The Fact-Value Distinction by David OderbergFebruary 4, 2011
The widespread skepticism about ethics today — the idea that there is no knowledge to be had about right and wrong (any more than about art –`there is no disputing matters of taste’, it is often said) — owes much to the empiricist tradition in philosophy. Empiricism places the foundations of knowledge wholly in experience, observation, feeling and sensation. If a discipline cannot be understood — at least ultimately — in those terms, it does not have a claim to objectivity. It is arguable whether, understood correctly, an empiricist philosophy should rule out ethics on this count, but it would take us too far afield to discuss empiricism in general. What I want to focus on is the so-called fact-value distinction which arises out of it and which is responsible for much modern skepticism.
The distinction finds its classic statement in the philosophy of David Hume. He famously remarked: `In every system of morality, which I have hitherto met with, I have always remarked, that the author … makes observations concerning human affairs; when of a sudden I am surprised to find, that instead of the usual … propositions, is, and is not, I meet with no proposition that is not connected with an ought, or an ought not.’ And he considered it `altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it’.’ In other words, how is it possible to get from statements of fact — statements about what is — to statements about what one ought to do, or what ought to be? Is there not an unbridgeable logical gap here? If so, then it seems morality is really detached from fact and must be about something else: feeling, taste, emotion, or whatever.
The so-called fact-value distinction embodies various confusions. but only the most important will be mentioned here. The first concerns facts. Proponents of the distinction argue as follows. Facts are the elements of reality that make declarative sentences true. To take a simple example, the sentence `Charles ate a sandwich for lunch’, if it is true, is made true by the fact that a certain individual, Charles, ate a particular thing, namely a sandwich, for his lunch. If reality did not contain this fact (say, because Charles did not exist, or ate ham and eggs for lunch, or are a sandwich, but for dinner rather than lunch), the sentence would not be true.
A similar analysis can be made of any fact, they say. But what about moral facts? What possible elements of reality would make true the sentence `Breaking a promise is wrong’? Ethical sentences — about what is right or wrong, good or bad — are in a different class from statements of fact, since there is no realm of reality in virtue of which they are true or false. Where is this `ethical realm’? How do we know about it? How can we check the truth of ethical statements against it? Are we to say that just as the predicate ‘ … is green’ refers to a property of the real world, namely the color green, so there is a property of the real world, namely wrongness, to which the ethical predicate ‘ … is wrong’ refers?
The above line of argument is a familiar one. Although it has been refined and complicated in various ways, and more subtle versions have been developed, the basic idea is always the same. It is also easily countered in several ways, but only one need be proposed here. The proponent of the fact-value distinction is in the dilemma of having to provide an account of `fact’ that neither begs the question against the believer in ethical facts nor allows such facts to be philosophically respectable after all.
Suppose the Humean (which I shall call the believer in the fact-value distinction) understands facts to be elements of concrete, observable reality that can be used to verify (or falsify) propositions. On such an understanding, it seems ethical facts do not get to first base. One would be hard pressed to say that we could check the statement `Breaking a promise is wrong’ against the fact that breaking a promise is wrong in anything like the way we could check the statement `Charles ate a sandwich for lunch’ against the fact that Charles ate a sandwich for lunch. The Humean is right: Where do we check? How? What do we look for?
The problem for the Humean, though, is that this account of facts simply begs the question against the ethical realist (as I shall call the believer in moral facts). All it amounts to is the insistence that if there were moral reality it would have to be just another part of empirical reality, which is supposed to close the debate. But the ethical realist does not share this insistence. If the only philosophically respectable facts are observable elements of concrete reality, then moral facts seem to be ruled out from the very beginning and nothing else the realist can adduce in favor of moral facts will count. But by claiming that there are moral facts, the realist is not thereby assuming that the empiricist view of facts is correct. If the skeptic about moral facts wants to use the notion of a fact to cast doubt on realism, then, he must not rely on a conception that the moral realist does not share in the first place.
Of course, the Humean is bound to justify his belief that the only genuine facts are observable elements of concrete reality, and he may well try to do so. This would involve a full-blown justification of a view of empiricism that supports such a belief, a view which there is no space to canvass here. The point is simply that, short of doing this, he will need to rely on a less controversial conception of facts, one which does not foreclose the debate against the realist. Again, without going into detail, it should be noted that the more deeply we look for a substantial view of facts as elements of reality, the more we realize how complicated our conception must be and the less probable becomes the claim that there is no space for moral facts.
Where in reality would we look for facts to make true or false statements such as: `If Alan had bet on horse number five, he would have won’; ‘One clay the universe will come to an end’; `Mountain climbing is a dangerous sport’; `Brian is a peculiar sort of chap’; `The joke Alan told me is very funny’; and so on? The more one thinks about such statements, the harder it is to point to an element of concrete observable reality that verifies or falsifies them. Let us leave it to the Humean to try to perform the impossible task of accounting for such cases while retaining a conception of facts that straightforwardly rules out the ethical.
A different way of understanding facts is possible, however. Instead of thinking of them as elements of concrete reality, we might plausibly think of them as true propositions. Part of the motivation for this approach is the notorious difficulty of individuating facts. One theory is as follows: sentences express propositions (for instance, a proposition is what is expressed by two sentences, in different languages, that say the same thing); it is propositions that are the primary entities that are true or false, with sentences deriving their truth or falsity from them; and a fact is whatever it is to which a true proposition corresponds.
We can individuate facts, then, by means of the true propositions that correspond to them. For instance, the fact that snow is white is distinct from the fact that snow is cold because the propositions that snow is white and that snow is cold are themselves distinct. But the problem then is that in order to know whether a proposition is true, you need to know if it corresponds to a fact; but in order to know whether there is a fact to which it corresponds, you need to know whether the proposition is true in the first place! So, in order to know whether the proposition that snow is white is true, you need to know if there is a fact that snow is white. But to know the latter, it seems you need first to know the former. It appears, then, that one cannot really separate facts from propositions.
This might make one move to a view of facts that does not take them to be anything over and above true propositions. (Let us leave aside false propositions, since the point can be made without considering this complication.) The fact that Charles ate a sandwich for lunch, then, just is the true proposition that he did so. If we take this less metaphysically robust notion of facts to be correct, however, moral facts begin to look philosophically respectable. For the fact that breaking a promise is wrong just is the proposition that breaking a promise is wrong, which is true. Moral facts, on this account, suffer from no special metaphysical problems. Just as the proposition about Charles can be asserted, denied, believed, entertained, and argued over, and conclusions can be drawn from it, the same goes for the proposition about promise-breaking. These are intrinsic features of all propositions.
The second point concerning the fact-value distinction can be dealt with more briefly. Proponents of the distinction often argue that morality is about reasons for action. For instance, the truth of `Pornography is morally objectionable’ is a reason not to promote it in any way, say by selling it or financing it. We can look further at the facts on which this reason is based, such as that pornography is a cause of sexual crime; but what about the reason itself: isn’t it over and above the facts that give rise to it? When talking about reasons, are we not also talking about norms or rules of behavior, and not facts? To the extent that this objection is merely a repetition in a different guise of the first one, namely the supposed strangeness of moral facts, we need not go over the same ground. What should be added, however, is that there is nothing peculiar in something’s being both a fact and a reason.
Reasons are about explanation and justification. The reason why Charles got food poisoning, for instance, is that he ate contaminated meat. The fact that he ate contaminated meat is also an explanation of his becoming ill. The fact that the meat was contaminated and the fact that he ate it give a justification for claiming that there is a causal connection between them and his becoming ill. The fact that the meat was contaminated can be used as a premise in an argument to the effect that he became ill because of what he ate. These are all familiar characteristics of the enterprise of giving reasons, but they are no less important for being familiar.
Facts can be reasons for saying certain things, for putting forward certain arguments, for drawing certain conclusions, for holding certain beliefs, and so on. A fact is no less a fact, as it were, for being able to do duty in these different ways. On the contrary, something would not be a fact if it did not have such features. Why then, are moral facts to be discounted because they have precisely the .same features? To say that morality is about reasons for action as opposed to thought, belief, assertion, and so on, is spurious; for these are all kinds of action as well. In short, the so-called duality of facts and reasons is an illusion.
This brings us to the central problem with the fact-value distinction, of which the above points are really just aspects. Taken one way, the distinction does imply an unbridgeable conceptual gap between facts and values — but the cost of forging it, for the Humean, is that he loses his grip on reality. Taken another way, however, there is still a distinction that can be made — this time one we are bound to accept, but not one the Humean can make any use of in supporting his skeptical position.
Consider the following description. Charles picks up a shiny metal object. The object contains several movable parts, one of which can be rotated. Charles rotates it. He then places his finger on another, curved movable part, holds up the object in front of him, and points at another object further away. He then moves his finger towards him while keeping it in contact with the carved movable part. This, is a perfectly coherent factual description. Does it enable us to know what Charles is doing? Not completely; but then no description of any situation can be complete. So why should we ask for any more information? Well, Charles might be doing a number of things — for example, he might be playing Russian roulette with a friend. “To say that this is not important extra information is absurd: it is information of the highest importance, which allows us to realize that Charles is doing something dangerous.
Consider another description. Alan digs a hole in a meadow. Is there anything else of any importance that we need to know? It might be wondered — what else is there to know? But suppose I tell you that the meadow in which Alan is digging a hole is an ancient religious burial site considered sacred by the local inhabitants, and that digging a hole in it is a gross profanation. Does that extra information not enable us to realize that Alan has offended the local inhabitants?
The point that such examples make clear is that there are various ways of describing a situation, ways which are more or less precise and more or less inclusive of important information — important because it may enable us to know what is really going on. It might be thought that no philosopher has cast in doubt such an obvious and vital point, but it is precisely this that was famously denied by Hume in his discussion of whether morality is something which can be recognized by the intellect in factual situations. He supposes:
Let us choose any inanimate object, such as an oak or elm [by ‘inanimate’ he means `not an animal’ as opposed to `not alive’; and let us suppose, that by the dropping of its seed, it produces a sapling below it, which springing up by degrees, at last overtops and destroys the parent tree: I ask, if in this instance there be wanting any relation, which is discoverable in parricide the murder of a parents or ingratitude? Is not the one tree the cause of the other’s existence; and the latter the cause of the destruction of the former, in the same manner as when a child murders his parent?
And he concludes that as the `relations are the same’ in both cases, and `as their discovery is not in both cases attended with a notion of immorality, it follows, that that notion does not arise from such a discovery’.’
At one level of description, Hume’s analogy between a sapling killing its parent tree and a child murdering his parent is correct. Indeed, Hume himself does not describe the case at the right level since he uses the word `murder’, which itself connotes wrongdoing. But if we think of them simply as cases of one living thing killing another, it is hard to see where morality can be `discovered’, to use Hume’s word. The problem is that the only levels of description at which Hume’s point has any force are those that involve a radically impoverished apprehension of reality: not an impoverished conception of morality, but of what exactly is going on. Only if we ignore crucial facts, properties and features of Hume’s analogy, or of the examples I gave above, can we hope to divorce fact from value. Of course there is nothing unusual about such examples — they can be multiplied indefinitely.
Any situation can be redescribed, or described at such a level of incompleteness, that it leaves out precisely those aspects that enable one to ‘discover its moral characteristics. We need not speak of murder — we can instead talk about killing, or perhaps termination. We need not talk about cowardice — we can instead talk about the conscious avoidance of situations involving minimal risk to life and limb. We need not speak of lying – instead we can speak of being `economical with the truth’, as a high-ranking British civil servant told the court in a famous legal case in the 1980s. Indeed, it is bureaucrats who are specially trained to describe situations at such a level of abstraction and incompleteness that they do not actually say anything false — but neither do they tell the whole truth, how things really are.
The proponent of the fact-value distinction, then, can purchase his distinction, but at the high cost of having only a partial grasp of reality. A more complete grasp is called for, but it is not one according to which the Distinction is remotely plausible in the sense that the Humean requires it to be to support skepticism. With a more complete appreciation of reality, there will still be a distinction between facts and values; there will still be a way of describing the world that only pays attention, say, to microphysics, to chemistry, to the movements of particles, to the interaction of objects, to pure cause and effect, and so on. But these descriptions will only capture a segment of reality, one which has a definite but limited place in ethical theory.
More generally, ethics is about good and bad. But goodness and badness are fundamental features of reality as much as any other. Why is it good to look both ways before crossing the road? After all, how can there be any objective goodness in the mere turning of the head while crossing a piece of bitumen? And why is it good to eat green vegetables? Where is the objective goodness in the mere ingestion of organic matter containing chlorophyll which contributes to cell repair and regeneration? The questions need only he posed for their answers to be self-evident. Both actions are good for human beings, because they prevent bodily harm and promote health. And the prevention of bodily harm and promotion of health are themselves good for human beings, which is where the actions which realize them derive their goodness. In this sense, then, the Humean can forge no relevant distinction between fact and value. Indeed, it is value `all the way down’.