In the early twentieth century, a peculiar problem began to haunt analytic philosophy. Philosophers had become very good at breaking down complex claims into their component parts — at clarifying what propositions meant before arguing about whether they were true. But one category of claims resisted this treatment in an uncomfortable way: ethical claims.
When you say "stealing is wrong," what exactly are you saying? You're not describing a physical fact, like the boiling point of water. You're not reporting a psychological state, exactly, since it doesn't feel like saying "I dislike stealing." You seem to be making a claim about reality — but what kind of reality?
This puzzle became known as the is-ought problem, first articulated by David Hume in the 18th century, and later formalized into what G.E. Moore called the naturalistic fallacy. Together, they form the backbone of modern metaethics — the branch of philosophy that asks not "what should we do?" but "what kind of thing are moral claims?"
Hume's Fork
David Hume noticed something that seems obvious once you see it: descriptive claims and normative claims are logically different in kind.¹
A descriptive claim tells you how things are: "People generally seek pleasure and avoid pain." A normative claim tells you how things ought to be: "You ought to help others." Hume observed that no amount of descriptive facts — no accumulation of statements about what is — can logically entail a statement about what ought to be, unless you already smuggle a normative premise into your argument.
This is the is-ought gap. You can't derive an "ought" from an "is" alone. If you try, you've made an implicit logical leap that needs to be made explicit and justified.
Consider a common-sounding argument: "Humans evolved to be social creatures. Therefore, we ought to help each other." The premise is empirical (and roughly accurate). But the conclusion doesn't follow without a hidden bridge: something like "we ought to act in accordance with our evolved nature." And that bridge itself is a normative claim — one that needs its own justification.
Hume's point was not that ethics is impossible. It was that ethical reasoning requires its own foundations, which cannot be borrowed from science or factual description alone.
Moore's Open Question
G.E. Moore extended Hume's insight in his 1903 Principia Ethica with what he called the open question argument.² Moore was targeting a family of views he called ethical naturalism — the idea that moral properties (like "good" or "right") can be identified with natural properties (like "pleasurable," "evolutionarily advantageous," or "socially approved").
Moore's challenge was simple: for any natural property N you propose as the definition of "good," you can always meaningfully ask, "But is N actually good?" If the question is still open — if it makes sense to ask it — then N cannot be the same thing as "good," because you can't meaningfully ask "is good good?"
For example: if "good" just means "what produces pleasure," then "Is pleasure actually good?" should be a nonsensical question, like asking "Is a bachelor unmarried?" But it's not nonsensical. We can genuinely wonder whether certain pleasures are worth pursuing. Therefore, Moore argued, "good" cannot just mean "pleasurable." The same argument applies to any natural substitute.
Moore concluded that goodness is a non-natural property — real, but not reducible to anything science could in principle measure. He called this position moral realism paired with non-naturalism: moral facts exist, but they're grasped by something more like moral intuition than empirical observation.
The question "But is that actually good?" never loses its grip. That grip is philosophy's oldest proof that ethics resists reduction.
Why This Matters: The Ongoing Debate
The is-ought gap and the open question argument don't resolve the debate — they inaugurate it. Modern metaethics has produced a remarkable range of responses.
Moral naturalists have refined their positions, arguing that Moore's argument only shows that moral terms aren't analytically equivalent to natural ones — not that they're not metaphysically identical. Frank Jackson and others argue that moral properties may supervene on natural ones even if we can't define them reductively.³
Non-cognitivists — following A.J. Ayer and later R.M. Hare — argued that Moore asked the wrong question. Moral sentences don't describe facts at all; they express attitudes or prescriptions. "Stealing is wrong" is more like "Boo, stealing!" than like "The sky is blue." This view sidesteps the naturalistic fallacy by denying that moral claims are in the fact-stating business at all.
Moral realists of various stripes — including contemporary philosophers like Derek Parfit, Russ Shafer-Landau, and David Enoch — defend the view that moral facts exist independently of minds and cultures, and that our moral intuitions are evidence (imperfect, revisable) about them.
Error theorists, following J.L. Mackie, accept that moral claims purport to describe objective facts, but argue that there are no such facts — so all moral claims are technically false.⁴ We're not reporting facts; we're projecting something onto a fact-free reality.
The Practical Stakes
This might all seem like a game philosophers play before lunch. But the metaethical question has serious practical consequences.
If moral realism is true, there is something to get right in ethics — moral progress is possible, moral disagreement can be reasoned about, and appeals to conscience and intuition carry epistemic weight. If error theory or strong non-cognitivism is true, moral language may be useful but is ultimately a kind of organized fiction.
For Christians, the debate intersects with divine command theory and natural law traditions that have their own answers to the is-ought question. Aquinas, for instance, grounded the normative order in a teleological account of human nature oriented toward God — a move that tries to bridge the is-ought gap theologically rather than logically.
Whether or not that move succeeds, it's philosophically serious — and understanding why Hume raised the challenge makes the theological answer more precise.
Conclusion
The is-ought gap and the naturalistic fallacy are not problems to be solved and discarded. They are permanent features of the philosophical landscape — clarifications that force every ethical thinker to be more explicit about where their normative foundations come from. Science can inform ethics, but it can't generate it. That task requires a different kind of reasoning: one that takes seriously the irreducible structure of moral claims and refuses to pretend the gap isn't there.
Sources ¹ David Hume — A Treatise of Human Nature, Book III (1739) ² G.E. Moore — Principia Ethica (Cambridge University Press, 1903) ³ Frank Jackson — From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford, 1998) ⁴ J.L. Mackie — Ethics: Inventing Right and Wrong (Penguin, 1977) ⁵ Russ Shafer-Landau — Moral Realism: A Defence (Oxford, 2003)
