Non-Reductive Physicalism and Emergence

Non-reductive physicalism holds that the world is wholly physical in its concrete constitution while mental properties and explanations remain real, distinctive, and not reducible to physical types. Minds do not float free of bodies: mental differences require physical differences, and every mental state is physically realised. Yet a complete neurological vocabulary need not replace belief, pain, intention, or emotion.

The position tries to combine the empirical discipline and causal completeness of physicalism with the multiple realisation and explanatory autonomy emphasised by functionalism. It is attractive because this pattern appears throughout science. Organisms, economies, programs, and storms are physically constituted without their useful kinds being identical with microphysical kinds. The difficulty is whether autonomy amounts to more than convenient description. If the physical base already causes every effect, what makes a distinct mental property causally efficacious?

What makes the view physicalist?

At minimum, non-reductive physicalism accepts:

1. Constitution: every concrete minded being is wholly physically constituted.
2. Dependence: no mental difference occurs without an appropriate physical difference.
3. Completeness: physical events do not require independent non-physical causes.
4. Reality of the mental: mental properties support true explanations and are not merely shorthand to be eliminated.
5. Non-reduction: mental kinds are not identical with, definable as, or derivable from one set of physical kinds.

The last two claims distinguish the position from identity theory and eliminativism; the first three distinguish it from robust property dualism. The challenge is showing that the package is stable. If physical constitution fully determines causal powers, the mental appears reducible or idle. If mental properties add powers, completeness appears false.

"Physical" itself is not simple. Defined by current physics, physicalism is probably false because current physics is incomplete. Defined by ideal future physics, it risks saying only that everything is whatever the final theory describes. This is Hempel's dilemma, discussed as an ontological issue in Fundamentality, Levels, and Priority. Non-reductive views often understand physicalism through causal and constitutive dependence rather than a fixed list of particles.

Supervenience

Mental properties supervene on physical properties when no mental difference occurs without a physical difference:

$$\square \forall x\forall y(P(x)=P(y)\to M(x)=M(y)).$$

Supervenience captures determination while allowing multiple realisation: different physical states may realise one mental state, but identical total physical states may not differ mentally. Local, global, weak, and strong versions differ over objects and possible worlds.

The relation is too weak to constitute physicalism by itself. Moral facts may supervene on natural facts without being reducible; a dualist God could ensure a necessary match between physical and phenomenal facts. Supervenience says that variation is constrained, not why the constraint holds or which side is prior. The view therefore needs realisation, constitution, or grounding.

Realisation, constitution, and grounding

Realisation connects a role property to what occupies it. Neural state $P$ realises mental state $M$ when $P$ has the causal organisation constitutive of $M$. Different states can realise the same role.

Constitution says that the physical parts and organisation compose the minded system without being numerically identical to every higher-level property it has. A statue is constituted by matter; an organism is constituted by cells. Constitution is not automatically causal explanation.

Grounding says the mental facts obtain in virtue of physical facts. It adds direction and explanatory priority to supervenience:

$$P\text{grounds}M.$$

These notions explain dependence in complementary ways, but none is magic vocabulary. A grounding claim must identify a sufficient base; a realisation claim must specify the role; a constitution claim must explain why this organisation constitutes a mind. The hard cases reappear as questions about sufficiency rather than correlation.

Multiple realisation and the special sciences

Higher-level sciences classify systems by patterns that cut across microphysical differences. A heart is a pump across variations in tissue; memory may be implemented through different neural mechanisms; a program survives changes in hardware. Such kinds support counterfactuals, interventions, and explanations unavailable at a list of microstates.

This is the autonomy of the special sciences. Autonomy need not imply independent substances or fundamental laws. It means that higher-level generalisations are stable under lower-level variation and track variables at the grain relevant to prediction and control.

Mental kinds appear to fit. Beliefs are grouped by content and inferential role, pains by experiential and motivational role, not exact microphysics. However, explanatory autonomy does not establish ontological irreducibility. Temperature explanations are autonomous in practice while temperature is reductively explicable statistically. The non-reductive physicalist needs to say whether mental non-reduction is principled or a temporary limit on derivation.

Kim's exclusion argument

The general exclusion problem and the alternatives to non-reduction are developed in The Problem of Mental Causation; the present discussion isolates its force against this particular form of physicalism.

The central objection can be stated for a mental property $M$, its physical realiser $P_1$, and physical effect $P_2$:

1. $M$ causes $P_2$.
2. $P_1$ realises $M$ and is sufficient for $P_2$.
3. $M$ is not identical with $P_1$.
4. $P_2$ is not systematically overdetermined.
5. Therefore $M$ does not contribute a distinct cause of $P_2$.

If mental properties are causally individuated, losing causal efficacy threatens their reality. Kim's conclusion is that stable physicalism tends toward reduction: higher-level properties either inherit the causal powers of their realisers and are reducible, or possess no powers of their own.

The argument assumes a competition model on which sufficient causes at different levels exclude one another. Many replies reject that model. A person's pressing a key and a neural-muscular process are not two independent causes; one event satisfies two descriptions. The question then becomes whether properties, rather than events, compete.

Same instance, different levels

One reply treats mental and physical explanations as descriptions of the same causal process at different grains. "The glass broke because it was fragile" and a molecular explanation do not posit two breakings or necessarily two causes. Fragility identifies a stable dependency useful across varied microstructures.

Applied to mind, a belief causes an action by being realised in the physical process that causes it. The mental cause is not an extra force added after the neural cause. It selects the feature proportional to the range of counterfactual changes: changing the belief across its possible realisers changes the action.

The objection is that this may secure explanations without distinct properties. If all causal powers belong to the physical realiser, calling the realised property a cause may be semantic generosity. The reply is that causation itself is often defined through difference-making variables rather than a metaphysical transfer of power; on that account, the higher-level variable can be causally apt precisely because it omits irrelevant physical detail.

Intervention and proportionality

An interventionist account asks whether changing $M$ in an appropriate causal model changes $P_2$, holding relevant alternatives fixed. Intervening on a belief means changing any suitable physical realiser so that the subject holds a different belief. If action changes robustly across realisers, the mental variable supports causal claims.

Proportionality favours causes neither too specific nor too general for their effects. The exact microstate may contain details irrelevant to the action, while the belief that the route is blocked captures the difference that predicts choosing another route. Higher-level causation is then not second-best shorthand; it can be more explanatory.

These accounts face two concerns. Interventions on mental variables are implemented physically, so critics say the physical variable still does all the work. And content properties may be externally individuated, making surgical intervention conceptually difficult. Still, the framework demonstrates that causal relevance does not require an independent force or a one-to-one fundamental property.

The subset-powers response

Another proposal identifies a realised property's causal powers with a subset of the realiser's powers. A neural state may cause many effects; as a pain it has the subset involving attention, aversion, learning, and withdrawal. The mental property is distinct because it selects a stable power profile, yet adds no power beyond the physical base.

This avoids overdetermination: one power is not exercised twice. It also makes mental kinds causally individuated. Critics argue that subset selection reduces the mental property to a functional physical property, yielding a sophisticated reduction rather than non-reduction. Whether that is a defeat depends on whether non-reductive physicalism sought distinct powers or only irreducible explanatory kinds.

Weak emergence

Weak emergence occurs when higher-level patterns follow from lower-level rules but are difficult to derive except through simulation, aggregation, or new concepts. Flocks, cellular automata, and thermodynamic patterns are standard examples. Nothing fundamentally new enters the causal inventory.

Mental properties may be weakly emergent if complex organisation produces capacities that are unpredictable from local components yet wholly realised by them. This fits physicalism and explanatory autonomy. It does not by itself answer why the organisation is conscious. Calling phenomenal character weakly emergent is informative only when a derivation or mechanism connects the levels.

Weak emergence is often epistemic: it concerns our methods of prediction. A property may be surprising to us while reducible in principle. Non-reductive physicalism needs a stronger claim if its non-reduction is metaphysical rather than practical.

Strong emergence

Strong emergence attributes novel, irreducible properties or causal powers to organised systems. The base necessitates or lawfully produces the higher property, but lower-level facts and laws do not exhaust its nature or powers.

If strongly emergent consciousness has downward causal effects, physical completeness must be qualified. If it lacks such effects, the view approaches epiphenomenal property dualism. Strong emergence therefore sits near the border of physicalism. Calling the emergent bearer physical does not settle whether the emergent property belongs to the physical domain.

A strong-emergence law could be fundamental and empirically serious. It would need to specify when the property arises, what causal difference it makes, and how those claims are tested. Without such content, "emergence" records surprise rather than explains it.

Comparison with other higher-level entities

The ontology section notes that derivative entities can be real and explanatorily indispensable. A phonon is a collective excitation grounded in lattice behaviour; biological fitness and economic inflation likewise depend on lower-level facts while supporting causal explanation. These cases weaken the inference from derivative to unreal.

The analogy does not settle consciousness. In accepted cases, the higher-level property's relation to the base can be modelled and derived, and no further phenomenal datum appears omitted. Non-reductive physicalism uses the cases to show that levels can coexist, not to prove that phenomenal character is one more ordinary higher-level pattern.

The right comparison is disaggregated. Belief and computational capacity may resemble functional higher-level kinds; phenomenal character may present a different problem. A pluralist physicalism can accept reduction or realisation on a kind-by-kind basis rather than promise one relation for all mentality.

Assessment

Non-reductive physicalism is less a comfortable midpoint than a demand for a precise dependence relation. Its best case comes from the ordinary legitimacy of higher-level science. Its hardest case is consciousness, where the relation between explanatory grain and phenomenal character remains exactly what needs to be shown.

Selected references

• Davidson, Donald. "Mental Events" (1970).
• Fodor, Jerry A. "Special Sciences" (1974).
• Kim, Jaegwon. Mind in a Physical World (1998).
• Putnam, Hilary. "The Nature of Mental States" (1967).
• Shoemaker, Sydney. Physical Realization (2007).
• Wilson, Jessica M. Metaphysical Emergence (2021).