Descartes’s world-view included bodies or material things, whose essence is to be extended in space, and minds, which are immaterial things whose essence is thinking. According to Cartesian dualism, the mind is not literally housed within the body, because spatial properties belong to matter and not to mind. But, when he talked about the way we experience the states of our own body, Descartes sometimes spoke of the mind being ‘mixed up with’ the body. Early theories of the brain as the seat of the mind assigned an important role to the ventricles. On Descartes’s view, mechanical operations involving the release of animal spirits in the ventricles were adequate to explain animal behaviour but intelligent human action required something more. He postulated that the immaterial mind could modulate processes in the material brain by way of a causal interaction operating through the pineal gland.

Dualist interactionism is challenged by theories about the motion of bodies. According to Descartes’s own theory, quantity of motion (defined as mass times speed) is conserved. Because motion is not a directional notion, this conservation law allowed that the immaterial mind could bend the trajectory of a physical particle in the pineal gland. But Gottfried Leibniz’s superior theory, with conservation laws for momentum (a directional notion) and kinetic energy, had the consequence that only impacts with other bodies could cause changes in the direction or speed of physical particles. This left no room for immaterial causes of material changes and, while Leibniz was a dualist, he was not an interactionist dualist but believed, instead, in a pre-established harmony between the material and immaterial worlds.

By departing from the idea that impact was the only force on bodies, and allowing action at a distance, Isaac Newton reopened the possibility of distinctively mental forces affecting the trajectory of bodies. These forces were not even ruled out by the law of conservation of energy, which was widely accepted by the middle of the nineteenth century, but advances in biochemistry and neurophysiology during the first half of the twentieth century made appeal to vital and mental forces seem increasingly unmotivated. Since around 1950, the dominant theories of the mind-brain relation have been compatible with a broadly physicalist world-view and with the completeness of physics: physical effects have wholly physical causes.⁽²⁾

From the 1920s to the 1950s, particularly in the United States, behaviourism was a dominant approach within psychology. This was not just methodological behaviourism, which is a restriction on the kinds of evidence that can be used, but a radical reconception of psychology as the science of behaviour rather than the science of the mind. In philosophy, analytical behaviourism was a doctrine about the meaning of our mental discourse. The idea was to analyse or translate our mental talk into talk about patterns of behaviour.

Gilbert Ryle promoted behaviourism as a response to what he called ‘Descartes’s myth’ of ‘the ghost in the machine’.⁽³⁾ A dualist would regard talk about being in love, or wanting to visit Paris, as talk about an immaterial mind whose states lie hidden behind observable bodily behaviour. Ryle proposed to analyse this mental talk as being about the observable behaviour itself. He did not, however, aim to replace all mental terms by terms appropriate to the science of material bodies moving through space. He analysed believing that the ice is thin as, in part, being ‘prone to skate warily’ and it was enough, for his purposes, that skating warily is an observable and recognizable kind of behaviour, even if it is not readily defined in terms of the trajectories of body parts.

Because action is explained in terms of what the agent believes and what the agent wants, analytical behaviourism faces a major objection of principle. There is no pattern of behaviour associated with a belief, by itself. Someone who believes that the ice is thin but has an unusual desire to be immersed in ice-cold water may not skate warily. So there is no prospect of analysing any belief in terms
of behaviour. We might elevate this point into a general requirement on the description of any creature as having beliefs. Attributions of beliefs are not warranted if they merely summarize the creature’s dispositions to exhibit patterns of behaviour. A belief is a mental state that can figure in the explanation of indefinitely many different actions in pursuit of different goals.

Ryle’s behaviourism involved a clear rejection of Descartes’s duality of material and immaterial substances, but central state materialism (also known as the identity theory) encapsulated a more thoroughgoing commitment to the physicalist world-view. If the physical effects of our experiences, thoughts, and volitions have wholly physical causes then there is no causal work left for distinct mental items to do. To avoid epiphenomenalism, mental states, processes and events were to be identified with physical states, processes and events, and mental properties with physical properties. Place advanced a precursor of the identity theory, restricted to the case of conscious experiences,⁽⁴⁾ and this was generalized by Smart, who identified beliefs and desires, intentions and hopes, as well as sensations and experiences, with brain states or processes.⁽⁵⁾

The identity theory defends the idea of mental causation by identifying each mental state with a physical state that is a locus of causal powers. But, taken literally, the identity theory is bound to seem chauvinistic. No being with a physical constitution radically different from ours could be described as feeling anything, or thinking anything, or wanting anything.

The functionalist response to the identity theory is that what a system does is more important than what it is made from. Physically different computing machines can run the same software and one version of the functionalist theory of the mind-brain relation is that the mind is the software of the brain.⁽⁶⁾

In an early version of functionalism, Hilary Putnam proposed that mental states are functional states like the states of an abstractly defined Turing machine rather than physical states like the states of a human brain.⁽⁷⁾ This machine functionalism had the advantage of not tying mental states to a particular physical substrate but also a disadvantage. Since a Turing machine is in only one state at a time, machine states are not analogous to mental states like being in love or wanting to visit Paris.

The dominant version of contemporary functionalism, attributable to David Lewis, is analytical functionalism.⁽⁸⁾ The leading idea is that commonsense specifications of the interconnected causal roles of mental states can be taken as interlocking analyses of mental state terms. For any physical being with a mind, there will be physical states playing each of the mental state causal roles but different physical states may play the same causal role in physically different minded beings—in human beings and Martians, for example.

Functionalism thus avoids the apparent chauvinism of the identity theory by allowing that a human being may be in the same mental state as a being with a very different physical constitution. But functionalism faces the opposite problem of apparently being too liberal. It seems to be possible to make up examples in which physical states play the causal roles that are supposed to define mental states, yet where, intuitively, there is no intelligence and no mental life.⁽⁹⁾