Morality Is Overrated

Hanging out with moral philosophers last week at Oxford reminded me of the old complaint that economists neglect morality.  Actually, I think the real problem is the reverse!  Let me explain.

Many people advise us on what to do.  Some discuss personal actions, while others suggest how groups could better coordinate.  And, crucially, some advise us on what we should do, while others advise us on how to get what we want. 

At the personal level, parents, teachers, preachers, and activists tend to tell us what is morally right, while friends, mentors, lawyers, doctors, therapists, and financial planners tend to tell us what will achieve our ends.  At the level of social policy, pundits and wonks give a mixture of rationales for their suggestions.  Moral philosophers, for example, tend to emphasize policies we should pick, while economists tend to emphasize policies to better get us what we want.   

All else equal, we may each prefer to do what is right, but when all else is not equal we often allow other considerations to weigh against morality.  After all, morality is only one of the many ends we pursue.  Yes we want to be moral, but we also want other things, and we each choose as if we often care about those other things more than morality.  (Some say moral beliefs directly cause us to be moral even if we don't want that, but I prefer to describe this as a revealed preference for moral ends, i.e., for "wanting" to be moral.) 

Economic analysis tries to infer what people want, largely from actions, and then tries to suggest policies to get people more of what they want.  (In particular, it suggests good deals - policy packages which should be better for most everyone.)  Yes people often make mistakes, are ignorant, and have conflicts with the wants of others, but economists have many reasonable fixes for such problems.  Critics, however, say economic analysis is untrustworthy because it is incomplete, since wants are only one of many moral considerations.  But this complaint seems to me backwards.

Yes, we "should" (morally) prefer to analyze policy in moral terms, and we should choose what moral analysis recommends.  For example, perhaps we should immediately and drastically cut CO2 emissions because we have no right to pollute natural purity, and we should care greatly about wildlife and distant future generations.  If so, economic analyzes advising only modest CO2 taxes are a moral travesty, reflecting an unconscionable neglect of "non-economic" considerations.  We should thus condemn these analyzes and the economists who support them. 

But in fact we care only moderately about what we "should" do.  We do not want immediate drastic CO2 cuts because we do not in fact care much about natural purity, wildlife, or distant generations, even if we should care more. Economic analyzes suggest modest CO2 taxes not because they ignore "non-economic" considerations - there are no such things - but because such analyzes give morality only as much weight as people do.   

What we humans want is policy that considers our wants overall, without giving excess weight to morality.  So we want policy advisors, like economists, who suggest actions that better get us what we want, even if those actions are immoral.  We do not want to just do what we should, but we instead want to achieve all our ends, including immoral and amoral ends.  So we mostly do not want to just do what moral philosophers suggest. 

Unfortunately, all this is clouded by our tendency to want to appear to care more about morality than we actually do.  We want to take the moral high ground and be seen as supporting highly moral policies, even if we don't actually want those policies implemented.  So we publicly support moral policies when our support seems unlikely to change the outcome.  But it is amoral advisors, like economists, who help us the most. 

Bottom line: We want to get what we want, not just do what we should, and so we want advisors like economists who tell us how to get what we want.  But we'd rather be seen as following advisors like moral philosophers who tell us to do what we should.

Thanks to Nicholas Shackel for stimulating discussions on this. 

Comments

[Scott Scheule]

Robin,

Michael F, morality is not the only or best way to deal with conflicts between individual wants.

This statement is paradoxical. To say that there is "best" means of settling conflicts is to introduce a moral element. Alternatively, you are using a definition of "morality" I don't recognize, a meaning distinct from "normativity."

So far as economists phrase their findings as "If we want, this we should do this" then they are being amoral. But if the "if" clause is missing, economists are making moral judgements. You apparently would prefer the if-clause left in more often. For instance, when a while ago you suggested adoption of the "tall tax," if you had wanted to be amoral, you should have said "if we want to maximize utility as economists define it, we should tax the tall."

[Peter Turney]

Robin,

... some advise us on what we should do, while others advise us on how to get what we want. ... Moral philosophers, for example, tend to emphasize policies we should pick, while economists tend to emphasize policies to better get us what we want. ... After all, morality is only one of the many ends we pursue. ...

Morality is a means, not an end. Consider the Iterated Prisoner's Dilemma. We tend to view cooperation as morally superior to defecting. Many people believe that tit-for-tat is a good moral rule. Note that tit-for-tat is an algorithm -- a means. Morality does not involve altering the payoff matrix in the Iterated Prisoner's Dilemma -- it is not an end; it is not another value to put in the matrix.

Moral rules encode hard-won wisdom about how we should best go about getting what we want. Consider a few familiar moral rules: tit-for-tat, do unto others as you would have them do unto you, watch your karma, what goes around comes around. Moral algorithms are algorithms that work better than immoral algorithms in the long run, averaged over the long term. Immorality is about short-term thinking and ignoring probabilities and risks (gambling).

some advise us on what we should do, while others advise us on how to get what we want

Once you see that morality is a means, not an end, this false dichotomy dissolves. The best, wisest advice about how to get what you want is also moral advice on what you should do.

For more on this, see A Scientific Approach to Morals and Ethics.

[Robin Hanson]

Scott, "best" is a flexible word which can refer to a wide range of metrics depending on the context. Here I meant "best at getting us what we want."

Peter, I'm using the word "moral" the way moral philosophers do, which is not as a means.

[Peter Turney]

Peter, I'm using the word "moral" the way moral philosophers do, which is not as a means.

Robin, I am a philosopher (PhD Philosophy, University of Toronto). You are not using "moral" the way I use it.

If I understand you correctly, you agree with me that morality is really a means. So your essay above might be summarized as follows: "Let's define morality as an end. When we do this, the conclusion is that morality is overrated."

If this summary is correct, then let's take the next step, which is to conclude that we should not define morality as an end; rather, we should define it as a means.

[Scott Scheule]

Robin,

To recap then, what you mean when you say "morality is not the... best way to deal with conflicts between individual wants" is "morality is not how most people want to settle conflicts between individual wants."

I'm not sure what to make of that. I'm certainly not sure it's true. You're now introducing meta-wants. If you're going to be so free-wheeling with what counts as a "want", one might as well define our predilection to engage in moral philosophizing as satisfaction of a want itself, a want to engage in moral philosophizing (and presumably a subsequent want to ignore it).

When we decide whether or not to adopt an economic plan, we're making a moral decision. I have no idea how you can extricate morality from any policy prescription--once an economist or anyone says we should do something, that's a moral judgment. I see no way to lessen that.

Your argument, as I see it, is not that we should be amoral or more amoral, but that we should not reevaluate our wants, and instead just take them as given. There are problems with this. One is the deonstruction-esque "What if we want to reevalute our wants?" The second is you don't give a reason why we shouldn't reevalute our wants. Your other point is, I take it, that we should be honest about the policies we want. That's uncontroversial, but of course, "one should be honest" is a moral opinion.

[Eneasz]

Phil, I believe the error you are making is in assuming that DU is a simplistic "Maximize the amount of U" type of theory. This is the same sort of error that makes 50 years of torture seem preferable to 3^^^3 dust specks. DU is based on facts observed in the real world. Unlike many ethical theores that simply postulate rules and entities, DU attempts to explain observed phenomena and predict future results. It works off facts and observation. Thus it recognizes that some desires simply cannot be changed beyond a certain point, and/or that the effort expended to change certain desires to such a degree would cost more than would be gained.

Any assessment that leads one to think DU would ultimatly result in Satanism or Buddhism fails to grasp real facts about the real world.

Satanism would lead to greater sickness, injury, and death. Sickness and injury are bad things because they prevent the fullfillment of present desires, and possibly future ones as well. Death prevents even the possibility of fullfillment of the vast majority of desires.

Buddhism in the way you describe would lead to death almost immediatly. Of course for a person who has no desires, death wouldn't be a bad thing. However a world without desires (aside from being an impossibility) is also a world in which there isn't any need for any ethical theory because there is no morality without desires.

[Lake]

@ Peter: philosopher or not, you're using the word "morality" in a peculiar way. If morals were mere means to an end, one would be free to disregard them provided one didn't aspire to that end. But morals are supposed to apply regardless of what you want. I also suspect that the history of the metaphysics of morals would become pretty unintelligible if you were to find and replace "morals" with "means" - another clue that you may be on a different page to your colleagues.

Your point that morals are frequently useful may well supply the basis for an ev. psych. explanation of why we have them. But that doesn't mean that the moral and the useful are coextensive, any more than the tasty and the nutritious are.

@ Robin and Scott: the problem seems to be that Robin is treating morality as just one lot of preferences among many, all of which must be traded off against one another to maximise some further quantity. Yet the only quantity that you could conceivably be obligated to maximise categorically is that of morality itself.

[Scott Scheule]

Lake,

I agree. I think Robin intends something different from "morality" in its typical sense. Perhaps, "non-economic goals" would be closer to when he means.

[Eliezer Yudkowsky]

I don't think you can get away with saying, "I don't need to define how I'm using the word 'morality', I'm using it the standard way philosophers use it." What standard? Which philosophers?

People talk about what 'should' be done as if that word has a distinct referent from what is done - for example, Robin Hanson thinks we should rate Hanson::morality less than we do.

For me, 'morality' refers to the mysterious shape of that strange word, 'should'.

[Unknown]

Richard Hollerith: Actually that was more or less my reaction when I read Eliezer's response. I thought, "Are you sure that's such a bad thing?"

I'm not sure I would mind being a means, if the end were important enough.

[Robin Hanson]

Many people keep trying to rephrase me as saying "we should" when I'm being very careful to avoid that. I'm saying "we want."

Scott, I have no objection to people re-evaluating their wants. I'm talking about what their current wants point to.

Lake, I'm not saying morality is a preference, but I am saying we have preferences about how moral to be.

Eliezer, moral philosophers seem to take "I should do act X now" as the same as "Act X is moral (for me now)."

[Constant]

I agree that economists should not answer questions about morality. But the reason is not that morality is overrated. The reason is that it falls outside of the scope of economics, and almost certainly outside the competence of most economists. The appropriate picture here is not of locking morality into the closet while we get down to the business of doing what we want. The appropriate picture is herding the economists back into the pen.

[Richard]

Robin - 'I'm not making "normative" claims, just claims about what we want. Call them "merely descriptive" if you want - I and many others find such claims persuasive regarding the actions we will take.'

I'm not sure you can avoid the normative so easily. I mean, presumably the only reason you'd bother drawing attention to a merely descriptive claim is that you think it has some normative upshot, i.e. it speaks to the practical question what to do (or 'how to live', as the ancient ethicists put it). Economists are rather notorious for offering normative claims by stealth, under the guise of the purely descriptive. Maybe you don't want to discuss the normative principles you're assuming in this post. But it seems strange to pretend that there aren't any there. (Especially when you've invoked so much evaluative language along the way.)

[Scott Scheule]

Robin,

In that case, what you mean isn't clear. Seeing as it's not reevaluated goals you're protesting against, I repeat, the best reading of what you mean by "moral" is "non-economic values."

Indeed, you come close to saying so when you write:

... If so, economic analyzes advising only modest CO2 taxes are a moral travesty, reflecting an unconscionable neglect of "non-economic" considerations.

So you think we should (or we really want to?) neglect--or at least neglect more than we do--"non-economic" considerations. So just say that--there's no need to bring in the bigger concept of "morality."

Also:

Many people keep trying to rephrase me as saying "we should" when I'm being very careful to avoid that. I'm saying "we want."

Yes, but you're saying morality is overrated. If you're not comparing that to what we should rate it as, then what are you comparing it to? What we deep down really want to rate morality as? Is your argument thus that we're under a false consciousness, and don't know what we really want (and you do)?

[conchis]

Robin,

"I'm not making "normative" claims, just claims about what we want...I and many others find such claims persuasive regarding the actions we will take."

(1) Robin::wanting something doesn't seem to be the same thing as conchis::wanting it, and I think the normative persuasiveness of your descriptive claim depends quite a bit on exploiting the ambiguity here.

Your descriptive claim seems obviously right if you define Robin::wants as revealed preferences (although you won't have said anything we didn't all already know). But I think most other people's working definition of wants is closer to what Unknown was getting at above: there's a structure to desires and desires-about-desires that seems relevant here (plausible DU accounts usually recognise this; perhaps that's what Enneasz was getting at?). Anyways, I'm pretty comfortable claiming that people conchis::want to be moral rather more than they Robin::want to.

To the extent that that's true it's no longer obvious what we should do to help people get what they "want". To help people get what they Robin::want we downweight moral concerns in our advice. But to help people get what they conchis::want, not so much.

(2) Neither you nor many others can get from descriptive statements about either Robin::wants or conchis::wants to conclusions about what you and said others will do without importing some sort of normative/moral standards (at the very least you need to balance these wants against each other somehow).

(3) Even if you could, I don't see why it should follow that because others are imperfectly moral you should strive to emulate their imperfection. Do you think your advice should be biased because people are biased?

[Robin Hanson]

Richard, one can speak to the practical question of what to do without speaking to the normative question of what one should do. One can instead speak to the question of what will get you what you want. For most people, this is in fact more persuasive.

[Unknown]

Robin, are you sure those are two different questions? Once you have properly evaluated your conflicting wants? It may be that acting morally simply means correctly weighing the the things that you want, so that the moral thing to do is the thing that really gets you the most of what you most want.

[Scott Scheule]

One can instead speak to the question of what will get you what you want.

Yes, but to weigh the wants of multiple people involves questionable normative claims. If you want to avoid that, you have to speak of what will get you what you want, assuming people's wants are comparatively weighed in manner X. Agreed?

[Richard]

Robin, I don't see two questions here. The question what to do just is the question what one should do; just as the question what to believe just is the question what one should believe. Now, one might think that the thing (one ought) to do, all-things-considered, is whatever 'will get you what you [already, unreflectively] want'. That sounds like a dubious normative principle to me. But that's what you're really presupposing here.

(Perhaps you are using the term 'should' in the so-called "inverted commas sense", to refer merely to what haughty moralists *say* you should do. But philosophers use the term 'should' to denote what one really has most reason to do.)

[Robin Hanson]

Richard, I see the question "what to do" as meaning "what would I choose to do, given as full a consideration as possible of that choice." I want to do what I would do if I considered the choice fully. This is not the same as what I "should" do. I can be fully aware of what I should do and have thought carefully about my choice and yet still choose something else.

[Richard]

Granted, there's the phenomenon of "weakness of will", whereby we act against our better judgment. But that seems to be a special case. Did you have something else in mind? (That is, do you endorse your divergent answer to the question "what to do" as your better judgment as to what act is warranted, or do you take yourself to be going wrong - by your own lights, even - in such a case?)

[Robin Hanson]

Richard, I'm not sure how I can be any more direct or clear about this: we all knowingly make choices contrary to what we "should" choose. Yes sometimes this is due to mistakes, but it mainly reflects the fact that we do not want only to be moral.

[Dr. Zeuss]

Robin, yes people tend to do what they want to do, and many people overestimate (I think this is where you're getting "overrate") how much weight people place on what they should do in deciding what they will do.

But that doesn't mean that people should place more weight on what they want to do then on what they should do. Maybe people will do what they want to do, but they should do what they should do.

[Richard]

Robin, you keep putting "should" in inverted commas. Philosophers are interested in what we should do, not what we "should" do. So I'm trying to work out whether we are just talking past each other. Shifting to a neutral, unambiguous vocabulary might help. Hence my question: are the "amoral" decisions you're talking about ones that the agent can endorse on reflection, or judge as what they have most reason (all things considered) to do?

If so, they're assuming normative principles. If not, the agent is irrational by their own lights. Either way, you can't avoid the fact that decision-making -- and the practical reasoning that underlies it -- is essentially normative.

[Peter Turney]

Robin,

I'm not sure how I can be any more direct or clear about this: we all knowingly make choices contrary to what we "should" choose. Yes sometimes this is due to mistakes, but it mainly reflects the fact that we do not want only to be moral.

I disagree. I believe that enlightened self-interest is a sufficient basis for morality. In this view, when we make choices contrary to what we "should" choose, it is always due to mistakes (lack of enlightenment). A fully enlightened being (if such a being could exist) would never knowingly make choices contrary to what "should" be chosen.

Sometimes we feel a conflict between what we want and what we know we should do. You seem to believe that this conflict must be due to competing values (wanting to do the moral thing versus wanting to do the immoral thing). I believe that the conflict is due to bugs in our algorithms for making decisions (lack of enlightenment).

What would it mean to want to be moral (to do the moral thing) purely for the sake of morality itself, rather than for the sake of something else? What could this possibly mean to a scientific materialistic atheist? What is this abstract, independent, pure morality? Where does it come from? How can we know it? I think we must conclude that morality is a means, not an end in itself.

[Robin Hanson]

Richard, it is not clear to me if what we should do is the same as what we have the most reason to do. But it is clear that we often knowingly choose acts other than what we believe is the act we have the most reason to do, and other than the act we think we should do. Call us "irrational" if you will, but we expect to and intend to continue this behavior.

[tobbic]

So a value is a rule of thumb that helps make decisions so that we get what we want in the long run (=means)?

Is it that a value is a separate term in our total utility function? Like "don't deceive people" = "i believe deception is wrong" = "when making decisions choose an action in which degree of deception is as low as possible. Any deviation from this will cause you huge penalty". So you just take a function (e.g. a sum) over the separate values and preferences/wants and voila there's your total utility. In this sense I still stick with the claim that in some contexts is useful to consider values and wants equivalent.

[tobbic]

The fact that value is not an end in itself distinguishes values from wants/desires. It's a heuristic to help us make better decisions (e.g. stick with social norms)?

[Unknown]

Robin, I expect to, but do NOT intend to, continue to choose to do things I shouldn't do and that I don't have good reasons for doing. What this this say about me, in comparison to someone who intends to continue acting in this way?

[Robin Hanson]

Dr, yes folks should do what they should, but don't want to.

Unknown, not expecting to do what you intend must feel very frustrating.

[Nick Tarleton]

Unknown, not expecting to do what you intend must feel very frustrating.

Are you saying you don't? You never expect to break some resolution due to weakness of will or simple forgetfulness or whatever?

[Rebecca Roache]

I wonder whether this is really true: 'But in fact we care only moderately about what we "should" do'. In my experience, when asked to justify their actions, many people will attempt to argue that they do act morally - and, if pointed out that they do not act in the most moral way, will claim that they are unable to do so because of certain constraints (e.g. the people who say, 'I know that buying meat that is produced cruelly isn't ideal, but it's cheaper and my money is limited'). Or, more rarely, they will disengage and agree that they don't act morally, but that they don't care. I haven't ever come across someone who concedes that they do not always act morally, but justifies this by claiming that they care less about morality than about other considerations. Perhaps this comes down to an impoverished conceptual scheme, or to social pressure to be seen to be moral. But there is social pressure to be seen to be moral for at least one good reason: being moral, at least in part, involves being considerate of others. For that reason, the claim that '[w]hat we humans want is policy that considers our wants overall, without giving excess weight to morality' is misleading. On one view of morality - namely liberalism - a policy that considers our wants overall just is a policy that gives primary weight to morality, provided that this policy recognises that people's various wants often conflict, and that measures are sometimes necessary to prevent the wants of one person or group causing significant harm to another person or group. Having said that, I think you're partly right: in liberal societies, people don't see why they shouldn't satisfy their wants providing that doing so doesn't harm anyone else. That this is true can be seen from the fact that legislation against acts that don't cause significant harm to others - such as homosexuality between consenting adults - is generally frowned upon in most Western societies.

A disclaimer: I have only skim-read the other comments, apologies if I'm rehashing something that's already been said!

[Robin Hanson]

Rebecca, when people say "they don't care" or when they cite "constraints", we can usually clearly see the other considerations that weighed against morality. For example, yes money is limited, but we can see other discretionary items in their budget and infer that they preferred those items to local meat. People like to talk as if they had no choice to excuse their choices, but we know better.

I agree there is social pressure to be moral and that one liberal ideal limits morality to dealing efficiently with conflicting wants, but most people's concepts of morality go well beyond this.

[Rebecca Roache]

I completely agree that it's a rationalisation, but the point is that they feel pressure to be seen to be acting morally, which - however you analyse it - involves (perhaps reluctant, in some cases) recognition that there are more important things than their own wants. Those recognised other things might involve morality, or they might involve a selfish desire to avoid being seen as a certain type of person. But even the latter indirectly involves moral concerns: in this case, it is the desire not to be seen as a selfish person who always prioritises their own wants. And our condemnation of selfish people who prioritise their own wants, if such condemnation is reflective and intelligent, is primarily moral. (If it is unreflective, there may be an evolutionary explation for it.)

It may be that most people's conceptions of morality are more substantial than what liberalism lays out, but my point is that even the liberal conception of morality is a conception of morality, and it captures quite well the desire you attributed to humans for a policy that considers our wants overall. Therefore, humans do want a moral policy. I think that to claim that what humans want is a policy that prioritises our wants in a way that does not conform to *any* plausible conception of morality is going to be implausible. For example, the claim that all (or most) humans want a policy that prioritises their own wants at the expense of everyone else's wants is, I believe, implausible. Some humans may want this, and perhaps many of us occasionally daydream about how nice such a policy would sometimes be, but (I hope that) few mentally healthy individuals seriously want it.

[Robin Hanson]

Rebecca, yes there is a possible position on morality which says morality is exactly getting everyone what they want, in which case there would indeed be no conflict between morality and getting everyone what they want. But you seem to accept that most people's concept of morality seems different from this, and thus can conflict with wants.

I said in the post that people want to appear more moral than they want to be, and I don't see that this theory is undermined by your observation that people publicly condemn immorality.

[Bill Liles]

Your comments seem reasonable. A cursory study of recent history surely suggests Republicans like to wrap their greed in blankets of religiosity. Yes I agree.

[Mathew Crawford]

I suspect we define morality differently. To me, morality defines distinction between good and bad. Of course, this distinction is not universally judged - it may be judged differently by each individual. In that sense, I tend toward a position of egoistic utilitarianism.

However, I also tend toward a position that scales. If we behave in a way that would cause us all misery should we behave that way, then we are shooting (not a collective foot) all our feet. In other words, we need to be mindful to disentangle ourselves from prisoner's dilemmas which pool into tragedies of the commons.

When there is no such conflict, acting to achieve our ends is perfectly moral. I suspect both that (a) this is true most of the time -- for most of our actions, and (b) that these are not the interesting decisions in life...