The fact-value distinction is, unfortunately, widespread wherever there is serious thinking about science, ethics or aesthetics in the Western world. It is based on the logical argument used by Hume[1], that a value claim cannot be implied by a factual claim – you cannot derive an ‘ought’ from an ‘is’. For example, according to this approach even the claim “Everyone believes murder is wrong” would not, if true, imply “murder is wrong.”

This logical distinction is based on the assumption of different justifications for believing factual and value claims. Factual claims are taken to be justified either through observation or through a priori reasoning, whilst value claims are taken to be conventionally agreed within a group or society. By this means, ‘facts’ become in principle absolute – verbal claims that have been tested for their precise correspondence to reality – whilst ‘values’ become in principle relative because they lack this correspondence. Even many thinkers who do not accept the relativism of Hume’s account of ethics have accepted the fact-value distinction, though they have offered different kinds of justifications for a universal ethics, such as revelation or intuition.

The effect of the fact-value distinction has been disastrous for Western ethics, because it has consigned all thinking to the dualism between absolutism and relativism. If one accepts Hume’s account of the matter (as many analytic philosophers have) ethics can consist only in the analysis of what people happen to think about ethics in your context, with there being no further way to show that this ethics has any universal validity. If, on the other hand, you challenge Hume’s relativism but still accept the fact-value distinction, the justifications for ethics that you can offer can only be based on the dogmatic over-extension of intuitions or a priori reasoning. Ethics has become either the preserve of mere conventional analysis or of extravagant metaphysics.

The fact-value distinction has also been disastrous for the status of science. Although it might seem that in the short term the distinction gives scientific claims a superior status separated from ‘subjective’ values, in the longer-term this has undermined appreciation of the degree of objectivity offered by science, by making its claims appear merely dogmatic when its pretensions to represent reality were punctured. Once sceptical arguments are applied to science, and the relativity of scientific claims starts to be appreciated, the dualism created by the fact-value distinction leads to the sudden loss of the credibility of science rather than an incremental calibration of that credibility in accordance with the evidence, and the falseness of the distinction, being revealed, leads to scientific ‘facts’ being seen as mere relative ‘values’. Those who now believe that creationism is a scientific theory to be treated on equal terms with evolution theory, or that homeopathy is a medical treatment to be given parity with mainstream Western medicine, do so by assuming that science, too, is ‘just’ a value judgement. Science is now getting (if you can excuse the pun) a dose of its own medicine, as an effect of previous support for the fact-value distinction under the illusion that the distinction helps to maintain objectivity. To get rid of the fact-value distinction would be in the interests of science, by enabling the public to grasp the basic point that justification is a matter of degree.

There are some philosophers who have attempted to challenge the fact-value distinction, but these have mainly done so using analytic methods, and by trying to find exceptions to the general rule, where it is claimed that our intuitions tell us that a value claim is implied by a factual claim. Thus Searle, for example, claimed that factual claims about promises implied value claims about the normativity of keeping them[2], and MacIntyre argued that the context of a ‘practice’ such as football, farming or sociology was one in which facts about certain kinds of effective actions within that practice implied values[3]. However, the ways that this approach is circumscribed make it impossible for it to challenge the absolutism vs. relativism dualism Hume created. Searle and MacIntyre were merely identifying occasions when people assume, disputably, that certain kinds of actions are good, not identifying any justification for believing that they are good in a more universal sense. Analysis by itself will only ever provide you with a description of what people believe, not with grounds for prescription. The very distinction between description and prescription is one established by the fact-value distinction, so that accepting analysis as the only legitimate (‘factual’) method of philosophical investigation into this problem is itself to implicitly accept the fact-value distinction.

Instead, we need to question Hume’s assumptions and look at the problem more widely. The prime questionable assumption made by Hume is that facts and values are justified in entirely different ways. This in turn assumes that value claims are subject to sceptical argument in ways that facts are not. Hume’s response to scepticism about empirical facts was naturalistic – he thought that we could not defeat the sceptical argument, but that we were simply not able to take it seriously, and should rely on observation as the most informative basis of judgement available. I have already commented on this argument in 1.b, pointing out that Hume has falsely assumed that the difficulties we would have in maintaining actual disbelief in our everyday perceptions should stop us taking a sceptical argument seriously, when sceptical arguments only deny us certainty about them.

If Hume’s assumptions about scepticism are wrong, so is his assumption about naturalism. We do not have to ignore sceptical arguments about factual claims in order to develop and use science, just to take factual claims provisionally. Although in practice many scientists do take their claims provisionally, the interpretation of the status of factual claims is inconsistent because of the influence of Hume’s arguments, because they can be taken to be either absolutely or provisionally correct. Furthermore, when factual claims are compared to value claims, sceptical arguments which apply just as much to the factual claims as to the value claims are not usually taken into account. “The dog is hungry” is just as subject to sceptical doubt as “You ought to feed the dog.” given all the possible mistakes we could make in interpreting the dog’s behaviour as well as issues of responsibility for animal care.

So, if we compare the justification of factual claims to that of value claims in the light of the full range of sceptical arguments, we find that they are not very different. The last of the ten modes of Pyrrhonism, for example, points out the relativity of moral claims, but all the others are concerned with empirical claims. All the other sceptical arguments considered in 1.a apply equally to factual and value claims – for example, both require further justification giving rise to a possible infinite regression, both are subject to the error argument, and both are subject to linguistic scepticism. There are no clear examples here of types of sceptical argument that apply to value claims but do not apply to factual claims. Both, therefore, are equally uncertain in principle, and both equally open to provisional assertion.

The fact-value distinction is also a result of the abstracted turn discussed in 1.e. Hume’s argument cannot be contested in its own terms in the abstract, but it does not apply to any actual examples of people’s judgements about facts or values. In the concrete situation, facts and values are never found in isolation. Whenever we assert a fact, we have a motive for doing so, which implies the value of doing so. Even a claim as apparently neutral as a mathematical claim, in the context of a mathematics classroom, comes with values attached about the value of studying mathematics (or, for younger students, perhaps the value of conforming to society’s expectations that every child will study basic maths). Facts do not exist in isolation from people who believe them. Whenever we assert a value, on the other hand, there will also be implied background facts that are assumed for that value to make sense. For example, “You should eat green leafy vegetables every day” is a value claim that assumes the existence of green leafy vegetables and the capacity of the auditor to eat them (apart from, very likely, beliefs about the relationship between the nutrients found in green leafy vegetables and health).

Without the abstracted turn, the fact-value distinction is irrelevant to us, and yet it is still assumed in many arguments about ethics, particularly in the widespread perception in the West that ethics is a personal matter, a mere matter of opinion. However, if we were to apply provisionality and incrementality to ethical claims just as much as to factual claims, it could be appreciated that ‘personal’ opinions can be justified to a greater or lesser extent. Philosophers bear a lot of the responsibility for not thinking harder about this topic in its concrete context, and for the social results of their failure.

The fact-value distinction makes even less sense if we are prepared to reform our account of meaning in the way I will be proposing in part 3. If the meaning of a claim does not just consist in a representation of the circumstances in which it would be true (or even of the social rules surrounding the language in the claim), but rather consists in the related experience of a physical organism, both cognitive and emotive, then there is no clear distinction between the type of meaning of factual statements and that of value statements. One can no longer argue, as A.J. Ayer did on the basis of his truth-conditional theory of meaning, that moral statements are strictly meaningless, because they do not correspond to any possible state of affairs[4]. Given that the meaning of a factual statement is not purely representational in its meaningfulness to us in the first place, Ayer’s emotivism is built on a set of unnecessary narrow assumptions.

One objection to such criticisms of the fact-value distinction attempts to distinguish between strong and weak versions of it, claiming that these criticisms only apply to a strong version of it, perhaps a ‘fact-value dichotomy’ as opposed to a ‘fact-value distinction’. A weaker distinction, it can be argued, can still be made on the basis of analysis of our everyday distinctions between facts and values. Even if our claims involve a mixture of facts and values, we still distinguish them in practice as fact or value claims. A physicist’s claim about the properties of hydrogen would not normally be thought of as a value claim, and “We should always tell the truth” would not normally be thought of as a factual claim.

In analytic terms this is obviously correct. We make a conventional distinction in practice between these different kinds of claim. However, when we consider the justifications for this distinction rather than merely the convention, there is not even a weak justification for the convention, given that in all possible concrete examples, facts and values remain mixed. There is also a pragmatic argument for not hardening this convention into a supposed philosophical truth, even of a weak kind. We will always recognise what is happening in practice more fully, and thus respond to it more adequately, if we take into account the ways in which facts and values are inextricably combined, and thus respond to both in any claim, rather than limiting the conditions we address by only considering one. If I recognise that the physicist’s paper about hydrogen is not just about hydrogen, but also the value of researching hydrogen, then I will respond better to the physicist if I take this point into account than otherwise. If we are even trying to move beyond a merely descriptive ethics, there is no point in trying to defend even a weak fact-value distinction. Claims to have weakened the fact-value distinction or to have gone beyond it are also not fully convincing unless we can develop an alternative to it, and show how moral objectivity is not strengthened but weakened by the abstraction of values to separate them from facts. So, this chapter needs to be read in conjunction with the account of objectivity found in section 4, and the account of ethics found in section 7 – indeed, in a broader sense, in conjunction with the whole of the rest of the volumes of this book.

---

[1] Hume (1978) p.469

[2] Searle (1964)

[3] MacIntyre (1985)

[4] Ayer (1946) ch.6