PSYCHOANALYSIS AND COMMONSENSE PSYCHOLOGY
J. A. Brook
Abstract
The purpose of this paper is, first, to explore the relationships between psychoanalysis and
commonsense psychology, and second, to examine some of the implications of these
relationships for the nature of psychoanalytic reasoning. We find that psychoanalysis depends on
commonsense psychology and is an extension of it in at least three ways: in vocabulary, in some
doctrines, and in the type of psychological explanation it uses. We conclude with an examination
of Grünbaum's critique of psychoanalytic case studies. Answers to some of his charges can be
found in these dependencies and extensions.
In addition to its unique, highly specialized theories, clinical psychoanalysis depends on and makes use of a lot of non-psychoanalytic knowledge. Knowledge of language is the most obvious example, but one could also cite knowledge of the physical world, of human biology, of the use and effect of body language and tone of voice, and so on and so forth. So far, few would disagree. It would be more controversial to suggest that psychoanalysis also depends on non-psychoanalytic psychological knowledge. And it would be even more controversial to urge that, as well as depending on non-psychoanalytic psychological knowledge, psychoanalysis is also an extension of it. Let us call the first claim the dependency hypothesis, the second the extension hypothesis. Both have had their advocates, including Edelson (1988) in the United States, and Hopkins (1986) in England. The non-psychoanalytic psychology in question is the commonsense psychological knowledge of everyday life, the knowledge that allows us to interpret others' noises as expressions of ideas, attitudes, intentions, etc., others' facial gestures and body language as expressions of affects and desires, etc., and so on. In the last decade, this commonsense psychology has sometimes been referred to as folk psychology (Stich, 1983, Dennett, 1987a, Churchland, 1984), but this term is radically misleading. We are not talking about something on a par with folk art; we are talking about the stuff of great novels, the raw material of Shakespeare. Even the term commonsense psychology is too weak for this store of interpretive skill and wisdom, but it is the term I will use. The first two-thirds of this paper will be taken up with exploring the two just-mentioned hypotheses about its relation to psychoanalysis.
By commonsense psychology, I do not have in mind primarily the conscious beliefs of everyday psychological life. Psychoanalysis has shown many of these to be false. Indeed, at this level, if commonsense and psychoanalysis coincide, it is probably by accident, as a colleague once said to me. Rather, I mean the interpretive, sense-making apparatus that commonsense psychology provides -- the vast resource of largely unconscious principles, precepts, rules and skills for recognizing and inferring psychological states from behaviour and other psychological states, the body of knowledge every competent language-user has available to make sense of his- or herself and others. This resource is what allows us to infer that someone is angry when we observe him or her shouting and red in the face, that someone wants to be comforted when he or she cries or moans, that someone is experiencing pleasure when he or she laughs, becomes animated, etc., and so on. We have at least many thousands of such recognitional and inferential tools at our disposal. Given how adept even very young babies are at recognizing forms as human and at inferring humans' emotional states, probably at least part of this resource is innate.
To illustrate both the kind of thing on which psychoanalysis depends and how this material enters clinical work, let me give a couple of brief clinical examples. The first is from the analysis of Mr. S., a 26 year-old graduate student. Near the end of a session during which he had anguished over his socially isolated childhood, I made the following interpretation:
AB: "Its a long way from that fringe kid hanging out with socially inadequate people to the young man with an honours degree in graduate school."
S: "I can't fit my image of myself and my life now together."
That is to say, Mr. S. took my interpretation to be about him feeling marginalized now. His reaction told me a number of things: how he sees himself now, what the transference was like at that moment, and so on. But that is not what interests me here. What interests me is the fact that I was able to make these inferences. Though I made them instantly and effortlessly, I did not use anything specifically psychoanalytic to do so. To spot the misunderstanding, I used semantic and pragmatic knowledge ('pragmatic' in the linguistic sense). To spot the implications for how he sees himself, I used precepts for inferring representations of self from utterances and their contexts. And so on. I did not use anything specifically psychoanalytic.
Let us now consider a second example, from the analysis of Ms. A., a 38-year-old government manager. Early in her analysis, the way leisure time and even the simplest pleasures caused her to feel uneasy and distracted from something more important became a point of urgency. Knowing something by then of her deprived, traumatized and intensely religious childhood, I once said, "One reason ordinary happiness leaves you feeling uneasy may be that it strikes you as shallow, a betrayal of the high mission of your childhood." This example has clear connections to a specifically psychoanalytic theory, of course, the theory of the super-ego. But how was I able to make the inferential jump from pleasure leaving her feeling uneasy and distracted, via information I had about her childhood (as she experienced it, of course), to the conclusion that she was feeling an unconscious sense of betrayal? Not, I think, by applying distinctively psychoanalytic knowledge.
Examples like these illustrate how clinical practice depends on commonsense psychology. They could be multiplied endlessly; indeed, new ones could be found in virtually every session. How do we spot that a patient raised by a depressed mother feels that he will never have the strength he needs until someone gives him what he feels he did not get from his mother? How do we spot in another patient that she always experiences a disappointment as proof of her inferiority? Or, to take a more 'classical' example, how do we suddenly infer that a third patient is immobilized by feelings of fear, anger and concern for a damaged mother that are of roughly equal strength and pulling in different directions? The fact that some of these examples are narcissistic, some to do with classical conflict is not significant in the present context. Indeed, that these conclusions make contact with psychoanalysis at all is not what interests me. What interests me is the inferential processes by which we reach them. So far as I can see, they are not based on much that is specifically psychoanalytic.
In addition to its intrinsic interest, the kind of relationship between psychoanalysis and commonsense psychology just introduced may be important for another reason. As is well-known, the last two decades have seen some far-reaching critiques of the scientific credentials of psychoanalysis, both by the hermeneutic reinterpretation of it as a kind of narrative art, and by philosophers of science, Adolf Grünbaum (1984, 1988, 1990) in particular. The relation between psychoanalysis and commonsense psychology may allow us to answer these charges from an unexpected direction. Commonsense psychology consists largely of factual claims and inference-rules about human psychological states, claims and rules that allow us to make and test predictions about future human behaviour, utterances, etc. Thus, even if these claims generate narratives, as the hermeneutical critics claim, they are also something more than narratives. They are factual claims about a particular kind of thing in the world, namely human beings. By factual claims, I do not mean the old empiricist idea of perceptions unsullied by any conceptualization. It is now generally recognized that every factual claim is saturated with theory and concepts. I mean claims that can be supported or disconfirmed by observation, even though these observations will also be theory- and concept-saturated. It is factual claims in this sense that I contrast with the view advanced in some quarters that psychoanalysis is in the business of constructing internally coherent narratives, not making and testing factual claims about events and causes. This view of the relation of the factual underpinnings of psychoanalysis to its concepts and theory has been receiving increasing attention (Goldberg (1988, esp. 1 and 2) gives a good account of it). However, I want to explore a different idea.
One of the most important features of commonsense psychology is that we know it works, in everyday contexts at least. Our great success in interpreting one another's noises as expressions of intentions, affects, wishes, etc., is proof of that. (Here I am talking about the interpretive apparatus of commonsense psychology, not its explicit beliefs.) Moreover, commonsense psychology has been massively tested over the course of human history, so we can consider it exceedingly well-confirmed. That raises an interesting possibility. Could we use the relationship of psychoanalysis to this body of well-tested commonsense knowledge to answer the sceptics? Put carefully enough, I think we can. I will take up this idea in the last section of the paper. Interestingly, Adolf Grünbaum flatly rejects this approach. "I no more think that psychoanalytic theory is an extension of commonsense psychology than I think theoretical physics is an extension of commonsense 'physics'" (quoted in Edelson 1988, 330); this stance may be connected to his scepticism.
The idea that psychoanalysis is dependent on or, even worse, an extension of commonsense psychology can threaten our sense as analysts that we see further, indeed, may even seem bizarre. Clearly psychoanalysis contradicts a lot in commonsense psychology and contains much that the latter could never have dreamt of. In general, I see the relationship between the two of them this way. The psychoanalytic theories of instincts, development, dreams, psychic structure (topographical and structural), defence, interpersonal relations (transference, counter-transference, projective identification) 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 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 to these claims that run directly counter to commonsense psychology, psychoanalysis has also proposed causal connections quite unlike anything it contains, notably those between infantile sexuality and narcissistic development and the vicissitudes of adult life, and it has rejected some of commonsense psychology's dearest beliefs, for example the belief that conscious ideas largely control one's life. It also has a method, the method of free association, that at the very least vastly expands the range of phenomena that can be interpreted in terms of reasons, and it has a whole range of special techniques to assist it in doing so. Clearly, psychoanalysis contains much that commonsense psychology has never dreamt of, indeed that is flatly inconsistent with it. That is why Grünbaum rejects the idea that psychoanalysis could possibly be an extension of commonsense psychology. Nevertheless, the two have a close relationship. (A further question, which I cannot go into, is how far the various theories of psychoanalysis are consistent or even comparable within themselves; on this, cf. Goldberg (1988, Ch. 3).)
Like all psychological theories, psychoanalysis began with and grew out of commonsense psychology. Some psychologies grew by way of rejection. Freud tried to do this in the 1890's with his mechanistic metapsychology (a term he actually coined only later) and behaviourism tried the same thing in a different way much later. What the two theories have in common is the attempt to replace the vocabulary and explanatory style of intentionality with something else. Despite officially beginning in mechanism, psychoanalysis did not follow this course; it grew by way of development, not rejection. For the most part, it has not replaced commonsense psychology but extended it. The same is true of practically all psychological theories: all the various 'dynamic' psychologies, most clinical psychologies and most of contemporary experimental cognitive psychology are all extensions of commonsense psychology. The failure of Grünbaum and others to see that psychoanalysis has roots in commonsense psychology is a serious one.
It is important to understand that psychoanalysis is in no way diminished by being built on a
foundation of commonsense psychology. What it has taken over is extremely basic and in no
way detracts from what it has gone on to add. In any case, it had no alternative. Virtually all the
inferences made by anyone from any point of view about people's psychological states and
behaviour must rely in some way on the resources of commonsense psychology. We could no
more give up our psychological vocabulary and interpretive resources than we could give up our
ability to speak. And we have no more reason to want to do so. Moreover, as Morton (1982)
points out, the traffic has been anything but one-way. If psychoanalysis grew out of
commonsense psychology, it has also had a deep, structural effect on the latter. For example,
morally-loaded character terms ('rigid', 'self-absorbed') have largely disappeared, in favour of
more neutral terms of personality ('obsessive', 'narcissistic'). Let us now turn to the first of the two
hypotheses about the relationship between the two introduced earlier, the dependency hypothesis.
The Dependency Hypothesis
The resources of commonsense psychology, and similar ones contained in our commonsense physics, biology, etc., have a special relationship to psychoanalysis. Though not themselves psychoanalytic, psychoanalytic thinking cannot take place without them. Edelson (1984, 83; 1988, 245, 254-5) calls them non-theoretic with respect to psychoanalysis. The non-theoretic for a given discipline is everything (1) that the discipline takes for granted in order to proceed, and (2) that we do not need the discipline to understand. Thus, for example, physics is non-theoretic in biology. In just the same way, much of commonsense psychology and virtually the whole of its mundane apparatus of psychological inference is non-theoretic for psychoanalysis. Psychoanalytic reasoning takes it for granted. Just as biology relies on chemistry and astronomy relies on physics, so psychoanalysis relies on commonsense psychology. As we will see, it also relies on commonsense syntax, semantics, physics, biology and perhaps even sociology.
In some measure, Freud himself was aware of this dependency. As he said repeatedly (though perhaps ambivalently), psychoanalysis is not a general psychology (preface of 1914 to 1905a, 130; 1912, 99 n.2; 1914, 50 ("psychoanalysis has never claimed to provide a complete theory of human mentality in general"); 1924, 209; 1937, 226 n.2). It is a psychology of a special class of phenomena (Edelson discusses the grander ambitions of Hartmann and others in 1986a and 1988, Ch. 7). If so, in addition to distinctively psychoanalytic ideas, psychoanalysis has no choice but to use a lot of non-psychoanalytic psychology, too. As Freud knew, psychoanalysts have to rely on their store of 'latent memories', as he sometimes called them, of people and their feelings and behaviour as much as anyone else does.
A psychoanalyst uses the resources of commonsense psychology in many places. One crucial one is in establishing and maintaining the intersubjective, communicative field between analyst and analysand. When we interpret another's words and understand the simple, immediate intentions they are trying to communicate, we do not use anything specifically psychoanalytic to do so. Yet if we could not rely on this communicative field being there and quietly working away, we could not bring the special knowledge of psychoanalysis to bear at all. Consider cases where this field is not secure, where the analyst is not sure an analysand understands him or her and communication is uncertain and insecure. In these cases the analytic experience is totally different from what it is in cases where analyst and analysand are both secure in the conviction that they understand each other's other immediate, conscious communications. What makes this understanding possible is mutual interpretation of the words each utters. This work of interpretation could not be using anything special to psychoanalysis because, for one thing, the analysand has no such knowledge. What makes it work is the commonsense psychological knowledge analyst and analysand share.
Indeed, a good deal of other commonsense knowledge is involved, too. Before we can interpret one another's communicative intentions, we must interpret the noises being made. This in itself is a multi-stage process of extraordinary complexity. First we must interpret the noises as phonemes and words. Next we have to interpret the syntactic structure of the words. Then we have to find a semantic interpretation, an interpretation of the linguistic meaning, which fits the syntactic structure (and everything else we know that is relevant to the situation). Only then are we in a position where we can begin to interpret the communicative intentions behind the linguistic meaning. Of these various stages, the only one that has been at all well-studied is the syntactic one, by Chomsky and his followers. The following is one of Chomsky's standard examples of syntactic interpretation. Compare the sentence,
John is easy to please.
to the sentence,
John is eager to please.
Physically, these sentences could hardly be more similar. The only difference is that for the letters 'sy' in the first, we find 'ger' in the second. Yet their syntactic structure is totally different, and this is 'intuitively' obvious to anyone who knows English. Indeed, most three-year-old children would respond to these sentences in a way that recognizes the difference. To the first, the response might be 'So I don't have to be so good, then?'. To the second, the response might be 'I'll ask him to buy me something, then'. Given the great physical similarity between these two sentences, clearly the rules that allow us to react to them as totally different must be complex and highly discriminating. They are also entirely unconscious. It seems clear that they have nothing to do with anything specifically psychoanalytic, yet we rely on rules like these by the thousands in every session.
Any communication with patients in the course of any kind of therapy makes use of our ability to interpret one another's speech and therefore makes use of the commonsense syntactic, semantic and psychological knowledge that underlies this ability. Unlike many other therapies, however, psychoanalysis does not immediately turn around and ignore or implicitly deny the special knowledge and skills that its intersubjective communicative practices presuppose. From the mechanistic neurology of the nineteenth century to the behaviourism of today, psychological treatments have ignored or denied in their theories what they assume and take for granted in their communicative practices. Not psychoanalysis. Once it freed itself of nineteenth century mechanism and energetics, its theory has been of a piece, most of the time, with its communicative practices.
Another place where commonsense psychology looms large is in the area of technique (Edelson 1988, 335). How are we able, via acts of inference, to conclude that unconscious psychic causes are at work behind what we see and hear? Though training can make analysts better at it, there is also much that is not specifically psychoanalytic in inference-making of this kind. One indication of this is that being good at it is in part a matter of talent; if talent is lacking, any amount of training may not make up for the lack. Another indication is that there are people who are good at psychological inference and yet have never had psychoanalytic training, novelists for example.
To see how commonsense psychological skills enter the activity of psychological understanding, think of the huge difference between reading a transcript of a therapeutic encounter and hearing a tape of the same encounter. Where the affects and transference are often quite opaque in the transcript, they become clear when we hear the tape. The reason, I think, is that we have a richly discriminating capacity to 'read' psychological states off tone of voice. In the transcript, we no longer have the voices of the participants. The skills we are using here are complex in the extreme and pervade all interpersonal life, psychoanalysis included. Yet there is little that is distinctively psychoanalytic about interpreting tone of voice, even though being able to do so is essential to psychoanalytic practise.
Illustrations such as these of the role of background commonsense psychology in psychoanalysis could be multiplied endlessly. Inferences from the manifest to the latent content of a dream, for example, bear some resemblance to those used to decode the real structure of a sentence from its often misleading surface structure. Inferences about how environment or interpersonal situation has affected someone rely very heavily on our commonsense knowledge about how people are affected by their environment and other people. So does the ability, as essential in everyday life as in psychoanalysis, to form a picture of a whole narrative from the briefest sketch of it and a few bits of personal history. Perhaps the clearest example of the role of commonsense psychology in psychoanalysis is the ability analysts have to reconstruct the role of significant figures in an analysand's childhood from the briefest, most fleeting bits of transference. In all these cases, commonsense psychological knowledge plays an essential role.
Certain striking clinical abilities may also have commonsense psychology at their root. I have empathic insight particularly in mind. Empathic insight is an activity, obviously, of grasping an analysand's psychological states on the basis of his or her words and actions (and sometimes our reactions to them); yet the competences that allow us to do so are not contained in any explicit theory, psychoanalytic or otherwise. (Some theories may give us the beginnings of an explanation of the phenomenon, but none of them are anything like detailed enough to give us any guidance as to how to do it.) If the competences that allow empathic insight do not come from commonsense psychology, I do not know where else they could come from. Having commonsense psychological ability to an unusually penetrating degree may also be what allows an analyst 'to turn his unconscious to the unconscious of the patient'. That people differ in the sensitivity and insight required to do so and differ to a striking degree is no argument against the idea that these clinical skills rely on commonsense psychological knowledge; people differ in commonsense psychological talent just as they do in everything else.
If empathic insight and unconscious understanding use more commonsense than specifically psychoanalytic psychology, that is to be expected. Commonsense psychology has not been well explored yet, but even our commonsense ability to use language, which has been explored by Chomsky and his followers, is vastly richer and more complicated than anything in any explicit body of theoretical knowledge. The complexity of the unconscious system that allows us to decipher the syntax and the semantics of the words we hear is truly staggering. A child as young as three can recognize and produce, if Chomsky (1959) is right, about ninety million well-formed sentences. Even in the earliest days of their speaking, children make very few grammatical mistakes, and only of certain sharply limited kinds. There is every reason to expect that commonsense psychology will turn out to be far more complicated than language. Certainly the corpus of our commonsense psychological beliefs and competences is orders of magnitude larger and more sensitive than anything we have ever come close to capturing in conscious psychological theory, psychoanalytic or otherwise. Great novelists, most of whom had no access to psychology at all except for commonsense psychology, attest to that. Nor do we need to look to novelists and poets to see its richness. Even young children have an astonishing capacity to identify complicated patterns of intentionality. Given the richness and subtlety of our background commonsense psychology, it is not surprising that the discriminations it allows us to make are much, much finer than those that any conscious psychological theory could generate. Indeed, the superior sensitivity of commonsense psychology may be one reason why it is better to listen to analysands with evenly-hovering attention, rather than with specific theories in mind, psychoanalytic or otherwise. Any adequate critique of psychoanalysis must take into account the role of commonsense psychology in it.
This idea that commonsense psychology plays a central role in psychoanalysis has been greeted with scepticism, even hostility. Freud would probably have been appalled by the idea (though, as we shall see, he himself made extensive use of it). As we have seen, however, the idea is not very radical. Nor are commonsense syntax, semantics and psychology our only stores of commonsense knowledge. We depend on a great variety of commonsense beliefs and competences. Every time we ascribe a psychological state to someone, conscious or unconscious, we must use not just commonsense psychology but also commonsense physics, to pick out and locate the body, and commonsense biology, to recognize that it is living, conscious and of our species. Only then do we get to the point where we can even begin to apply our commonsense syntax and semantics, to interpret the noises the body is making, and our commonsense psychology, to interpret the psychological states behind the noises. Only then does the specialized knowledge of psychoanalysis get a chance to go to work. If there is anything to the work of Levi-Strauss, Lacan and their followers, we may even have a store of shared, commonsense social and cultural beliefs and competences, too, what Lacan called the Symbolic.
The philosopher Daniel Dennett has described the role of commonsense or what he calls folk knowledge in our lives this way:
Thanks to folk physics we stay warm and well fed and avoid collisions, and thanks to folk psychology we co-operate on multiperson projects, learn from each other, and enjoy periods of local peace. These benefits would be unattainable without extraordinarily efficient and reliable systems of expectation-generation [1987a, 11, cf. 17-18].
In short, commonsense psychology is not only ubiquitous in almost all psychological thought,
psychoanalysis included, but it also works.
The Extension Hypothesis
To understand the extension hypothesis, the crucial point is that commonsense psychology is an intentional psychology. An intentional psychology is a psychology about states and events that are about something, that is to say, have intentionality. Clearly, as it has in fact developed, psychoanalysis ended up as an intentional psychology, too, whatever Freud may have originally intended. (We will return to this point.) Since commonsense psychology is the ultimate ground out of which all intentional psychologies grew, this alone is enough to show that the two are closely related. What defines an intentional psychology is its vocabulary. The vocabulary of intentional psychology is utterly unlike the vocabulary of any other explanatory activity. First, psychological vocabulary is a language of representations: of perception, belief, desire, affect and psychic reality (dreaming and imagination). Secondly, psychological vocabulary ascribes intentionality to whatever it describes. This makes it unique. Intentionality here does not mean 'intention', though intentions do have intentionality. (The etymological link between 'intentionality' and 'intention' can be misleading, because many other psychological states have this property of being about something, too.) It means to have an object, to be about something (Searle, 1983; Dennett 1987b). 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. The technical term for what a mental state is about is intentional object. In this usage, events and states of affairs, including psychological events and states, can just as well be intentional objects as objects strictly defined. Moreover, intentional objects need not exist, correspond to anything real. Some psychoanalysts and philosophers also refer to intentional objects, what a psychological state is about, as content.
Contrary to the dreams of the metapsychologists, the working vocabulary of psychoanalysis is and always has been intentional in this way. The psychological vocabulary of beliefs, desires, emotions, attitudes, motives, fantasies -- the language of representations in psychic reality -- is also the language of psychoanalysis (on this, cf. Edelson 1988, xxii-xxiii, 102-109, Ch. 7). Psychoanalysis has supplemented our common psychological vocabulary in various ways, of course, introducing terms such as repression, super-ego, identification, selfobject and so forth. But nothing has ever come close to eliminating or replacing it. In fact, there have only been two serious attempts to eliminate psychological vocabulary from psychology. One was nineteenth century quantitative mechanism as expressed in metapsychology, the other was modern behaviourism. Neither was able to do the job. (On the failure of mechanistic metapsychology, cf. Schafer 1976 and the papers in Gill and Holzman 1976. On the failure of behaviourism, cf. Dennett 1978, Ch. 4).(1) If psychoanalysis is an intentional psychology, that is enough by itself to show that it is an extension of commonsense psychology, at least in respect of its vocabulary.
In a slightly different way, all the theorists who see psychoanalysis as an interpretationist, hermeneutic activity advance the same idea. They see procedures for interpreting 'meanings' as central to psychoanalysis, procedures for interpreting people's motives, etc., for thinking, feeling and acting as they do. These procedures use intentional psychology and are direct descendants of the everyday activity of making sense of ourselves; interpreting 'meanings' is a direct descendant and extension of interpreting one another in everyday life. If psychoanalysis is an activity of interpretation of 'meanings', then it is also an extension of commonsense psychology. For at least twenty-five years, a debate has raged over these issues. On the one side, there are those who think that any adequate theory of the human psyche must be developed in purely non-intentional terms, so that psychoanalysis either is or should become a non-intentional science. At the very least, anything intentional in it could make no special contribution to its status as science. Freud's metapsychology, the psychological theory built out of such notions as psycho-energetics, cathexes, barriers, discharges, etc. (economics, dynamics and topography/structure), is the best worked-out example of this point of view. The other side consists of the interpretationists. Taking their inspiration from Dilthey, Weber and Jaspers, they argue that psychoanalysis is an activity of psychological interpretation. It aims to uncover the often unconscious meanings (motives and other reasons) behind patients perceiving and believing and feeling and wishing and fantasizing and dreaming and acting as they do. (Rycroft (1966) was the first; after him came G. S. Klein (1969, 1976), Gill (1976), Ricoeur (1970, 1981), Schafer (1976, 1978), Spence (1982), Wollheim (1984), Hopkins (1986), Edelson (1988) and Goldberg (1988).) Edelson and Goldberg and others argue that being interpretive does not preclude being scientific, indeed that all science is interpretive, a point Goldberg particularly emphasizes. Among the many things that make this debate interesting is the fact that Freud's work richly represents both sides of it.
In his official pronouncements, Freud usually advocated the stance of non-intentional science;
but his actual reasoning, as we will see, was often pure intentional psychological, both in vocabulary and in the kind of explanations he gave. This duality in Freud may have had something to do
with his feelings about Brentano. Brentano is the father of modern intentionality theory (cf.
1874). Though Freud had attended his lectures for two years at exactly the time Brentano was
writing the famous 1874 book that reintroduced intentionality theory into modern thought, Freud
shut him out of his work completely (so far as the explicit text is concerned, at any rate) and
talked as though the only respectable scientific vocabulary for psychoanalysis was mechanistic
metapsychology (energy, energy discharge, etc.). (There is only one reference to Brentano in the
whole Standard Edition (1905b, 31 n. 6 and App.) and it is to a riddle Brentano devised, not to
intentionality. By comparison, Freud refers to Plato, Aristotle, Kant, Nietzsche, etc., dozens of
times each.) Yet, at the same time, Freud not only used intentional psychological language
constantly, he also modelled the special concepts of psychoanalysis upon it. Indeed, Freud was
as clear about the crucial concepts of psychoanalysis being extensions of related concepts in
commonsense psychology as he could possibly have been.
Extensions of Commonsense Vocabulary
Consider Freud's favourite argument for the existence of the unconscious, an argument he used repeatedly and over a period stretching from 1909 (175-6), through 1915 (166ff.) and 1923 (14-18), to 1938 (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. 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.
Indeed, this argument relates the concept of the unconscious to our standard concept of consciousness in more ways than one. Freud says that unconscious mental processes are like ordinary ideas; ideas have intentionality. Unconscious states, he says, are a way of "filling up the gaps in the phenomena of our consciousness" (1938, 196-7). "All the categories that we employ to describe conscious mental acts, such as idea, purposes, resolutions and so on, can be applied to them" (1915, 168). And so on. For Freud, unconscious mental processes are also quite different from conscious ones in some ways, of course. In particular, they are largely governed by the primary process. But the concept for them is still a concept for an intentional phenomenon.(2) In short, Freud's derivation of the unconscious depends essentially on commonsense psychology, indeed on commonsense psychology being correct as far as it goes. 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.
In addition, and this is a point of fundamental importance, when psychoanalysis is through
extending intentional vocabulary into its own special concepts, it does not discard this
vocabulary. Quite the reverse. Psychoanalysts continue to use concepts such as belief, desire,
perception, imagination (or cognates), dream, memory, thought, feeling, effort, trying, anger,
longing, etc., just as much as any lay person. Freud's dream of a non-intentional vocabulary for
psychoanalysis notwithstanding, it is hard to see how psychoanalysis could even begin to
dispense with its vocabulary of representations in psychic reality. To say that psychoanalysis has
extended and supplemented this vocabulary in various ways is a very different thing from saying
that it could do without it, as Freud may have dreamed.
Extensions of Commonsense Doctrines
In addition, and appearances perhaps to the contrary, a number of the most central doctrines of psychoanalysis are also extensions of commonsense psychology, extensions of some of its precepts (cf. Hopkins 1986; Edelson 1988, Ch. 15). We can find a number of examples in the theory of defence. The concept of repression seems to be an extension of the precept that people tend to forget or distort their recollection of troubling experiences (Edelson 1988, 337). The theory of splitting representations seems to extend 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 1992), seems to extend the commonsense notion that we can have inconsistent attitudes to things without being aware of it. The theory of resistance seems to extend the commonsense idea that we will try to avoid thinking about things that trouble us. And so on. Although the various parts of the theory of defence have been a signal contribution to intentional psychology, many of them are also fairly clear extensions of what we already know from commonsense.
The theory of transference is another example. The concept of transference is an extension of the 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 in psychological beings causes from the distant past can continue to operate in the present, via both memories and dispositions (Wollheim 1984). And it is tied to theory of neurosis by the idea that neurotic pathogens inevitably appear in the transference. Grünbaum (1990) finds this last idea both far removed from our normal intuitions and particularly badly supported. I think he is wrong on both counts. He seems to think that Freud's only reason for thinking that the causes of neurosis will be recapitulated in the transference was that earlier ways of reacting to people are recapitulated in the transference. This would indeed be grossly invalid reasoning. I think, however, that Freud's reasoning was more subtle. His conviction that pathogens will appear in the transference relies upon a general background conviction, a conviction as much a part of commonsense psychology as of psychoanalysis, that what is troubling people will appear in their relationships to significant others. This idea is intuitively plausible. The extension that psychoanalysis introduces is the idea that the repetition of childhood patterns will be particularly relevant, and it creates a setting that maximizes the chances that this repetition will occur. Thus even the doctrine that pathogens appear in the transference is an extension of commonsense psychology. Clinically, transference is generally held to be the most important field of events to which analysts apply their interpretive and explanatory skills, but that does not mean that it is unrelated to commonsense psychology.
Psychoanalysis has also made many contributions to intentional psychology that go far beyond commonsense psychology, of course. The innovative parts of the theory of defence and the theory of transference are two; it would take us too far afield to explore the many others. Methodological contributions have been particularly rich. Much psychoanalytic work, as I will argue shortly, is concerned with restoring the real intentionality of affects, fantasies, etc. The methods of free association, evenly-hovering attention and empathy are probably the first and only tools ever discovered for dealing with pathologies of intentionality. Other important contributions include the theory of drives as the biological basis of intentional reactivity, the theory of the developmental stages, the theory of interpersonal intentionality (transference, counter-transference, projective identification), 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 and contributions to intentional psychology in general would be very long.
Let me close this section with a look at the opposition. As we saw earlier, Grünbaum argues
vociferously that psychoanalysis is no more an extension of commonsense psychology than
theoretical physics is an extension of commonsense 'physics'. If psychoanalytic theory is an
extension of commonsense psychology, why, he asks, did it encounter so much disbelief? Many
of its doctrines are quite counter-intuitive. Indeed, the idea that horror dreams should be wish-fulfilling is "utterly incredible common-sensically" (quoted in Edelson 1988, 330). Well, maybe
there is also more to commonsense physics than Grünbaum thinks! But the real point is that he is
failing to make a distinction. That psychoanalysis ends up with results that are counter-intuitive
to commonsense psychology does not mean that it does not use, model itself on and extend the
methods and vocabulary of commonsense psychology (Edelson 1988, 330). Moreover, as I will
argue in the final section, it is a very good thing that psychoanalysis does so, because the
methods and vocabulary of commonsense psychology contain powerful scientific tools.
Extensions of Psychological Explanation
What makes it important that the phenomena of interest to psychoanalysis have intentionality and are extensions of commonsense vocabulary and precepts is that this allows it to use a distinctive kind of interpretation and explanation. Usually called psychological explanation, it is the methodology of interpreting and explaining actions and psychological states in terms of the reasons for them. Unlike phenomena that do not have intentionality (energies, neurons, etc.), we can explain intentional phenomena in terms of reasons. This feature of psychological explanation is unique to them. As I mentioned earlier, Edelson and Goldberg and others are now making the point that all science is interpretive. The observation that psychological explanation interprets reasons allows us to carry this point one step further. In interpreting people's reasons, psychoanalysis and psychological explanation in general are different from other sciences. (That does not mean that they are not science at all, but I will have to take that big issue up in another place (cf. Edelson (1988, Ch. 11), and Goldberg (1988, Ch's. 1 and 4).
It is almost impossible to fit interpretive explanations in terms of reasons and meanings into non-intentional metapsychology. Despite this, Freud made psychological explanation, and its central idea that people generally do what they do for a reason, one of the cornerstones of psychoanalysis. It is the basis of the crucial claim that when there is nothing conscious to be the cause of some behaviour, the cause will often still be psychological, there will often still be reasons, they will just not be introspectible. This idea was one of Freud's most important contributions. Thus, his official preference for metapsychology notwithstanding, psychological explanation as I have just defined it is the heart and soul of his work. So far as I know, he never tried to reconcile the two.
To illustrate just how central psychological explanation is in his work, consider this comment from 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].
Though Freud does not use the term, this is a remark about how psychological explanation
works. It is clear from the way Freud writes that he saw this kind of explanation as central to psychoanalysis (cf. Sachs 1982; for a marvellous example Freud giving an explanation of this kind,
cf. 1926, 117). Note that Freud does not argue that the Rat Man had reasons for feeling guilt; he
takes that for granted. So automatic was his recourse to psychological explanation, he does not
even feel a need to justify the assumption. His innovation was to see that often people's reasons
are unconscious. That is what made the very ordinary precept that there are reasons for what
people feel and do such a powerful tool in his hands. Suddenly there could be reasons for vast
ranges of thought, feeling, motive and behaviour that until then had seemed completely senseless.
The Alternative Theories of Hopkins and Edelson
To get a more detailed idea of how psychoanalysis makes use of and extends the basic principles of psychological explanation, we might examine the alternative accounts of this process given by Hopkins (1986) and Edelson (1988). To orient ourselves, let me begin with a brief clinical vignette.
When Ms. B., a 50-year old divorced counsellor, 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 much of the scene, which 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 during that month. As soon as mother returned home, she cut off Ms. B.'s beautiful long hair. I suggested to Ms. B. that this bloody scene followed by sinful pleasure and humiliating punishment had become her model of sexual pleasure.
This vignette contains material that is clearly psychoanalytic, material indeed that we probably could not have seen before the advent of psychoanalytic theory gave us the lens to see it. Nevertheless, the type of explanation it gives of Ms. B.'s adult ways of reacting to sexual pleasure, that the memories and fantasies she carried away from her earlier traumatic experiences contain her reasons for reacting as she does, is an extension of a pattern of explanation intrinsic to commonsense, intentional psychology. Hopkins and Edelson give different accounts of how this extension occurs and what it is like.
In outline, Hopkins begins with one of the basic interpretative patterns of commonsense psychology, inferring a pattern of desires (or other motives) and beliefs (or other cognitive states) such that, given what we infer a person to desire and believe, it would have been rational to act as he or she did. This pattern was first articulated by Aristotle and has been much studied by philosophers and cognitive scientists of late (Davidson (1970, 1974), Dennett (1978, 1987a), Fodor (1988) and others). Freud took over the pattern, argues Hopkins, but extended it in two important ways. First, Freud was primarily interested, not in actions, but in psychological states, especially states that do not issue directly in intended actions, such as dreams, fantasies and parapraxes. Secondly, parallel to the way that actions are explained when we see how the person doing them could have taken them to satisfy a desire, these psychological states are explained when we see how they could have been a way of imagining the fulfilment of a wish. (Hopkins, following Wollheim, carefully distinguishes wishes from desires, but we need not follow him in that here.) Thus, psychoanalysis is not a direct application of the commonsense psychological principle that, all other things being equal, people will act to get what they believe will satisfy their desires. It is, however, an extension of that principle. To be sure, there are also important differences. If a wish causes a dream, fantasy or parapraxis, there need be no belief that any of the latter will be a means to satisfying the former, for example. But the lineage is still clear.
Hopkins puts this argument to good use, and it certainly seems to have merit. For example, he uses it to show that an interesting observation of Grünbaum's does not mean what it seems to mean. The observation (1984, 75) is that important classes of psychoanalytic explanation cannot be assimilated to the desire/belief/action pattern. Grünbaum is right about this, Hopkins (1986, 55) argues, but he draws the wrong conclusion from it. Contrary to Grünbaum, these explanations are still psychological explanations. But the pattern they follow is an extension of the belief/desire/action pattern, not the pattern in its original form. However, Hopkins' account also faces a fairly obvious problem. Contrary to Hopkins, explanations of fantasies, day-dreams, etc., as wish-fulfilling fantasies are not an extension of commonsense psychology, they are part of it. Our everyday activities of making sense of our own and others' psychological states use the wish/imaginary wish-fulfilment pattern of explanation just as much as the belief/desire/action pattern.
Edelson's (1988, Ch. 15) approach to the extensionist argument avoids this problem. Unlike Hopkins, he does not argue that psychoanalysis picked one form of commonsense psychological explanation and extended its logic. Rather, he argues that psychoanalysis extended all the forms of explanation found in commonsense psychology. However, it did not extend their logic, it extended their range of application by applying them to new things: infantile sexuality and stages of psychosexual development as well as adolescent and adult ones; unconscious wishes and fantasies as well as conscious ones; primary as well as secondary processes; latent as well as manifest content; bizarre as well as acceptable wishes, desires and affects; and so on (1988, 327-338).
In my view, as well as allowing Edelson to avoid the objection just mentioned, this difference gives his approach two further advantages over Hopkins'. First, it allows for a more realistic picture of the richness and diversity of the explanatory strategies of commonsense, everyday psychological interpretation. Explanations of actions in terms of beliefs and desires hardly scratch the surface. Secondly, Edelson's approach presents a more accurate picture of how psychoanalysis grew out of and now relates to commonsense psychology. It grew out of many forms of everyday psychological discourse and many interpretive patterns of commonsense psychology, not just one of them. It is as closely related to our fictional and mythological activities, to poetic speculations about sexual love, to all the products of the human imagination, as it is to our everyday activities of explaining our own and others' actions. (Freud refers to Shakespeare far more often than to Herbart!) Psychoanalysis grew out of many different forms of human activity and reflection on it, not just one of them.
Nevertheless, Edelson's account has a weakness, too; it contains far too little detail. Nowhere does he attempt to lay out how many kinds of interpretive patterns there are in commonsense psychology and nowhere does he explore just how psychoanalysis extends them. In fact, his account is more like a suggestion for where a theory might be developed than a completed theory. Edelson's basic idea is that psychoanalysis uses what is confidently known in commonsense intentional psychology "as a homely model" (1988, 329), as an analogy. He sees commonsense psychology as providing useful analogies at two or more levels. At the most general level, a lot if not all of the explanatory patterns of commonsense psychology find the causes of psychological states by studying how content has been conveyed down a causal chain. A desire for x, a belief that doing y will secure x, an intention to secure x, and an action that attempts to secure x by doing y all have x as their common content. Likewise, a wish for x, a wish-fulfilling fantasy of getting x, and a guilty feeling of not being entitled to x all have the same content. This process of conveying content results in what has come to be called thematic affinities between cause and effect, affinities sometimes spanning many years or even whole lifetimes. Edelson emphasizes that "reasoning by appeal to thematic affinities is ... central in the explanatory strategies of psychoanalysis" (1988, 332). This claim will play an important role in the final section, because Grünbaum has argued that such appeals to thematic affinities are 'unavailing' as a way of establishing causal connections.
At a slightly less general level, commonsense psychology provides two analogies, one far more useful than the other. The less useful one is its way of finding the causes of actions (also intentions) by finding beliefs and desires that would have made it rational to do them. This pattern is not very important in psychoanalysis, Edelson believes. Far more usefully, commonsense psychology has a pattern of explanation in which various imaginings, dreams, etc., are explained as expressions of wishes. In Edelson's view (1988), the latter pattern is also the distinctive explanatory pattern of psychoanalysis.
In my view, all this is an extremely good start. However, these two points of contact between the explanatory patterns of psychoanalysis and commonsense psychology do not exhaust the field. Recall the vignette with which I began this section. Explanations in terms of wishes are forward-looking: we explain the fantasy, etc., as a surrogate for going out and getting something. However, the psychological states displayed in the vignette were backward-looking: Ms. B.'s way of experiencing sexual pleasure is really about, or partly about, the bloody scene she witnessed as a child and the feelings she felt then. The same was true in the vignette from Ms. A. we examined much earlier: Ms. A. felt she was betraying an ideal -- an ideal she has held for a very long time. Likewise, the interpretations I offered appeal to past psychological states in both cases: retained memories (or repressed, memory-like states), and retained ideals. Many of the psychological states of most interest to psychoanalysis have this character. To explain them, we must refer to earlier psychological states -- to retained dispositions, to memories (and repressed memories) and, of course, to fantasies that are really about earlier experiences (including earlier fantasized experiences).
Yet Hopkins and Edelson deal only with forward-looking psychic states like desires and wishes, states aimed at securing or avoiding something, and the imagined, wish-fulfilling surrogates for action they generate. The examples of psychoanalytic explanations we have examined are not like this. They do not work from wish-fulfilments to the wishes they fulfil in fantasies or dreams, but back from fantasies, anxiety and ways of experiencing to earlier experiences that the fantasies, etc., are really about. As I put it in another paper (Brook, in progress), many psychoanalytic explanations are centred around restoring the intentionality of psychological states where processes of defence have rendered their true intentionality obscure or invisible, not forward-looking and in terms of desire-satisfaction or wish-fulfilment. Fortunately, adding the restoring of the intentionality of psychological states to our list of types of psychoanalytic explanation is no problem for the extensionist hypothesis about psychological explanation. Backward-looking explanations in terms of earlier traumatic memories, fantasies, etc., are just as much part of commonsense psychology as forward-looking ones in terms of desire-satisfaction and wish-fulfilments. Put another way, the explanatory direction of thematic affinity can run either way.
In short, Edelson's suggestions about how the explanatory patterns of psychoanalysis extend
those to be found in commonsense psychology is promising, but there is still work to be done. In
addition to extending the explanatory patterns of commonsense psychology, we have also seen
that psychoanalysis has extended and continues to use its vocabulary, and that some of the most
distinctive doctrines of psychoanalysis are extensions of precepts of commonsense psychology.
Furthermore, as we saw even earlier, specifically psychoanalytic interpretations and explanations
require a sub-structure of ordinary, commonsense psychological interpretations before they can
even get started. (We must first interpret the direct, literal meaning of an expression or report, for
example, before we can interpret its unconscious meaning). Neither Hopkins nor Edelson has
much to say about any of these other connections.
Commonsense Psychology in Case Studies
Let us now apply the results of our investigation to one of the great sceptics about the possibility of establishing psychoanalytic hypotheses, Adolf Grünbaum. Grünbaum thinks that the primary method of investigation in psychoanalysis, the case study, cannot establish claims about either fact or cause. As we saw earlier, he also rejects the idea that psychoanalysis is an extension of commonsense psychology. Yet, as we have seen, it extends commonsense psychology in no less than three different ways: in vocabulary, doctrines, and type of explanation. In addition, psychoanalysis uses a lot of commonsense psychology that it does not take up into its own corpus at all. In my view, Grünbaum's scepticism about case studies as a method of uncovering causal connections and his rejection of the idea that psychoanalysis is an extension of commonsense psychology are connected. Commonsense psychology gives psychoanalysis a special power to uncover fact and cause that Grünbaum completely overlooks.
The distinctive home of psychoanalytic thinking is the case-study built out of psychological explanations. Grünbaum's (1984, 1988, 1990) reservations about the extent to which case-studies can identify causes and establish causal claims, claims like the big hypotheses of psychoanalysis at any rate, are far-reaching. Because of the possibility of suggestion, confusion of fantasy with memory, etc., he argues, case-studies cannot establish what happened: the facts about such things as temporal sequences, correlations and thematic affinities in content. Even if they could, however, they could not uncover the causal relations in what happened. To find out what caused what in these data, we need to find out what made a difference to the occurrence of what. But a case study, he argues, could never tell us when a prior or correlated event or something with a similar thematic content made a causal difference, changed the probability of something occurring. A main reason for this (1984, 280) is that claims about causal difference must be tested against a control, and case studies provide no independent control. Grünbaum's charges cannot be sustained (Edelson (1988, Ch. 15) agrees). Grünbaum is particularly interested in the ability of case studies to establish the big theses of psychoanalysis: the generalizations meant to hold true of everyone, such as that repression is a central factor in neurosis, or that neurotic pathogens will appear in the transference, and especially the more counter-intuitive ones, such as that paranoia in males has a homosexual etiology.(3) But it seems to me that if his charges apply anywhere, they ought to apply just as much to other uses of case-studies, too.
Most of the time, analysts use case studies to figure out individual lives, not to uncover or test general hypotheses. As used for this purpose, we can see immediately that Grünbaum's critique of case studies cannot be entirely right. The problem he identifies, lack of a control, for example, would vitiate the use of case studies to understand individual lives just as much as it would vitiate their use to establish big, generalized hypotheses. Yet it could not possibly be right to claim that case studies are impotent to establish causal claims about individual lives. We are doing case studies practically every waking moment of our lives. When I understand even something so simple as the words a person utters, I have to interpret what the person means by the noises he or she makes -- a miniature case-study. Similarly, every time I judge what a person is believing, desiring, feeling, hoping, etc., on the basis of what the person is saying and doing, I have done a miniature case study. And so on. (The theory behind the points I am making here was first developed by Davidson (1984) in the 1970's.) These little case studies involve psychological interpretation and explanation just as much as big, psychoanalytic studies of a life do.
Moreover, the small case studies and psychological explanations of everyday life work. We do understand one another, and we do assess each other's thoughts and feelings correctly, most of the time and in simple situations, at least. Or if we do not understand one another completely, we come close enough for practical purposes. It is a simple matter of fact that not just psychoanalysts but everyone can establish fact and cause using case studies. We can all judge when something prior to a given phenomenon might or might not have been causally relevant to its coming to be. We can decide when the probability that a thematic affinity is the result of a causal connection is high or low. And so on. And we do these things by interpreting reasons. Whether case studies can establish the more complicated and generalized hypotheses of psychoanalysis is another matter, to which we will return. It is also true that people sometimes render some of their thoughts, feelings and actions pretty opaque, by defensive, ideological or other distortion. But any suggestion that case studies cannot discern causes of behaviour in principle would have to be wrong; for vast reaches of ordinary human existence and most of the time, they can. If untutored commonsense can do this well, surely psychoanalytic case studies will often do much better.
I am not maintaining, of course, that case studies relying on psychological explanations can do everything. Clearly, they cannot. A lot of the causal sub-structure of any thought, feeling or action has nothing directly to do with psychological factors at all (appropriate processing of oxygen, for example). Freud was well aware of this limitation:
So long as we trace ... development from its final outcome backwards, .. we feel we have gained an insight which is completely satisfactory or even exhaustive. But if we proceed the reverse way, ... we notice at once that there might have been another result, and that we might have been just as well able to understand and explain the latter [1920, 167-8; my emphasis].
But the fact that case studies using psychological explanations cannot do everything does not mean that they can do nothing.
For an example of how commonsense psychological explanation enters a psychoanalytic case study, go back to the vignette from Ms. B.'s analysis. In the context of Grünbaum's critique, this vignette is particularly interesting; it is centred on a thematic affinity of just the sort to which Grünbaum has given special attention, in this case the similarity between Ms. B.'s early experience of a bloody miscarriage followed by Oedipal guilt and punishment, and her current way of experiencing sexual pleasure. As I related earlier, I inferred that we could find the reasons for the current reaction in the earlier experience. In making this connection, I was relying on commonsense psychology, and specifically on its widely-instantiated and well-supported precept that such earlier experiences would give a person, in particular a female, good reason to react to later experiences that are thematically related in just the way Ms. B. now reacts.
As we will see shortly, Grünbaum accuses Freud of making inference like this on the basis of the thematic affinity alone. As a charge against analytic reasoning in general, that could not be right. Rather, when analysts come across a striking thematic affinity of the sort I found in Ms. B. and interpret a causal connection, I think, following Hopkins and, by implication, Edelson (1988, 332-3), that their reasoning is both more complicated and more plausible than Grünbaum allows. To reach a conclusion about a causal connection, they call on the background psychological knowledge of how a normal person would probably react in a situation like this. At any rate, that is what I think I did. I referred, as I said, to a widely-shared belief that early experiences such as Ms. B. had would live on and give any normal female excellent reason to view later experiences having a similar thematic content in the way Ms. B. now views sexual activity. This belief, it seems to me, is widely instantiated in ordinary human experience.
Note that in my inference to the cause, the thematic affinity played no direct role. It merely alerted me to the possibility of a connection. What allowed me to infer that a connection actually existed was my background commonsense psychological knowledge of the conditions under which thematic affinities are apt to be the result of causal connections in normal persons. It is this background that allows "reasoning by appeal to thematic affinities [to be] ... central in the explanatory strategies of psychoanalysis" (Edelson 1988, 332; cf. 331-5 and Hopkins 1986, 50 for related arguments). Let me pursue this point a bit further. Grünbaum thinks that Freud repeatedly made the error of inferring causal connections from thematic affinities alone. I am not convinced. To the contrary, I am inclined to think that what this criticism provides, rather, is an unwitting example of what Grünbaum is overlooking. Take his critiques of Freud's case studies of the Rat Man and the Wolfman. Repeatedly, Grünbaum accuses Freud of drawing causal conclusions from mere thematic affinities. Here is part of what Freud had to say about the Rat Man.
The notion of a rat is inseparably bound up with the fact that it has sharp teeth with which it gnaws and bites. ... [They] are cruelly persecuted and mercilessly put to death by man, ... . But he himself had been just such a nasty, dirty little wretch, who was apt to bite people when he was in a rage, and had been fearfully punished for doing so. ... It was almost as though Fate, when the captain told him [of the anal rat punishment], had been putting him through a ... test: ... he had reacted to it with his obsessional idea [that his dead father was to undergo the same punishment] [1909, 215-16].
Now consider Grünbaum's response. He suggests that Freud draws an etiologic inference connecting the Rat Man's experience of himself as rat-like when a child and his current obsession with the idea of his father undergoing the monstrous punishment of anal rat penetration on the basis of a mere thematic affinity (1988, 631).
As I said, I am not convinced. When Freud reasoned from affinity to cause, as he seems to have been doing here, I think it was always, though perhaps not explicitly, via a reference to some background principle about how human beings act. In this case, the principle might be something like this: 'Feeling that one is like a rat in early childhood makes one susceptible to obsessions about rats doing evil things to loved ones in adulthood, especially if one would like to do such evil things oneself but is not aware of it.' I am not saying that such a principle is true, only that Freud was appealing to some such principle when he drew his etiological conclusion, not to a bare thematic affinity. I have not been able to find a single place where Freud clearly and unambiguously used nothing but a thematic affinity as his grounds for inferring a causal connection.
In general, Freud's evidence for his claims concerning the Rat Man and the Wolfman in his case studies of them does indeed often look weak, as Grünbaum reconstructs them. But perhaps Grünbaum has missed something. For Freud's own words convey quite a different impression. Even though some of Freud's causal claims might now be questioned, his own words have an interest, and sometimes even a plausibility, that is lost in Grünbaum's reconstructions. What seems to make the difference is this. Freud weaves a whole web of inferences connecting behaviour in the present to reasons for that behaviour from deep in the past. When he does not actually weave such a web, he creates in us a sense of what the web might have been like. These inferences are drawn from both commonsense and psychoanalysis. All that remains of them in Grünbaum's reconstructions are the bare bones of the claims themselves, starkly isolated from their context, and the strictest construal of Freud's representation of the evidence on which he made them (cf. Fine and Forbes (1986, 238)). Grünbaum misses what really gives Freud's conclusions their appeal, namely the implicit web of background psychology, both commonsense and psychoanalytic, that was Freud's real basis for his inferences.
I would make the same response concerning the other false principle of causal inference that Grünbaum claims to find in Freud's work, the tally principle or NCT (necessary condition thesis). This is the thesis that improvement in a patient is evidence for correctness of interpretation, because only correct interpretations can cure. Again, to the extent Freud did reason from improvement to causal accuracy, he did so, I think, via implicit reference back to his general knowledge of the conditions under which change occurs in analytic patients. He may not always have been entirely clear about this, and he did say things inconsistent with this reading (e.g. 1916-17, 452, justly quoted by Grünbaum in 1984, 138, and Grünbaum et al 1986, 222), but I know of nothing in his actual causal practises that is inconsistent with this reading. (Moreover, as Grünbaum relates (1984, 152; cf. Grünbaum et al 1986, 222), Freud himself came to harbour reservations about the principle (1926, 154) and said things inconsistent with it; on Freud's final view of the general relation between insight and cure, cf. (1937).)
Nothing I have said so far touches Grünbaum's most important objection, that case-studies lack an independent control. Without a control, we cannot distinguish the merely 'post hoc' from the truly 'propter hoc', so we cannot distinguish causal connections from mere co-incidences. Edelson and Hopkins both respond to this argument by urging that in normal psychological reasoning, no such controls are required. Edelson (1986b; reprinted in 1988, Ch. 13) thinks we can do the job they do, and more, with causal mechanisms. Hopkins (1986, 37) argues that inferring motive from action or wish-fulfilling fantasy by coincidence of content needs no control. By contrast, I think that Grünbaum is right; case-studies do need an independent control. I also think we have one.
The control we have is our background commonsense psychological knowledge of how people think, feel and behave. The whole of a psychoanalyst's psychological competence and experience is the control against which he or she tests causal judgments. This is not a formal control, consisting as it does of the collective knowledge of human causality built up over human history, but what has been called a historical control (Lasagna 1982). Nevertheless, it is a control. Far from being negligible, these background beliefs, principles, precepts, rules and skills for recognizing and inferring psychological states are a huge control and one that has been thoroughly tested. It is against this that we assess observed correlations, sequences and thematic affinities for their causal content.
To be worth much, this body of background psychological knowledge and the special knowledge that psychoanalysis has added must:
(1) generate generally good predictions and explanations;
(2) allow principled generation and resolution of disagreements (Putnam 1983), and;
(3) do so by using its own method of interpretation and explanation in terms of reasons, in terms of meaning in psychic reality.
Both commonsense psychology in general and psychoanalysis in particular usually meet these tests. The explanations they generate yield fairly accurate predictions, we can generally recognize when they are apt to be good and when they are not, and we are able to do so without straying outside psychological interpretation and explanation.
As accounts of individual lives, I have been arguing, case studies not only can but do work fairly
well. Let us now turn to the applications where Grünbaum's scepticism is higher, namely the
general hypotheses of psychoanalysis, such as that unconscious fantasies exist and repressed
conflicts are pathogenic, and especially the counter-intuitive ones, such as that paranoia in males
has a homosexual etiology. Can case studies using commonsense psychology and its psychoanalytic extensions give us what we need to scale up from accounts of individual lives to
hypotheses applicable to whole classes of lives? As a science, psychoanalysis has to be able to do
so. I will call this the problem of scale.
The Problem of Scale
At first blush, Grünbaum's concern here appears to be easy to meet. Surely, case studies relying on psychological explanation can address the big theses just as well as the small ones. Indeed, the only obvious way to test the big theses is by reducing them to multiple smaller theses about individual patients. Each case study of an individual may yield results with some degree of uncertainty, but so long as the uncertainty has a different source in each case and yet the results they yield are consilient, progress can be made. In short, at this level of abstraction, the idea that the big theses of psychoanalysis can be tested by case studies, in principle at least, seems untroubled. If case studies work for individuals, the usual canons of induction should be able to scale up their results to classes of individuals without too much difficulty.
These points, however, only touch the surface. Not just psychoanalysis but all forms of intentional psychology have signally failed to achieve the theoretical unity, breath and depth of even the most modest natural science. An increasingly influential school of thought in cognitive science now argues that they are unlikely ever to do so (P. M. Churchland 1984, 46; P. S. Churchland 1986). For one thing, no one has ever succeeded in quantifying phenomena described in the language of intentional psychology. Yet quantified relationships are the hallmark of science. Moreover, despite attempts to get rid of the economic point of view, it is hard to see how we could do without quantified relationships in psychology. Surely, for example, we will always need to compare desires by their strength: 'he did it because he wanted x more than he wanted to stay out of trouble'. It is hard to see how cases studies -- or anything else -- could help us with this problem of quantifying psychological states.(4) Note that the problem I am raising here is not just a problem of 'reducing' intentional psychology to some non-intentional discourse such as metapsychology or neuroscience. It is a problem of organizing psychological observations and inferences into any kind of unified general theory -- even a unified general theory in the language of intentional psychology!
The problems I have just sketched are extremely serious, so far as the hopes of any intentional psychology to become a mature science are concerned. Will psychoanalysis ever be able to arrive at and justify quantified, unified theories? Will it ever generate long-term, detailed predictions and retrodictions that are as successful as its explanations and predictions about individual lives? At this point, we do not know. However, two points should be made. First, the problems just discussed apply to all intentional psychology, not just psychoanalysis. Secondly, there is nothing specific to case studies behind them.
If case studies relying on commonsense psychological explanations can do so well, why has psychoanalysis had such a hard time marshalling convincing evidence for its larger theses? Here I will make just two points. Historically, psychoanalysis has gotten into the most trouble when it has veered farthest from the vocabulary and methods special to psychology. Think of the quantitative mechanistic notions Freud played with in the 1890's and their residue in metapsychology. The legacy of psychic energies and 'cathexes' of energy and energy barriers persisted and interfered with scientific progress in psychoanalysis for a long time. Where psychoanalysis has stuck to intentional psychology, it has done much better. Secondly, psychoanalysis has often paid too little attention to evidence. A recent study by psychoanalysts of sixty often-cited papers from major psychoanalytic journals found that not one of them offered any direct evidence for the claims they were making (Klumpner 1989). Sometimes ideas become influential in psychoanalysis not because of the strength of the evidence for them but because they capture the imagination. Psychoanalytic research, of course, does not have to be like this.
REFERENCES
Brentano, F. (1874), Psychology from an Empirical Point of View, trans. D. B. Terrell. London: Routledge and Kegan Paul, 1971.
Brook, J. A. (1992), Freud and splitting. International Review of Psychoanalysis, 19:
-----------. (in progress), A hermeneutic science.
Clark, S. and Wright, C., eds. (1986), Psychoanalysis, Mind and Science. Oxford: Basil Blackwell.
Churchland, P. M. (1984), Matter and Consciousness. Cambridge, MA: MIT Press.
Churchland, P. S. (1986), Neurophilosophy. Cambridge, MA: MIT Press.
Davidson, D. (1970), Mental events. In: D. Davidson (1980).
-----------. (1974), Psychology as philosophy. In: D. Davidson (1980).
-----------. (1980), Essays on Actions and Events. London: Oxford University Press.
-----------. (1984), Inquiries into Truth and Interpretation. London: Oxford University Press.
Dennett, D. (1978), Brainstorms. Montgomery, VT.: Bradford Books.
----------. (1987a), The Intentional Stance. Cambridge, MA.: MIT Press.
----------. (1987b), Intentionality. In: R. Gregory, ed. The Oxford Companion To the Mind. Oxford: Oxford University Press.
Edelson, M. (1986a), Heinz Hartmann's influence on psychoanalysis as a science. Psychoanalytic Inquiry, 6: 575-599.
----------. (1986b), Causal explanation in science and in psychoanalysis: implica- tions for writing a case study. Psychoanalytic Study of the Child, 41: 89-127.
----------. (1988), Psychoanalysis: A Theory in Crisis. Chicago: University of Chicago Press.
Fine, A., and Forbes, M. (1986), Commentary on Grünbaum (1984). Behavioral and Brain Sciences, 9:237-8.
Fodor, J. (1988), Psychosemantics. Cambridge, MA.: MIT Press.
Freud, S. (1905a), Three essays on sexuality. Standard Edition, 7:125-248. London: Hogarth Press, 1953-1974.
--------. (1905b), Jokes and their relation to the unconscious. Standard Edition, 8. London: Hogarth Press, 1953-1974.
--------. (1909), Notes upon a case of obsessional neurosis. Standard Edition, 10:153-318. London: Hogarth Press, 1953-1974.
--------. (1912), The dynamics of transference. Standard Edition, 12:97-108. London: Hogarth Press, 1953-1974.
--------. (1914), On the history of the psychoanalytic movement. Standard Edition, 14:3-66. London: Hogarth Press, 1953-1974.
--------. (1915), The unconscious. Standard Edition, 14:159-216. London: Hogarth Press, 1953-1974.
--------. (1916-17), Introductory lectures on psychoanalysis. Standard Edition, 15 & 16. London: Hogarth Press, 1953-1974.
--------. (1920), The psychogenesis of a case of homosexuality in a woman. Standard Edition, 18:145-174. London: Hogarth Press, 1953-1974.
--------. (1923), The ego and the id. Standard Edition, 19:3-68. London: Hogarth Press, 1953-1974.
--------. (1924), A short account of psychoanalysis. Standard Edition, 19:191-212. London: Hogarth Press, 1953-1974.
--------. (1926), Inhibitions, symptoms and anxiety. Standard Edition, 20:77-178. London: Hogarth Press, 1953-1974.
--------. (1937), Analysis terminable and interminable. Standard Edition, 23:209-270. London: Hogarth Press, 1953-1974.
--------. (1938), An outline of psychoanalysis. Standard Edition, 23:141-204. London: Hogarth Press, 1953-1974.
Gill, M. and Holzman, P. (1976), Psychology versus Metapsychology. New York: International Universities Press.
Goldberg, A. (1988), A Fresh Look at Psychoanalysis: The View from Self Psycho- logy. Hillsdale, NJ: The Analytic Press.
Grünbaum, A. (1984), The Foundations of Psychoanalysis. Berkeley, CA: University of California Press.
------------. (1988), The role of the case study method in the foundations of psychoanalysis. Canadian Journal of Philosophy, 18: 623-658.
------------. (1990), 'Meaning' connections and causal connections in the human sciences: the poverty of hermeneutic philosophy. Journal of the American Psychoanalytic Association, 38: 559-578.
Grünbaum et al (1986), Précis of The Foundations of Psychoanalysis, with peer commentary (including by analysts and analytic writers such as Eagle, Holt, Edelson, Luborsky, Marmor, Pollock, Reiser, Spence and Storr). Behavioral and Brain Sciences, 9: 217-284.
Hopkins, J. (1986), Epistemology and depth psychology: critical notes on The Foundations of Psychoanalysis. In: Clark and Wright (1986).
Klein, G. S. (1969), Freud's two theories of sexuality. In: Gill and Holzman (1976).
-----------. (1976), Psychoanalytic Theory. New York: International Universities Press.
Klumpner, G. (1989), (Interview with, Chair, Committee on Scientific Activities). American Psychoanalytic Association Newsletter 22: 13.
Lasagna, L. (1982), Historical controls. New England Journal of Medicine, 307: 1339-1340.
Morton, A. (1982), Freudian commonsense. In: Wollheim and Hopkins (1982).
Putnam. H. (1983), Realism and Reason. Philosophical Papers, Vol. 3. Cambridge: Cambridge University Press.
Ricoeur, P. (1970), Freud and Philosophy. New Haven, CA: Yale University Press.
----------. (1981), Hermeneutics and the Human Sciences. New York: Cambridge University Press.
Rycroft. C. (1966), Psychoanalysis Observed. London: Constable.
Sachs, D. (1982), On Freud's doctrine of emotions. In: Wollheim and Hopkins (1982), 92-105.
Schafer, R. (1976), A New Language for Psychoanalysis. New Haven, CN: Yale University Press.
----------. (1978), Language and Insight. New Haven, CN: Yale University Press.
Searle, J. (1983), Intentionality. Cambridge: Cambridge University Press.
Spence, D. (1982), Narrative Truth and Historical Truth. New York: Norton.
Stich, S. (1983), From Folk Psychology to Cognitive Science: The Case Against Belief. Cambridge, MA: MIT Press
Wollheim, R. (1984), The Thread of Life. Cambridge, MA: Harvard University Press.
-----------. and Hopkins, J., eds. (1982), Philosophical Essays on Freud. Cam- bridge: Cambridge
University Press.
2364 Orlando Ave.
Ottawa, CANADA
K1H 7K1
1. Schafer's (1976, 1978) attempt to do analysis in the language of action may seem to some like a third attempt, but it is not so much an attempt to replace psychological vocabulary with something else as to force a part of that vocabulary (action language) to serve for the whole (the full range of psychological terms for belief, desire, affect and so on).
2. Edelson (1988, Ch's 6 and 7) gives a good account of the idea that the concept of the unconscious, indeed psychoanalysis in general, is based on filling causal gaps in consciousness with unconscious psychological states.
3. The only big hypotheses Grünbaum considers are the ones Freud offered. However, the question of whether analysts hold these views in any unqualified form nowadays is not important for present purposes.
4. In an excellent discussion of the relation of psychoanalysis to the neurosciences (1988, Ch. 7), Edelson agrees that intentional psychology will not reduce to neuroscience. Following Fodor, he thinks that this is because psychological concepts group things so differently from the neurosciences that they will not usefully reduce (cf. Davidson's anomalous monism (1970, 1974)). The argument we are examining claims the concepts are so flawed that they are empirically useless in the first place.