Monday, January 9, 2012

Mailbag Monday: Is-Ought Problem

Mailbag Monday: A weekly segment that covers readers' questions and concerns about all things Philosophy, Bro, and Philosophy Bro that don't quite fit anywhere else. Send your questions to with 'Mailbag Monday' in the subject line.


Alex writes,

will you explain the is-ought problem and its implications?
Thanks, bro.

Aw yeah, the is-ought problem. Shit's classic, bro, goes back to Hume, and it goes something like this:

Sometimes, when I'm talking ethics with my bros, they describe the way the world is. They're like, "That chair is over there. The bar closes at 4AM. The sky is blue." and I'm like, "Yup. Yup." How could I disagree? These are obviously true things. But then suddenly they're like, "Therefore, we should go on a roadtrip tomorrow." And I'm like, "woooooooooah!" How the fuck did they get from how things are, to how things should be? Those aren't the same at allIf I asked, "Where should we go?" and you told me where we already are, that wouldn't answer my fucking question.

Let's say my bro Ice and I are drinking. Ice is about to do some stupid drunken shit (as he's wont to do) and I say the following:

"You are so drunk that you could die if you tried that."

That's an is-statement. It just describes something about Ice and the thing he's about to do. Maybe you're thinking he should not do this fucking thing. But why not? "Because he'll die." So what? "Well, you shouldn't do stuff that will kill you." AHA! See, that's not an is-statement. It's an ought-statement, which describes how Ice should proceed. Before we could say that Ice shouldn't do this stupid fucking thing, we needed to say he shouldn't do stupid fucking things in general.

People typically accept that statement, and in fact, so many people accept that statement that it seems obvious, and you might miss that it's hiding in there. You might just gloss right over it.

Sometimes, though, the hidden ought isn't so widely accepted.

Next morning, Ice and I are lying around on couches, nursing hangovers, and we have this conversation:

PB: "Dude, I am so hungover." (Is-statement) 
Ice: "uuuuuugh."
PB: "I seriously feel like I'm going to throw up." (Is-statement)
Ice: "nnnnn..."
PB: "Also, I spent all the cash I had last night." (Is-statement)
Ice: "Shhh..."
PB: "Dude, we ought to never get that drunk again." (Ought-statement)
Ice: "...Wait, what?"

How did I get from those first three things to that last one? I've only described my hangover. That's all I've done. No matter how fucking miserable my hangover is, it doesn't support my conclusion that I should never drink that much again; maybe I love being hungover. (I don't.) But I accept something like the following premise, the way-too-drunk premise: "You should never get so drunk that you spend all your cash and end up super hungover the next day." BOOM. Another ought-statement. Ice is confused because he loves getting that drunk, and is willing to put up with the hangover and relative poverty. He does not accept the way-too-drunk premise, so we disagree on how we ought to act, even though we totally agree on how things are right now.

The is-ought problem arises because of a particular property of logic called conservation, which (roughly) says that whatever your conclusion is, it has to be in your premises in some form or another. It doesn't have to be obvious - in fact, often it isn't - but it is, in some form, lurking down there. (This is a really rough gloss on conservation, so bear with me here.) It's a feature that makes logic really fucking great - it tells us exactly what is and isn't supported by our premises. We want conservation in logic. Conservation is our friend. When someone objects, "Wait, your conclusion isn't supported by your premises!" she is actually objecting that I've snuck something into my conclusion that isn't anywhere in my premises, which is a no-no.

So if you start with only descriptions of how things are right now, then you can't decide anything about the way things should be. You need oughts in your premises to get them in your conclusions.

Occasionally, bros invoke the is-ought problem to argue that there is no morality at all or some similar such claim. Those bros are fucking doing it wrong. It turns out that ethical theories just are attempts to provide ways to move from is to ought. For example, utilitarianism says, at bottom, that we ought to act such that pleasure (or whatever is "good") is maximized, something like, "Whichever act creates the most pleasure is the act that you ought to do." See how that gets you from is to oughtSo you know some things about which act creates the most pleasure, your is-statement; then you've got the utilitarian rule, which is an ought-statement, and that moves you to your conclusion, that you should do that act. So the is-ought problem doesn't mean there is no morality.

It is, however, a problem for deciding which moral system is the right one. The utilitarian rule is an ought-statement, too; so it helps us choose other ought-statements, but how did we get it in the first place? Usually, through some hand-waving. When J.S. Mill argues for utilitarianism, he's mostly like, "I mean, c'mon! Who doesn't like pleasure, amirite? Obviously we should maximize pleasure and minimize pain. Duh." And when Kant argues for the Categorical Imperative, he's like, "I mean, it comes from a good will, and obviously a good will can't do anything bad! So, you know, act with that shit." Some ethical theories are even more upfront about their hand-waving: W.D. Ross argues for a bunch of duties like charity and justice and stuff, and he argues for them by just saying that justice and charity and stuff are obviously things we should value.

So the is-ought problem is a problem of metaethics, the philosophy of how ethics works. Some philosophers have tried to cross the divide by giving arguments about how we can get from an is to an ought; some just argue for certain baseline oughts like charity and go from there. For example, maybe it just is the case that killing is bad. Immoral, even. Then we ought not kill. That was simple! "But why shouldn't we be immoral?!" Because that's what "immoral" means. It means don't do this fucking thing; it should not be done. (To be honest, the meaning of "x is immoral" is also hotly debated, but I'm not going to get into that here.) The point is, it might be that "x is a thing that ought not be done" might be an is-statement that also gives us an ought. That's the route that tries to cross the divide.

It's a super-interesting problem, and lots of really, really fucking cool work is being done on it. Ultimately, the problem shows us that finding out the right thing to do isn't as simple as we'd like it to be. Damn.


As usual, the Wikipedia page is pretty helpful, both in explaining and knowing where to go next.
Hume originally proposes the problem in Book 3.1 of the Treatise


  1. Enjoyed reading this. Glad to have you back, PB.

  2. Oooh, nice. I've always kinda dismissed the Is-Ought problem out of hand with a simple, "the fuck is Hume talking about, man?"... So I guess that's what Hume was talking about. Pretty awesome stuff :D

  3. Nice intro to a tricky subject. Keep it up.

  4. i think this is the example i will use when trying to describe this problem to my 2nd year Kant students!!!
    thanks again philosophy bro

  5. Rawrawr

    Yeah, it's easy to dismiss Hume for thinking that the problems he brings up aren't really problems but once you get into the meat of his argument it's like "Whoa, that's crazy." His Is-Ought distinction is merely one of his empirical dismantlings. His problems are the problems that other big name philosophers (such as Kant) try to solve.

    To list off a few for other people:
    The problem of induction
    The uncertainty of cause-and-effect
    Sufficient cause (I'm not sure of the actual name so I stuck one on it)

    The problem of induction and the uncertainty of cause and effect are two big problems that haunt science in general.

  6. Right on!
    Now what about the evolutionary psych view?

    IS: Humans have certain evolved traits.

    IS: These traits include some vague sense of social ethics and of pleasure and empathy and lots of other shit I don't have time to go through and that we don't even understand yet.

    IS: We are human.

    IS: We will have innate tendencies toward certain judgments of what we ought to do as a matter of our evolution.

    OUGHT: We should accept that our system of ethics needs to be compatible with our biological make-up, because we can't change that anyway. So we just need to balance the different features of our biological system such as social vs personal, empathy, pleasure, etc. And we could, in principle, study what balance seems most overall compatible with most people's biological reactions...

    Ok, that's all a stretch, but isn't this sorta an ought from an is without first positing an ethical judgment?

    1. The ethical judgement occurs right when you said "we should accept our system of ethics... because we can't change that anyway". It is not necessary that just because we "can't change it anyway" that we should accept our ethics. One could just as easily try and change things even if they seem futile.

      That Sisyphus will never get the boulder to the top of the hill does not mean that he "should" stop. It is our own "passions" which drive us to say what we "should" or "should not" do according to Hume.

    2. and we "OUGHT to imagine him happy."

    3. @Yoo-jin
      I think Aaron was saying that you cant change our biological make-up, not our system of ethics.

  7. I haven't found it especially convincing to others, but my take on the is-ought problem is that it is making an assumption I find unsupported - that our 'ought' experiences should be considered less empirically respectable than our 'is' experience.

    This results from the false ideas at the time about how the eyes and brain work - the idea that light hits our eyes kind of like light hitting a camera, creating an image that our brain studies and identifies items from.

    In fact, the way our brain builds our world is much more complex and organic than that, much closer to how a video game works tha a movie - our brain appears to pick out relevant objects individually, build a landscape out of those, and then fill in the background with a vague wash.

    Under such a picture, there is no reason to find our moral experiences any less respectable than our other experiences - everything has gone through the same filters, and has been picked out by our brain as relevant and important, or we wouldn't have evolved the capacity to have that experience.

    So, if we feel that something is wrong, that should be considered sufficient justification to say people ought not to do that. They can disagree, and such disagreements are pretty intractable - but anyone who's ever argued about whether a paint colour is brown or green knows that arguments about basic, empirically derived facts are always intractable.

    1. It's not saying that our "ought" experiences are less empirically respectable than our "is" experiences. It's claiming that we don't get our "oughts" from our "is's". As Hume famously stated:

      "Reason is, and ought only to be the slave of the passions."

      To say something "is" the fact is different from saying something "ought" to be X and/or Y. He finds that "ought" judgements are, actually, pre-"rational". Now the thing is that where you seem to say that morals are derived from simple experience-input, Hume says that morals are derived from our psychology. He isn't saying that we cannot have oughts. He's saying that if we were to limit ourselves to reason or logic, we would never be able to derive ethics.

    2. Yes, but there is an implicit premise that we can't get ought statements empirically - that our ought claims are not as respectable empirically as our is claims. Otherwise, his whole argument really doesn't seem to amount to much.

      I guess what I'm really saying is that ought statements are just another kind of is statement - that the whole distinction between the two types of claim is a mistake. Ought claims are just claims about what it is that we ought to do, and you may as well talk about the animals/cats distinction. It's certainly broadly true that animal claims that aren't about cats will tell us nothing about cats without the addition of a cat centered claim, but that fact does not create an animal/cat distinction of the type Hume seems to intend.

    3. I suppose that is possible if you really want to divide it in that way. He wasn't saying, however, that morals and ethics are useless or that they don't make sense. It's just that you cannot observe a specific occurrence and derive an ought simply out of reason.

      Perhaps a better analogy would be that two or more people can read the same book and have different interpretations of what is important. Not a perfect analogy, however, since there is the matter of author's "intent" in writing, but it's basically the same matter. We humans have differing "passions". How would we determine, exactly, which passions are more valid than others?

      For instance, I can debate whether or not to eat food and make it a moral case. Should I eat food or should I not? I suppose I should eat the food if I want to survive, but why should I wish to survive? All sorts of arguments for and against can pop up in such a consideration without any definitive case so that it ultimately comes down to how someone feels about the matter. There is no "fabric of reality" which necessitates that I have to survive or pursue my goals even if I take them into account.

    4. "So, if we feel that something is wrong, that should be considered sufficient justification to say people ought not to do that."

      This is exactly the reason why emotivism sucks as an ethical theory. An example: Suppose I have a very strong feeling of repulsion towards homosexuals. Also, suppose I have a strong feeling of repulsion towards white people (I'm half hindustani). These feelings are widely shared by a lot of people - I'm pretty confident that the large majority of humanity doesn't like homosexuality and given our colonial history, I can also safely assume that the largest part of humanity detests white people.

      If we are to believe you, then we would have to accept that judgment, because empirically speaking, that judgment is what most of humanity automatically feels.

      For the record: I don't really detest homosexuals or whites, this is just an illustration.

  8. Surely authorial intent stands in the place of ontological reality in your metaphor. Actually, it's a pretty much perfect fit. The world of our experiences is like a novel in a lot of ways, really - both that different people take different things away from it, and that none of them can directly access the author's intent through reading.

    As to which passions are more valid, you may as well be asking which is the more valid position - my belief that the dining room table is appropriately sized for the dining room, or my wife's belief that it is too large and "takes over the room." Both of us (probably) experience the same reality, yet we clearly experience it differently. And don't get me started on "That's brown!" "No, it's green!" "You're blind!" "No, you're crazy!"

    We can not assign truth values to such things. There are basically two places we can look - consequence and consensus. If we can build a strong consensus on something, particularly if that consensus crosses cultural boundaries, then that is a good reason to feel confident about it. If believing something produces strong negative consequences, particularly if those consequences are delivered without human intervention, then that is a good reason to re-consider it very carefully.

    My point is simply that we tend to treat the passions as if they are internal, and most of our other experiences as if they are external, and to priviledge the latter over the former. I am asserting that that is simply a false view - that all of our experiences are internally generated, and that we have no way of conclusively determining veridicality, and therefore no reason to consider ought statements any less empirically respectable than is statements.

    1. The reason why there would be such a divide between ought and is statements is that, as you said, any sort of situation can have different opinions upon what "ought" to be done.

      What "is" the case is, for the most part, irrespective, at least in the empirical view, of the way in which we wish to perceive it. That a table is a table, for instance, is an empirical fact. Now, however, if we want to get into the realm of ethics, we rely upon the passions which are a personal matter. That I might get hurt if I run into the road with cars oncoming does not mean that I shouldn't run into the road. In fact, it could be the very fact that I feel that I want to be hurt, maimed, or killed and I feel that I should do what I wish to do. This can be diametrically opposed do someone who wishes to preserve their own life. Thus, for these two people, they have the same experience but two different possible "ought" statements.

      And as an empiricist, Hume would probably reject "internal" experiences as the thing that we can be certain about is that experiences exist. Whether or not there is a self perceiving these experiences is a different matter. But that's opening another can of worms.

    2. You seem to think that I'm arguint that 'ought' statements are being underrated, and should be promoted to equivalence with 'is' claims. I'm saying the opposite, however - that as a matter of fact, the apparent superiority of 'is' statements is entirely illusionary, and that they are actually no better than 'ought' statements.

      Experiences of passions are the same as our other experiences - they may have something to do with ontological reality, or they may not, but they certainly are constrained by our natures.

      As a matter of fact, only people with severe brain damage seek to be maimed or killed. They may accept it as the price of something they desire, but otherwise they will not. Why not? Presumably because all the humans whose brains were so configured that they wanted to die, died without leaving offspring.

      Of course, we have no real access to other people's experiences - it's theoretically possible that everyone around me is doing their utmost to kill themselves horribly, and is just REALY BAD at it. That seems like a pretty poor hypothesis, however.

      What I'm saying is that my judgements about morality are exactly as respectable as my judgements about shape, colour, odour, taste, etc. It all comes from the same place - my brain - and there is no good reason to priviledge some of the things my brain tells me over others (or at least no good reason that does not depend on induction from previous patterns of experience. (In that obviously, if one of myexperiences has proven to bring about negative consequences when Iacted upon it in the past, that is a good reason to accord it less weight in influencing my future actions.))

    3. The problem is that judgements about colour, odour, shape, taste, et cetera, are not wholly dependent on how you feel about them. While they are dependent on your brain, it is not your judgement (unless we're talking about how you feel about them) which determines the experience for the most part.

      "Ought" statements, however, are *totally* reliant on our judgements. Even if we are not severely brain damaged, it does not mean that I could not just decide that I should be maimed or killed. Perhaps I see myself as a danger. Perhaps I do not see purpose in life. Why are these judgements not as valid as ones which seek to preserve life?

      And even if negative consequences occurred in the past, Hume also has the problem of induction. Just because something has occurred in the past in a certain way does not mean that it shall continue to eternity in that way. And even then, that it has a negative consequence doesn't mean I "ought" to avoid it. It only means I ought to avoid it if I believe "I ought to avoid that which brings negative consequences to me." There is nothing *necessary* about having to preserve one's self.

    4. All of your experiences interact with one another to create your world. Colours seem duller when we're sad, and brighter when we're happy. A soft chair makes the same price seem more generous than if you considered it in a hard seat. A resume is better if it's attached to a heavier clipboard. Hell, we form preferences between identical items, based solely on their relative locations.

      Yeah, I just flatly disagree with your last paragraph. It is intrinsic to the whole concept of a negative outcome that we want to avoid it. If we wanted it, we wouldn't consider it negative. Hume makes the mistake of trying to separate things that by their nature form an organic whole.

      As for the problem of induction, I'm well aware that induction can not be derived deductively from experience. It does, however, work really well. Plus, of course, we aren't capable of ignoring it even if we try.

    5. Thank you both for the debate. I found myself agreeing with stuart but it was very interesting to read both of your sides.

    6. "Want" to avoid. It doesn't mean that we shouldn't avoid it, though. You have to take it as an axiom that "I shall avoid what seems negative to me." However, just that it's something that we wish to avoid doesn't mean that we shouldn't do it. That there are certain negative consequences like the "use/waste/passage" of time when doing a certain action (such as writing a book) doesn't mean that I should avoid doing such actions. That I dislike eating broccoli doesn't mean I should avoid eating it. There is nothing "intrinsic" about an action which makes it bad. It is what we take as axioms based from our passions that make it so.

      This is what he means by "ought" being distinct from "is". While colours seem duller, brighter, and such things, it doesn't mean that you cease seeing colours entirely. And I don't see how your "identical items" objection goes against Hume. Wouldn't that be an instance of our passions?

    7. Stuart, are you trying to say that we should do whatever feels good or brings positive consequences, and avoid whatever feels bad or brings negative consequences? Because that seems like just another attempt (unsuccessful, imo) to bridge the is-ought divide.

      It seems your reason for supporting this moral theory is that it is true by definition, as when you say, "It is intrinsic to the whole concept of a negative outcome that we want to avoid it." But as Yoo-jin mentions above, that doesn't explain why we should avoid negative consequences. It only equates a negative judgment with a tendency to avoid.

      But why should someone act according such tendency as opposed to going against it, assuming we have the free will to act by avoiding or approaching things we judge negatively or positively? Or, as I suspect, do you not believe we have such free will?

    8. People with some kinds of brain damage do stop seeing colours entirely. Others are unable to discern some colours, but see others normally. Some are blind, but can still respond to thrown objects as if thety could see. And of course even in a normal person in a constant emotional state, colours can be pretty sketchy, our eyes and brain working together to figure out what colour things are from a remarkably limited amount of data and often making mistakes or guesses. Colour based optical illusions are easy to find.

      Our colour vision, like all of our senses, evolved to be useful - not to discern objective reality. It might approximate the latter (we have no way of knowing) but we are damned near certain it doesn't accurately depict it.

  9. THis post came just in time. I'll borrow your drunken bro example for my intro to ethics course. Great work PB!

  10. Finally, a real-world application that addresses the moral implications of is-ought! These "ought" conclusions are not ends but simply means to other moral/ethical interpretations. Solid post.

  11. Excellent! Love you summaries, bro and glad to have you back! I only recently discovered your site and have been glued to it ever since. I am a history major who has found a love of philosophy and its applications in larger historical contexts. More specifically, how guys like Nietzsche and Hegel affected the thought processes and actions of the people living during their particular time with the philosophical bombs they dropped on society. Great stuff and please keep it coming!

  12. Bitchin' read, bro.

    I think the philosopher of economics Mark Blaug (1976: 354/4) has something interesting to say about the question:

    "No doubt, Hume's Guillotine tells us that we cannot logically deduce 'ought from is' or 'is from ought'. We can, however, influence 'ought' by 'is' and vice versa: moral judgments may be altered by the presentation of facts and facts are theory-laden so that a change of values may alter our perception of the facts."

  13. Whoa, wait a minute bro. You say the is-ought problem doesn't mean there's no morality, then go on to show how everyone who tries to overcome it resorts to hand-waving? If there's morality, none of them seem to have shown it yet. I'll wait until these bros can sort it out. Could be a long time, since they've been trying for millennia.

  14. It just so happens that I just had an argument with my girlfriend about some random shit. Both of our arguments were made up of is-ought statements. The problem is that when a couple can largely agree on the "is" statements, but can fundamentally disagree on the "ought" statements, do the ought statements become less valuable in favor of relationship preservation? (I mean, we do love eachother... (Is statemment. haha.)) Is this a subjective decision which has to be made, (whether or not to stay together based on fundamental disagreements on oughts) or is there a process/structure for forming is statements that contain ought statements so we can facilitate the merging of our ideas about what is, and ultimately, what ought?

  15. Is Philosophy Bro on another hiatus? Almost two months bro without new content.

  16. you are my life blood. where is the new content!?

  17. welcome back!

    there's also the epidemiological question that is attached to it. sure, the main focus is on the moral, ethical and linguistic side. but some terribly thorny question have always been: what kind of warrants would support such a transfer, what kind of knowledge is produced by an is-ought conditional, how reasonable should they be seen in conversation etc.

    keep up, bro!

  18. "x is a thing that ought not be done" Surely this is just a combination of the two premises "X is this" and "X ought not to happen". It's like soft determinism's pussy cop out of "it looks X to me", it isn't helpful bro, its just messing about with language.

  19. This is a great posting I have read. I like your article.

  20. Our beliefs, moral or otherwise, in order to be knowledge, must be ‘both’ justified (ought) by reasons ‘and’ true (is) to reality, satisfying both Plato and Hume.