The basic claim of behaviourism is that any sentence about mental states can be reduced to statements about behaviour. So the statement “Peter is in pain” can be reduced to statements such as: “Peter clutches his right foot”, “Peter hops around and says ‘ouch’”, “Peter avoids walking on his right foot” etc. Behaviourism therefore favours a strategy of reduction similar to phenomenalism. With phenomenalism all statements about existence are thought to be translatable into statements about actual or possible sense data. Behaviourism makes the parallel claim that all statements about mental states can be translated into statements about actual or possible behaviour.
It is worth keeping in mind that behaviourism was fashionable at the same time as logical positivism which influenced linguistic phenomenalism in theory of knowledge and emotivism (all moral statements can be translated into statements about emotions) in ethics. Also psychology was beginning to develop as a new and independent science seeking for an empirical basis as an alternative to Freud’s method of psychoanalysis. Descartes notion of a mind totally different and independent of the body was no fit object for scientific analysis and so psychologists like B.F. Skinner (1904-1990) and J.B. Watson (1878-1958) set about investigating behaviour as the content of psychological theory. Skinner and Watson supposed that only what is publicly observable is a fit subject for science and so excluded from ‘scientific’ psychology all reference to private mental states.

Watson had been inspired by the famous work of Russian physiologist and psychologist Ivan Pavlov (1849-1936) who while studying dogs’ saliva developed the idea of conditioned relex. He had shown by experiement with dogs how the secretion of saliva can be stimulted not only by food but also by the sound of a bell associated with the presentation of food.

Not all philosophers labelled as “behaviourist” followed the ideas of Skinner and Watson closely. It is common to distinguish between two types of philosophical or analytical behaviourism. With ‘hard’ behaviourism represented by Carl Hempel (1905-1997) and ‘soft’ behaviourism is represented by Gilbert Ryle (1900-1976). Ryle was not so scientifically minded and saw in Descartes “ghost in a machine” a category error solvable through a more careful analysis of terms. He likened it to a foreigner visiting Oxford or Cambridge and being shown a number of colleges, libraries, playing fields and scientific departments and then asking to now see the University. It will need to be explained to him that the university is not some additional building separate from and comparable to those he has already seen, it is all those buildings and organised relation to each other. The same applies to the mind which is not a thing additional to the body (reification fallacy) but just the way that the body is organised in terms of the actions it performs. Ryle developed an analysis of “dispositions” so that to say of someone that they were clever is to indicate the kind of behaviour that they are disposed to exhibited under appropriate conditions. In this way Ryle took overt intelligent actions to be the very workings of the mind and not clues to a secret world inaccessible to normal perception. There are not two worlds, a private world of thoughts and a public world of actions. Once you have perceived the intelligent behaviour you have perceived an intelligent mind.

Carl Hempel was instead part of the logical positivist movement and held that psychology must emulate the ‘hard’ sciences. Basing itself on the verification principle psychology must only employ a non-mentalistic vocabulary. Since other peoples minds are not-observable and our own minds cannot be publicly verified statements about human psychology must only concern observable behaviour. This also implied a more stringent proscription whereby statements cannot refer to actions as actions imply intension which is not observable. One argument often used in support of this view is that if we think about the word “pain” we learn to use this word by observing other people using it. This means that what the word “pain” signifies must be something that is observable and shared by different people. Shared language assumes shared experience and this excludes the possibility of a language that referrers to a purely private or subjective experience. This is called the “private language” argument which was originated by Ludwig Wittgenstein who nevertheless resisted being labelled a behaviourist. Still by showing that such a language is impossible, some believed that this shows that the idea of purely subjective “mental phenomenon” is also impossible and must instead have a public and observable character i.e. behaviour.

The problem with this is that what if Peter wants to be perceived by others as a hard guy and so doesn’t want to show his pain. His behaviour will then need to be explained by his mental states. For example “Peter doesn’t say ouch because he believes this would make him seem weak to others” etc. and so behaviour is explained in terms of mental states rather than the other way around. This is sometimes referred to as the intensional circle. For example Caroline wants to stay dry so she takes an umbrella when she leaves the house, but according to the behaviourist Caroline’s desire to stay dry is constituted by her taking the umbrella. This is because on a behaviourist analysis mental events are not causes of behaviour, or reasons for acting, instead the distinction between mental events and dispositions to behave in certain ways is collapsed.

A second problem is similar to that which plagues all attempts at analytic reduction is whether statements about mental states can really be translated successfully into language about behaviour. Donald Davidson noted that any attempt to translate statements such as “Fred believes there is life on Mars” into behaviour statements cannot escape relying on the language of mental states. So to say that Fred in the context of certain conversations is disposed to make statements along the lines of “I believe there is life of mars” can only serve as a translation of “Fred believe there is life on Mars” if Fred is not being ironic, if Fred understands English and means to make an assertion and so on. The italicised words above show how ‘mentalistic’ language cannot be eliminated. Another way to put this is to say that while mental states typically cause behavioural effects they always do so in relation to other mental states. This state of affairs is sometimes called the holism of the mental.
Related LinksPhilosophy of mind