Does “ought” imply “can”?

Many analytic philosophers think that when we use the word ‘ought’, we express a desire (typically, a desire for everyone to behave in a particular way). The object of a desire is something considered valuable or worth striving for — a goal. By contrast, when we use the word ‘is’, we express a belief. If the belief is true, it corresponds to an actual state of affairs — a fact. Any fact exists independently of how desirable or undesirable it happens to be — it isn’t affected by what “ought” or “ought not” to be the case.

This distinction between “is” and “ought” (or if you prefer, between facts and values) originated with David Hume. It’s really part of his larger theory of action.

When we act (as opposed to merely twitching involuntarily, say) our mental states cause bodily movements. These mental states include both beliefs and desires. In other words, we do things not simply because we think the world is actually arranged in a particular way, but primarily in order to bring about an as-yet unrealized arrangement that we regard as valuable. All acts have goals, and are caused by goal-directed states — in other words by desires. According to Hume, the purpose of belief is secondary to the purpose of desire. We have an internal “map of the world” so that we can realise our goals. Reason is the slave of the passions, as Hume put it.

Now philosophers often talk about belief, typically as part of the analysis of the concept of knowledge. But they hardly ever talk about desire. I think this is a terrible oversight. The narrow fixation on belief promotes the misleading idea that when people do bad things, it is their beliefs that are at fault. And the most obvious way beliefs can be at fault is if they are held without sufficient evidence. In this way, lack of sufficient evidence (or uncertainty) can come to look like the root of all evil. Thus “skeptics” are engaged in a moralistic crusade against “lack of evidence”. Richard Dawkins identifies religious faith as the main culprit in barbarous cultures that mutilate children and oppress minorities. And so on.

I submit that we should focus our attention on desires rather than beliefs. Instead of dwelling on the irrationality of faith, or suchlike, we should consider desirous states such as the willingness to be cruel. Acts are bad when they are deliberately aimed at — or inadvertently result in — bad things. These are usually the product of bad goals.

We can get confused about the difference between “is” and “ought”, because sometimes we describe psychological facts about ourselves, including such facts as that we desire something. If we’re not careful, it may seem as if a description of such a desire expresses it, and so there may seem to be a blurring of the distinction between “is” and “ought”. But the difference can be put in quite stark terms. An “is” is true or false. But an “ought” is neither true nor false — beneath it lies a “pro attitude”, which has more in common with a command, prescription, exhortation, recommendation or endorsement than a description. An “ought” expresses “to-be-done-ness” rather than describing any sort of fact.

There is another reason why we get confused about the difference. Moral “ought”s are usually disguised as “is”s. This is no accident. A moral “ought” expresses ways we would like everyone to behave. It isn’t meant to express a desire that the speaker alone happens to have, but one that everyone is meant to share as a matter of necessity. Its scope is meant to extend beyond the subjective first-person singular, and apply “inter-subjectively”. The ambition of inter-subjectivity takes the form of an appearance of objectivity. And objectivity involves facts. So moral “ought”s are dressed up to look like true descriptions of moral facts. A moral “ought” is a desire masquerading as a belief.

One of the best-known consequences of the sharp distinction between “is” and “ought” is the idea that it is a fallacy — a mistake in reasoning — to try to derive one from the other. Hume himself complained about writers who slid imperceptibly from sentences that purport to describe facts to sentences that prescribe behavior. The currently-popular way of putting it is that you can’t derive an “ought” from “is”s alone. But the converse is also true: you can’t derive an “is” from “ought”s alone. The claim that “ought implies can” is an attempt to do just that. Why do so many philosophers make that claim?

The answer, I think, is that they are thinking parochially about moral “ought”s. They assume an “ought” prescribes behavior like a rule, and it is an agent’s duty to comply with such a rule. No reasonable person would charge an agent with a duty he could not be expected to perform by asking him to follow a rule he cannot comply with. So it is assumed that he must be able to comply any duty he is charged with as a matter of fact.

This is to think parochially about “ought”s, because in effect it assumes that all “ought”s are like those of a deontological moral theory. But there are other kinds of “ought” than those of ethics, and other ways of thinking about ethics than Kant’s.

Consider chess. Several rules prescribe how the pieces move — the bishops should move diagonally, and so on. There is also an “object of the game” — to get your opponent’s king into checkmate. Any of these could be understood as an “ought” of playing chess. If a person sitting at a chessboard moving pieces around does not observe every such “ought”, he isn’t really playing chess at all. I submit that philosophers who claim “ought implies can” are thinking only of the first sort of “ought” — one of the “rules of the game”. The second sort of “ought” expresses an ideal that might or might not be achieved. Depending on who is playing, it might be practically impossible to achieve it, although it continues to guide the behaviour of the players.

There are various activities in life — amoral or moral — in which impossible ideals successfully guide behavior. A “Gordon Gekko” type whose only motive is profit might do everything in his limited power to maximize profit. Yet it remains impossible for every decision he takes to actually result in the greatest profit. The “object of the game” here is expressed by an amoral “ought” that does not imply “can”. There are examples in ethics too: utilitarians are committed to maximizing an independently-specified good such as happiness or the satisfaction of desires. There is no conflict between striving to achieve such a goal, and its being practically impossible to actually achieve it. Again, “ought” does not imply “can”.

But even with the first sort of “ought”s — the “rules of the game” rather than the “object of the game” — we can still imagine how they need not imply “can”. It would have been very unreasonable for the inventor of chess to write rules that can’t be complied with, such as that pawns move forwards ten squares at a time. (Unreasonable because there’s isn’t room on a chessboard for a pawn to move that way.) But unreasonable and all though that would have been, no logic prevents it. They ought to move that way, but they can’t. This is a possibility because of the independence of “is” and “ought”.