The Mind-Brain Identity Theory

The identity theory holds that mental states are physical states, ordinarily states of the brain or nervous system. Pain does not accompany a neural process as a distinct effect and is not merely correlated with it: the pain and the process are one state, described through different concepts. If true, identity removes the need for a causal bridge between mind and body. A pain's causing withdrawal is a physical process causing movement, not two causes competing for one effect.

The theory rose with mid-twentieth-century neuroscience as an alternative both to substance dualism and to behaviourism. It accepts inner episodes that behaviourism tried to analyse dispositionally, while denying that they belong to a separate substance or irreducible property. Its strongest form identifies mental types with physical types. The difficulty is explaining how one physical type could unite the diversity of creatures and systems capable of the same mental state.

Identity, not correlation

U. T. Place, J. J. C. Smart, and Herbert Feigl proposed that consciousness and sensation are brain processes in the same broad sense that lightning is an electrical discharge and temperature is mean molecular kinetic energy. Scientific inquiry can discover that two descriptions concern one phenomenon even when the descriptions are not synonymous.

This is stronger than correlation:

$$M=P,$$

not merely

$$M\leftrightarrow P.$$

If pain is identical with physical state $P$, there is no possible instance of this pain without $P$ and no instance of $P$ without the pain. By contrast, correlated events remain numerically two and could in principle come apart.

The distinction matters for explanation. Correlation invites the question why the states travel together. Identity blocks that question in one form: nothing explains why lightning accompanies electrical discharge because there are not two occurrences. It leaves another question: what justifies the identity, and why does the state known physically also present a phenomenal character?

Type and token identity

Type identity associates each repeatable mental kind with a repeatable physical kind:

$$\text{pain}=\text{physical type }P.$$

This supports a strong reduction of psychology. Laws stated in mental vocabulary can in principle be connected to neurological laws through type identities, much as thermodynamic properties connect with statistical mechanics.

Token identity makes only a claim about particulars:

$$\forall e(Mental(e)\to Physical(e)).$$

Every particular pain, belief, or decision is a physical event, but different tokens of one mental kind may be physically heterogeneous. Token identity secures an all-physical inventory and avoids causal competition at the event level. It does not by itself reduce mental properties or explain what makes heterogeneous physical events tokens of one mental type.

The distinction prevents an invalid inference. Evidence that every observed mental episode has a neural realisation supports token physicalism. It does not show that pain, belief, or perception corresponds to one physical type across organisms.

Contingent scientific identity

Smart denied that mental and physical expressions have the same meaning. "I am in pain" and a neurological description differ in how speakers identify the state, just as "lightning" and "electrical discharge" once differed. The identity can therefore be an empirical discovery rather than a conceptual truth.

This answers a simple objection from meaning: a person can know they are in pain without knowing neuroscience because co-referring expressions need not be cognitively equivalent. It raises a modal issue. If true identities are necessary, then pain could not have existed without its physical identity, even though that separation appears conceivable. Kripke argued that the lightning analogy breaks down. We can explain the appearance of lightning without electrical discharge by imagining a different cause of the same visual appearance; pain, however, is the appearance itself. Imagining the feeling while removing the neural state appears to preserve pain rather than only its surface sign.

The identity theorist can reject the claimed possibility, distinguish phenomenal from physical concepts, or identify only tokens. Each response preserves identity at a cost: modal intuition is declared misleading, a theory of concepts is added, or type-level reduction is abandoned.

Bridge laws and reduction

Classical intertheoretic reduction derives laws of a higher-level theory from a lower level plus bridge laws connecting their vocabularies. If the bridge says "pain occurs if and only if $P$ occurs," it may state a correlation. If it says "pain is $P$," it states identity.

Identity theorists prefer identities because correlations leave two domains and an unexplained connection. Yet the evidence often establishes only functional dependency, intervention, and covariance. An identity claim earns its explanatory role when it supports unification, prediction, and integration better than rival relations such as realisation or grounding.

Reduction also comes apart at several levels. Ontological reduction says no mental entities exist in addition to physical entities. Property reduction identifies mental kinds with physical kinds. Explanatory reduction shows how higher-level regularities follow from lower-level mechanisms. Token identity may deliver the first without the second or third.

Multiple realisation

Putnam's decisive challenge asks why pain should be tied to one human neural type. Mammals, cephalopods, possible extraterrestrials, and artificial systems may share the causal profile of pain while differing physically. If so, the mental kind is multiply realisable.

Three identity-theory replies are available.

1. Species-relative identity: human pain is neural type $P_H$, octopus pain is $P_O$, and so on. This preserves local identities but fragments the generalisation that both are pain.
2. Disjunctive physical type: pain is $P_H\vee P_O\vee P_A\vee \cdots$. The disjunction appears to be a physical kind only because its members share the mental role, so the reduction explains unity circularly.
3. Empirical restriction: perhaps genuinely shared mental kinds are less widespread than intuition suggests. This avoids premature liberalism but risks chauvinism, excluding minds because their implementation differs from ours.

Multiple realisation does not refute token identity, and empirical neuroscience may find more conserved organisation than early functionalists expected. Its force is against the claim that each useful mental kind is one microphysical kind. It motivates functionalism, which locates unity in causal role instead.

Grain and physical kinds

The objection depends on the level at which physical types are specified. At a very fine grain, even two human pains differ in their exact neural microstate, making type identity implausible. At a coarser grain, organisms may share dynamical or network properties despite cellular differences.

An identity theory formulated at the right organisational level can therefore absorb some multiple realisation. But if the relevant physical type is defined as "whatever plays pain's causal role," the view has become functionalist in substance. A non-circular identity theory must identify the physical kind independently and show that it tracks the mental kind across relevant cases.

Phenomenal and semantic objections

Identity with a physical type must cover more than causal role.

Phenomenal character. A complete description of neural structure and function appears to leave open what the state feels like. The identity theorist may treat this as an epistemic gap between concepts rather than an ontological gap between properties. The challenge is to explain why the concepts are so unlike if they concern one property.

Intentional content. Beliefs are individuated partly by what they are about, and content may depend on the subject's environment or linguistic community. Physically identical internal states could have different wide contents. A simple internal neural identity therefore cannot capture every content property, though it may identify the vehicle that carries content.

These objections suggest pluralism about mental kinds. Sensations, computational states, and externally individuated beliefs may require different physical relations. That is compatible with token physicalism but weakens one uniform type-identity theory.

Causation and explanatory levels

Identity gives the cleanest response to causal exclusion. If mental event $M$ is physical event $P$, citing $M$ and $P$ does not add two sufficient causes. The mental description may still explain why an action occurred: "she left because she believed the building unsafe" captures a rational pattern omitted by a neural description.

The difficult question is whether the mental property is causally relevant as mental. If the effect follows from the event's physical properties while its content plays no role, reasons explanations may be true descriptions without tracking causal structure. Type identity answers by identifying the relevant properties; token identity alone cannot. This is one reason token identity often develops into non-reductive physicalism rather than standing alone.

Assessment

The identity theory establishes a benchmark: if mental and physical properties can be identified, dependence and causation cease to be separate mysteries. Its strongest version is threatened by physical diversity; its weaker token version preserves the inventory while leaving the unity and causal relevance of mental kinds to another theory.

Selected references

• Feigl, Herbert. "The 'Mental' and the 'Physical'" (1958).
• Kripke, Saul. Naming and Necessity (1980), lecture III.
• Place, U. T. "Is Consciousness a Brain Process?" (1956).
• Putnam, Hilary. "The Nature of Mental States" (1967).
• Smart, J. J. C. "Sensations and Brain Processes" (1959).