The Ghost in the Machine
The Philosophy of Ratomorphism
Behaviourism started as a kind of puritan revolt against the excessive use of introspectionist methods in some older schools of psychology which held -- in James's definition -- that the business of the psychologist was 'the description and explanation of states of consciousness'. Consciousness, Watson objected, is 'neither a definable nor a usable concept, it is merely another word for the "soul" of more ancient times. . . . No one has ever touched a soul or seen one in a test-tube. Consciousness is just as unprovable, as unapproachable as the old concept of the soul. . . . The Behaviourists reached the conclusion that they could no longer be content to work with intangibles and unapproachables. They decided either to give up psychology or else to make it a natural science. . . . ' [21]
This 'clean and fresh programme', as Watson himself called it, was based on the naive idea that psychology could be studied with the methods and concepts of classical physics. Watson and his successors were quite explicit about this; their efforts to carry out their programme became a truly procrustean operation. But while that legendary malefactor merely stretched, or cut off, the legs of his victim to make him fit his bed, Behaviourism first cut off his head, then chopped him up into 'bits of behaviour in terms of stimulus and response'. The theory is based on the atomistic concepts of the last century, which have been abandoned in all other branches of contemporary science. Its basic assumptions -- that all activities of man, including language and thought, can be analysed into elementary S-R units -- were originally founded on the physiological concept of the reflex arc. The newborn organism came into the world equipped with a number of simple, 'unconditioned' reflexes, and what it learnt and did in its lifetime was acquired by Pavlovian conditioning. But this simplicist schema soon went out of fashion among physiologists. The greatest among them in his time, Sir Charles Sherrington, wrote already in 1906: 'The simple reflex is probably a purely abstract conception, because all parts of the nervous system are connected together and no part of it is probably ever capable of reaction without affecting and being affected by various other parts. . . . The simple reflex is a convenient, if not a probable, fiction.' [22]
More recently, a leading neurologist, Judson Herrick, summed up the situation:
During the past half-century an ambitious programme of reflexology was elaborated, notably by Pavlov and the American school of Behaviourism. The avowed objective was to reduce all animal and human behaviour to systems of interlocking reflexes of various grades of complexity. The conditioning of these reflexes by personal experience was invoked as the mechanism of learning. The simple reflex was regarded as the unit of behaviour, and all other kinds of behaviour were conceived as brought about by the linkage of these units in successively more complicated patterns. The simplicity of this scheme is attractive but illusory. In the first place, the simple reflex is a pure abstraction. There is no such thing in any living body. A more serious defect is that all the information we have about the embryology and phylogenetic development of behaviour shows clearly that local reflexes are not the primary units of behaviour. They are secondary acquisitions. [23]
With the decline of the reflex, the physiological foundations on which S-R psychology was built, had ceased to exist. But that did not unduly worry the Behaviourists. They shifted their terminology from conditioned reflexes to conditioned responses, and kept manipulating their ambiguous terms, in the manner we have seen, until responses became controlled by stimuli still in the womb of the future, reinforcement turned into a kind of phlogiston, and the atoms of behaviour evaporated in the psychologist's hands even as the physicist's hard little lumps of matter had evaporated long ago.
Historically, Behaviourism started as a reaction against the excesses of introspective techniques, as practised particularly by German psychologists of the so-called Würzburg school. At first its intention was merely to exclude consciousness, images and other non-public phenomena as objects of study from the field of psychology; but later on this came to imply that the excluded phenomena did not exist. A programme for a methodology, which had its arguable points, became transformed into a philosophy which had no point at all. One might as well tell a team of land surveyors that for the purpose of mapping a limited area they could treat the earth as if it were flat -- and then subtly instil the dogma that the whole earth is flat.
Behaviourism is indeed a kind of flat-earth view of the mind. Or, to change the metaphor: it has replaced the anthropomorphic fallacy -- ascribing to animals human faculties and sentiments -- with the opposite fallacy: denying man faculties not found in lower animals; it has substituted for the erstwhile anthropomorphic view of the rat, a ratomorphic view of man. It has even re-named psychology, because it was derived from the Greek word for 'mind', and called it the 'science of behaviour'. It was a demonstrative act of semantic self-castration, in keeping with Skinner's references to education as 'behavioural engineering'. Its declared aim, 'to predict and to control human activity as physical scientists control and manipulate other natural phenomena' [24], sounds as nasty as it is naive. Werner Heisenberg, one of the greatest living physical scientists, has laconically declared: 'Nature is unpredictable'; it seems rather absurd to deny the living organism even that degree of unpredictability which quantum physics accords to inanimate nature.
Behaviourism has dominated the stage throughout the dark ages of psychology, and is still, in the 1960s, dominant in our universities; but it never had the stage all to itself. In the first place there have always been 'voices in the wilderness', mostly belonging to an older generation which had come to maturity before the Great Purge. In the second place, there was Gestalt psychology, which at one time looked like a serious rival to Behaviourism. But the great expectations which the Gestalt school aroused were only partly fulfilled, and its limitations soon became apparent. The Behaviourists managed to incorporate some of their opponents' experimental results into their own theories, and continued to hold the stage. The interested reader can find this controversy outlined in The Act of Creation, and there is no need to go into it here.* But the net result was a kind of abortive Renaissance followed by a Counter-Reformation. Lastly, to round off the picture, there is a younger generation of neurophysiologists and communication theorists who regard orthodox S-R psychology as senile, but are often forced to pay lip-service to it, if they want to get on in their academic careers and get their papers published in the tight sort of technical journal -- and who become in varying degrees infected in the process by the doctrines of flat-earth psychology.
* Particularly in Book Two, Chapter Twelve, 'The Pitfalls of Learning Theory', and Chapter Thirteen, 'The Pitfalls of Gestalt'.
It is impossible to arrive at a diagnosis of man's predicament -- and by implication at a therapy -- by starting from a psychology which denies the existence of mind, and lives on specious analogies derived from the bar-pressing activities of rats. The record of fifty years of ratomorphic psychology is comparable in its sterile pedantry to that of scholasticism in its period of decline, when it had fallen to counting angels on pin-heads -- although this sounds a more attractive pastime than counting the number of bar-pressings in the box.
II
THE CHAIN OF WORDS AND THE TREE OF LANGUAGE
On an occasion of this kind it becomes more than a moral duty to speak one's mind. It becomes a pleasure. Oscar Wilde
The emergence of symbolic language, first spoken, then written, represents the sharpest break between animal and man. Many social animals have some system of communication by signs and signals, but language is a species-specific, exclusive property of man. Even 'mongolian' idiots, incapable of looking after themselves in the most primitive ways, are capable of acquiring the rudiments of symbolic speech but not dolphins and chimpanzees, highly intelligent as they are in other respects. Nor rats and pigeons.
Language, then, one would expect, is a phenomenon whose study more than any other would show up the absurdity of the ratomorphic approach. It not only does that; i
t also provides the best opportunity for introducing, by way of contrast, some of the basic concepts of the new synthesis in the making. This contrast between the orthodox and the new approach can be summed up by two key words: the chain versus the tree.
The Chain
The long extract which follows is representative of the orthodox Behaviourist approach to language. It is taken from a textbook for college students to which various professors at distinguished American universities have contributed. [1] The author of the extract is himself chairman of a psychology department. It was published in 1961; the dialogue featured in the extract is adapted from an earlier textbook. I mention these details to show that this text, fed to thousands of students, is in the most respectable academic tradition. It is headed 'Complex Activities' and it is the only passage devoted to the glories of human language in this entire textbook:*
We have said that learning either may be of the respondant [classical Pavlovian] or of the operant [Skinner, Hull] conditioning type. . . . The experimental data that we have presented in connection with our conditioning studies have, however, been limited to rather simple responses such as salivation [in dogs] and bar-pressing [by rats]. In our everyday life we seldom spend much time in thinking about such isolated responses, usually thinking of more gross activities, such as learning a poem, carrying on a conversation, solving a mechanical puzzle, learning our way around a new city, to name only a few. While the psychologist could study these more complicated activities, as is done to some extent, the general approach of psychology is to bring simpler responses into the laboratory for study. Once the psychologist discovers the principles of learning for simpler phenomena under the more ideal conditions of the laboratory, it is likely that he can apply these principles to the more complex activities as they occur in everyday life. The more complex phenomena are, after all, nothing but a series of simpler responses [sic.] Speaking to a friend is a good example of this. Suppose we have a conversation such as the following: He: 'What time is it?' She: 'Twelve o'clock.' He: 'Thank you.' She: 'Don't mention it.' He: 'What about lunch?' She: 'Fine.' Now this conversation can be analysed into separate S-R units. 'He' makes the first response, which is emitted probably to the stimulus of the sight of 'She'. When 'He' emits the operant, 'What time is it?', the muscular activity, of course, produces a sound, which also serves as a stimulus for 'She'. On the receipt of this stimulus, she emits an operant herself: 'Twelve o'clock', which in turn produces a stimulus to 'He'. And so on. The entire conversation may thus be diagrammed as: In such complex activity, then, we can see that what we really have is a series of S-R connections. The phenomenon of connecting a series of such S-R units is known as chaining, a process that should be apparent in any complex activity. We might note that there are a number of sources of reinforcement throughout the chaining process, in this example the most obvious being the reinforcement of 'She' by receiving an invitation for lunch and of 'He' by having the invitation accepted. In addition, as Keller and Schoenfeld point out, there are such sources of reinforcement as the hearer 'encouraging' the speaker to continue, the use that the conversationalists make of the information received (he finds out what time it is), etc. This example of the analysis of a complex activity is but one of numerous activities that we could discuss. You should continue to think of others yourself and try to diagram the chaining process for them. For instance, what would a diagram look like for a football end running downfield and catching a pass, for a pianist playing a piano, or for a girl knitting a sweater? [2] * An extract from this text also appeared in The Act of Creation, p. 603.
And this is the end of what the student learns about 'complex human activities'. The rest of this chapter, entitled 'Learning, Retention and Motivation', is concerned, in the author's own words, with 'salivation and bar-pressing'.
Reading this dialogue one has the vision of two cute automatic slot-machines facing each other on the college campus, feeding each other with stimulus coins and popping out pre-packaged verbal responses. Yet this inane exchange between He and She is not a random improvisation by the author -- he adapted it reverently from another textbook, Keller and Schoenfeld's Principles of Psychology, and other writers have done the same, as if it were a classic example of human conversation.
The diagram represents the application to language of the Behaviourist credo: that all human activities can be reduced to a linear chain of S-R units. At a first glance, the diagram might impress one as a simplified but plausible schematisation -- until one takes a closer look at it. It is based on Skinner's book Verbal Behaviour -- the first large-scale attempt to tackle human language in terms of Behaviourist theory. According to Skinner, speech sounds are emitted as any other 'bits of behaviour'; and the process of conditioning which determines verbal behaviour (including thinking) is essentially the same as the conditioning of rats and pigeons; the methods of these experiments, Skinner claims, 'can be extended to human behaviour without serious modification'. [3] Thus when our author speaks of the psychologist's preference for studying 'simpler responses', he means the responses of salivation and bar-pressing, as the context shows. But what on earth have the S-R symbols in the diagram in common with bar-pressing? What justification is there to call 'Don't mention it -- What about lunch?' a 'conditioned response unit'? A conditioned response is a response controlled by the stimulus; and a 'unit' in experimental science must have definable properties. Are we to believe that He was conditioned to answer each 'Don't mention it' with a lunch invitation? And in what conceivable sense are we to call 'Don't mention it -- What about lunch?' a unit of behaviour?
I seem to be labouring points which are obvious to the non-psychologist, but the purpose will soon become apparent. Obviously, then, the phrase 'Don't mention it' might also produce the response 'Well, goodbye' or 'You have got a ladder in your stocking' or a number of alternative 'bits of verbal behaviour', according to whether She uttered the phrase lingeringly with a sexy smile, or as a brisk brush-off, or hovering between the two; and further depending on whether or not He finds her attractive, whether He is free for lunch, and if so whether He has the cash to pay for it. The simple S-R unit is neither simple nor a unit. It is difficult for the layman to believe that the textbook author is not aware of the complex, multi-levelled mental processes which go on in the two people's heads during and in between the emission of sounds. Surely these 'private processes' must be implied, taken for granted, in what the author is saying? Perhaps they are; but by denying that private events have a place in psychology, he has denied himself the possibility, and even the vocabulary to discuss them. The Behaviourist's way to get around this difficulty is to lump all these unmentionable private processes together in the nondescript term 'intervening variables' (or 'hypothetical mechanisms') which 'mediate between stimulus and response'.* These terms are then used as a kind of garbage bin for the disposal of all embarrassing questions about the intentions, desires, thoughts and dreams of the organisms called He and She. An occasional reference to 'intervening variables' serves as a face-saving device, since everything that goes on in a person's mind is covered by it, and need not be discussed. Yet in the absence of any discussion of the mental events behind the dialogue, the comments of our textbook author are reduced to utter triviality, and the neat diagram is empty of meaning. A diagram is meant to give a graphic representation of essential aspects of a process; in this case both text and diagram pretend to do so, but in fact give no indication of what is really happening. The same dialogue could have taken place between casual acquaintances, or shy lovers, or it could record the picking up of a prostitute. The pseudoscientific balderdash: 'When He emits the operant, "What time is it?", the muscular activity produces a sound which also serves as a stimulus,' and so on, is totally irrelevant to the episode it pretends to describe and explain. And this applies generally to any attempt to describe the language of man in terms of S-R theory.
* See Appendix Two.
The Tree
The strategic advantage gained from lab
ouring the obvious absurdity of a theory is that it makes the proposed alternative appear as almost self-evident. The alternative, set out in the pages that follow, proposes to replace the concept of the linear S-R chain by the concept of multi-levelled, hierarchically ordered systems, which can be conveniently represented in the form of an inverted tree, branching downward:
We find such tree diagrams of hierarchic organisation applied to the most varied fields: genealogical tables; the classification of animals and plants; the evolutionist's 'tree of life'; charts indicating the branching structures of government departments or industrial enterprises; physiological charts of the nervous system, and of the circulation of the blood. The word 'hierarchy' is of ecclesiastical origin and is often wrongly used to refer merely to order of rank -- the rungs on a ladder, so to speak. I shall use it to refer not to a ladder but to the tree-like structure of a system, branching into subsystems, and so on, as indicated in the diagram. The concept of hierarchic order plays a central part in this book; and the most convenient way to introduce it is by means of the hierarchic organisation of language.
The young science of psycholinguistics has shown that the analysis of speech presents problems of which the speaker is blissfully unaware. One of the main problems arises from the deceptively simple fact that we write from left to right, producing a single string of letters, and that we speak by uttering one sound after the other, also in a single string, along the axis of time. This is what lends the Behaviourist's concept of a linear chain its superficial plausibility. The eye takes in a whole three-dimensional picture, embracing many shapes and colours simultaneously; but the ear only receives linear pulses one at a time, serially, and this fact may lead one to the fallacious conclusion that we also respond to each speech-sound, bit by bit, one at a time. This is the bait which the S-R theorist has swallowed, and on which he has been dangling ever since.