Sunday, February 9, 2014

The real is-ought problem

I think most people who have heard of the is-ought problem would say that it is about the fact that you can not derive an 'ought' from an 'is': nothing follows from facts about the way the world is, concerning the way the world ought to be; nothing follows from the way people do behave, about the way we ought to behave.  Of course, this is not a problem unless people try to produce such derivations, and it may seem like we do. For example, people often say silly things like "That would be cruel, therefore you shouldn't do it."

But that is not a derivation; it is ordinary enthymematic reasoning. If it is to be a derivation, then we have to add a principle, say, "If the purpose of an act is to derive enjoyment from the suffering of another being, then it ought not to be done." This was Hume's point when he filed his original complaint.

In truth, deriving an ought from an is is no more problematic than deriving an is from an is or an ought from an ought. No statement, by itself, implies any other statement. P does not even imply P except in light of the principle: If A, then A (where A is any proposition, e.g, P.)

Now I told you that so I can tell you this:  There is an is-ought problem, it's just not that one.  The real is-ought problem is this:

We often fail to distinguish the ought from the is.

In the simplest and least interesting case, this is just because we are liars. We are asked what is the case, and we say what ought to be the case instead. When James says No when Molly asks whether he is having an affair, he is affirming the way the world ought to be, not saying how it is.

Things get more interesting when the questions challenge us epistemically. We can have a hard time telling the difference between is and ought, and sometimes we just can't tell at all.

Why?  Let's start at the bottom.

Perhaps there is a way the world fundamentally is, but that world is not simply Given to us in perception. Kant was the first to fully penetrate this illusion, teaching that our shared empirical reality is something our minds construct using a shared set of  rules, what he called "apriori intuitions." Quine later naturalized the Kantian insight as the scientific project of discovering the rules by which the "meager input" of sensory irritation is amplified into the "torrential output" of our theoretical vocabulary. And cognitive neuroscience now aims at exactly this: producing a materialistically adequate account of this process, the mathematical rules by which the brain sustains and deploys an internal model of the world on the basis of meager electronic pulses generated by specialized nerve cells in our sense organs.

Whether you think of them as mental categories or as regulated patterns of neural activity, the rules of cognition and perception are fundamentally tools for norming our sensory input. We expect the world to behave as it ought to, and that is how we always initially try to represent it. Now, of course, we survive because our default expectations can be swiftly overridden when things go obviously awry. But usually the world appears just slightly off. And in such cases we naturally and unconsciously resolve ambiguity in the direction of the normal. We have to. If we didn't, we'd go nuts.

Fortunately, the requirements of sanity are usually compatible with the way the world works. Weirdness is almost always just random fluctuation from the norm. Strange sounds in the night, strange feelings in the body, strange looks from other people are usually meaningless. (We just remember the ones that aren't.)

But even when it's not just noise, denial can be useful.  Honey, is something the matter? No dear, why do you ask? But that's not true, is it? The truth, if you could attend to it, is that your partner coming home from work with a nasty cold has made you upset, and in fact upset with her. Now there will be no movie; now there will be no sex. But these are feelings you are ashamed of and your morals swiftly denihilate them.

The use of fundamentally moral rules to resolve epistemic ambiguity is a very common, but largely hidden phenomenon.  Some of my favorite examples are from sports. Umpires and referees are paid to make the call, even when it is just too close to call. In these cases, officials must see the world as it ought to have happened. Was he really tagged out at second? He deserved to be, he had no business stealing. Was that ball really wide of the line?  It ought to have been, she shanked it.

This is an example of what Daniel Kahneman calls substitution. We answer a different question than the one being asked, because it is more available, and because it is typically a satisficing substitute.

But I called this the real is-ought problem. What makes it a problem? The answer is just that our instinct to norm the world can go too far and we recognize only too late that we have a major problem on our hands.  It is too easy to be glib here: The signs of a failing marriage or a weakening economy or a climate disaster or an imminent terrorist attack were all around us. How could we have missed them? Someone must have been asleep at the wheel. Well, that's called hindsight bias, and it is typically a gross exaggeration.  Sure, the signs were all around us, right there with all the signs that everything was just dandy.

The real is-ought problem, then, is not an error, it is a task.  It is the task of getting better, both as individuals and as a group, at distinguishing signal from noise, at discerning when the world is not behaving as it ought.

G. Randolph Mayes
Professor, Department of Philosophy
Sacramento State University


  1. Randy, I’m not entirely convinced by you opening analysis of the is-ought problem as bequeathed to us by Hume. I agree that any derivation of one proposition (Z) from another (A), if it is to be logically watertight, requires a justifying principle of derivation (B). I believe this point is made by Lewis Carroll in ‘What the tortoise said to Achilles.” He then argues that this leads to an infinite regress, since to accept that A plus B implies Z, we need a further justifying principle (C), which say that A plus B plus C implies Z, etc. etc.
    I don’t think Hume’s problem is with that. I think it’s more directly with statements like “If the purpose of an act is to derive enjoyment from the suffering of others, then it ought not to be done.” This statement says that a descriptive statement (“The purpose of my action was to derive enjoyment from the suffering of others”) implies a normative statement (“I should not have performed the action”). Hume is aware that the conditional statement expresses moral common sense. As a good guy, he’d doubtless agree with it. But he doesn’t see how to justify it.
    Contrast that with this conditional statement: “If all men are mortal and Socrates is a man, then Socrates is mortal.” I don’t think Hume would see this as problematic (although Lewis Carroll would). He’d say that recognizing its truth is a hallmark of rationality. Anyone who doubts it shows themselves to be intellectually deficient. But in he is-ought case this isn’t so. Perhaps they show themselves to be morally deficient according to conventional moral norms. But they aren’t being irrational. For “It is not against reason that I should prefer the destruction of half the world to the pricking of my little finger.”


  2. Hi Em, thanks for the comment. That all sounds right to me. I think Hume's issue is about where our moral principles come from. Where do we get principles that are antecedently descriptive and consequentially normative? This is essentially the same problem as his problem of induction, where we are divining the oughts of nature rather than the oughts of morality.

    I am inclined to think that Hume would have profited from Caroll's sophistication, because he would not (like everyone else at the time) have labored under the false belief that “If all men are mortal and Socrates is a man, then Socrates is mortal." is a hallmark of rationality. If he could have appreciated that it is just as much in need of logical justification as the sorts of claims he was concerned with, then he might not have regarded is-ought principles as singularly problematic..

    Put differently, he would have seen that his naturalistic solution of characterizing moral principles and the principle of induction as features of our cognitive and emotional architecture applies equally well to the sorts of principles he was unperplexed by.

  3. In her October 10th talk at CSUS on Judging others: the limits of perspective taking, Jennifer Zamzow pointed out that in the Milgram experiments everyone predicted that they would never shock the subject at the highest level, but 65% did. This might be explained as above. Subjects substitute a normative judgment for a prediction because that is what is available.