Neuroscience versus Psychology in Freud

Andrew Brook

In the 1890's, Freud attempted to lay out the foundations of a complete, interdisciplinary neuroscience of the mind. The conference that gave rise to this collection of papers, Neuroscience of the Mind on the Centennial of Freud's Project for a Scientific Psychology, celebrated the centrepiece of this work, the famous Project (1895a). Freud never published this work and by 1896 or 1897 he had abandoned the research programme behind it. As he announced in the famous Ch. VII of The Interpretation of Dreams (1900), he would thereafter restrict himself to psychology proper, i.e., what could be done within the ambit of psychological descriptions. The task of characterizing the neural implementation of the psychological was impossible to carry out given the state of knowledge in his time. As Pribram and Gill (1976), Kitcher (1992) and others have demonstrated, Freud's attempt to sketch an interdisciplinary model of the mind using the language of neurons, quantities of energy, etc., was extremely advanced for its time and was probably about as good as could have been done with what was known in 1895. Knowledge of the brain, evolutionary biology, etc., was too limited to allow more.

When Freud narrowed his focus to psychology proper, i.e. to what could be described in the language of psychology, and specifically to the psychology of the unconscious, what happened to his old neuroscientific preoccupations? In my view, they did not disappear. The old preoccupations seldom appear in their explicitly 1895 form of Q and Q energy in , , and systems; Freud had transformed this framework into what he called metapsychology. But the transmutation was a modest one and it is easy to see the old framework at work in the new. The old types of energy and Greek-lettered systems are mirrored in the new metapsychological framework by the dynamics and economics of energy and the topological model of unconscious, preconscious and conscious systems, the latter including perception as one of its subsystems. Metapsychology spelled some things out in more detail than the old neuroscience of the Project, but the neuroscience is taken up into it and in this form, Freud's preoccupations and ways of thinking of the 1890's regularly reappear in his work for the rest of his life. They are very clearly present, for example, in The Interpretation of Dreams (1900), Freud's own explicit pronouncements notwithstanding, the 1915-16 papers on metapsychology, Beyond the Pleasure Principle (1920), and the New Introductory Lectures (1933)--in short, in most of the major works on psychoanalytic theory.

What characterizes this transmuted form of the old neuroscience model? What characterizes the psychological model on which Freud came to focus? And how do the two co-exist in Freud's thought? --These are the topics of this paper.

From the Project to the Interpretation of Dreams

To situate the two kinds of thinking in Freud's work, let us begin with Mark Solms' interesting analysis in his contribution to this collection. Solms follows the development of Freud's thought from his monumental work of neuroscience, the monograph On Aphasia (1891), through the Project (1895a), to what might be thought of as his first work fully in the sphere of psychological theory, Interpretation of Dreams (1900) (though Studies in Hysteria [1895b] would also be a strong candidate for that title).

In Aphasia and the Project, Freud offers neuroscientific, structural and quantitative models of, in the former, the language function and, in the latter, cognition as a whole. Clearly at this stage in his work, he viewed the language of neuroscience as the canonical language in which to do science of the mind. By Interpretation, he had abandoned this view, in some part at least. Solms hypothesizes that the reason was this. Up to and including the 1895 work, Freud equated the psychological with the conscious. Since consciousness is full of gaps, a unified, non-gappy account of the mind is possible only if we switch to a non-psychological account. The only plausible candidate is the "mechanical explanations" (Freud 1954, p. 126) of neuroscience. (We will take up the question of what characterizes these gaps later.)

With the discovery of the unconscious, the range of states that we can describe psychologically greatly expands. Suddenly accounts laid out purely in the language of psychology can be much more complete and much less gappy--suddenly we have the prospect of a complete or far closer to complete account of the mind done in the language of psychology alone. Once we see this, we can abandon neuroscience and do pure psychology. This, in Solms' view, is the position Freud had reached by 1900.

One might quarrel with some aspects of this story. For example, surely Freud was well aware of unconscious psychological states well before 1900, indeed well before 1895. Studies in Hysteria and other psychological works of the period offer one kind of evidence for this; his studies with Charcot and his fascination with post-hypnotic suggestion are another. But on the main point Solms is surely right: by 1900 and thereafter without exception, Freud thought that he had to "remain upon psychological ground" (1900, p. 536), i.e., stick to the language of psychology in his research on the mind.

Is it also true that Freud stopped giving mechanistic, neuroscientific explanations of the mind thereafter? Not at all. Rather, as I said, Freud transformed them into `the witch metapsychology' and regularly had recourse to them throughout his life. The discovery of the unconscious did not lead him to abandon mechanistic models of the mind.

The idea that discovering the impossibility of neuroscience and finding a way to give a more complete picture at the level of psychology should lead to the abandonment of the former suggests that the two are talking about two different things: the brain and the psyche--in an older discourse, the brain and the mind. But they are not; they are talking about one thing, the brain/mind, in two different ways. Freud is very clear about this in 1900 ("the mental apparatus ... is also known to us in the form an anatomical preparation" [p. 535]) and repeatedly thereafter. The two different ways of talking are very different indeed, of course. But they are still talking about one thing. It is rather like the difference between talking about the colours and the shapes of a sculpture. Colour is very different from shape; but they are both properties of a single thing. Likewise, psychological states are very different from neural circuits; but they are both properties of a single thing, viz., the system of roughly 100 billion neurons housed in our cranium.

So what did change between 1895 and 1900? I think the picture is something like this. The heroic speculations of 1895 convinced Freud that he could not unravel the neural implementation of psychological states. Thus, if he was going to study psychological states at all, he had to give up his interest in how they were implemented neurally and concentrate on the states themselves. However, he continued to think that these states could be studied in two ways: psychologically, and metapsychologically. The former is concerned with things like meaning, the relationship between states that are reasons for something and the states that ensue, and so on; the latter are concerned with things like the structure of the systems in which these states are found, the mechanical, energistic relationships among them, and so on. One system, two kinds of interest in it.

One way to see how closely metapsychology is to the neuroscientific speculations of the Project is to note that metapsychology develops everything in terms of flows, blockages, transformations, etc., of energy. Thus the drives are sources of energy, cathexes are investments of energy, the defences are mechanisms for preventing painful or morally unacceptable investments of energy, narcissism is investment of energy on oneself (Freud's view of narcissism varied at various periods, however), the unconscious is a particular relationship of certain residues of energy management to the system Cs (the system Consciousness), primary and secondary process are processes for managing energy and energy discharge, and so on. About all that had changed was that Freud had given up on the hope of developing any account of how these functions were implemented neurologically.

In fact, the change was not even as big as this might make it appear. As we have just seen, Freud's accounts of the metapsychology of structure, energistics, etc., are entirely at the level of function: even the structures of the topographical and structural model were characterized only functionally, not in terms of anything remotely resembling a brain structure. Thus, the system Cs or the system Ego is characterized by how psychological states behave in it, not by anything to do with the brain. However, as Solms and others have remarked, even Freud's early, explicitly neurological model characterizes the neural purely functionally (`there has to a neural system that does this, and one that does that', etc.). Thus, the switch from neuroscience to metapsychology was no more than a switch from a mechanistic functionalist account of the psychological functions of assemblies of neurons to a mechanistic functionalist account of psychological states. Since the latter are merely the former described psychologically, this was not a very large switch.

When thinking about Freud, most commentators focus exclusively on one side or the other. Hartmann and the early American ego psychologists, Kitcher 1992, and others focus exclusively on his mechanistic side and ignore his psychological side almost entirely. Rycroft 1966, Ricoeur 1970, 1981, Habermas 1971, Klein 1969, 1976 and, in this collection, Solms focus on his switch to the language of psychology so exclusively that they ignore the continued role of mechanistic explanation in this thought. The trick is first to give both their due, then understand the shape each took in Freud's thought, and then come to some understanding of how they co-existed. I turn now to the second of these tasks.

Neuroscience and psychology: two images of the person

If the switch from explicitly neurological speculations to the language of metapsychology was a relatively small one, the difference between either of these mechanistic explanations and the kinds of explanation generated when we use the language of psychology is anything but small. If we are to understand the form these two approaches to the mind took in Freud, we have to lay out the fundamentals of this difference.

The difference is no less than the difference between two entirely different images of the human person. Here the philosopher Wilfred Sellars can help us out. In a widely-cited 1963 paper, Sellars distinguishes two radically different images of the human person, the manifest image and the scientific image. The manifest image is the image of the human person that we find in ordinary moral, social, and interpersonal life. The scientific image is the image of the human person as a vast assemblage of some tinier unit: neurons, stimulus-response reflexes, or whatever. In this image, the person is seen as a system made up of a vast number of these postulated tiny units. The leading theory in this image nowadays is that persons are a vast assemblage of neurons tied together in complex biochemical and informational relationships. Not just Freud but all of us have and make use of both images constantly and they are both deeply embedded in our thought.

Freud's psychology and metapsychology are simply one expression of these two images. When Freud did psychology in the language of psychology, he starts from the manifest image of the person; his metapsychology, in his view, was one way of doing proper science of the person (the psyche). That is to say, it is one expression of what Sellars calls the scientific image.

The two images are quite different. The crucial difference is that whereas the manifest image is built out of objects and events that are manifest in everyday life, the scientific image has as its basic building blocks unobservable entities and processes postulated to explain features of the manifest (Sellars 1963, p. 19). The scientific image of the human person has had various building blocks at different times; some of the most prominent include: neurons and quantities of energy (Freud), conditioned stimulus-response arcs (behaviourism), symbolic structures in computational relationships (computational cognitive science), neurons in electro-chemical, informational relationships (recent neuroscience), and, in the last decade, weighted nodes (neural networks). An interesting historical example is the neural hydraulics of Descartes. All these versions of the scientific image have a similar general structure, one that makes them all quite different in much the same way from the manifest image. The idea behind them all is that the large basic unit of the manifest image can be understood as vast numbers of the smaller basic unit of the scientific image. Thus it does no harm to talk of `the' scientific image.

By contrast, in the manifest image, the basic unit of analysis is not some tiny postulated entity. The building block of the manifest image is nothing less than the whole human person, a being that can observe, make decisions, identify itself with things, enter into relationships with others, govern itself by standards, and so on. As Sellars tells us, it is the framework within which we encounter ourselves (p. 6). That is to say, it is the framework within which a person experiences, reflects on, relates to, and interacts with him- or herself and other people. What is built into our manifest image conception of the human subject is one of the great intellectual mysteries. Probably nothing has more stood in the way of achieving a rapprochement between manifest-image psychological theory and the scientific-image theories of the neurosciences.

These two images are images of the human person as a whole; they are meant to encompass every important aspect of the person.

The governing idea behind the scientific image is that by postulating entities simpler and smaller than the manifest object, entities in some complex pattern of unobservable relationships, we can explain what manifest objects are like. How? One of our most basic ways of explaining something is to find out how it is built. Figuring out how something is built tells us not only what the subsystems making the thing up are like but also, usually, how it works. If a clock is malfunctioning, for example, our first approach will be to find out how it is built, because this will usually tell us enough about how it works to be able to identify where the malfunction is occurring. Clearly both Freud's early neurological speculations and the metapsychology that he developed in 1900 and thereafter fit this picture. They were both designed to reveal the deeper hidden structure of the states studied by psychology.

To be sure, if Freud's psychology grew out of the manifest image, it also extended it. The manifest image is centred on consciousness. But Freud's great discovery was that much of psychic activity is not conscious. Here, in rough terms, is how I think the extension went. Freud began, as we all begin, with the conscious human person of the manifest image. He then extended and deepened this picture in a variety of clinically essential ways. The contents of the human subject were extended to include vast ranges of unconscious content. The processes of human subjectivity were extended to include the primary processes. The principles by which psychic content are managed were extended to include the defences. The sources of psychic content were extended to include infantile sexuality, the psychosexual stages, and infantile traumatic fantasies. And so on and so forth. (For a more extended discussion of this issue, see Hopkins 1986, Edelson 1988, and Brook 1992a.) But these additions are all still extensions of the manifest image.

His metapsychology was very different from this, however. Metapsychology does not extend the manifest image, it attempts to replace it, or at the very least to reduce the contents of the manifest image to it. Another, more recent expression of the same scientific-image urge in psychoanalysis is Peterfreund's (1971) information-processing model. By contrast, Freud's psychological theories attempt neither to replace nor to reduce the manifest image. Rather, they takes the manifest image as given and attempt to extend it.

Psychology and metapsychology

The difference between Freud's psychology and his metapsychology is profound. His metapsychology talks about psychic life in terms of patterns of energy flows, channelling, blockages, barriers, investments (cathexes), and discharges. The flows, etc., of energy vary in three dimensions, labelled by Freud the dynamic, the economic and the topographic (1915). The dynamic dimension concerns the relationships among different energy patterns and their effects on one another as each is channelled, blocked, transformed, invested, etc. The economic dimension concerns the strength of different energy patterns. It has been much criticized but it is hard to see how we could do without it. No understanding of wishes and desires could be complete, for example, without taking into account the fact that wishes and desires can have widely varying strengths. The third dimension refers to a central structure of the system within the energy processes take place, not the energy processes themselves. Freud called this structure topographic, because it consists of three vast subsystems of psychic states that can be viewed as stacked on one another. There is the system of unconscious states governed by primary processes of association, displacement, condensation, etc. Then there is the system of preconscious states, states that are introspectible at will but are not currently being introspected. And finally there is the system of conscious states, the states currently being introspected. Unlike unconscious states, the conscious and preconscious ones are mostly governed by secondary processes of rational and partly rational association.

Later, of course, Freud added a structure of id, ego and superego to the topographic structure just described (1923). Id consists entirely of unconscious states and the primary process, so it covers the same ground as the old topographical unconscious. Ego and super-ego, however, both have conscious and unconscious states and, in its primitive forms, the superego can evidence primary as well as secondary process patterns of association, so these two structures are quite different from the old preconscious and conscious layers of the topographical model.

Freud's psychological accounts could hardly be more different from his metapsychology. Now Freud expresses himself not in the language of energy flows and discharges but in the language of motive, beliefs, affects, fantasy--in short, in the representational language of goal-directedness, the language of what Freud's teacher Brentano called intentionality. Consider this bit of theorizing in the Rat Man case study:

When there is a mésalliance ... between an affect and its ideational content (in this instance, between the intensity of the [Rat Man's] self-reproach and the occasion for it) ... the analytic physician says 'The affect is justified. The sense of guilt is not in itself open to further criticism. But it belongs to some other content, which is unknown (unconscious), and which requires to be looked for' [1909, 175-6, emphasis in original].

'The affect is justified. The sense of guilt is not in itself open to further criticism.' That is to say, the patient has reasons for feeling guilty, it is just that they are not what he or she thinks they are. An explanation such as this could hardly be more different than an explanation in terms of flows and blockages of energy. This fundamental divide has continued in psychoanalysis from his day to ours; in the past twenty years, psychological theories in terms of reasons have become more and more dominant and mechanistic metapsychology has gradually been eclipsed.

The crucial difference between mechanistic metapsychology and psychology proper is that, whereas metapsychology explains people's thoughts, affects, fantasies, behaviour, etc., in terms of blind forces and barriers, psychology proper explains them by finding the patient's reasons for thinking or feeling or fantasizing or acting as he or she does. Often of course these reasons will be unconscious--the patient will not know what they are.

To find someone's reasons for something is to find out what the thing means to the person, what significance it has in their life, how they interpret it. Thus there is a central interpretive element to finding people's reasons for things that is altogether missing in metapsychological explanations in terms of blind forces.

Examples of psychological theories in Freud laid out in the language of people's reasons for things, the language of intentionality, include: the theory of defense and resistance (with its multiple subparts to do with repression, projection, introjection, isolation, disavowal, splitting, etc.); the theory of infantile sexuality with the subtheories of the psycho-sexual stages, penis envy, castration anxiety and the Oedipus complex; the theory of infantile determinants of neurosis; the theory of narcissism; the theory that dreams are meaningful, though their meaning is disguised; the theory of transference and counter-transference; and so on.

It is sometimes suggested that the difference between metapsychology and psychology in Freud is primarily a difference of level. Metapsychology is more abstract, because it is further removed from everyday observation and interaction, it is broader--it is a theory of all psychic life, not just this or that part of it, couched in terms of deep, postulated processes. Clearly this is correct as far as it goes; metapsychology is both more abstract and more all-encompassing than Freud's properly psychological theories were. But there is a good deal more to the difference between metapsychology and his psychology than that. They are radically different kinds of theory.

To begin to bring out the full dimensions of the difference, first note that the two have totally different roots. The roots of metapsychology are the language of science, and specifically the nineteenth century science of energies, masses, channels, barriers, etc. This language traces its history back to the Newton's vision of the universe as a massive system of matter in motion.

The roots of Freud's psychological theories, by contrast, are in the commonsense, everyday activity of making sense of ourselves and other people that has come to be called commonsense or, by some, folk psychology. There is no deep science buried in this alternative vision, just the accumulated wisdom of the centuries about why people do what they do. That this wisdom is not scientific does not mean it is unimportant, as autism amply demonstrates. There is a lot of evidence that the people with this illness cannot develop this commonsense understanding of people: in the jargon of the research, they do not have a theory of mind, either their own or other people's. The devastation this causes is ample proof of the importance of our commonsense psychology.

Freud's psychology as an extension of the manifest image

The evidence that Freud's properly psychological theories are extensions of the commonsense wisdom of the manifest image comes in three forms (I explore the general question of the relationship of psychoanalysis to commonsense psychology in Brook 1992a). First, many of Freud's psychoanalytic concepts are extensions of commonsense psychological concepts. Here are some examples.

The concept of repression is an extension of the precept that people tend to forget or distort their recollections of troubling experiences (Edelson 1988, p. 337). The theory of splitting representations is an extension of the precept that people tend not to integrate feelings, motives, fantasies, etc., when they conflict or are inconsistent. The theory of the splitting of the ego, that people will isolate inconsistent attitudes to something (Brook 1992b), is an extension of the commonsense notion that we can have inconsistent attitudes to things without being aware of it. The theory of resistance extends the commonsense idea that we will try to avoid thinking about things that trouble us. And so on.

Even the concept of transference is an extension of the commonsense precept that people tend to repeat ways of reacting to people learned in childhood. There is much more to the doctrine of transference than that, of course. It also reflects the fact that causes from the distant past can continue to operate in the present in beings like us, operate in two ways, in fact, via memories and via dispositions (Wollheim 1984). And it contains the idea that neurotic pathogens appear in the transference, though this too may be no more than an extension of a commonsense idea. But even the idea of transference is an extension of commonsense psychology.

As I said, Freud went beyond commonsense psychology in many directions. His theory of defence and theory of transference are both illustrations of this; others can be found in the methods of psychoanalysis, especially the method of free association, of evenly-hovering attention, and the method of empathy. Other important contributions include the theory of drives as the biological basis of intentionality, the theory of the developmental stages, the theory of interpersonal intentionality (transference and counter-transference), the theory of the real content of dreams and how to uncover it, and so on. A complete list of psychoanalysis's extensions of commonsense psychology would be very long.

Freud's work also contains much that commonsense psychology would never have dreamt of. The psychoanalytic theories of instincts, development, dreams, defence, interpersonal relations (transference and counter-transference), and psycho-pathology all go far beyond anything in commonsense psychology. The same is true of that most basic of psychoanalytic concepts, the concept of the unconscious, the concept of motives, beliefs, fantasies, etc., which cannot be introspected and are governed by a strange primary process. When we add ideas contained in the newer theories, such as the concepts of projective identification, unconscious fantasies of infantile helplessness, murderous aggression and uninhibited lust, fantasies populated by infantile 'internal objects', various part-objects, one-sided pre-ambivalent affects, etc., in object-relations theory, and the concepts of selfobject, narcissistic deficit, the developmental line of the self, empathy connection, selfobject transferences, transmuting internalizations, etc., of self psychology, we end up with something very remote from commonsense psychology.

In addition, Freud proposed causal connections quite unlike anything contained in commonsense psychology, notably those between infantile sexuality and narcissistic development on the one hand and the vicissitudes of adult life on the other. Indeed, some of his ideas flatly contradict beliefs of commonsense psychology, for example the belief that conscious ideas largely control one's life. He also expanded the range of phenomena that can be interpreted in terms of reasons vastly, and gave us a whole range of special techniques to assist us in doing so. Clearly, psychoanalysis contains much more than commonsense psychology does. Nevertheless, most of Freud's properly psychological theories have their roots in the commonsense intentional psychology of the manifest image, the psychology of reasons and meanings.

Extensions of commonsense vocabulary

Two further kinds of evidence that Freud's psychology proper is rooted in commonsense psychology can be found in the language he uses. First, like commonsense psychology, the language of Freud's psychological theories is intentional language, as is most recent psychoanalysis as a whole (Edelson, 1988). As we will see, that is not true of the language of metapsychology, with one notable exception. Second, when Freud introduces new psychological terms, he introduces them as extensions of terms of commonsense psychology--sometimes quite explicitly.

The nature of intentional language

As I said, the states and events treated in Freud's properly psychological theories have intentionality, while metapsychology treats states and processes independently of whether they have intentionality. Indeed, the language of the psychological theories is very different from the vocabulary of metapsychology and the natural sciences generally. First, it is a language for representational stages: perception, belief, desire, affect, and psychic reality (dreaming and imagination). Second, and this follows, the states it treats are viewed as being about something, having an object. A belief is about whatever is believed, a desire is about whatever is desired, a fantasy is about whatever is imagined in it, an affect is about whatever it is directed at, and so on. This property of being about something is what is meant when a state or process is said to have intentionality ( Brentano 1874, Searle 1983, Dennett 1987b).

The kinship of the term `intentionality' to the English word `intention' is largely accidental, though like almost all psychological states and activities, intentions also have intentionality. `Intentionality' is a technical term invented in the middle ages for the relation of something being about something else. There is also a technical term for the object of this relationship, namely, intentional object. It is important to note that intentional objects need not correspond to anything that exists; this feature of intentionality is important because it is what makes fantasies, mistaken beliefs, dreams, wishes, etc., possible. They all have intentional objects that do not correspond to any real object, event or state of affairs. Events and states of affairs, including psychological events and states, can be intentional objects just as well as objects strictly so called. Some theorists also call the intentional object of a psychological state its content.

It will be obvious that virtually all the states of interest to psychoanalytic interpretation and explanation have intentionality as just defined. Moreover, whatever Freud's hopes for metapsychology or a neuroscientific foundation for psychoanalysis might once have been, the clinical vocabulary of day-to-day psychoanalytic practise is now entirely an intentional vocabulary. But it largely was in his own work, too. With the exception of metapsychology, the language of psychoanalysis is the language of beliefs, desires, emotions, attitudes, motives, fantasies--representations in psychic reality that have intentionality.

Intentionality had--or should have had--little or no place in metapsychology. Metapsychology attempts to explain human thought, affect and behaviour in terms of movements, blockages, transformations and discharges of energy. But energy as such does not have intentionality. When the oxidation of gasoline in an internal combustion motor produces intense heat and pressure, for example, the heat and pressure are not about anything, though of course they have various effects. The same should be true, then, of quantities and vicissitudes of psychic energy in the psychic apparatus. Moreover, Freud would have failed in his project of `reducing' psychology to (capturing psychology in) a mechanistic account, the kind of account required by the scientific image, if he had let intentionality creep into metapsychology.

That Freud used intentional language needs little argument; barring the Project, almost all his works are written in the distinctive vocabulary of intentional psychology. That is no criticism of him. He knew that he could no more complete the reduction of psychology to metapsychology than to the earlier neuroscience of the Project, so he had no choice but to use the language of commonsense psychology. However, in the transformation of the neuroscience of the Project into metapsychology, he let intentionality creep into metapsychology itself. This is more problematic. The first mature statement of his drive theory, the famous schema of source, object and aim, occurred as early as (1905). We can already find elements of intentionality here. How could blind flows of energy have an object or aim?

They can certainly have a source. But how could anything purely mechanistic have an object or an aim? Consider the energy released by burning gasoline in an engine. It does not have either object or aim. Nor does any other release, flow or transformation of energy. Something more than simple mechanical energy has crept in: the kind of directedness to an object that is distinctive of psychological states. There may be some way to understand this something extra as a feature of energy flows of a special kind. Indeed, if psychological states are neural states and neural states function by flows and transformations of energy, there would have to be such a way to understand them. But energy flows as such do not need to have intentionality. Whatever he may have thought, Freud did not keep his metapsychology as a pure exemplar of a natural scientific framework. Intentionality crept into it.

The distinction between states and processes that have and do not have intentionality further clarifies the difference between metapsychological and psychological theorizing in Freud. If Freud had kept his metapsychology pure, metapsychology would theorize over neurons and energy flows while his psychology proper would theorize over intentional, goal-directed agents who have conscious and unconscious reasons for the various things they think and feel and do. In metapsychology the basis of human functioning is mechanisms for generating, channelling, blocking, reshaping, investing, and ultimately discharging somatic energy: energy having its source in the drives, governed initially by primary process, later by the secondary processes of the ego via mechanisms of defense and other mechanisms, and finally discharged in behaviour. In the psychology proper, goal-directed states that give the subject reasons for thinking, feeling and acting as he or she does are the basic principle of human functioning.

Thus we see two things. First, there was a deep-running duality in Freud's thought, one marked by the split between metapsychology and proper, i.e., intentional, psychology. Second, the latter is an extension of commonsense psychology.

Freud's new concepts as extensions of commonsense concepts

The third reason I mentioned for viewing Freud's psychological theorizing as an extension of commonsense psychology is that the new, properly psychoanalytic concepts that Freud introduced are an extension of concepts of commonsense psychology. Consider Freud's favourite argument for the existence of the unconscious, an argument he used repeatedly and over a period stretching from 1909 (pp. 175-6), through 1915 (pp. 166ff.) and 1923 (pp. 14-18), to 1938 (pp. 196-7). In it, Freud quite explicitly models this most central notion of psychoanalysis on consciousness. In outline, the argument goes as follows. Our inferences as to the psychic states that caused a person to behave in a certain way usually correspond, Freud begins, to what the person him- or herself is conscious of and believes to be the cause. But sometimes a person either grossly misidentifies what seems to us to be the clear cause of some behaviour, or he or she is just not aware of any cause at all. Either way, there is a gap. (We will return to the nature of this gap in the next section.) At this point, Freud adds something new. In these cases too, he tells us, the cause of the behaviour is something psychological. This thought almost instantly generates the concept of the unconscious: since there is nothing conscious to be the cause, Freud infers, the cause must be some psychological state that is unconscious. As he put it,

[w]e have found--that is, we have been obliged to assume--that very powerful mental processes or ideas exist ... which can produce all the effects in mental life that ordinary ideas do (including effects that can in their turn become conscious as ideas), though they themselves do not become conscious [1923, 14].

This momentous conclusion, of course, changed forever both psychology and the way we view ourselves. Momentous as it was, note that Freud's new concept of the unconscious is entirely modelled on the standard concept of consciousness. The latter is a concept, in fact the central concept, of commonsense psychology. The idea of the unconscious does not replace commonsense psychology, it extends it. If even the most distinctive of psychoanalytic concepts, the concept of the unconscious, is such a close kin of a concept of commonsense psychology, surely the same will be true and even more true of many of the others.

To sum up the story so far: like commonsense psychology, the language of Freud's psychological theories is a language for states and processes that have intentionality, while the language of metapsychology treats the states and processes of its domain as not having intentionality. A belief is about whatever is believed, a desire is about whatever is desired, a fantasy is about whatever is imagined in it, an affect is about whatever it is directed at, and so on. Thus the intentional language of Freud's properly psychological theories is very different from the natural science language of his metapsychology. Freud made extensive use of both kinds of theorizing.

Explanation in Freud's psychology and metapsychology

If the language of Freud's psychology proper is very different from the language of his metapsychology, the same is true of the styles of explanation used in them.

The explanations of metapsychology are mechanistic; they explain psychic states and human behaviour in terms of causal forces such as quantities, channels, investments (`cathexes'), blockages, conversions and discharges of somatic energy. In this form of explanation, things are without direction or goal; even the tendency to discharge in Freud's drive discharge model is initially just a tendency, not something directed at achieving a goal. By contrast, in psychology proper, we explain things in term of people's reasons for them; explanations here are intentional.

The idea at the heart of intentional explanation is the familiar one that things have meaning for people. As a result, people can think and feel and do things for reasons, not just as the result of mechanistic causes. Now reasons themselves may very well be causes of a kind (see Davidson 1963, Brook 1995), but if they are, their causal action is very different on the face of it from the causal action of forces, energy flows, channels, blockages and the like.

To illustrate how different intentional explanations in terms of reasons are from mechanistic explanations, it might be helpful to have an example. Here is a vignette, a very typical vignette I think, from the analysis of a woman I will call Ms. B.

When Ms. B. was seven years old, she witnessed her mother having an almost fatal miscarriage. The miscarriage was quite possibly self-induced and the baby was probably not her father's. Ms. B. witnessed pretty well the whole scene, and it was extremely bloody. Mother was then hospitalized for a month. Ms. B.'s mother was difficult and Ms. B. enjoyed having her father to herself for that time. As soon as her mother returned home, she cut off Ms. B.'s beautiful long hair. I suggested to Ms. B. that this bloody scene and its aftermath of sinful pleasure followed by humiliating punishment became her model of all experiences of sensual pleasure.

As I said, the explanation offered in my interpretation seems typical of the kind of explanation we use in psychoanalysis. The aim of the interpretation was to explain Ms. B.'s experience of sexual pleasure by finding the reasons for it--in this case, reasons derived from the experiences and fantasies of her childhood. Note that these reasons may also very well be the causes of her adult experience of sexuality; in fact, I would argue that they were causes. If they were causes, they are a quite different kind of cause from mechanistic ones. To see this more clearly, think of trying to explain Ms. B.'s way of experiencing sexual pleasure in adult life in terms of blockages, diversions and conversions of energy, and compare such an explanation to the one I actually gave to her in the interpretation just reported.

Explanations in terms of people's unconscious reasons for something abound in Freud's work. Think, for example, of his explanation of the development of the super-ego, that the child adopts (internalizes) its parents' moral code. Freud's theory of what causes this internalization, that the child develops the superego to cope with its various Oedipal wishes and the attendant prohibitions and anxieties, assumes that the child has reasons for developing a super-ego and that these reasons are what leads it to do so. Likewise, when Freud spoke of gaps at the level of consciousness, to return to that topic, what he meant was that if we restrict ourselves to reasons of which a person is conscious, there will be gaps in the person's reasons--there will be much in his or her thinking, feeling and doing for which there is no conscious reason. In short, intentional explanation played a huge rule in Freud's thinking.

The duality in Freud between the kind of explanation typically contained in interpretations, explanations aimed at uncovering the often unconscious `meanings' and reasons behind patients' perceptions, beliefs, feelings, wishes, fantasies, dreaming, and action, and explanations in terms of a mechanics of somatic energy, its generation, management, and discharge is clearly very important. And it is equally clear that he continued to make use of both kinds of explanation throughout his career as the first theorist of psychoanalysis.

This duality in kinds of explanation came to be expressed in European thought in a division of even wider application, the division between Verstehen knowledge of the human sciences, the kind of knowledge gained by understanding or interpretation, and Erklären knowledge in the natural sciences, or causal explanation. The idea that there is a distinctive kind of knowledge appropriately called Verstehen knowledge was first articulated by Dilthey in 1911 and developed by Weber and Jaspers and others in the early twenties. Their notion of Verstehen knowledge--interpretation, understanding--is the original source of the idea that theorizing about human thought, feeling and action proceeds by finding the meanings behind them, people's reasons for them. Dilthey and the others contrasted this Verstehen style of theorizing with Erklären (explanation) or kausal Erklären (causal explanation) theorizing, whose aim is to find generalizable theories about patterns of causal correlation.(1) The entities and relationships postulated by metapsychology and the ideas of flows and transformations of energy in neurons, neuro-transmitters, etc., are clearly examples of the latter. Verstehen theorizing, by contrast, is concerned with function or role, and with intentional states such as meanings, purposes, projects, and ends--in short, with teleology. More recently, the notion of Verstehen knowledge has also come to be associated with the method of intuition or empathy, as advocated for example by Kohut, but the original idea was simply that we get a distinctive kind of knowledge by interpreting and understanding what things mean to people, people's reasons, a kind of knowledge different from the kind we get by uncovering underlying causes. The view of psychoanalysis as an activity of finding meanings clearly has close affinities with and probably grew out of this work on Verstehen knowledge.

A great deal more could be said about the kind of psychological theory we have been discussing, where theories are build out of intentional language and grow out of and extend the commonsense psychology of the manifest image. I hope I have said enough to make two points: (1) it is very different from the scientific image language and theorizing that characterized Freud's early neuroscientific theorizing as found in the Project and other works of the time; and (2) if Freud never gave up metapsychology, he also made extensive use of the kind of purely psychological theorizing found in commonsense psychology.


If both kinds of theory are to be found in Freud's work throughout his career, both metapsychology and psychology proper, how did they co-exist? This is a big question; it would require a paper of its own. Here just let me say this. At one level, the two kinds of psychological theorizing co-existed quite happily--because Freud did not see any tension between them. Once he gave up the specifically neuroscientific thinking of the Project and replaced it with the more general model of metapsychology, a model that made not even functional claims about how things are at the neural level, he seems to have lost any sense that there might be a problem reconciling the two. He could switch from an account in terms of agents' reasons to an account in terms of generation, distribution, transformation, and discharge of psychic energy on a single page and not be troubled. We find him doing just this in Beyond the Pleasure Principle (1920) and in other works. The ontological commitments of the two kinds of theory are very different, so he probably should not have been so sanguine about their mutual compatibility, but it seems that he was. Since to this day we do not know how to reconcile the two, we should not hold this to his discredit.

This divide between mechanistic models out of the scientific image and intentional explanations out of the manifest image continues to this day, and not by any means only in psychoanalysis. In psychoanalysis it has come in the past three decades to take the form of the division in psychoanalytic theorizing between clinical theory and metapsychology (Rubinstein 1975, Gill and Holzman 1976, Brook 1995). In this battle, clinical theory has won, hands down. Indeed, metapsychology is rapidly ceasing to pay any important role in psychoanalysis at all. But the duality also appears in cognitive science more generally.

In cognitive science, the division between mechanistic, scientific-image approaches to the mind and intentional, manifest-image approaches is at work very clearly in the debate over the future of commonsense psychology as science. On the one hand, traditional functionalists such as Jerry Fodor (1987) and others of similar ilk argue that `folk psychology is here to stay', Daniel Dennett (1987a) for example. For these people, intentional psychology will never be replaced by explanations from the neurosciences. Such a replacement is about as likely as replacing explanations of colour with explanations of shape. The two kinds of explanation do different kinds of job and neither will ever replace the other.

On the other hand are the so-called eliminative materialists. These theorists, headed by Paul and Patricia Churchland (1984, 1986), argue that intentional psychology deriving from commonsense psychology has never got anywhere as science, indeed has made little progress since the time of Sophocles 2,500 years ago, and is unlikely to do much better in the future. What we have to do is to return to Freud's dream of the 1890's, the dream of a complete neuroscientific understanding of the mind in terms of patterns of neural activity, and replace psychology proper with the results of this new science.

It will be some time before we know who is right. In the meantime, it is a fact of no little interest that Freud worked both sides of the explanatory fence throughout his career and had considerable sympathy, at different times, with each side of the debate over the future prospects of commonsense intentional theorizing in psychology.


What can we conclude from all this? Two things, I hope. First, Freud did not give up the neuroscientific model developed in the Project; rather, he transformed it, with some modifications, into metapsychology, when then became a feature of his work for the rest of his life. Second, if first neuroscience, then metapsychology embodied his hopes for psychoanalysis as a science, such scientific-image thinking was always accompanied by and co-existed in some tension with a more properly psychological style of theorizing using intentional language and beginning from the concepts and wisdom of commonsense psychology.



Bonaparte, M. et al. 1954. The Origins of Psychoanalysis: The Freud/Fliess Correspondence. New York: Basic Books

Brentano, F. 1874. Psychology from an Empirical Point of View, trans. D. B. Terrell. London: Routledge and Kegan Paul, 1971.

Brook, A. 1992a. Psychoanalysis and commonsense psychology. Annual of Psychoanalysis. Hillsdale, NJ: The Analytic Press.

--------. 1992b. Freud and Splitting. International Review of Psychoanalysis, Vol. 19:3, pp. 335-50.

--------. 1995. Explanation in the hermeneutic science. International Journal of Psychoanalysis, Vol. 76:3, pp. 519-32.

Churchland, P. M. 1984. Matter and Consciousness. Cambridge, MA: MIT Press/Bradford Books.

Churchland, P. S. 1986. Neurophilosophy. Cambridge, MA: MIT Press/Bradford Books.

Clark, S. and Wright, C., eds. 1986. Psychoanalysis, Mind and Science. Oxford: Oxford University Press.

Davidson, D. 1963. Actions, reasons and causes. In: D. Davidson 1980.

-----------. 1980. Essays on Actions and Events. London: Oxford University Press.

Dennett, D. 1987a. The Intentional Stance. Cambridge, MA: MIT Press/Bradford Books.

----------. 1987b. Intentionality. In: R. Gregory, ed. The Oxford Companion to the Mind. Oxford: Oxford University Press.

Edelson, M. 1988. Psychoanalysis: A Theory in Crisis. Chicago: University of Chicago Press.

Fodor, J. 1987. Psychosemantics. Cambridge, MA: MIT Press/Bradford Books.

Freud, S. 1891. On aphasia

--------. 1895a. Project for a scientific psychology. S.E. 1.

--------. 1895b. Studies in hysteria. S.E. 2.

--------. 1900. The interpretation of dreams, S.E. 4 and 5.

--------. 1905. Three essays on sexuality. S.E. 7.

--------. 1909. Notes upon a case of obsessional neurosis. S. E. 10.

--------. 1915. The unconscious. S.E. 14.

--------. 1920. Beyond the pleasure principle. S.E. 18.

--------. 1923. The ego and the id. S.E. 19.

--------. 1933. New introductory lectures. S.E. 22.

--------. 1938. The outline of psychoanalysis. S.E. 23.

Gill, M. and Holzman, P. 1976. Psychology versus Metapsychology. New York: International University Press.

Habermas, J. 1971. Knowledge and Human Interests, trans. J. Shapiro. Boston: Beacon Press.

Hopkins, J. 1986. Epistemology and depth psychology: critical notes on The Foundations of Psychoanalysis. In: Clark and Wright 1986.

Kitcher, P. 1992. Freud's Dream. Cambridge, MA: MIT Press/Bradford Books.

Klein, G. S. 1969. Freud's two theories of sexuality. In: Gill and Holzman 1976.

--------------. 1976. Psychoanalytic Theory. New York: International University Press.

Peterfreund, E. 1971. Information, Systems and Psychoanalysis. New York: International University Press.

Pribram, K. and Gill, M. 1976. Freud's Project Reassessed. New York: Basic Books

Ricoeur, P. 1970. Freud and Philosophy. New Haven, CN: Yale University Press.

----------. 1981. Hermeneutics and the Human Sciences. New York: Cambridge University Press.

Rubinstein, B. 1975. On the clinical psychoanalytic theory and its role in the inference and confirmation of particular clinical hypotheses. Psychoanalysis and Contemporary Science. 4, pp. 3-57.

Rycroft. C. 1966. Psychoanalysis Observed. London: Constable.

Searle, J. 1983. Intentionality. Cambridge: Cambridge University Press.

Sellars, W. 1963. Philosophy and the scientific image of man. In: Science, Perception and Reality, pp. 1-40. London: Routledge and Kegan Paul.

Wollheim, R. 1984. The Thread of Life. Harvard University Press.

1. A terminological awkwardness arises at this point. I call the activity of accounting for thought, feeling, and action in terms of reasons intentional explanation; Dilthey and others in his tradition say that this activity is not explanation, that what it achieves is understanding, not explanation. The resolution of the difficulty is to note that what they mean by explanation is mechanistic explanation; their understanding and what I am calling intentional explanation are much the same thing. I prefer the term `explanation' to `understanding' because when we find someone's reasons for something, we seem to be explaining it in a perfect good sense of the term `explanation'.