Is it possible to get an ‘ought’ from an ‘is’? That is, can we derive normative claims about what ‘should be’ from descriptive claims about what ‘is’? David Hume, and many others since, have argued that there exists an insurmountable gap between these two types of claims. In this piece I will argue that this view is mistaken. I begin by outlining a theory of morality, based on the reductive naturalistic account developed by Peter Railton. I will then outline how, given such a conception of morality, the ‘is-ought’ gap rests on a conceptual confusion, since ‘ought’ claims simply are a certain type of ‘is’ claim, and as such there is no intrinsic problem in deriving one from the other. I then proceed to consider in turn the issues of moral motivation and moral relativism, arguing that the moral framework I have presented is sufficient to provide objective, non-relativist reasons for action for any morally-competent person. In doing so I argue that the notion of a rationally compelling reason for action that is totally independent of the desires or objectives of the agent in question is incoherent, and so the inability of my metaethical theory to provide one does not constitute a limitation of the account, but merely reflects a constraint on what itself reason can do.
Railton’s Reductive Naturalism
First of all, let me present an account of what an ‘ought’ claim is. The account I present here is a condensed version of that outlined by Peter Railton, and is a form of reductive naturalism. Railton constructs his account of the good in two stages. First, he defines a concept that he calls ‘non-moral good’:
“An individual’s (non-moral) good consists in what they would want themselves to want, or to pursue, were they to contemplate their present situation from a standpoint fully and vividly informed about themselves and their circumstances, and entirely free of cognitive error or lapses of instrumental rationality.”
The basic idea of this account is that what is (non-morally) good for a person is what would satisfy their idealised preferences. By ‘idealised preferences’, I mean the preferences they would want themselves to hold if they were ideally rational, and fully informed about all possible considerations and alternative courses of action, ways of living, consequences of different alternatives, etc. This additional element of abstraction enables the account to incorporate the fact that we often don’t know what we actually want, or what would be good for us, and so frequently hold mistaken beliefs about what will satisfy us or be helpful for us.
Next, Railton generalises this conception of the ‘non-moral good’ into ‘moral good’ by introducing a social dimension:
“X is morally right if and only if X would be approved of by an ideally rational and fully informed agent considering the question ‘How best to maximize the amount of non-moral goodness?’ from a social point of view, in which the interests of all potentially affected individuals were counted equally.”
So, by Railton’s account, the moral good consists in maximising non-moral good from a social point of view, treating the interests of all individuals equally. I think this captures a very significant aspect of morality, which is its intrinsic social, relational dimension. Morality is about incorporating the concerns of others into one’s own decision making, and not privileging one’s own interests over those of others. I think this is key aspect of morality is captured very well in Railton’s account, and this acts as a powerful consideration in its favour.
‘Nasty Preferences’ Objection
What if someone has nasty preferences? What of the person whose preference is to rape little children, or torture animals, or whatever other depraved action we may care to contemplate? If they have a preference to do these things, does that mean that such actions are good by this account?
I believe this objection rests on a twofold failure to understand Railton’s account of moral goodness: an initial failure to understand his notion of ‘non-moral goodness’, and a subsequent failure to understand its extension into ‘moral goodness’. In regard to ‘non-moral goodness’, this objection presumes that what an individual’s preferences are now is the same as what a fully informed and instrumentally-rational version of that individual would want their preferences to be. This assumption is unwarranted, because precisely the point of introducing such a concept is because the two are often considerably different. If we consider, for instance, the cat torturer or the child molester, we must ask the question: what sort of life would they want themselves to pursue, and what sort of preferences would they desire themselves to have, if they had access to all relevant information about possible alternative lifestyles, approaches to living, sources of meaning, etc? If they could somehow consider all the alternatives and choose carefully which they most preferred, would they still choose child molestation or cat torturing? I consider this to be very unlikely; I just don’t think it is the case that many people would actually find those to be the good life for themselves, if they had proper access to alternatives and the time and mental clarity to consider them. To an extent, this is an empirical question, and one I would welcome further research into. Nonetheless, it seems to me that it is at least highly plausible to argue that at least a sizeable proportion of cases of ‘objectionable preferences’ would be eliminated by simply considering idealised, rather than actual, preferences.
That being said, let us say for the sake of argument that there is some subset of people who, even after considering all the other possible modes of living and their comparative virtues, nonetheless would still prefer a lifestyle which we would view as abhorrent or undesirable in some way (e.g. cat torturing, child molesting, etc). Does Railton’s account require us to say that such a lifestyle (and its constituent actions) are morally good? Not at all, for such an inference rests on a misunderstanding of the second aspect of Railton’s account, the distinction between ‘non-moral good’ and ‘moral good’. Even if we accept the premise that, for some small number of depraved persons, their ‘non-moral good’ consists in doing things which we would consider to be repugnant, when determining the moral good we must also take into consideration the interests of others. If, from a social point of view, these interests outweigh the non-moral good of the individual in question (as I think they inevitably will for the sorts of abhorrent acts mentioned above), then it will still be the case that Railton’s account declares such actions to be immoral. In essence, it is not the ‘non-moral good’ of a particular person which is the most important; what is fundamental to this account are the interests of all individuals considered from a social point of view. And needless to say, the abhorrent preferences of a depraved few will be easily outweighed by the former. Thus, the ‘nasty preferences’ objection fails as a substantive critique of Railton’s account of morality.
Deriving ‘Ought’ from ‘Is’
Dissolving the Problem
Having outlined Railton’s reductive naturalistic account of morality, I now wish to return to the main question with which I began this essay: how to derive an ‘ought’ from an ‘is’. My answer, in essence, is that there is actually no real problem, because ought claims just are ‘is’ claims. By Railton’s reductive naturalistic account, moral claims constitute a particular subset of factual claims about the way the world is. Specifically, moral claims are statements about what would satisfy the idealised preferences of individuals in a society, treating every person equally. By this account, therefore, moral claims are both ‘ought’ statements and ‘is’ statements – the former is in fact a subset of the latter. The alleged difficulty of deriving an ‘ought’ from an ‘is’ thus rests on a conceptual confusion, namely presupposing ex ante that the two claims are of a fundamentally different sort – a view which I see no reason to accept. I believe that a perfectly sensible reductive naturalist account of the nature of moral values can be provided (viz Railton’s reductive naturalism), and as such there is simply no difficulty in deriving ‘ought’ from ‘is’, since ‘ought’ claims just are a particular type of ‘is’ claims.
But how can I make the leap from the statement “X will maximise idealised preference satisfaction from a social point of view” to the statement that “X is morally good”? Isn’t the first simply a claim about facts of the world – an ‘is’ statement – while the latter is an evaluative claim – an ‘ought’ statement? How to we get from the one to the other? Asking the question in this way is, I think, betraying precisely the sort of conceptual confusion which I am here attempting to defuse – namely, that there is some unbridgeable gap or ultimate chasm between ‘is’ and ‘ought’. This is precisely what I am denying. I am arguing, in fact, that the set of actions and states of affairs which satisfy the predicate “maximises idealised preference satisfaction from a social point of view” is the same as the set of actions and states of affairs which satisfy the predicate “is morally good” (caveat: I am not actually claiming the sets are precisely equal. More carefully, I am saying that the sets are close enough for this account of morality to be a very good one).
Moore’s Open Question Argument
It is important to understand that I am not saying that these two things (‘moral goodness’ and ‘maximising social preferences’) are the same by definition (this point formed the basis of Moore’s famed ‘Open Question Argument’ against ethical naturalism). Rather, what I am saying is that we begin with this concept called ‘morality’, which is rather fuzzy and confused in some ways, but which we nonetheless seem to posses some basic intuitive understanding of. We then think very hard, look around at the world, and ask the question: “does this concept of ‘morality’ refer to anything real?” That is, we ask ourselves whether we can give a sensible account of this thing we call ‘morality’, or, alternatively, will we find that no sensible account can be given, and that like so many other naive concepts (such as bodily humours, phlogiston, absolute space and time, élan vital, and many more), we must consign morality to the dustbin of history? What I am arguing is that in thinking hard about our concepts and looking around at the world, we do find that there is in fact something which matches up quite well to our notion of morality, and as such our conception of morality is validated – not by definitional fiat, but by conceptual analysis and empirical investigation.
It is true that the account of morality that Railton provides is (as he calls it) a ‘reforming definition’, by which he means that it does not, absolutely and in all respects, match up perfectly to our naive, unreflective conceptions of morality. However Railton also argues, as do I, that the account is sufficiently close to that of ‘naive morality’, and also has sufficient explanatory power, such that we are adequately justified in maintaining the concept of morality and declaring that we now have a reasonable account of what it refers to. This contrasts, say, with the concept of a ‘vital life force’, which was not found to match up sufficiently with any real-world phenomenon, and as such the idea was discarded.
Water and H2O are not defined to be the same thing – we know they are the same thing because we went out into the world and looked, and discovered that they are the same. I am saying that something analogous can be done with morality: we begin with a naive concept, we go out and see if there is anything that is a reasonably close match for this concept that we can give a sensible account of. Since we are able to do this, the concept of morality is validated, and we are affirmed in our use of it (and indeed our understanding of it is heightened). Morality refers to idealised preference satisfaction from a social point of view. We have not defined it to mean that – we examined our concepts, we examined the world, and we matched the two together. We constructed an empirical explanatory account of the concept of morality, just like we have done in the sciences for many other entities.
So, bringing the threads together, how do we make the leap from the positive to the normative? I am saying that there is no leap. We have constructed an account of the normative by which we understand that it just is one class of claims about matters of fact. The problem is simply dissolved; there is no ‘is-ought’ gap under a reductive naturalistic metaethic.
Reasons for Action
Rationality and Moral Motivation
It is important to understand that there is one thing which this account of morality does not do: by itself, it does not provide us with any universal reason to be moral, or even to care about morality at all. At this point, I will say that if what one expects morality to provide is some reason for action which is in some way ‘binding’ or persuasive to all rational agents of any sort, regardless of their particular goals or objectives or preferences, then one will be disappointed, because this is something this account does not provide. I do not, however, consider this to be any real problem, because I think that the notion of something which is universally rationally compelling to all agents, regardless of their particular preferences or goals, is simply incoherent. I understand ‘rationality’ to mean “taking effective means to one’s ends” (as borrowed from D. A. Lloyd Thomas). By this understanding, there is simply no fact of the matter as to what is ‘rational’ for an agent to do or pursue absent some predefined ‘end’ or ‘purpose’ or ‘objective’.
Appeals to Teleology
Taking an Aristotelian perspective on the matter, one might say that there is a fact of that matter as to what a rational agent should do, even absent any particular end or objective which they hold to, and that relates to the agent’s underlying ‘innate’ teleology, or purpose in being. For example, Aristotle famously argued that humans are rational creatures, and so our teleology lies in the pursuits of reason, wherein man achieves his greatest excellence. Though I think there are some valuable insights to be gleaned from this approach, overall I find it unpersuasive as a metaethical theory as it fails to provide any clear notion of what constitutes an innate teleology or whence such a notion can be derived. Perhaps a notion of teleology could be constructed via some reductive account of the moral good, for example by using aspects of Railton’s account. However in that case the notion of teleology is doing no explanatory work, as it is merely being defined in the same manner as morality, rather than providing a more ‘ultimate’ basis out of which the latter can be defined or understood. As such, ultimate teleology is either a mistaken concept (as it does not refer to anything real), or an unnecessary one (because it is merely parasitic on the concept of morality, instead of underpinning it as we had hoped).
Though there is much more to say on the notion of teleology, I will forebear at present, and merely summarise this portion of the essay by restating that I do not believe there is any fact of the matter as to what it is rational for an agent to do sans any account of their particular interests or desires. As such, I do not believe there are any universally-compelling reasons to be moral. On the other hand, I also do not think there are any universally-compelling reasons to believe in the truths of mathematics or logic or science, or indeed to care about truth or good arguments at all. In short, I do not think there are any universally-compelling reasons for anything at all. Every reason of any sort must, at bottom, make some sort of appeal to a pre-existing goal or value or presumption; reason always has to begin somewhere. As such, I think a notion of objective-independent reasons which are rationally compelling to all rational agents independent of their goals or desires is simply incoherent, and is not something we should demand from an account of morality.
Reasons to be Moral
I must emphatically emphasise, however, that I am not saying that we do not have good reasons to be moral. I believe there are immensely good reasons to be moral, none of which are particularly innovative on my part: acting morally helps others, it allows us to reduce suffering and promote human flourishing, it brings us joy and helps us to become better people (in either a moral or non-moral sense), etc. Are these not good enough reasons to care about being moral and doing good? I think they should be perfectly good reasons, perfectly sufficient reasons for action, for any morally competent person. They will not, of course, be compelling to a person who does not already possess some fundamental, underlying ‘moral competence’, but why is this a concern? Do we really expect that the reason people lack moral motivation is that they are actually irrational; that there exists some ‘killer reason’ or ‘overwhelming argument ‘ which would cause them to become morally motivated by the sheer force of reason? To me that sounds absurd on the face of it, and even more absurd in the light of the analysis I have given about the nature of rationality. I think we all have ample reasons to be moral. If those reasons are uncompelling to a particular person (assuming they understand them), it is because said person is simply not morally competent. That is unfortunate, but is a defect of that person’s character, and not of the account of morality which I am here defending.
Relativism and Subjectivity
A Non-Subjective Morality
In metaethical discussions, the question of relativism inevitably rears its ugly head. Is the account that I have here presented of the meaning of and basis for moral claims a relativistic one? That is, when I say “it is wrong to murder” or “you ought not murder”, am I merely expressing a subject personal preference against murder? Is it simply a question of taste – some people like murder and some people don’t, and we just happen to be fortunate enough to live in a society in which most people share my distaste for it?
To answer this question: no, I do not think that moral claims are subjective or relative in this way. By the account that I have defended, moral claims are statements of fact which can be true or false. Their truth or falsity depends upon the manner in which the world is constituted, including certain properties of human beings, their desires and preferences, how these preferences are socially distributed, etc. These facts are objective, in the sense that they are not dependent upon the minds of those making the judgement. It may be the case that peanut butter is delicious ‘for me’ and disgusting ‘for you’, but by Railton’s account of morality, it cannot be the case that X is morally right ‘for you’ (say because you dislike murder), but not morally wrong ‘for me’ (say because I’m not bothered by murder), since what is morally good is determined by facts of the world which are invariant to the perspective of the person considering them. Of course, different people may disagree about what these moral facts are, but disagreement does not imply that there is no fact of that matter at all.
What maximises idealised preference satisfaction from a social point of view does so regardless of whatever your or my attitudes about murder or anything else might happen to be. As such, this account of morality is not relativist, and does not degenerate into mere preferences. (Note: the concept of ‘non-moral good’ is subjective in this way, because it depends on the agent’s idealised preferences, however the notion of ‘moral good’ is not subjective in this way, because it depends upon the preferences of all individuals, and not merely the subjective attitude of the person making the ethical judgement).
When I make the injunction “you ought not murder”, am I merely expressing a personal, subjective attitude towards murder, which at base is no more or less justifiable than any other possible attitude? I do not think so. Rather, what I am doing is making a statement concerning what would maximise the idealised preferences of individuals from a social point of view, which is an objective claim that is not dependent upon any single person’s attitudes towards the claim. Now, of course, whether or not the person whom I am addressing cares about what is right or wrong is completely separate question; they may accept that murder does not maximise the idealised preferences of individuals from a social point of view, but simply not give a damn. That is unfortunate, and I would of course try to persuade them that they should care (by appealing to the ‘reasons to be moral’ I outline above), but as I previously argued, if my interlocutor does not share at least some fundamental concern about the welfare of others, or a desire to do good (or something like that), then there is simply no reason I can give them which will rationally ‘compel’ them to care about doing good. As I argued, the notion of ‘bootstrapping’ any motivation from ‘pure reason’ in this way, without any reference at all to pre-existing desires or concerns, is simply incoherent. This, however, is not a limitation of my account of morality – it is imply a fact about the limits of reason. Nor does it follow that moral injunctions are mere statements of preferences. They are statements of facts about the world, which one may or may not happen to care about. If one does not care about these moral facts, that does not necessarily make one irrational. But it does make one immoral, in an objective, mind-independent sense.
What of objective moral obligations? Do they exist, and do they have any normative force? Can we make sense of them within the framework I have outlined? I would argue that the notion of ‘moral obligation’ is mostly (if not entirely) redundant, as to say that someone has a ‘moral obligation to do X’ is just to say that ‘it is morally good to do X’ or ‘you ought to do X’, which, in turn, simply means that ‘doing X will maximise the satisfaction of idealised preferences from a social point of view’. As such, I do not believe that the concept of ‘moral obligation’ offers any addition insights or provides any additional moral or motivational force beyond that which existing concepts already possess, and so while I think objective moral obligations are perfectly compatible with a reductive naturalist account of moral value, I also think they are a rather superfluous addition.
In this essay I have argued that there is no fundamental problem with deriving ‘ought’ from ‘is’ claims, because ‘ought’ statements are in fact merely a special type of ‘is’ statements. In making this argument I first outlined Peter Railton’s reductive naturalist conception of moral good, and defended it against the objection of ‘nasty preferences’. I then argued that on the basis of this account, the ‘is-ought’ gap can be dissolved as resting on a conceptual confusion, because ‘ought’ claims simply are a special type of ‘is’ claims. Morality is validated, I argued, on the basis of our ability to construct a plausible naturalistic account of its relationship to objects in the real world. I then considered the issue of moral motivation, arguing that this account of morality will not motivate those who lack even the most basic moral competence, but that this does not constitute a limitation of the account, as the notion of a reason that is rationally compelling to all agents regardless of their personal preferences is simply incoherent. Finally, I considered the issue of subjectivity, arguing that the account of morality I outlined is not subjective, and does not rest merely on the attitudes of those making moral statements.