The Concept of Supervenience: When One Level Rests on Another

Philosophers regularly need to say that one set of facts "depends on" another without claiming they are the same facts. Supervenience is the technical tool for that job, and understanding it clarifies arguments across philosophy of mind, ethics, and the theory of meaning.

What supervenience actually says

The core idea is asymmetric covariance: B-properties supervene on A-properties when no two situations can differ in their B-properties without also differing in their A-properties. Put plainly, fix all the A-facts and you fix all the B-facts too. Change a B-fact and something in the A-base must have changed to account for it.

The canonical example is mental and physical states. If your brain were in precisely the same physical state on two occasions, you would be in the same mental state on both. Any difference in mood, belief, or sensation requires some corresponding physical difference. This is a claim most people find plausible without being committed to the much stronger claim that mental states just are physical states — that mental talk can be translated, term by term, into neuroscience.

Supervenience thus carves out a middle position: dependence without reduction. That gap is where most of the philosophical action takes place.

The different strengths of the relation

Not all supervenience claims are equal. Philosophers distinguish at least two important grades.

Weak supervenience says that within a single possible world, indiscernible A-bases guarantee indiscernible B-properties. If two people in this world have identical neural states, they have identical mental states. But weak supervenience permits two people in different possible worlds to have identical A-bases while differing in B-properties. That is a surprisingly loose connection.

Strong supervenience tightens the screws: if X has property B in virtue of having property A, then any individual in any possible world that has A also has B. The B-property tracks the A-property across worlds, not just within ours.

There is also global supervenience, which operates at the level of entire worlds rather than individuals. Two possible worlds that are indistinguishable in every physical respect are indistinguishable in every mental (or moral) respect. This is weaker than strong supervenience in some ways — two worlds could be globally identical in mental character while individual object-to-property pairings differ — but it is a useful formulation for certain arguments about consciousness.

The distinctions matter because a physicalist about mind wants strong supervenience at minimum. Weak supervenience would let a philosophical zombie world — one physically identical to ours but with no inner experience — count as genuinely possible, which is exactly what most physicalists want to deny.

Supervenience in ethics

Moral philosophy borrowed the concept from philosophy of mind and put it to equally heavy use. The standard claim is that moral properties supervene on natural properties: two acts or agents that are identical in every natural respect — in their physical, psychological, causal, and relational properties — cannot differ in moral status. If it is wrong for you to break a promise under some precise description of circumstances, it is wrong for anyone under those same circumstances.

This is sometimes called the universalisability thesis, and Kant's categorical imperative trades on something like it. But supervenience is actually weaker than full universalisability: it says sameness of natural base entails sameness of moral verdict, not that you can always give a neat natural description of what makes an act wrong.

The relationship between moral supervenience and moral realism is subtle. A realist thinks moral properties are genuine features of the world. A non-realist (say, an expressivist) thinks moral claims express attitudes rather than describe facts. Both can, in principle, accept moral supervenience. The expressivist simply has to explain why our attitude-expressions are constrained to covary with natural facts — and several sophisticated expressivists, notably Simon Blackburn, have offered accounts of this.

What supervenience does not settle is whether moral properties are reducible to natural ones. That is the question behind G. E. Moore's open-question argument and what he called the naturalistic fallacy — the mistake of identifying goodness with some natural property like pleasure or evolutionary fitness. Even if goodness supervenes on pleasure (where there is more pleasure and no countervailing factors, there is more goodness), it does not follow that "good" means the same thing as "productive of pleasure." The supervenience can hold necessarily while the concepts remain distinct.

Why supervenience is not the same as reduction

This is the point most worth pressing, because the two ideas are easily conflated.

Reduction requires that higher-level facts be derivable from or identical to lower-level facts, at least in principle. Water is H₂O: that is a reduction. Temperature in a gas is mean molecular kinetic energy: that is a reduction. In each case, the higher-level property is nothing over and above the lower-level one; they are the same thing picked out by different descriptions.

Supervenience requires only covariance. The mental supervenes on the physical without mental properties being identical to physical ones. Pain might supervene on C-fibre firing without "pain" meaning the same as "C-fibre firing," because pain could in principle be realised by different physical states in different creatures — silicon-based systems, radically alien biologies — while still being pain. This is the idea of multiple realisability, which was developed by Hilary Putnam against type-identity theories of mind.

The gap matters enormously for questions about emergence. When physicists talk about emergent properties in complex systems — the wetness of water, the rigidity of a lattice, the coordinated behaviour of a flock of birds — they usually mean supervenience-style dependence, not irreducible ghostly additions to the physical inventory. The higher-level patterns are real, they are not separately caused, and they are fully dependent on the lower level. But describing them at the higher level is not idle: the higher-level description often captures patterns that cross-cut the lower-level one in ways that cannot be expressed from below without infinite complexity.

The explanatory gap and where supervenience runs out

Supervenience can show that mental states depend on physical ones without explaining why that particular physical configuration gives rise to that particular experience. This is the explanatory gap identified by Joseph Levine and sharpened by David Chalmers into what he calls the hard problem of consciousness.

Even granting strong supervenience of mind on brain, we can ask: why does this neural pattern produce the particular qualitative character of seeing red, rather than some other experience, or no experience at all? Supervenience tells us the correlation holds necessarily; it does not tell us why. The explanatory gap is not a gap in the supervenience relation itself — it is a gap in our ability to derive experiential facts from physical ones by way of any intelligible necessity we can grasp.

This is why supervenience, though useful, is not a complete theory of the mind-body relation. It describes the shape of the dependence without explaining the mechanism or the metaphysical ground. Some philosophers — type-identity theorists, eliminativists, certain functionalists — think that closing the explanatory gap requires abandoning the appearance of a gap by refining our concepts. Others think the gap points to something genuinely irreducible about consciousness, without that irreducibility implying dualism in the Cartesian sense.

What supervenience teaches us about layered explanations

The broader lesson is that reality tolerates multiple, non-competing levels of description. Economics supervenes on psychology and sociology, which supervene on biology, which supervenes on chemistry, which supervenes on physics. Each level is fully dependent on the levels below it — there is no top-down causation that violates physical law. But the higher levels introduce patterns, generalisations, and causal explanations that are invisible from below.

This has direct consequences for debates about religion and science. When critics of naturalism argue that minds or moral facts cannot be physical because they are "too different" from neurons or particles, the supervenience framework offers a disciplined response: difference in kind at the descriptive level is compatible with dependence at the metaphysical level. Resistance to reduction is not evidence for an immaterial substance; it may simply reflect the genuine autonomy of higher-level sciences.

Equally, when defenders of moral realism face the challenge that moral facts are spooky additions to the physical universe, they can appeal to the fact that mathematical facts, logical facts, and dispositional facts face the same challenge — and we do not typically conclude that mathematics is therefore fictional. Supervenience, carefully applied, shows how moral facts can be objective and mind-independent while being grounded in the natural world.

None of this dissolves the hardest questions. The explanatory gap in consciousness remains genuinely open. The grounds of moral supervenience — why natural differences always carry moral weight — are contested. But the concept gives us a precise vocabulary for what we are asking, and in philosophy, that precision is often the first real step forward.