Is Psychotherapy an Autonomous Scientific Discipline?

A Preliminary Investigation

By Professor Emmy van Deurzen-Smith and David Smith

Although psychotherapy, in its modern form, has existed for some two hundred years, it has only recently begun claiming recognition as an autonomous scientific discipline independent from psychology and medicine (psychiatry), the professions within which it has traditionally been imbedded. The present paper seeks to critically examine this claim drawing on the resources of cognitive science, the history and philosophy of science, psychology, psychiatry and psychotherapy.

1. What is psychotherapy?

Psychotherapy, in its modern form prohably dates from the latter part of the eighteenth century (Ellenberger, 1970). Reil’s Die Psychische Curmethode, published in 1803, seems to have been the first systematic discussion of the subject (Decker,1977). Psychotherapy was first recognised as an auto­nomous discipline in 1811 with the appointment of Heinroth as Professor of Psychotherapy in Germany (Ibid).

Since that time, a large number of psychotherapeutic  modalities have come into existence. There are probably several hundred specific types of psychotherapy. However, virtually all of these can be subsumed under five major categories:

  1. Psychoanalytic psychotherapy includes those forms of psychotherapy derived from the depth psychological theories of Freud, Jung and others. Psychoanalytic psychotherapy considers problems in living to stem from unconscious conflict and developmental arrest.

  2. Cognitive/behavioural therapy includes those forms of psychotherapy based on learning theory and cognitive psychology. It considers problems in living to stem from erroneous learning and thinking.

  3. Humanistic psychotherapy includes those forms of psychotherapy emphasizing human potential and developed in reaction to the alleged reductionism of psychoanalysis and behaviourism. It considers problems in living to stem from blocked feelings.

  4. Systematic psychotherapy includes those forms of psychotherapy based on general systems theory. It considers problems in living to be generated by malfunctioning in the system or group to which the person is affiliated.

  5. Existential psychotherapy includes those forms of psychotherapy based on existential and phenomenological philosophy. It considers problems in living to be generated by a lack of clarity and understanding of the human condition.

Because of the diversity of psychotherapeutic modalities, it is unlikely that a single definition can adequately capture all of them. The various forms of psychotherapy are united not definitionally, but rather through their “family resemblances” (Wittgenstein, 1953). Nonetheless the following definition will be found to do considerable justice to the major modes of psycho­therapy.

Psychotherapy is that discipline that seeks to alleviate internally gener­ated or sustained problems in living, primarily although not exclusively through verbal means, by the implementation of methods drawn from a distinctive body of theory dealing with the origins, nature and ameliora­tion of such problems in living.

2. Psychotherapy as a pre-paradigmatic science

Whilst it may seem to beg the question of the scientificity of psychotherapy, which will be discussed in Section 4, it is perhaps useful to point out that the very diversity of psychotherapeutic approaches does not supply adequate a priori grounds for the denial of its scientificity. Kuhn (1962) has shown that every developed science undergoes a lengthy phase of diversity and con­fusion which Kuhn characterises as its “pre-paradigmatic” phase. Kuhn’s approach to the philosophy and history of science famously claims that once a science has reached a highly developed stage it adopts a “paradigm” – a core conception of the domain under consideration and the investigative procedures proper to that domain – to which virtually all scientists adhere. From that point onwards, what Kuhn calls “normal science” takes place within the boundaries of the paradigm, whilst “revolutionary science” in­volves the radical replacement of the paradigm for another. Kuhn’s pers­pective is widely misrepresented by popular writers on psychotherapy, who confuse Kuhnian paradigms with theories, models of “approaches”.1 Para­digms are global and all-encompassing: models and theories are nested within them (it may be that “approaches” – a more global conception – are characteristic of the pre-paradigmatic phase). Kuhn’s model rules out the simultaneous existence of a multitude of paradigms.

Masterman (1970) distinguishes between two varieties of pre-paradigmatic science: non-paradigm science and multiple-paradigm science. In modifying Kuhn’s use of “paradigm” as setting the horizons for an entire discipline, Masterman introduces some linguistic confusion. Notwithstanding this, her account of “multiple-paradigm” science may provide some insight into the present state of the field of psychotherapy.2

Here, within the sub-field defined by each paradigmatic technique, tech­nology can sometimes become quite advanced, and normal research puzzle-solving can progress. But each sub-field as defined by its tech­nique is so obviously more trivial and narrow than the field as defined by intuition, and also the various operational definitions given by the techniques are so grossly discordant with one another, that discussion on fundamentals remains and long-run progress (as opposed to local pro­gress) fails to occur. This state of affairs is brought to an end when some­one invents a deeper, though cruder … paradigm, which gives a more central insight into the nature of the field, though restricting it and making research into it more rigid, esoteric, precise … either by causing rival, more shallow paradigms to collapse, or alternatively, by attaching them somehow or other to itself, triumphs over the rest, so that advanced scientific work can set in, with only one total paradigm (74).

Without at this point wishing to beg the question of the scientificity of psychotherapy, it is interesting to note that the growing professionalization of psychotherapy, which has been accompanied by an increased willingness to soften theoretical and ideological boundaries, suggests that the field is becoming ripe for the emergence of a single paradigm.

3. The question of scientificity

Prior to any examination of the subject of the scientificity of psychotherapy it will be necessary to briefly examine the general issues surrounding scien­tificity per se, as well as the concept of hermeneutics.

(a) Science is defined methodologically rather than on the basis of its domain

A discipline is not scientific by virtue of what sort of thing it investigates, but rather by virtue of the way in which it investigates – or proposes to investigate – its subject matter (domain). For example, population genetics is not scientific because it studies the proliferation of inherited characteristics. It is scientific because it investigates such phenomena by means of methods designed to objectively discriminate between true and false ideas about the proliferation of inherited characteristics. Science seeks to obtain reliable, objective knowledge about the world through the disciplined implemen­tation of investigative procedures intended to facilitate the evaluation of the degree of evidential support (or falsification) for hypotheses about items within its domain. There is probably no single method for doing this (Feyerabend, 1970) although there are logical norms that any sound scientific activity should satisfy.

(b) Approaches to scientific rationality

There are of course many diverse characterisations of science in the relevant philosophical literature, as well as considerable debate about the relative merits of these characterisations. I will confine myself to delineating three broad approaches to scientific rationality that have played an important role in the debate over the scientificity of psychotherapy. These three approaches are enumerative inductivism, falsificationism and eliminative inductivism.

Enumerative inductivism, holds that every positive instance of a universal generalisation (every “confirmation”) increases the probability that that generalisation is indeed true. The best-supported hypotheses are therefore those with the largest number of observed and reported positive instances. It is the injunction to accumulate large numbers of positive instances that gives this approach its enumerative appellation. Imagine a man wishing to test the universal generalization “all swans are white”. Using the enumerative inductivist strategy he would attempt to accumulate a large number of posi­tive instances of white swans. He would regard each positive instance as increasing the probability that his universal generalization is in fact true.

Falsificationism, an approach to scientific rationality developed by Karl Popper, argues that enumerative inductivism falls foul of the “problem of induction” identified by the Scottish philosopher, David Hume, in the eight­eenth century. Hume demonstrated that no amount of positive instances can logically underwrite an inductive inference. Popper concluded that scientists should therefore attempt to discover negative instances (“falsifications”) of hypotheses, because just one negative instance logically refutes a universal generalisation. According to Popper, science is distinguished from non-science in that its hypotheses are, in principle or actuality, falsifiable. Scientists should therefore (a) ensure that their claims about the world are structured in such a way as to permit falsification, (b) specify just what observations would supply such refutation and (c) seek out or encourage others to seek out just such observations. Falsificationism therefore entails the claim that it is not possible to establish positive knowledge. Knowledge is reliable only to the extent that it has not (yet) been refuted. Using the same example as above, the investigator choosing the falsificationist strategy would seek out a single instance of a non-white swan which he would take as decisively refuting the universal generalization.

Eliminative inductivism. There are several notorious problems inherent to falsificationism. Among these are (a) the fact that Popperian norms do not in fact correspond with events in the real history of successful scientific activity, (b) that they cannot cope adequately with probabilistic hypotheses, and (c) that they cannot always successfully demarcate science from prima facie non-science. A fourth very powerful criticism of falsificationism stems from the work of the French philosopher/physicist Pierre Duhem. The thesis, known as Duhem’s Thesis (or sometimes the Duhem/Quine Thesis) states that the failure to confirm a prediction deduced from a scientific theory may stem from the scientist adhering to a false auxiliary hypothesis or from an incorrect assessment of initial conditions rather than being attributable to the falseness of the theory taken as a whole.

Eliminative inductivism is an approach to scientific rationality that attempts to establish positive knowledge not through accumulating in­stances of verification, but rather through falsifying an hypothesis with respect to some rival hypothesis by adverting to evidence appropriate for pro­viding rational grounds for this procedure. Again using the same example, the investigator deploying the method of eliminative inductivism would test two competing hypotheses (for example, the claim that all swans are white and the claim that all swans are black) against a decisive body of evidence.

4. The six canons of eliminative inductivism

According to the canons of eliminative inductivism as adapted from Edelson (1984) some observation provides probative evidence that hypothesis x is true if and only if:

1) The outcome is deducible from hypothesis x.
2) The outcome is predicted by hypothesis x.
3) The outcome might, in principle, not occur.
4) The outcome supports hypothesis x over rival hypothesis y (sharing the same domain) if the outcome falsifies y (is not deducible from y), neither the occurrence of the outcome nor its failure are deducible from y, or, given the truth of y, the outcome is extremely unlikely to occur.
5) Research is designed in such a way that, if the outcome deduced from hypothesis x occurs, then alternative explanations for the outcome can be excluded.
6) Research is designed in such a way that, if the outcome deduced from x fails to occur, explanations other than the falsity of x can be excluded.

Psychotherapy must, in principle or in fact, be able to honour these criteria for hypothesis evaluation if it is to be considered a scientific discipline.

5. Hermeneutics

The term “hermeneutics” was taken into philosophy from its source in theology by Wilhelm Dilthey, as part of his Verstehenphilosophie. Hermen­eutics is that discipline that seeks to interpret the meaning of human activity rather than specify its causes.3 This distinction underpins Dilthey’s demarca­tion between Geisteswisenschaft and Naturwissenschaft. It is arguable that psychotherapeutic research adverts to meaning-structures rather than to causal relations, and therefore that psychotherapy is a hermeneutic rather than a natural scientific discipline.

Although the term “science” has been construed in the present report as Naturwissenschaft rather than Geisteswissenschaft, it is important to emphasize that the hermeneutic conception of psychotherapy has gained considerable currency and has generated its own literature. There also exists a substantial literature on hermeneutic and phenomenological methods of research appropriate to the study of psychotherapy, a comprehensive survey and evaluation of which would require a separate investigation but which may offer considerable scope for the establishment of a science of psychotherapy.

6. Psychotherapy as science

Practically all of the debates about the scientificity of psychotherapy have focused on just one form of psychotherapy: psychoanalysis. This has prob­ably been due both to the cultural prominence of psychoanalysis and Freud’s explicit natural scientific claims on its behalf. Be this as it may, most of the issues that have been raised with respect to psychoanalysis are also applic­able to competing psychotherapeutic modalities. I will therefore review the main questions that have been raised with respect to the scientificity of psychoanalysis as points about the scientificity of psychotherapy.

The logical positivist critic, Ernest Nagel (1959) has criticised psycho­therapeutic knowledge-claims as unverifiable on the grounds that there are no clear, disciplined correspondence rules linking theoretical terms with obser­vation statements. There are consequently no principled methods for deter­mining just when an observation is a positive instance of an hypothesis. Nagel’s work challenges psychotherapists to make their theoretical language more perspicuous.

Popper (1962) has criticised psychotherapy on falsificationist grounds, claiming that the form of psychotherapeutic theoretical propositions is such that they are, in principle, unfalsifiable and that psychotherapy is therefore non-scientific. Psychotherapeutic claims are, in essence, compatible with too wide a variety of possible observations. For example, there is no conceivable observation capable of refuting the idea, propounded by Melanie Klein, that all people unconsciously suffer from psychotic anxieties. Popper’s work challenges psychotherapists to make sure that their claims are in principle refutable on evidential grounds, to specify just what observations would supply such refutation and to seek out such refuting evidence.

Eliminative inductivist Adolf Grünbaum (1984, 1993) has taken issue with Popper not on the grounds of the scientificity of psychotherapy but on the basis of the specific defects responsible for this status. According to Grünbaum, at least some of the claims offered by psychotherapists are falsifiable (for instance, Freud’s claim that paranoia is caused by repressed homosexual desires). However, if successful therapeutic outcome is used effectively to underwrite the truth of the theory from which the curative interventions have been deduced, it is necessary to neutralise the powerful competing explanation of placebo effects in a fashion conforming to the six canons of eliminative inductivism itemised above. Grünbaum argues that this is probably impossible in the clinical situation, and that research evaluating the theoretical claims of psychotherapy should therefore rely on experimental and epidemiological strategies.

Hermeneutic critics (e.g. Habermas, 1970; Klein, 1976; Ricoeur, 1970; Schafer, 1983) have sharply criticised any attempt to situate psychotherapy as a natural scientific discipline. Amongst the first of those to accuse psychotherapy of “scientistic self-misunderstanding” was Wittgenstein (1966). Wittgenstein claimed, when discussing psychoanalysis, that (a) Freud confuses explanation adverting to causes with explanation adverting to reasons, (b) meaningful human activity is properly explained with reference to reasons rather than causes (c) explanations adverting to causes are proper to natural science and (d) psychoanalysis is therefore not a natural science, although it confusingly presents itself as such.

Wittgenstein’s critique is rooted in the notion of a radical distinction between reasons and causes, a view with the status of a philosophical ortho­doxy for several decades. However, during the nineteen-sixties this ortho­doxy was overturned in the world of analytic philosophy by the work of Donald Davidson (1963). Davidson demonstrated that reasons are a species of cause. In spite of this, Davidson preserves the gist of Wittgenstein’s criticism by arguing that psychological explanation possesses special properties that are absent from explanation in the natural sciences.

7. The psychology of propositional attitudes

Psychotherapeutic theory is normally couched in a form of theoretical discourse described by philosophers as “propositional attitude psychology”.4 The term “propositional attitude” was coined by Bertrand Russell to denote psychological attitudes (referred to by psychological verbs such as “believe”, “hope”, “fear”, “love” and so on) towards propositions. Statements like “John wanted to kiss Mary” are statements within the purview of proposi­tional attitude psychology. As we shall see, adjacent disciplines such as psychiatry and psychology characteristically eschew propositional attitude psychology in favour of forms of discourse invoking sub-personal items (neurophysiological processes, units of behaviour, and so on).

Davidson (1970) and others have argued that propositional attitude psychology possesses several special characteristics:

(a)  Propositional attitude psychology is constrained by norms of rationality. In order to successfully attribute propositional attitudes to persons we must assume that a basic rationality informs their behaviour.

(b) Propositional attitudes are organised holistically in complex networks rather than atomistically. A single psychological item cannot be abstracted from the web of meaning in which it participates.

(c)  Propositional attitudes cannot participate in strict scientific laws, either psychological or psychophysical. The description of human beings in terms of propositional attitudes is therefore incompatible with the de­scription of human beings in terms of strict nomologicity, although it does not entail denying that mental life is in some sense nomologically ordered.

According to Davidson, human activity described in terms of intentional psychology cannot give rise to, or be incorporated into, anything like a natural science. By implication, psychotherapy – which relies on folk-psychological patterns of explanation – cannot be a natural science.

8. A conception of “science” appropriate to psychotherapy

Debates about the scientificity of psychotherapy have been centred on the examination of actual, specific psychotherapeutic modalities. Philosophers of science have not found grounds for arguing that psychotherapy is unscien­tific tout court. Although hermeneuticians have argued that psychotherapy cannot in principle be a natural science, this view has been challenged as being based on inaccurate conceptions of science (Eagle, 1987; Grünbaum, 1984; Grünbaum, 1993). In any case, hermeneutic writers consider psycho­therapy a science in the sense of one of the Geisteswissenschaften. Also any hermeneutic construal of psychotherapy must come to grips with the eliminative inductivist requirement that some objective method be deployed to adjudicate between rival theoretical claims.

Given the varied and diverging views about the actual or potential scien­tificity of psychotherapy, it may be more productive to consider in what sense psychotherapy might actually or potentially be considered to be a science whilst satisfying the minimal requirements of both natural scientific and hermeneutic investigators. I suggest the following:

Psychotherapy is a potential science in that it is a clearly demarcated discip­line with a clearly defined domain which seeks to make justified knowledge claims about that domain based on investigative procedures the results of which are independent of the perspective of any individual investigator. Psychotherapy seeks to deduce from its body of reliable knowledge effective means of applying this knowledge to preventative, palliative and therapeutic ends and to use pragmatically established methods as the basis for inductions contributing to the fund of theoretical knowledge. Psychotherapy strives to deploy objective methods for evaluating the relative merits and demerits of rival hypotheses adverting to its domain.

9. Criteria for disciplinary autonomy

Any consideration of whether or not psychotherapy can be considered an autonomous discipline must be preceded by a consideration of just what characteristics distinguish autonomous disciplines. Any approach to discip­linary autonomy must consider the following criteria.

  1. 1. The discipline under consideration must be theoretically distinct from any adjacent discipline.
  2. 2. The discipline under consideration must possess a theory which is irreducible to any adjacent theory.

Even if, for example, obsessive-compulsive symptoms can be given a true neuroscientific characterisation, this does not entail that a true psycho­analytic account of such symptoms is therefore reducible to a neurophysiol­ogical account, any more than a functional description of the operation of a pocket calculator is reducible to a physical description of the same device.

A problem arises in that these two points seem to apply to a paradigmatic rather than pre-paradigmatic sciences. As a matter of fact, the various forms of psychotherapy employ diverse although frequently overlapping theore­tical terms. It might be argued that for any pre-paradigmatic sub-discipline, autonomy is secured if and only if its theory is neither identical nor reduc­ible to the theory of an adjacent discipline. This approach to the problem entails that:

  1. 1. The autonomy of psychotherapy taken as a whole is just the autonomy of each of its constituent modalities.

  2. 2. The non-autonomy of specific forms of psychotherapy has no implica­tions for the autonomy of rival forms of psychotherapy which do not share its theoretical terms.

I submit that, under the terms just presented, the psychotherapies are autonomous disciplines in that their theoretical terms are neither identical to those of adjacent disciplines nor are they reducible to them (no such inter-theoretic reduction has to date been achieved).

10. Relations between disciplines

There are, of course, numerous ways in which scientific disciplines can be related apart from those of identity or reducibility (Kitcher, 1992). Other relations include constraint, a relation that holds when it is stipulated that a discipline may not assert theses incompatible with some other discipline. The relation of constraint holds, for example, between medicine and physiol­ogy, i.e. medicine cannot assert anything that is incompatible with the laws of physiology. A discipline may also seek support and inspiration from another discipline (with or without being constrained by it). Bowlby’s (1969) use of ethology as a basis for reconceptualising psychoanalysis is an example of this. A discipline may also provide explanatory extension for another discipline. Matte Blanco’s (1985) use of set theory to illuminate Freud’s con­cept of the unconscious is an example of the use of mathematical logic to provide explanatory extension for psychoanalysis.

On the whole, psychotherapists have shown little interest in situating their discipline in the context of other disciplines. Freud was an exception to this trend, and his model of the relationship between psychoanalysis and neuroscience, evolutionary theory, sexology, literature, the history of civilisation, ethnic psychology (Volkerpsychologie), mythology, linguistics, anthropology, sociology, child psychology, general psychology and psychi­atry is carefully delineated by Kitcher (1992). It is perhaps worth noting that Freud held psychoanalysis to be autonomous with respect to each of these adjacent disciplines.

11. Psychotherapy and medicine

It is often asserted – implicitly or explicitly – that psychotherapy should be subsumed under medical science. The association between psychotherapy and medicine is so pervasive that it is often taken for granted that psycho­therapy is appropriately practised by psychiatrists, in the context of psychi­atric departments situated within medical hospitals, that psychotherapy should be covered by medical insurance, that psychotherapy should be regulated by governmental Departments of Health, and so on.

Historically, modern psychotherapy began as an outgrowth of philosophy rather than medicine (Decker, 1977). There has been detailed work on the rise of modern psychiatry and its relationship to psychotherapy in the German context which can probably be generalised with reasonable accur­acy. In Germany, psychotherapy was originally the province of the “roman­tic psychiatrists”. This term is, however, rather misleading as these men were not psychiatrists in the contemporary sense of the word.

The romantic psychologists were strongly influenced by naturphilosophie, and attempted to explain mental disorders as mentally, rather than soma­tically, caused (Decker, 1977). The rise of modem psychiatry was inaug­urated by the publication of Wilhelm Griesinger’s Pathologie und Therapie der psychischen Krankheiten (1945). Griesinger’s shibboleth “Geisteskrankheiten sind Gehirnkrankheiten” (“Illnesses of the spirit are mental illnesses”) became a rallying-cry of the new organic psychiatry. By mid-century, psychiatry had become wedded to the idea that mental disorders were reducible to anatom­ical or physiological pathology, often of an innate nature, and could only be allowed to run their course (Ibid).

Now [psychiatry] … was carried on by people who no longer shared their whole life from morning to night with their patients. It found its way into the laboratories for brain anatomy or for experimental psychopathology; it became more coldblooded, detailed, impersonal and less humane. It lost itself in endless particulars, measurements, statistics and findings. It lost imagination and design (Jaspers, 1964: 847).

From its inception, then, modern psychiatry ignored the distinctively psychotherapeutic principles of (a) explaining mental disorders in mentalistic terms and (b) intervening verbally to ameliorate mental disorders. For the psychiatrists, psychotherapy was tainted by unscientific “philosophy” and “seemed to offer minimal contributions towards the understanding and care of the mad and the insane” (Decker, 1977: 61).

In the main, the development of psychotherapy occurred outside main­stream psychiatry, often incurring condemnation by psychiatrists as a throwback to naturphilosophie (Decker, Ibid). Those psychiatrists, like Jung and Bleuler, who also practised psychotherapy often found their inspiration outside psychiatry. I am not aware of a single approach to psychotherapy that was deduced from psychiatric theory. This is hardly surprising. Psychi­atry qua psychiatry offers causal and functional explanations of disorders of the nervous system, and physicalistic modes of treatment deduced from these (or inductively related to them). Psychiatric theory does not properly advert to the intentional level of description characteristic of psychotherapy. Of course, a number of psychiatrists also practice psychotherapy, but this fact has no bearing on the conceptual relationship between the two disciplines (Szasz, 1961).

It is clear, however, that psychotherapy is to some extent constrained by psychiatry. Psychotherapy should not, for example, make claims incompatible with psychiatric work on the neurophysiological basis of some forms of mental disorder (unless, of course, it is prepared to rigorously test such claims against evidence).

12. Psychotherapy and psychology

The term “psychology” is used in everyday parlance to refer to anything pertaining to the human mind. Of course, in this broad and non-technical sense of the word, psychotherapy would naturally be subsumed under psychology. However, this everyday usage must be distinguished from the use of “psychology” to denote a special scientific discipline with its own institutional structures and methodological norms. When we enquire into the relationship between psychology and psychotherapy, it is the latter, technical meaning of the word that is of concern to us.

Modern scientific psychology arose during the 1850’s. The new approach, which repudiated earlier more “philosophical” forms of psychology, was christened “experimentelle Psychologie” by its founder, Wilhelm Wundt. Psychology was henceforth to ally itself with physiology, and as the German psychologist Störring put it, no longer consider itself “the science of the soul” (Decker, 1977: 194). The psychotherapies were therefore seen as hostile to these efforts to define psychology as an experimental science. In 1879 Wundt established the first university department of experimental psychol­ogy at Leipzig, and by 1900 more than one hundred such departments had been established world-wide (Fancher, 1990). Although Wundt’s “introspectionism” is rejected by contemporary psychology, it was Wundt who estab­lished psychology as a scientific discipline relying on experiment, observa­tion and quantification. Robinson (1981) summarises this tradition, which characterises contemporary scientific psychology, as follows:

First, there is wide agreement that a scientific psychology must be an experimental discipline rather than a deductive or rationalistic one. Secondly – and because of this agreement – there is a general adoption of the view that the subject matter of psychology must be confined to obser­vable and quantitatively describable phenomena. Thirdly, to the extent that modern psychology is attentive to philosophy of science, there is widespread support for the (neo-positivist) view that a developed science can rise no higher than the level of reliable empirical laws (423).

Robinson’s second point restricts psychology to the study of behaviour. [5] Given the concern of psychology with strict nomologicity and quantification, this virtually always entails that behaviour is explained with reference to theoretical terms other than the networks of prepositional attitudes normally invoked by psychotherapists.

I have already described how psychotherapy developed under the influ­ence of romantic philosophy, and was largely rejected by the new discipline of medical psychiatry. Somewhat paradoxically, the psychologists largely regarded psychotherapy as the province of medicine, because of its “thera­peutic” component. In the case of psychoanalysis “The psychologists’ texts and monographs virtually neglected Freud; a single reference by Wundt was the only exception” (Decker, 1977: 222). The psychologists’ reception of psychoanalysis was considerably more negative than that of the physicians.

As is the case with psychiatry, scientific psychology is set off from psychotherapy in its rejection of intentionalistic modes of theoretical dis­course. If intentionalistic “propositional attitude” discourse is characteristic of psychotherapy, then the mode of theoretical discourse characteristic of scientific psychology is strictly incompatible with that characteristic of psychotherapy. Conceptually, therefore, psychotherapeutic theories are neither identical with nor reducible to psychological theories. Two approaches to psychotherapy, namely behaviour therapy[6] and cognitive therapy, are directly derived from systems of psychological theory. It could therefore be said that, according to the criteria presented in the present paper, cognitive and behavioural approaches to psychotherapy are not auto­nomous disciplines and should be subsumed under the umbrella of psychol­ogy. However, to the extent that these two approaches have incorporated principles from other psychotherapeutic approaches they are no longer strictly reducible to their psychological roots and should be considered to be sub-disciplines within the autonomous discipline of psychotherapy.

Psychotherapy is thus clearly separate from but constrained by psychology. The current debate about “false memory syndrome”, to give one example, demonstrates that psychotherapeutic claims about apparent recollections of early experiences should not be incompatible with psychological knowledge about the operation and limits of memory. Psychological knowledge is therefore able to offer guidance to psychotherapy insofar as the latter addresses entities and processes which have been researched by the former.

It should be added that it is conceivable that a special sub-discipline might be developed within the science of psychology that is specifically concerned with psychotherapy (indeed the relatively new sub-discipline of “counselling psychology” would appear to be a step in this direction). This sub-discipline might be called “psychology of psychotherapy”. The existence of “psychology of psychotherapy” would not be identical with psycho­therapy proper, and would not provide even a prima facie case for subsuming psychotherapy under the umbrella of psychology.

13. Redrawing professional boundaries

Recognition of psychotherapy as an autonomous discipline is a new phen­omenon and will inevitably eventuate in some re-adjustment of existing professional boundaries. As it stands, both psychiatry and psychology continue to lay claim to aspects of both theory and practice which, on the present analysis, are properly within the domain of of psychotherapy. It is, however, clear from this investigation thus far that training in psychiatry or psychology does not provide an appropriate basis for psychotherapeutic work. The demarcation of psychotherapy as an autonomous discipline will entail that psychiatrists and psychologists will no longer be regarded as qualified to practice psychotherapy simply by virtue of having undergone psychiatric or psychological training. Of course, any such professional re-alignment will arouse opposition, as vested interests become threatened. Opposition to the professionalisation of psychotherapy stemming from those professions that have something to lose must therefore be given close and critical scrutiny.

14. Psychotherapy and the intentional stance

Dennett (1986) demonstrates that intentional systems can fruitfully be viewed from three distinct perspectives or “stances”, each of which involves a distinct form of theoretical discourse. An intentional system is “a system whose behaviour can be – at least sometimes – explained and predicted by relying on ascriptions to the system of beliefs and desires” … (3). This kind of prediction and explanation is characterised as intentional prediction and explanation. Human beings are obviously intentional systems. Understand­ing an intentional system with reference to intentional idioms such as “belief” and “desire” is to take an intentional stance towards the system. Intentional explanation is basically the same as explanation adverting to propositional attitudes.

Intentional systems can also be understood from non-intentional perspec­tives. The design stance enjoins us to analyse the system’s competencies. Design-stance analysis is functional and homuncular, breaking down the behaviour of the system into relatively simple interacting units. When taking the physical stance, we examine a system neither with respect to its inten­tional properties nor with respect to its functional units. Instead, we examine it as a purely physical system explicable in strictly physicalistic terms.

As I have noted in other terms, the intentional stance towards human beings characterises psychotherapeutic modes of theoretical discourse. Arguably, the design stance and physical stance are the proper theoretical modes for psychology and psychiatry respectively.

27 p16

In light of these considerations it seems warranted to insert an addendum to the definition of psychotherapy given above.

Psychotherapy is that discipline that seeks to alleviate internally gen­erated or sustained problems in living, primarily although not exclusively through verbal means, by the implementation of methods drawn from a distinctive body of theory dealing with the origins, nature and ameliora­tion of such problems in living addressing the intentional dimension of human life and couched within an intentionalistic theoretical idiom.

15. Criteria for a general science of psychotherapy

We can now turn to those features of psychotherapy which might in prin­ciple provide the basis for a unified science of psychotherapy.

All of the psychotherapies potentially possess a common data base. Con­sensus about what counts as valid psychotherapeutic data a consensus which has not yet been achieved) would permit psychotherapists to test competing hypotheses against mutually accepted data. This would only be possible, of course, if there is consensus about methodological norms (broadly conceived) for evaluating hypotheses. In particular, psychotherapeutic approaches should spell out their criteria for falsification, using these to generate testable predic­tions and retrodictions deduced from the theory. The acceptance of these two proposals would make it possible to test their theories against clinical data generated by practitioners advocating rival theories, thus neutralising the charge that confirmations are due to suggestion.

It might be argued that the psychotherapies cannot constitute a general science for non-cognitive reasons. Given the purpose of psychotherapy, it can be cogently argued that each psychotherapeutic approach advocates a moral or ideological vision of the “good life”, and that not all of these visions are compatible.7 We can, however, grant such diversity on non-cognitive levels whilst (in principle) retaining broad consensus with respect to the broad methodological norms specified above.

16. Conclusions

Issues pertaining to disciplinary autonomy, theoretical reducibility and scientificity are widely accepted as being highly complex. This is particularly true in the psychotherapeutic context for, as I have noted above, psycho­therapy is an umbrella term encompassing a number of distinctive “para­digms”. It has however been possible to reach several fundamental conclus­ions in this investigation.

  • Psychotherapy is, generally speaking, an autonomous discipline with its particular theoretical domain, norms of theoretical discourse and modes of intervention.
  • Psychotherapy cannot be subsumed under medicine, although medical science constrains and may provide guidance to psychotherapeutic theory.
  • Psychotherapy cannot globally be subsumed under psychology (although there are possible exceptions to this dictum). Psychological science constrains and may provide guidance to psychotherapeutic theory [8]
  • Although there is not yet a general science of psychotherapy, there are no clear reasons why such a general science cannot in principle be developed.

These conclusions present clear implications for the development of psycho­therapy as a discipline.

  • Psychotherapy should be free-standing rather than being incor­porated within some other discipline.
  • Psychotherapy should be self-regulating rather than be subjected to methodological, institutional, administrative and educational norms that are alien to it.
  • Training in psychiatry or psychology is not an adequate training for psychotherapists. Psychotherapists should undergo a specifically psychotherapeutic training.


Barkham, M. (1990) Research in individual therapy. In W. Drained (ed) Individual Therapy: A Handbook. Milton Keynes: Open University Press.

Bowlby, J. (1969) Attachment and Loss: Vol. I. Attachment London: Hogarth Press.

Davidson, D. (1963) Actions, reasons and causes. In Essays on Actions and Events. Oxford: Clarendon Press, 1980.

Davidson, D. (1970) Mental events. In Essays on Actions and Events. Oxford: Clarendon Press, 1980.

Decker, H. (1977) Freud in Germany. Madison, CT: International Universities Press.

Eagle, M. (1987) Recent Development in Psychoanalysis: A Critical Evaluation. Cambridge, MA: Harvard University Press.

Edelson, M. (1984) Hypothesis and Evidence in Psychoanalysis. Chicago: University of Chicago Press.

Ellenberger, H. (1970) The Discovery of the Unconscious. New York: Basic Books.

Fancher, R.E. (1990) Pioneers of Psychology. London: W.W. Norton & Co.

Feyerabend, P. (1970) Against method. In Minnesota Studies in the Philosophy of Science. 4.

Frank, J. (1983) Persuasion and Healing. Baltimore: Johns Hopkins University Press.

Griesinger, W. (1845) Patholgie und Therapie der psychischen Krankheiten. Stuttgart: Adolf Krabbe.

Grünbaum, A. (1984) The foundations of Psychoanalysis: A Philosophical Critique. Berkeley: University of California Press.

Grünbaum (1993) Validation in the Clinical Theory of Psychoanalysis: A Study in the Philos­ophy of Psychoanalysis. Madison, CT: International Universities Press.

Habermas, J. (1970) Zur Logik der Sozialwissenschaften. Frankfurt: Surkamp Verlag.

Jaspers, K. (1973) Allgemeine psychopatholgie. New York: Springer Verlag.

Kitcher, P. (1992) Trend’s Dream: A Complete Interdisciplinary Science of Mind. Cambridge, MA: Bradford/MIT.

Klein, G. (1976) Psychoanalytic Theory: An Exploration of Essentials. New York: International Universities Press.

Kuhn (1962) The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Masterman, M. (1970) The nature of a paradigm. In I. Lakalos & A. Musgrave (eds) Criticism and the Growth of Knowledge. Cambridge: Cambridge University Press.

Matte Blanco, I. (1977) The Unconscious as Infinite Sets: An Essay in Bi-Logic. London: Duck­worth.

Murphy, P.M., Kramer, D. & Lilley, F.J. (1984) The relationship between curative factors perceived by patients in their psychotherapy and treatment outcome: an exploratory study. British journal of Medical Psychology. 57: 187-192.

Nagel, E. (1959) Methodological issues in psychoanalytic theory. In S. Hook (ed). Psycho­analysis, Scientific Method and Philosophy. London: Transaction.

O’Leary, K.D. & Borkovek, T.D. (1978) Conceptual, methodological and ethical problems of placebo groups in psychotherapy research. American Psychologist, 33: 821-830.

Popper, K. (1962) Conjectures and Refutations. New York: Basic Books.

Ricoeur, P. (1970) Freud & Philosophy: An Essay on Interpretation. New Haven: Yale University Press.

Robinson, D. (1981) An Intellectual History of Psychology. New York: Macmillan.

Schafer, R. (1983) The Analytic Attitude. London: Hogarth Press.

Szasz, T.S. (1961)  The Myth of Mental Illness. New York: Hoeber-Harper.

Wittgenstein, L. (1953) Philosophical Investigations. Oxford: Basil Blackwell.

Wittgenstein, L. (1966) Wittgenstein: Lectures and Conversations. Oxford: Basil Blackwell & Mott.


1. Kuhn stated in the postscript to the 1969 edition of The Structure of Scientific Revolutions that scientific revolutions can occur within scientific micro-communities. However, the “paradigms” upturned in such episodes should not be confused with Kuhn’s standard concept of “paradigm”.

2. Masterman (1970:74) states that the psychological, social and information sciences are presently in a multiple-paradigm phase.

3. “Hermeneutics” also has a special technical meaning in Existentialist philosophy.

4. Sometimes called “folk psychology” or “desire/belief psychology”.

5. I use “behaviour” in a broad sense to include classes of non-intentional physiological events (e.g. neurophysiological events).

6. Of course, if it turns out that behaviour therapy is deduced from a false psychological theory, or is falsely deduced from a true psychological theory, then these considerations no longer hold. Dennett (1986) writes of behaviourism that “The apparent successes of Skinnerian behaviourism … rely on hidden intentional predictions”.

7. This raises some rather obvious difficulties for the comparative study of outcome.

8. With Freud, we should not assume that psychiatry and psychology are the only relevant adjacent disciplines. I would include anthropology, biology, linguistics, computing science and philosophy as well.

Professor Emmy van Deurzen-Smith is External Relations Officer of the European Association for Psychotherapy and teaches at the New School of Psychotherapy, Schiller International University, London.