Functionalism

Functionalism: mental states as functional roles#

StandardAQA 7172 3.4 Metaphysics of mind — FunctionalismDfE GCE Philosophy subject content

A mental state as a functional role

Fig. 1Pain is defined by its inputs (damage), outputs (avoidance) and relations to other states — whatever realises it.

Functionalism holds that what makes a state a particular mental state is not its physical make-up (as identity theory says) nor its behavioural manifestation alone (as behaviourism says), but its FUNCTIONAL ROLE — the causal role it plays in the system, defined by its typical CAUSES (inputs), its EFFECTS (outputs, including behaviour), and its relations to OTHER mental states. Pain, for example, is the state typically CAUSED by bodily damage, that CAUSES the belief that one is hurt, the desire to relieve it, and avoidance behaviour, and that interacts with other mental states. Any state that plays this causal role — whatever it is made of — IS pain. Mental states are thus defined RELATIONALLY, by their place in a causal network, like functional kinds elsewhere: a carburettor is defined by what it DOES (mixing fuel and air) not by its material, and a mousetrap by its function not its shape.

The crucial improvement over behaviourism is that functionalism defines mental states by their relations to OTHER MENTAL STATES, not just to behaviour. Behaviourism failed by the circularity objection (a belief's behavioural effects depend on the agent's desires, and vice versa) and by omitting inner states (the super-Spartan). Functionalism EMBRACES these interconnections: it defines the whole network of mental states together, by their mutual causal relations and their links to inputs and outputs, so that a belief IS in part its disposition to combine with desires to produce action. This lets functionalism posit genuine INNER states (unlike behaviourism) that mediate between stimulus and response, while still being definable by their causal roles.

Hilary Putnam's early version was MACHINE-STATE (or Turing-machine) functionalism: the mind is analogous to the software or 'program' of a computer, and mental states are like the internal states of a Turing machine, defined by a machine-table specifying, for each state and input, the output and the next state. Just as the same program can run on different hardware (a chess program on a PC or a Mac), the same mental 'program' can run on different physical substrates — human brains, octopus nervous systems, or (in principle) silicon. This is the celebrated software/hardware analogy: mental states are to the brain as software is to hardware. Later versions (psychofunctionalism; analytic/common-sense functionalism, Lewis) refine how the functional roles are specified (by scientific psychology, or by the platitudes of folk psychology), but the core thesis is the same: mental states ARE functional states, individuated by causal role.

Functionalism is officially NEUTRAL on the ontological question — it says what mental states ARE (functional roles) without dictating what REALISES them, so it is compatible with physicalism (roles realised by brains) and, in principle, with other substrates. In practice it is almost always paired with physicalism: the functional roles are realised by physical states, so every mental state is physically realised, but no mental TYPE is identical to any physical TYPE (since the role can be multiply realised). This is exactly what the multiple-realisability objection to identity theory demanded — a theory on which the same mental state can be realised by different physical states — which is why functionalism became the dominant theory of mind in the late twentieth century, as the natural successor to identity theory.

Worked example

The software/hardware analogy

Explain Putnam's software/hardware analogy and what it captures about the mind.

The analogy
Mental states are to the brain as a program's states are to the computer's hardware: defined by their role in the program, not by the physical stuff.
Multiple realisability
The same program runs on different hardware; so the same mental 'program' can run on human brains, octopus nervous systems, or silicon.
What it captures
It explains why the same mental state can have different physical realisers (answering multiple realisability) while still being physically realised (keeping physicalism).
The limit
But software has no felt QUALIA, which is exactly where the qualia objections (inverted, absent) will press — a running program need not feel anything.

Result: The software/hardware analogy captures multiple realisability and the relational nature of mental states, making functionalism the successor to identity theory — but it also foreshadows the qualia problem, since a program's functional states seem to leave felt experience unaccounted for.

Typical mistakes

Confusing functionalism with behaviourism; functionalism defines mental states by relations to OTHER MENTAL STATES and posits genuine inner states, not merely behavioural dispositions.
Saying functionalism identifies mental states with brain states; it identifies them with functional ROLES, which are multiply realisable, so no mental type = a single physical type.

Active revision

Explain functionalism and how it differs from logical behaviourism. (12-mark AO1 response)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Functionalism's advantages over behaviourism and identity theory#

StandardAQA 7172 3.4 Metaphysics of mind — FunctionalismDfE GCE Philosophy subject content

Functionalism among the theories

Fig. 2Functionalism keeps inner states (unlike behaviourism) and multiple realisability (unlike identity theory).

Functionalism is best understood as designed to keep the strengths of behaviourism and identity theory while avoiding their fatal weaknesses. Against BEHAVIOURISM, functionalism has two decisive advantages. First, it posits genuine INNER states that mediate between stimulus and response, so it is not vulnerable to the super-Spartan objection: the super-Spartan is in pain (has the inner functional state typically caused by damage) even though, by an unusual further state (iron self-control), the state does not issue in pain-behaviour on this occasion — the functional role includes typical outputs, but a competing state can override them. Second, functionalism escapes the circularity objection: behaviourism could not analyse a belief into behaviour without mentioning desires (and vice versa), which looked viciously circular; functionalism EMBRACES this interdefinition, defining all mental states together as a network of causally interrelated states, so the mutual reference is a feature, not a bug — the whole system is defined at once (as by a Ramsey sentence).

Against IDENTITY THEORY, functionalism's key advantage is that it accommodates MULTIPLE REALISABILITY, the objection that felled type identity. Because mental states are defined by their functional ROLE rather than by any specific physical realiser, the same mental state (pain) can be realised by different physical states in different creatures — human C-fibres, octopus neurons, silicon circuits — provided each plays the pain-role. Functionalism thus captures the intuition that creatures very different from us physically can nonetheless share our mental states (an octopus can feel pain), which type identity 'chauvinistically' denied. It keeps physicalism's advantage (mental states are physically realised, so no immaterial soul, no interaction problem) while shedding type identity's liability (the implausible tying of each mental type to one physical type).

Functionalism also has independent theoretical attractions. It fits the computational and cognitive-scientific approach to the mind, on which the mind is an information-processing system whose states are individuated by their role in computation — making functionalism the philosophical basis of much of cognitive science and artificial intelligence research. It explains mental causation straightforwardly (functional states, realised physically, cause behaviour via their physical realisers). And it captures the RELATIONAL, holistic character of mental life — that beliefs, desires and experiences form an interconnected web, each defined partly by its relations to the others — better than either behaviourism (too tied to isolated stimulus-response) or identity theory (too tied to isolated brain states).

So functionalism combines physicalism's respect for the dependence of mind on brain, behaviourism's insight that mental states are connected to inputs and outputs, and a solution to multiple realisability — which is why it became the dominant theory of mind. But its central strength contains its central vulnerability. By defining mental states PURELY by their functional/relational role, and abstracting entirely from their physical/intrinsic nature, functionalism seems to leave out the INTRINSIC felt quality of experience — the qualia. If a mental state is JUST a node in a causal network, then it seems a system could have all the right functional organisation and yet lack any felt experience (absent qualia), or have systematically different felt experiences (inverted qualia), because the functional role does not appear to fix the phenomenal character. The qualia objections of the next sections are precisely this worry, and they are the knowledge and zombie arguments (from the dualism topic) applied specifically to functional states — the recurring hard problem, now aimed at functionalism.

Worked example

Why functionalism succeeds identity theory

Explain how functionalism preserves physicalism while solving the multiple-realisability problem.

The problem
Type identity ties pain to one physical type (C-fibres), wrongly excluding octopuses and machines that share the mental state.
The functionalist move
Define pain by its ROLE (caused by damage, causing avoidance), so any state playing that role is pain, whatever its material.
Physicalism kept
The role is always PHYSICALLY realised (by brains, etc.), so there is no immaterial soul and no interaction problem — physicalism is preserved.
Multiple realisability solved
The same role has different realisers in different creatures, so the same mental state is multiply realisable — exactly what the objection demanded.

Result: Functionalism keeps physicalism's virtues (physical realisation, mental causation, no interaction problem) while solving multiple realisability by individuating states by role, not physical type — the reason it succeeded identity theory as the dominant view.

Typical mistakes

Presenting functionalism as simply 'better behaviourism'; its key advance over IDENTITY theory (multiple realisability) is equally important and often the examiner's focus.
Failing to link functionalism's abstraction from physical nature to the qualia objections; the very feature that solves multiple realisability is what makes qualia problematic.

Active revision

'Functionalism is the best physicalist theory of mind.' Assess this claim. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Inverted qualia#

AdvancedAQA 7172 3.4 Metaphysics of mind — FunctionalismDfE GCE Philosophy subject content

The inverted-qualia objection

Fig. 3Two functionally identical people could have inverted colour qualia, so functional role does not fix the quale.

The inverted-qualia objection argues that functionalism cannot account for the specific phenomenal CHARACTER of experience, because two people could be FUNCTIONALLY identical yet have systematically INVERTED qualia. Imagine two people, Norm and Invert, whose colour experiences are spectrum-inverted relative to each other: when both look at a ripe tomato, Norm has the experience you call 'red' and Invert has the experience you call 'green', and vice versa for grass. Crucially, they are FUNCTIONALLY IDENTICAL: both call tomatoes 'red' (they learned the word for whatever experience tomatoes give them), both stop at red lights, both group tomatoes with fire engines — every functional role, every input-output relation, every behavioural and cognitive connection is the same. Yet their inner qualia differ systematically.

The argument against functionalism runs: (P1) inverted qualia are possible — Norm and Invert are conceivable, functionally identical but phenomenally different. (P2) If functionalism were true, then two functionally identical states would BE the same mental state in every respect (functional role fixes the mental state completely). (P3) But Norm's and Invert's colour experiences are functionally identical yet phenomenally DIFFERENT (different qualia). Therefore (C) functional role does NOT fix the phenomenal character of experience; there is more to a colour experience (its quale) than its functional role, so functionalism is incomplete — it leaves out qualia. The objection thus targets functionalism's defining claim, that mental states are EXHAUSTIVELY defined by functional role.

The inverted-qualia argument is a cousin of the zombie and knowledge arguments: where zombies have functional organisation with NO qualia, inverts have the same functional organisation with DIFFERENT qualia — both aiming to show that functional/physical facts do not fix the phenomenal facts. The intuition it pumps is powerful and old (Locke raised the inverted-spectrum possibility): it seems coherent that the quality of your experience of red could be swapped with your experience of green without any functional or behavioural difference, which suggests qualia 'float free' of functional role.

Functionalists reply in several ways. First, they may DENY that inversion is genuinely possible: perhaps the qualia are not really independent of functional role, because colour experiences have functional connections to MORE than colour naming — to judgements of similarity, warmth, salience, aesthetic response, and to the fine structure of colour space (the space of colours is asymmetric — there are more discriminable warm than cool shades, and unique versus binary hues), so a genuine, undetectable inversion may not be functionally possible after all (the 'asymmetry' reply of Shoemaker and others). Second, they may bite the bullet with REPRESENTATIONALISM: qualia just ARE representational/functional properties (the quale of red is the property of representing surfaces as red), so functionally identical states cannot really differ in qualia — the supposed inversion is incoherent. Third, some functionalists accept that qualia may not be fully captured by NARROW functional role and adopt a wider or hybrid view. Evaluation weighs whether the inversion is genuinely possible (if the asymmetry reply works, it is not, and the objection fails) or whether it reveals a real gap between functional role and phenomenal character (in which case functionalism is incomplete about qualia, like the theories before it). A balanced verdict is that the inverted-qualia objection makes a serious prima facie case that functional role underdetermines phenomenal character, but that the asymmetry and representationalist replies give the functionalist a genuine response, so the objection is not clearly decisive — its force, like that of the zombie argument, turns on whether the conceived scenario is really possible, which is exactly what is disputed.

Worked example

The asymmetry reply to inverted qualia

Explain the functionalist reply that a genuine, undetectable qualia inversion may not be possible.

The objection's assumption
Colour qualia can be swapped while ALL functional relations stay identical (a functionally undetectable inversion).
The asymmetry
But colour space is asymmetric: there are more discriminable warm than cool shades, unique vs binary hues, and qualia connect to similarity, salience and warmth judgements.
The consequence
A true inversion would show up in these functional relations (Invert would misjudge similarities), so a functionally UNDETECTABLE inversion may be impossible.
The upshot
If inversion cannot be functionally undetectable, the objection's key premise fails, and functional role may fix qualia after all.

Result: The asymmetry reply challenges the possibility of a functionally undetectable inversion, blunting the objection; whether it fully succeeds depends on whether ALL phenomenal differences must show up functionally — the disputed crux, mirroring the zombie debate.

Typical mistakes

Describing inverts as behaving differently; the objection requires them to be FUNCTIONALLY and behaviourally IDENTICAL, differing only in the intrinsic quale.
Treating the objection as obviously decisive; the asymmetry reply argues a genuine, undetectable inversion may not be functionally possible, which must be engaged.

Active revision

'Functionalism cannot account for the qualitative character of experience.' Discuss. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Absent qualia: Block's China brain#

AdvancedAQA 7172 3.4 Metaphysics of mind — FunctionalismDfE GCE Philosophy subject content

Block's China brain

Fig. 4A nation implementing the brain's functional organisation seems to lack any felt experience, so function does not suffice for qualia.

The absent-qualia objection argues that a system could be FUNCTIONALLY identical to a conscious human yet have NO qualia at all — no felt experience — which, if possible, refutes functionalism (since functional organisation would then not suffice for consciousness). Ned Block ('Troubles with Functionalism', 1978) makes the case vivid with the 'CHINA BRAIN' (or 'Chinese Nation') thought-experiment. Suppose the entire population of China (about a billion people) is organised to simulate the functional organisation of a human brain: each person plays the role of a neuron, sending signals to others by radio according to the same rules by which neurons interact, so that the whole nation implements exactly the functional/causal organisation of your brain when you are in pain. The system receives the same inputs and produces the same outputs, with the same internal functional states.

Block's intuition — the 'absent qualia' intuition — is that this China-brain system would NOT have any conscious experience: there would be no felt pain, no 'what it is like' to be the system, even though it is functionally identical to a person in pain. It seems absurd (a 'prima facie doubt', as Block cautiously puts it) that a nation of people passing radio signals could, collectively, FEEL anything — the system would have the functional organisation of pain but no phenomenal pain. If this is right, then functional organisation does not suffice for qualia: a system can realise the pain-role and yet have absent qualia, so functionalism is false as an account of consciousness. This is the zombie argument in a specific, concrete form — the China brain is a functional duplicate with (allegedly) no experience.

Block's objection targets what he calls 'liberalism' in functionalism — the worry that functionalism is too LIBERAL, counting as conscious (as having qualia) systems that intuitively have none, precisely because it individuates mental states by abstract functional role, which strange systems (a nation, a vast look-up table, a system of beer cans and string) could implement without being conscious. (The mirror worry, 'chauvinism', is that identity theory is too RESTRICTIVE, denying minds to creatures with the wrong physical make-up; functionalism was meant to avoid chauvinism, but Block argues it falls into liberalism instead.) The China brain is designed to make the liberalism worry vivid: functional organisation, realised in an intuitively non-conscious way, without qualia.

Functionalists reply in several ways. First, they may CHALLENGE THE INTUITION: our confident sense that the China brain is not conscious may be unreliable — it reflects the system's strangeness and scale (we cannot IMAGINE a nation being conscious), but a genuinely faithful functional duplicate of a brain, at the right level of organisation and speed, MIGHT be conscious, and our incredulity is a failure of imagination, not evidence (compare: it is hard to imagine that mere neurons, individually mindless, collectively feel — yet they do). Second, they may argue the China brain does not really replicate the RELEVANT functional organisation (the right causal structure at the right level and speed, with the right internal integration), so it is not a genuine functional duplicate, and the case does not bear on properly specified functionalism. Third, some accept that qualia may require more than abstract functional role — perhaps a specific KIND of physical or organisational realisation — moving to a restricted functionalism. Evaluation weighs whether the absent-qualia intuition is a genuine datum (in which case functionalism, like its predecessors, fails to capture qualia — the hard problem again) or an unreliable reaction to a bizarre scenario (in which case functionalism survives). A balanced verdict is that the China brain, like the zombie argument it concretises, makes a strong intuitive case that functional organisation does not obviously SUFFICE for consciousness, but that the functionalist's challenge to the intuition (our incredulity may be mere imaginative failure, as with neurons) is a serious reply — so the objection is powerful but, once again, turns on whether the conceived possibility is genuine, which the hard problem leaves unresolved.

Worked example

Challenging the absent-qualia intuition

Explain the functionalist reply that our intuition about the China brain is unreliable.

The objection
The China brain is functionally identical to a person in pain, yet intuitively feels nothing — so functional organisation does not suffice for qualia.
The reply
Our confident intuition that the nation feels nothing may be an artefact of its strangeness and scale — we cannot IMAGINE a nation being conscious, but imaginative failure is not evidence of impossibility.
The parallel
It is equally hard to imagine that individually mindless neurons collectively feel pain — yet they do; so 'I cannot see how X could be conscious' is a weak guide to whether it is.
Assess
The reply blunts the intuition but does not dispel the worry: the China brain still seems relevantly different from a brain, and the hard problem (why any organisation feels) remains unanswered.

Result: The functionalist's challenge to the absent-qualia intuition (imaginative failure, the neuron parallel) is a serious reply that weakens Block's case, but the residual force of the hard problem means the objection is contained rather than dissolved — the same stand-off as the zombie argument.

Typical mistakes

Confusing Block's China brain (absent qualia — a functional duplicate) with Searle's Chinese Room (which targets intentionality/understanding, not primarily qualia); they are distinct objections.
Treating the absent-qualia intuition as decisive; the functionalist replies that our incredulity may be mere imaginative failure (as with mindless neurons collectively feeling), which must be weighed.

Active revision

'A functional duplicate of a human brain would have no conscious experience.' Assess this claim. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

The Chinese Room and the limits of functionalism#

AdvancedAQA 7172 3.4 Metaphysics of mind — FunctionalismDfE GCE Philosophy subject content

The Chinese Room

Fig. 5The man produces correct Chinese outputs by following rules but understands nothing, so syntax is not sufficient for semantics.

John Searle's Chinese Room argument ('Minds, Brains, and Programs', 1980) targets 'strong AI' and machine/computational functionalism — the claim that running the right PROGRAM (having the right functional/computational organisation) is sufficient for genuine understanding and mental states. Where Block's China brain concerns absent QUALIA, Searle's argument primarily concerns INTENTIONALITY and UNDERSTANDING: the claim that a system implementing the right program thereby understands. The thought-experiment: Searle, who knows no Chinese, is locked in a room with a rule-book (in English) that tells him, for any string of Chinese symbols passed in, which Chinese symbols to pass out. By following the rules, he produces responses indistinguishable from those of a native Chinese speaker — the room passes the Turing test in Chinese. Yet Searle, inside, understands NOT ONE WORD of Chinese; he is merely manipulating symbols by their shapes (syntax), with no grasp of their meaning (semantics).

The argument runs: (P1) the man in the room implements the program (follows the rules) and produces the right outputs — he is functionally/computationally equivalent to a Chinese-understanding system. (P2) But the man does not UNDERSTAND Chinese — there is no genuine comprehension, only rule-governed symbol manipulation. (P3) The man is doing everything the computer does (running the program), and there is nothing else in the room that could understand. Therefore (C) implementing the right program is NOT sufficient for understanding: 'syntax is not sufficient for semantics'. Since functionalism (in its computational form) claims that the right functional/computational organisation suffices for mental states like understanding, and the Chinese Room has that organisation without understanding, functionalism is false — running a program does not, by itself, produce genuine intentionality.

The most important reply is the SYSTEMS reply: it is true that the MAN does not understand Chinese, but the WHOLE SYSTEM (man + rule-book + symbols + room) does understand — understanding is a property of the whole system, not of the CPU (the man) alone, just as a single neuron does not understand English though the brain it is part of does. Searle's rejoinder: let the man MEMORISE the entire rule-book and do all the processing in his head, so he IS the whole system — and he STILL does not understand Chinese; there is no extra understanding anywhere. Other replies include the ROBOT reply (embed the program in a robot with sensors and effectors that causally connect the symbols to the world — grounding the syntax in the environment might give it semantics) and the BRAIN SIMULATOR reply (a program simulating the actual neural firing of a Chinese speaker's brain would surely understand). Searle answers each: the robot reply concedes that pure symbol-manipulation is insufficient (understanding needs causal connection to the world, which is not mere program-running); the brain-simulator reply fails because simulating the FORMAL structure of neural firing still leaves out whatever it is about the brain's actual CAUSAL POWERS that produces understanding — a simulation of a fire does not burn.

Evaluation weighs the Chinese Room against these replies. Searle's positive view is that intentionality and understanding are BIOLOGICAL phenomena, produced by the specific causal powers of brains, which mere symbol-manipulation (whatever the substrate) lacks — so he is NOT a dualist but holds that the RIGHT KIND of physical system (a brain, or something with equivalent causal powers) is needed, not just the right program. Against him, the systems and robot replies press that understanding IS a property of a suitably organised, world-connected whole, and that Searle's intuition (that the man does not understand) illegitimately focuses on the wrong part (the CPU) or ignores the need for causal embedding. A balanced verdict is that the Chinese Room makes a powerful case that PURE computation/syntax is not sufficient for understanding and intentionality (the 'syntax is not semantics' point is widely influential), which is a serious problem for computational functionalism specifically; but that the systems and robot replies show the argument may target a crude version — a functionalism that grounds symbols in the world and takes understanding as a property of the whole embedded system is not obviously refuted. Together with the inverted-qualia and China-brain objections, the Chinese Room shows that functionalism, like every physicalist theory before it, struggles with the two marks of the mental identified at the start of the unit: the China brain and inverted qualia press the QUALIA problem, and the Chinese Room presses the INTENTIONALITY problem. The recurring difficulty of consciousness and understanding — the hard problem and the problem of original intentionality — is thus the thread that runs through the whole metaphysics-of-mind unit, unresolved by dualism (which honours qualia but faces interaction), by physicalism (which honours the science but faces qualia), or by functionalism (which solves multiple realisability but faces absent/inverted qualia and the Chinese Room). A first-class judgement reaches a verdict on which theory best balances these demands while acknowledging that the felt and the meaningful character of mind remains the deepest open problem in the philosophy of mind.

Worked example

The systems reply and Searle's rejoinder

Explain the systems reply to the Chinese Room and Searle's response, and assess who is right.

The argument
The man follows the rules and outputs correct Chinese, but understands nothing; so running the program does not suffice for understanding.
The systems reply
The MAN does not understand, but the WHOLE SYSTEM (man + rules + symbols) does — understanding is a property of the system, as it is of the brain, not of any single component.
Searle's rejoinder
Let the man MEMORISE the rules and internalise the whole system; he still understands no Chinese — so there is no understanding in the system either.
Assess
Searle's rejoinder is forceful, but critics reply that internalising the rules does not make the man the RIGHT KIND of system (world-connected, at the right level), so the systems/robot reply is not clearly defeated.

Result: The Chinese Room powerfully argues that pure syntax is insufficient for understanding, but the systems and robot replies show it may target a crude, world-disconnected functionalism; whether a suitably embedded functional system could understand remains the live, unresolved question — the intentionality analogue of the hard problem.

Typical mistakes

Confusing the Chinese Room (intentionality/understanding; syntax vs semantics) with the China brain (absent qualia); they are different objections targeting different marks of the mental.
Presenting Searle as a dualist; he holds intentionality is a BIOLOGICAL product of the brain's causal powers — the right physical system is needed, not an immaterial soul.

Active revision

'The Chinese Room shows that a computer could never understand.' Is this claim true? Defend your answer. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Key points

The software/hardware analogy

Key points

Why functionalism succeeds identity theory

Key points

The asymmetry reply to inverted qualia

Key points

Challenging the absent-qualia intuition

Key points

The systems reply and Searle's rejoinder

Sources

Functionalism

Key points

The software/hardware analogy

Key points

Why functionalism succeeds identity theory

Key points

The asymmetry reply to inverted qualia

Key points

Challenging the absent-qualia intuition

Key points

The systems reply and Searle's rejoinder

Sources