentialism tells us that the correct answer to the question what to do is simply whatever's best. Critics object that attempts to calculate what's best often turn out badly. This is often taken to imply that consequentialism is self-effacing: it tells us to ignore its guidance, and forget that it is true. But this is too quick. While it's always an empirical possibility that it'd be most rational to induce irrationality in ourselves, it's less clear that our ordinary non-calculative dispositions are examples of this. That is, I would question whether they must be thought of as manifestations of irrationality (anti-consequentialism).

The crucial point here was developed in my recent discussion of 'Satisficing and Salience', with the distinction between how to answer a question vs. whether to ask it in the first place. To illustrate, begin with a moral datum: it's regrettable that the Great Pyramids were built with slave labour. A rational agent must be responsive to these reasons for regret. But that is not to say that she must always be actively responding to these reasons, for that might distract her from more pressing tasks. So the rational agent need not actively regret everything that is regrettable. Granted, she must be disposed to answer the question correctly when it is raised, so to speak. (She'll feel regret if she attends to this aspect of history.) But it's just a mistake to think that rational agents must constantly be raising the question. Most of the time we needn't think about the regrettable aspects of ancient history at all.

I now propose that we may find something similar with regard to the question how to act. Most of the time we act without conscious deliberation (let alone calculation). That doesn't make us implicit anti-consequentialists, any more than our everyday inattention to Ancient Egyptian slavery makes us implicit slavery apologists. You can't be accused of answering a question incorrectly when the question never arose in the first place. (This is a bit quick: we may also be interested in what sorts of considerations in your immediate environment you're disposed to respond to, for purposes of character evaluation. Someone disposed to neglect or downplay others' suffering - even when it's right before their eyes - could hardly be considered a utilitarian in their heart of hearts.)

Anyway, just as it would be a rational defect of sorts to counter-productively obsess over ancient slavery, so it would be defective to be constantly questioning yourself about what you ought to be doing right now, and now, and now, and now... A significant degree of default trust in your ordinary functioning is required if you are to have any hope of competently acting in the world at all.

So, nobody should think that constant deliberation about what to do is necessary for (or even compatible with) being a well-functioning rational agent. The excessively deliberate agent is defective. This fact about when (not) to deliberate holds independently of our substantive views about how it's rational to deliberate when deliberation is called for. So the mere fact that we shouldn't constantly be calculatively deliberating is not in itself any kind of argument against maximizing consequentialism.

Of course, we can develop a more promising argument out of this. There's no denying that sometimes deliberation about what to do is called for. So we can ask what sort of deliberation would (on those occasions) be rational according to consequentialism -- and then assess whether that answer is plausible. But that's a task for another post. For now, I simply want to establish that we shouldn't understand the consequentialist agent as always having "one thought too many". In most cases, they -- like anyone -- won't think at all before acting.

Acts and Meta-Acts

2009-09-16T14:14:00.004-04:00

In 'Can Consequentialism Cover Everything?', Bart Streumer claims that according to consequentialism "there is no normatively significant difference between performing an act and bringing it about that an act is performed." This isn't quite right. It would be more accurate to say that the pro tanto reasons for and against φ-ing count equally for or against bringing it about that one φs. But we must leave open the possibility of the latter ("meta") action being influenced by other reasons in addition. After all, this act may have additional consequences besides just the consequences of the downstream act of φ-ing, for there's no guarantee that all else is equal between the two acts in question.

Standard cases of "rational irrationality" illustrate this nicely. If faced with the prospect of blackmail in the near future, I may have reason to cause myself to become utterly nihilistic and indifferent to the welfare of my family, so that the blackmailer has no incentive to threaten them harm. It doesn't follow that my later nihilistic actions are themselves rational. There's a normatively significant (consequential) difference between φ-ing in circumstances C and bringing it about that I will φ in circumstances C.

Newcomb-like problems might provide another instance of this divergence. Prior to the Oracle making her prediction, I have most reason to bring it about that I one-box. (This increases my chances of winning the bonus million dollars.) But once faced with the actual decision, I arguably have most reason to take both boxes (since whatever is in the first box, taking the second in addition will net me an extra $1000).

Finally, for a more pedestrian example, contrast (i) A thief's freely deciding to return a stolen paperclip, versus (ii) The police department's expending resources to bring it about that the thief returns the paperclip. Obviously there's a consequential difference here, namely that the second action involves additional costs. Although it would be good for the thief to return the goods, it may not be worthwhile for anyone else to expend the resources necessary to bring this about.

Similar remarks apply to other evaluands besides acts. The best character to have might be too costly to be worth bringing about. This explains why, contra Streumer, we should not interpret Global Consequentialism as implying that we ought to bring about 'the right X' (whether X stands for acts, characters, climates, or whatever).

Reasons Deflate Global Consequentialism

2009-09-14T15:59:00.004-04:00

I'm beginning to think that global consequentialism isn't really a distinctive position in its own right. It may be better understood simply as act consequentialism that takes care to not make straightforwardly false claims about the consequential value of other evaluands (like rules, motives, etc.). Global consequentialists recognize that these other evaluands might bring about good or bad consequences by "external means" other than the agent's own downstream actions. But the denial of this is not anything so principled as an opposing position. It's just an oversight.

Here's the structure of my argument. I take as a premise the assumption that substantive normative claims must fundamentally concern normative reasons (of one kind or another). The core of my argument thus hinges on identifying what categories of practical reasons there are, about which consequentialist theories may then make substantive claims. When we do this, I argue, we find that there is nothing more for global consequentialism to be besides act consequentialism that avoids silly oversights.

Okay, let's begin. There are normative reasons for all sorts of judgment-sensitive states: beliefs, emotions, etc. But core moral theory is concerned with two categories in particular, the right and the good. 'The good' we may understand as a matter of desirability, or fitting reasons for desire. 'The right' concerns reasons for action. Consequentialist moral theories are those that treat the former class of reasons as having logical priority. Our reasons for action derive, in some sense, from considerations of desirability (value). Different consequentialist theories offer different derivations. Act consequentialists claim that our reasons for action derive directly from the value of so acting. Others (e.g. rule or motive consequentialists) suggest a more circuitous route, claiming that reasons for action instead derive from the value of possessing (say) the character traits that would dispose one to perform the action.

Global Consequentialists agree with Act Consequentialists about our reasons for action. (The right act is simply the one that's best.) But then they purport to make further claims, which is where matters become puzzling. As Pettit and Smith define the view:

Global consequentialism identifies the right x, for any x in the category of evaluands -- be the evaluands acts, motives, rules, or whatever -- as the best [available] x, where the best x, in turn, is that which maximises value.

We previously saw that 'the right act' can be understood in terms of reasons for action. But what does it mean to call other evaluands, e.g. character traits, 'right'?

It is important to note that by (say) 'the right character', Pettit and Smith do not mean 'character such that the agent has decisive reason to act so as to bring it about.' The claim is not about reasons for action. It is just a claim about reasons for desire: the 'right' character is the best (most desirable) of those available. (Note that it may be desirable that Bob have some character trait X, without Bob thereby having sufficient reason to act so as to bring about this state of affairs, since the act may have costs -- including opportunity costs -- which render it not the best act available to him at this time.)

But this claim about desirability is not substantive. It's just a stipulation about how to apply the term 'right' to other evaluands. The fundamental normative theory differs not a whit from act consequentialism. Consider: Act consequentialism presupposes a theory of the good (desirability), and makes one new substantive claim: we have most reason to perform the act that would have the most desirable outcome. Global consequentialism agrees with this, and adds a bunch of stipulative claims in addition: e.g. we can call those character traits 'right' that would have the most desirable outcomes. But there's no new substantive claim here.

Anyone can happily assent to a tautology. A fortiori, Act Consequentialists can happily assent to the claim that we have most reason to desire that people have those traits [of those available] which are such that we have most reason to desire that people have them!

For Global Consequentialist claims (e.g. about the rightness of character traits) to be substantive in the way that Act Consequentialist claims about the rightness of actions are, it would have to be the case that we have basic "reasons for character" like we have reasons for action. But we don't. You can't reason your way to having a character trait. At best, you can reason your way to wanting to have a certain character trait, or else to acting so as to bring it about that you have a certain character trait. There are reasons for action and for desire, in relation to character traits, but there are not reasons directly "for" character traits.

Global consequentialists want to be consequentialists about everything. Since consequentialists treat value - desirability - as fundamental, to be a consequentialist about everything would be to treat all other kinds of reasons as deriving from reasons for desire. But the only other kind of reasons (within core moral theory)* are reasons for action. Given this limited domain available for consequentialists to make substantive claims, we find that act consequentialists are already consequentialists about everything. There's really nothing more to be a consequentialist about.

* = Unless one extends the view beyond the domain of core moral theory. One might be a consequentialist about belief, for example, by claiming that our reasons for belief derive from the desirability of holding a belief. But that's an absurd view. (At least, once we take care to distinguish reasons for belief from reasons for acting upon oneself in a way that will causally produce belief.)

Painless Meat and Anthropomorphic Objections

2009-09-12T23:09:00.004-04:00

Discussed here (via Robin):

Might “pain-free” be the next sticker slapped onto a rump roast? … Progress in neuroscience and genetics in recent years makes it a very real possibility. …

One objection to the idea of knocking out pain in livestock is that it could mean they put themselves in harm’s way...

But why would such merely physical "harm" matter? It matters to humans, of course, (and any other rational beings) because we are reflective creatures for whom larger goals and life projects contribute to our well-being. But ("lower") animal well-being is, most plausibly, purely hedonistic in nature. Moreover, they lack the sorts of enduring personal identities that would make death bad for them (or so I argue). So, aside from the inconvenience (to humans) of an farm animal's pain-free premature death or injury, I don't see any real grounds for objection here.

Many New Scientist commenters argue from analogy that this must be wrong because farming and killing pain-free humans would be wrong. But again, the relevant difference is that reflective beings have a non-hedonic element to their welfare that mere animals lack.

Chancy Decisions and Actualist Obligations

2009-09-12T20:51:00.007-04:00

The Stakes of the Actualism-Possibilism Dispute

2009-09-12T17:49:00.008-04:00

I should say "disputes", since the issue of evaluating options is importantly different from that of selecting options. As Jackson & Pargetter put it in their seminal 1986 paper, 'Oughts, Options, and Actualism':

There are two matters that need to be borne in mind when considering what ought to be done in terms of the option approach. One is the evaluation of options. We have urged the plausibility of evaluating options in terms of what agents would do were they to adopt them. That is Actualism. The other matter is how to select the right set of options. That depends on exactly which question you want the answer to. If you want the answer for some action as to whether an agent ought to do it, look at the set consisting of the action and what the agent would do instead; if you want the answer as to what an agent ought to do at or during some time, look at all the maximally relevantly specific actions possible at or during that time. (p.255)

I whole-heartedly agree with their Actualism, for reasons articulated in my post 'Ignoring Reality Ain't So Ideal Either'. When possibilists say that Professor Procrastinate ought to agree to review the paper (since he could, even though he actually won't, do the job on time) they are making a substantive error in their evaluation of the token decision that Prof. P. faces at this time.

I have more difficulty grasping the point of the 'selection' debate. In a recent PEA Soup post, Ralph Wedgwood argues that it's most intuitively plausible to look at all the maximally specific actions available, not only when assessing what an agent ought to do at a time, but even when asking the more specific question whether they ought (at this time) to φ. If φ-ing is something awful, and the agent has a better option ψ available, surely they ought not to φ (even if it happens that if they don't φ, they'll do something even worse). Whereas Jackson & Pargetter see the relevant options in this case as either φ-ing or doing the worse thing, Wedgwood insists that we include ψ-ing as among the relevant options here too. This effectively collapses the question "ought I to φ?" into the question "what ought I to do? (Perhaps, φ?)"

I don't see much room for substantive debate here. Everyone agrees that, between the options of φ-ing or doing what they'd otherwise do (i.e. something worse), the agent should φ. Everyone also agrees that what the agent ought to do at this time is to ψ instead. What remains is a merely semantic question: whether to interpret the English phrase "Ought I to φ or not?" as questioning the former normative claim or the latter. Does this really matter?

Satisficing and Salience

2009-09-08T16:50:00.006-04:00

Michael Stocker (Plural and Conflicting Values, p.321) writes:

Maximizers hold that the absence of any attainable good is, as such, bad, and that a life that lacks such a good is therefore lacking. I disagree. One central reason for my disagreement stems from the moral psychological import of regretting the absence or lack of any and every attainable good. This regret is a central characterizing feature of narcissistic, grandiose, and other defective selves. It is also characteristic of those who are too hard on themselves, who are too driven and too perfectionistic.

This doesn't strike me as a good (truth-indicative) reason to disagree with the Maximizer. It may be bad, or even in a sense inappropriate ("defective"), to regret every little regrettable thing. But those things may be regrettable all the same.

How is this? Of course, it isn't news that rationally justified or warranted attitudes may be undesirable ("bad") in some circumstances. But one may find more puzzling the suggestion that warranted attitudes could characterize a kind of psychological defect. After all, aren't warranted attitudes precisely those that it would be (rationally) appropriate or fitting to have?

Well, sort of. But -- given that our minds are only finite -- we need to distinguish (i) the appropriate answer to a question, from (ii) whether a well-functioning agent would raise that question in the first place. To actively regret something requires first attending to it. So it may be the active attention, rather than the resulting regret per se, that is defective.

To illustrate: imagine that Bob is constantly regretting the slave labour that went into building the Great Pyramids. His obsessive historical sympathy interferes with the living of his life, to the point that he ends up absent-mindedly stepping in front of traffic and losing his legs to an oncoming vehicle. Years later, Bob's wife Sally just can't stop regretting that Bob lost his legs. Every time she looks at her husband, she sees his disability (and little else).

Clearly, Bob and Sally are defective moral agents. But nobody would conclude from this that slave labour and traffic accidents aren't really regrettable. Rather, the problem is that (like the immodest person whose achievements and status are excessively salient to his mind) they are attending to the wrong things. Likewise, I suggest, with Stocker's perfectionist. If it would have been better for us to do Y rather than X yesterday, then that's the correct answer to the question what we should have done. In this sense, our doing X instead was indeed regrettable. But that doesn't mean we should now regret it, because it may be that we shouldn't be raising that question at all. (Spilt milk, and all that.) This is perfectly compatible with the Maximizer's claims, for Maximization is a view about correct answers, not correct questions. (One might extend it to the latter by raising the question of "what question to raise next"; but obviously the answer to this needn't be "whatever question Maximizing Consequentialism has an answer for!")

Resolutions and Rational Bootstrapping

2009-09-06T18:02:00.010-04:00

The best rules (or dispositions) and the best actions may come apart in theory. Fortunately, a reflective agent may be able to limit the damage by means of what Ainslie (1975) calls "private side bets".

Consider Kavka's toxin puzzle. Many philosophers think that a rational person couldn't win the prize, because by the next day they would no longer have any reason to drink the toxin -- and, recognizing this, they will be presently unable to intend to so act. To be successful in situations like this, agents require the internal power to 'bootstrap' additional reasons into existence. (Note that the terms of the game specify that you're not allowed to employ external incentives to win the prize. So you can't make a bet with anyone else that would give you additional reason to follow through and drink the toxin.) Intuitively, the solution seems to be that you should just make a resolute commitment - a promise to yourself - to drink the toxin. The hope is that this internal act of self-promising will provide you with a new reason - an internal incentive - to go ahead and drink the toxin, thus allowing you to intend this action and to reap the rewards.

But can a backward-looking consideration, the fact that you made yourself a promise, really provide you with a reason to act when the time comes? Consequentialists will naturally be skeptical of this, and even deontologists may doubt whether promises to yourself can really be binding. (After all, the recipient of a promise is typically at liberty to release the promisee.) But even so, perhaps we can finagle a forward-looking reason out of this. Just consider how beneficial it is to be the sort of agent who can trust one's future selves to carry out earlier resolutions. Who knows how many more toxin puzzles or newcomb problems you'll face in future. If you renege on your self-promise now, you'll lose this valuable self-trust. Next time you sail past the Sirens, you'll have to tie yourself to the mast rather than steadfastly resisting their call. (Etc. etc.)

In short: being less trustworthy, even just to yourself, is costly. So you can bootstrap additional reasons into existence whenever you need them, just by promising yourself that you'll do this thing, thus putting your self-trust on the line. (A slightly different way to put it is that you're effectively bundling together a host of desirable actions into a single fragile disposition that you won't want to risk weakening -- depending on the precise details.)

There are obvious limitations to this approach. If the immediate gain of defecting outweighs the long-run expected cost of tarnishing your self-trust (and of weakening your general disposition to act on globally optimal rules), then you'll defect. So you won't be able to win Kafka's prize if, for example, you know that this is the last decision you'll ever make. (Cf. one-shot vs. iterated prisoner's dilemmas.)

One final move that might help further is to place some intrinsic value on being the sort of agent who sticks to resolutions and so is capable of inter-temporal coordination. I'm not sure how plausible this is, but it doesn't seem completely crazy at least. There's something appealing about this self-conception -- about seeing one's momentary selves as united in doing their bit to implement the agency of a temporally-extended whole -- rather than being self-contained islands of agency, perhaps causally related yet normatively independent and isolated from the others. [Update: In other words, the atomist adopts what Strawson calls "the objective attitude" towards himself: his other timeslices are "to be managed or handled or cured or trained", rather than to be reasoned with co-operatively.] That's not to say you should maintain your holistic self-conception at any cost, of course. But it might be a shame to give it up too cheaply.

Scanlon on Rational Revision

2009-09-06T14:06:00.004-04:00

What's the best way to introduce undergrads to the idea that one can reason about ethical questions? Suggestions welcome. In the meantime, here's a nice passage from Scanlon's What We Owe to Each Other (pp.66-67):

"What grounds can I have for deciding that I was mistaken to think that something was a reason? How can I be justified in calling this a process of correction rather than merely a change in my reaction? To answer this we need to see what such a process may involve.

Suppose, for example, that you are enraged by your child's defiant and insolent behavior. This behavior seems to you a reason to strike the child. But is it a reason? What kind of reason would it be? Is violence appropriate because it creates fear or shows your power? Why is that desirable? Is it supposed to be good for the child, or simply to demonstrate something about you? If the latter, what does that imply or signify about your relations with the child? If the former, why think that the effects will be good? What alternatives are there, and what would their effects be? The process here is first to clarify what kind of reason this is supposed to be and then to see whether the initial tendency to take this as a reason stands the test of reflection. ...

[H]ow does the process that you went through support you in taking the later judgment to be the one that is correct? Here are two reasons. First, the later conclusion is supported by a clearer and more detailed conception of what the reason in question might be -- of exactly what it is that is supposed to count in favor of striking the child. Second, in virtue of this reflection, it is less likely to be affected by distorting factors such as your rage."

Open Thread: critic's edition

2009-09-05T15:02:00.002-04:00

As before:

Sometimes, when reading a blog, you may get the feeling that the blogger's posts are infused by a fundamentally misguided assumption. But such deep-rooted disagreements can't typically be raised within the scope of any particular post. So consider this open thread an invitation. Do you find yourself raising an eyebrow at some of my basic presuppositions? Any disagreements that run so deep you wouldn't even know where to start? Try here!

I'd especially like to invite readers to offer two kinds of criticism:
(i) Identification of inconsistencies or other mistakes which you think I should (perhaps after a bit of argument) be able to recognize as such.
(ii) Intractable disagreements -- where we just have radically different starting points, such that you doubt you could bring me to agree with you even after protracted argument.

The first kind would probably be more useful to me, but even the second could be interesting.

Oh, and it's an open thread, so you're also welcome to discuss anything else (philosophy or blog-related) you like.