Skip to main content

What makes something a heart? Not the material: a plastic pump that circulates blood counts as an artificial heart. What makes something a mousetrap? Anything that reliably catches mice. Many philosophers think minds work the same way. This idea, called functionalism, became the dominant theory of mind in the second half of the twentieth century, and it lies behind most serious talk of machine minds today.

Mental states as roles

Functionalism holds that a mental state is defined by what it does: its causal role linking inputs, outputs and other mental states. Pain, on this view, is whatever state is typically caused by bodily damage and in turn causes distress, the wish for it to stop, protective behavior and the belief that something is wrong. What the state is made of does not matter, so long as it plays that role.

Functionalism grew from a problem with the identity theory. In 1967 the American philosopher Hilary Putnam argued that if pain simply is a particular human brain process, then an octopus, whose nervous system is built very differently, could not feel pain. That seemed absurd. Putnam concluded that mental states show multiple realizability: the same state can be realized in very different physical systems. A striking consequence follows. If the right functional organization is all that matters, then a mind could in principle run on silicon as well as on neurons.

Objections from experience

Critics argue that functionalism captures what minds do but leaves out how they feel. The oldest version comes from the English philosopher John Locke, who suggested in 1690 that the idea a violet produced in one person's mind might be the very idea a marigold produced in another's. Today this is called the inverted spectrum: two people could be functionally identical while their color experiences are swapped. Both call grass green and stop at red lights, yet what one experiences when seeing grass is what the other experiences when seeing a ripe tomato. If that is possible, experience is not fixed by function.

In 1978 Ned Block offered another challenge, often called the China brain. Imagine the whole population of China, each person given a two-way radio and a set of instructions, together imitating the functional organization of a human brain for an hour. Each citizen plays the role of a neuron. If functionalism is true, the nation as a whole would have experiences, perhaps a toothache. Block found this hard to believe, and so do many readers. Functionalists reply that our disbelief comes from the odd scale and speed of the system, not from a real flaw in the theory.

Searle's Chinese Room

The most famous challenge came in 1980 from the American philosopher John Searle, in a paper titled "Minds, Brains, and Programs." Searle imagined himself locked in a room. He knows no Chinese. Through a slot come cards covered in Chinese characters. Using a large rulebook written in English, he matches the shapes and passes back other cards. Outside, native speakers find the answers perfect: as far as they can tell, someone in the room understands Chinese. Yet Searle understands nothing; he is only shuffling symbols.

A computer running a program, Searle argued, is in the same position. It manipulates symbols by their shapes, according to rules. His slogan was that syntax is not sufficient for semantics: running a program is not enough for understanding, however well the program performs. Searle did not deny that machines could think. He held that brains are machines that do think, because of their specific biological powers, a view he called biological naturalism. Any artificial mind would need the causal powers of a brain, not just its program.

Replies

The most common reply is the systems reply: Searle is only one part of the system, like a single neuron, and the whole system of man, rulebook and cards understands Chinese. Searle answered that he could memorize the rulebook and do everything in his head, and still not understand a word. The robot reply says understanding needs a body: put the program in a robot with cameras and arms, linking symbols to things in the world. Searle answered that this only adds more symbols coming in from sensors. Note that Searle's argument was first about understanding, not feeling, though he held that the two are closely linked.

Worked example: replacing a brain one cell at a time

Chalmers proposed a thought experiment that pulls in the other direction. Suppose surgeons replace one of your neurons with a tiny chip that does exactly what the neuron did: same inputs, same outputs, same timing. They replace another, and another, over months, until your whole brain is silicon.

Your behavior never changes, because each chip does its neuron's job. So at every stage you would say, sincerely, "I feel fine; the sunset still looks glorious." If experience were slowly fading as the chips took over, you would be wrong about your own experience while your reports stayed confident. Chalmers argued that this kind of systematic error is very implausible, and concluded that experience depends on functional organization, not on the material that carries it. A follower of Searle would answer that the fading is possible precisely because the reports are produced by function alone.

Quiz

5 questions · Practice: scored, not saved

1. According to functionalism, what defines a mental state?
2. What did Putnam mean by multiple realizability?
3. What follows from functionalism about machine minds?
4. What does the inverted spectrum scenario imagine?
5. In Block's China brain, what does the population do?
6. How did Searle answer the systems reply?
7. What does the systems reply to Searle say?
8. What did Searle call his view that brains think because of their biological powers?

CivilSimian.com created by AxiomaticPanic, CivilSimian, Kalokagathia