The Mind-Body Problem

What It Is

A neuron firing in the visual cortex is a physical event: a change in membrane potential, a release of neurotransmitter, a propagation of an electrochemical signal. The same event, viewed differently, is part of seeing red: a subjective state with a felt quality, accessible to introspection, reportable in language. The mind-body problem asks how these two descriptions are related.

Three claims constrain the problem.

1. Causal closure of the physical. Every physical effect has a sufficient physical cause. The brain is a physical system; what it does is fixed by physics.
2. Mental causation. Mental states have physical effects: my deciding to raise my arm causes my arm to rise. If mental states are not physical, then either causal closure is false or mental states are causally inert (epiphenomenal).
3. Phenomenal consciousness. There is something it is like to be in a conscious mental state. The subjective character (qualia, what-it-is-likeness) is the distinctive feature of conscious experience, not reducible to behavior or function in any obviously satisfying way.

The positions in the philosophy of mind are largely classified by which of these three claims they emphasize, weaken, or deny.

This page assumes What Is Philosophy?. The functionalist position is given practical depth in The Chinese Room Argument, which is the canonical thought experiment against pure functionalism. The intersection with personal identity is in Personal Identity Over Time.

Substance Dualism

René Descartes in the Meditations (1641) argued that mind and body are distinct substances: the body is res extensa (extended substance, occupying space), the mind is res cogitans (thinking substance, not occupying space). The famous argument: I can clearly and distinctly conceive of myself existing without a body, but I cannot conceive of myself existing without thinking. So mind and body are not the same kind of thing.¹

Two contemporary forms.

Cartesian substance dualism (classical). Two distinct substances; the mind is non-extended and non-physical. Causal interaction occurs at the pineal gland (Descartes's proposal). The position is largely abandoned in contemporary philosophy because of the interaction problem: how does a non-physical substance causally interact with a physical one? Any answer either makes the mind quasi-physical (and so collapses the dualism) or violates the causal closure of the physical.

Property dualism. A weaker form. There is only one substance (the physical world), but it has two kinds of properties: physical properties and mental / phenomenal properties. The mental properties are additional to the physical: they supervene on physical states but do not reduce to them. David Chalmers's "naturalistic dualism" (1996²) is the contemporary version; it accepts causal closure of the physical but treats phenomenal consciousness as a fundamental feature of the world that we currently lack a science of.

Property dualism preserves the intuition that consciousness is something extra, but pays a price: the mental properties are causally idle (or there is overdetermination, with both the physical state and the mental state independently sufficient for the effect). Most physicalists treat the epiphenomenalism objection as decisive against property dualism.

Reductive Physicalism

U.T. Place (1956) and J.J.C. Smith Smart (1959) argued that mental states are identical to brain states.³ The seeing of red is identical to a specific pattern of neural activity; just as water is identical to H$_2$O, pain is identical to the firing of C-fibers (or whatever the empirically right neural correlate turns out to be).

The position is type-identity physicalism: type of mental state = type of brain state. The motivating analogy is reductive identities in science (water = H$_2$O, heat = molecular kinetic energy, lightning = electrical discharge).

The major objection is multiple realizability (Putnam 1967⁴). Pain is plausibly a state that humans, octopuses, hypothetical silicon-based aliens, and possibly future AI systems can be in. If pain is type-identical to a specific neural state, then beings without that neural state cannot be in pain. This is empirically wrong: pain is a state that beings with very different neural architectures share.

The reductivist response is token-identity physicalism: each particular pain event is identical to some particular physical event, but no general type-identity holds. This preserves the reduction at the token level but gives up the explanatory ambition of type-identity (it does not tell us what pain is across systems).

Functionalism

Functionalism, articulated by Putnam in the late 1960s and developed by David Lewis and others in the 1970s, holds that mental states are defined by their functional role: the pattern of causal connections between sensory inputs, behavior, and other mental states.

A mental state $M$ is the state that occupies a certain functional role $R(M)$, where $R(M)$ specifies: which inputs cause $M$ (e.g., tissue damage causes pain), which outputs $M$ causes (e.g., pain causes withdrawal behavior and a desire to remove the cause), and which other mental states $M$ interacts with (e.g., pain combined with the belief that one should not flinch causes effortful suppression).

The key claim: the same functional role can be realized by different physical substrates. Pain in humans is realized by C-fiber firing; pain in an octopus is realized by something else; pain in a silicon system is realized by something else again. What makes them all pain is that they all occupy the pain-role.

Functionalism inherits the spirit of identity theory (mental states are physical) while solving multiple realizability (the realizer can vary). It is the dominant position in contemporary philosophy of mind.

Three sub-varieties.

• Machine functionalism (Putnam, early). Mental states are states of a probabilistic automaton. The mind is whatever software runs on the brain hardware.
• Analytic functionalism (Lewis, Armstrong⁵). The functional roles are read off from the structure of folk-psychological commonsense generalizations about belief, desire, perception, and action.
• Psychofunctionalism (Block, Fodor). The functional roles are read off from the deliverances of scientific psychology, not folk psychology. The relevant generalizations are those that empirical cognitive science will validate.

Two persistent objections to functionalism.

The Chinese Room (Searle 1980⁶). A person who does not understand Chinese follows a rule book that allows him to respond to Chinese input with Chinese output indistinguishable from a native speaker's. The person, room, and rule book together are functionally equivalent to a Chinese-speaking mind. But there is no genuine understanding anywhere in the system. Functional equivalence is therefore insufficient for genuine mental states. See The Chinese Room Argument for the full treatment.

Absent and inverted qualia (Block 1978⁷). If functionalism is right, then any system that realizes the right functional roles thereby has the right mental states. But it seems conceivable that the United States or a giant lookup table could realize the functional roles of a mind without there being any qualia, any what-it-is-likeness. And it seems conceivable that two systems with identical functional organization could differ in qualia (one seeing red where the other sees green, with no behavioral difference). If absent / inverted qualia are conceivable, function does not fix qualia.

The Hard Problem of Consciousness

David Chalmers's 1995 paper "Facing Up to the Problem of Consciousness"⁸ distinguished the easy problems of consciousness from the hard problem.

Easy problems. Explain how the brain integrates information, focuses attention, reports its own states verbally, controls behavior, distinguishes wakefulness from sleep, and so on. These are tractable in principle: cognitive neuroscience will eventually explain the mechanisms.

Hard problem. Explain why any of this should be accompanied by subjective experience. Why is there something it is like to integrate information, focus attention, report a state? An explanation of the mechanisms leaves untouched the question of why those mechanisms are accompanied by phenomenology.

Chalmers's argument that the hard problem is genuinely hard rests on two thought experiments.

The zombie argument. A philosophical zombie is a being functionally and physically identical to a conscious human but lacking subjective experience. If zombies are conceivable (which Chalmers argues at length), then phenomenal consciousness is not entailed by the physical facts. Some non-physical fact distinguishes the zombie world from ours. Therefore physicalism is false (or at least incomplete).

The argument has the form of a modal argument: from conceivability to possibility to actual non-entailment. Physicalists respond by denying the conceivability step (zombies are not genuinely conceivable, just superficially imaginable) or by denying the conceivability-implies-possibility step. The literature here is vast and unresolved.

The knowledge argument (Jackson 1982⁹). Mary has been raised in a black-and-white room and has learned every physical fact about color vision. She knows the neuroscience, the optics, the wavelengths, and the brain states. She is then released from the room and sees red for the first time. Does she learn anything new?

The intuition is that she does: she learns what it is like to see red. If she had every physical fact in the black-and-white room and yet learns something new on release, then phenomenal facts are not entailed by physical facts. Physicalism is false.

Physicalist responses split. Ability-knowledge responses (Lewis, Nemirow): Mary acquires a new ability (to recognize, imagine, remember red), not a new fact. Phenomenal-concepts responses (Loar, Papineau): Mary acquires new concepts but no new facts; the same fact is now graspable under a phenomenal mode of presentation. Acquaintance responses: Mary now has acquaintance with red, which is not propositional knowledge at all.

Eliminativism

Paul and Patricia Churchland's eliminative materialism¹⁰ holds that folk-psychological mental states (beliefs, desires, sensations as folk concepts) do not exist. The folk theory of mind is a primitive, lay scientific theory that mature neuroscience will replace, not reduce: just as phlogiston theory was eliminated rather than reduced when modern chemistry arrived, the folk-psychological taxonomy of mental states will be eliminated when scientific cognitive neuroscience matures.

The argument runs through three stages. (i) Folk psychology is a theory (it makes generalizations: "people who want X and believe doing Y achieves X will tend to do Y"). (ii) The theory is empirically vulnerable: it has poor predictive power on disorders (schizophrenia, autism, addiction), it has stagnated for two thousand years, and its categories do not map cleanly onto brain processes. (iii) When a theory fails this badly, the right response is elimination, not reduction.

Objections. The most common is self-refutation: the eliminativist's claim that folk psychology is false presupposes that the eliminativist has beliefs about folk psychology, which are themselves folk-psychological states. The Churchlands respond that the self-refutation objection is weaker than it looks (it begs the question by assuming the contents of belief are exactly what folk psychology says they are), but the response has not convinced everyone.

Eliminativism is a minority position. Most contemporary philosophers accept that folk-psychological states are real but disagree on whether they reduce to (or are realized by) physical states.

Common Misconceptions

• "Physicalism is the same as eliminativism." No. Physicalism holds that mental states are physical (or supervene on the physical); eliminativism holds that mental states (as folk-psychologically conceived) do not exist at all. A reductive physicalist keeps mental states; an eliminativist discards them.
• "Functionalism implies any computer running the right program has consciousness." This is the worry, and it depends on the variety of functionalism. Coarse functionalism (any system with the right input-output profile) faces the Chinese Room and Block's homunculi worries. Finer-grained functionalism (the right kind of computation, at the right level, with the right kind of integration) is harder to dismiss but does not automatically extend consciousness to any pattern-matching system.
• "The hard problem is just an argument from ignorance." This is the deflationary reply, but it misstates the argument. The hard problem is not "we don't know how to explain consciousness physically, therefore we never will"; it is "we don't have an account of why any physical story should be accompanied by phenomenal experience, and the structure of physical explanation does not obviously deliver one."
• "Property dualism is just substance dualism in a hat." No. Property dualism keeps a single substance ontology (the physical world); it denies only that mental properties reduce to physical ones. The interaction problem is much weaker for property dualism (mental properties supervene on physical ones; there is no second substance to interact).

Comparisons to Related Views

Go Further

• Stanford Encyclopedia of Philosophy, "Dualism" by Howard Robinson. Survey of the dualist family.
• Stanford Encyclopedia of Philosophy, "Physicalism" by Daniel Stoljar.
• Stanford Encyclopedia of Philosophy, "Functionalism" by Janet Levin.
• Stanford Encyclopedia of Philosophy, "The Hard Problem of Consciousness" surveyed in the consciousness entry by Robert Van Gulick.
• Chalmers, David J. The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press, 1996. The canonical book-length defense of property dualism and the hard problem.
• Block, Ned, Owen Flanagan, and Güven Güzeldere, eds. The Nature of Consciousness: Philosophical Debates. MIT Press, 1997. Anthology with most major statements.
• Kim, Jaegwon. Philosophy of Mind. 3rd ed. Westview Press, 2010. The standard graduate-level textbook; precise and current.

For the canonical thought experiment against pure functionalism, see The Chinese Room Argument. For the related question of what a symbol-processing system is, see What Is a Symbolic System?.