What Is the Is-Ought Gap? Getting From Facts to What Should Be Done

Hume noticed that writers on morality proceed for some time with ordinary factual claims and then, without explanation, begin using ought rather than is, and he observed that this change requires justification since the new relation is entirely different from the old and could not be deduced from it. The point is logical, since a valid argument cannot contain in its conclusion anything not present in its premises, so a set of purely descriptive premises cannot validly yield an evaluative conclusion. That leaves anyone reasoning from facts to conclusions about what should be done needing at least one premise that is itself evaluative, and the observation is therefore a demand for that premise to be stated rather than smuggled in. Stated that way it is uncontroversial and its consequences are not.

Most practical reasoning does bridge it and the bridges are identifiable:

• •An explicit evaluative premise, which is the honest approach and simply states the value being relied on
• •A shared assumption the parties both accept, which functions as a premise without being stated
• •A definition connecting a descriptive term to an evaluative one, which relocates the gap rather than closing it
• •An appeal to what a person already wants, which yields a conditional conclusion about means rather than an unconditional ought
• •Institutional facts, where a rule already in force generates obligations, though the obligation to follow the rule is a further question
• •Simply not noticing, which is the commonest route and is what the observation was aimed at

The observation is frequently confused with a separate argument made much later by Moore, who held that any attempt to define goodness in terms of a natural property fails, since it always remains an open question whether something with that property is good, which would not be open if the definition held. The two claims are related and distinct, since one concerns what follows from what and the other concerns what can be defined as what. A third confusion attaches the label to any appeal to nature, as in arguing that something is right because it is natural, which is a different error again and is a genuine fallacy. Keeping the three separate matters because they have different force, and the appeal to nature is straightforwardly bad while the other two are substantive philosophical positions with serious defenders on both sides.

The gap appears constantly outside philosophy and recognising it improves arguments. Appeals to what is natural, to what is traditional, to what most people do and to what has always happened are all descriptive claims being used to support evaluative conclusions, and each requires a premise stating why that fact carries moral weight. Scientific findings are frequently presented as settling policy questions, and they constrain the options without determining the choice, since deciding what to do with a finding requires a view about what matters. Economic analyses reporting that a policy is efficient are making an evaluative claim that efficiency is what should be pursued. The useful habit is to ask, whenever a conclusion about what should be done follows a set of facts, what the missing premise is and whether the speaker would defend it if stated.

Philosophers have taken several positions and none commands agreement. Some accept the gap and conclude that moral claims are not factual at all, expressing attitudes or prescriptions rather than describing anything, which is a whole family of positions in metaethics. Some accept the gap and hold that moral facts exist as a distinct kind, not derivable from natural facts but real, which requires explaining how they are known. Some deny the gap, arguing that certain descriptive facts about function, flourishing or what a thing is for carry evaluative content already, which is the approach of several accounts drawing on Aristotle. Some argue that the gap is real and unimportant, since practical reasoning has always involved evaluative premises and nobody claimed otherwise. The dispute is live and the practical advice is the same on every view, which is to state the evaluative premise.

A valid argument cannot conclude what should be done from premises stating only what is the case, so at least one evaluative premise is required. That demand is uncontroversial and its consequences are not. The observation is distinct from the claim that goodness cannot be defined naturally, and from the plain fallacy of arguing that something is right because it is natural.