Psychoanalytic Theories, Logical Status of
PSYCHOANALYTIC THEORIES, LOGICAL STATUS OF
Since psychoanalysis fails to conform to currently accepted methodological models, its prominence on the contemporary scene constitutes a challenge to the methodologist. He must either revise his canons or show the psychoanalyst the error of his ways. Both tacks have been tried, but thus far the second has predominated. This entry will be confined to methodological problems raised by psychoanalytic theory, though as we shall see, such problems cannot be pursued very far without running into questions concerning the clinical interpretation of particular cases.
Content of Psychoanalytic Theory
Within psychoanalytic theory there are diverse strands, and the relations between them are by no means obvious. For one thing, there are theoretical ideas at different levels. Fairly close to actual clinical practice are found the concepts of repression, regression, projection, reaction formation, and transference. At a higher level there is a theoretical model of the mind in terms of psychic energy, which gets attached to various ideals, the transformations of which are governed by quasi-mechanical principles. This is, in fact, designed to be a perfectly general model of the mind, in terms of which, in the last analysis, all psychological processes and states may be conceived. At this level we have also the division of the psyche into the three systems—id, ego, and superego—together with an account of their properties and interrelations.
In addition to the distinction between levels, we have the distinction between developmental and dynamic theories. In the first group is the theory of psychosexual stages—oral, anal, genital—according to which there is a biologically determined order, beginning from infancy, in which first one, then another, area of the body is maximally sensitive to pleasurable stimulation and according to which certain personality traits predominate as one or another stage is prolonged or transcended only with difficulty. For example, passivity and lack of initiative are associated with the oral stage, during which sensuous pleasure comes mostly from taking things into the mouth.
By contrast, the dynamic theories have to do with processes that take place, or can take place, over a short span of time or at least within the same stage of a person's life. Under this heading we have, for example, the theory of defense mechanisms, according to which the person will defend himself against dangerous impulses by various devices—going to the other extreme (reaction formation), attributing the impulses to someone else (projection), and so on. One of the reasons that the distinction between developmental and dynamic theories is important is that many of the philosophical difficulties raised about psychoanalytic theory center on the notion of unconscious psychic processes, and such processes are more central in dynamic than in developmental theories.
In order to have something fairly definite to work with, let us take the following to be an oversimplified formulation of the psychoanalytic theory of psychic conflict, which is basic to all the dynamic theories.
(1) When it is very painful for a person to be aware of the fact that he has a certain desire, he represses it (prevents it from becoming conscious). The pain may stem from a severe conflict between the desire and the person's standards for himself, from fear of the consequences of attempts to satisfy the desire, or from both.
(2) Repressed psychic material exhibits primitive, infantile features. These include the lack of sharp distinctions, which is in turn conducive to the formation of strong associations between a certain desire and many other, often irrelevant, things and a tolerance for lack of realism and for incompatibility of one's desires and thoughts.
(3) A repressed desire (which continues to exist as a desire) can be partially satisfied by happenings, in actual occurrence or in fantasy, which are associated with the object of the desire.
(4) When the substitute satisfactions themselves arouse too much anxiety, the person seeks to ward them off, often in equally derivative ways.
This basic theory is then applied to the explanation of dreams, slips of the tongue, and neurotic symptoms by studying the ways in which such phenomena constitute substitute satisfactions of repressed desires and/or defenses against such satisfactions.
An illustration of these ideas is presented by Sigmund Freud in Lecture 17 of his General Introduction to Psychoanalysis. A girl has, for obvious reasons, repressed a strong desire for sexual intercourse with her father. In the unconscious, various things happen to this desire and the ideas involved in it. The dread of carrying out the act generalizes to a dread of sexual activity of any sort. An association is formed between sexual intercourse and breaking a vase. The bolster at the back of the bed is pictured as the girl's father and the back of the bed as her mother. The pressure of this repressed material becomes so great that the girl develops a compulsion to go through an elaborate ritual before going to sleep at night. She arranges the vases in her room so that breakage is impossible, thus symbolically guarding against sexual intercourse, and she takes care lest the bolster touch the back of her bed, thus achieving a substitute satisfaction for her desire to keep her father and mother apart.
Some of the philosophical objections to psychoanalytic theory can easily be shown to have little or no force. For instance, some philosophers object that the theory postulates unobservable entities; others believe that it is self-contradictory to speak of unconscious mental processes, for what is mental is, by definition, conscious.
In answer to the first objection, it can be pointed out that this practice is common in the most respectable parts of science. Electromagnetic fields and energy quanta are as unobservable as unconscious fantasy. They are, nonetheless, scientifically legitimate because of the functions performed by the theories embodying them, a point to which we shall return. In answer to the second objection, it may be admitted that psychoanalytic theory involves some stretching of such terms as "desire" and "thought" (as in the unconscious thoughts believed to underlie the conscious content of a dream). But, again, this is standard practice in scientific theorizing. The submicroscopic particles postulated in the kinetic theory of gases are modeled on familiar physical objects, like baseballs, except that they lack some of the properties of baseballs, like color and texture, and they possess perfect elasticity. One may as well say that it is a contradiction to speak of physical particles that have no color. Difference from familiar concepts is not in itself fatal. Again, the crucial question is what can be done with the concepts thus derived.
The serious difficulties emerge when we try to determine whether psychoanalytic concepts have the kind of status that is required for scientific validity and fruitfulness. This problem has two closely related parts. (1) Do psychoanalytic terms have any empirical significance, and if they do, how can it be exhibited? (2) How can theoretical principles couched in these terms be put to an empirical test? These questions become two sides of the same coin if we make certain assumptions that are widely shared by contemporary philosophers of science. First, a term has the kind of semantic status required for science if and only if statements in which it figures have implications for what would be experienced under certain circumstances. Second, one brings out a term's empirical or scientific significance, as contrasted with its pictorial associations, by tracing out such implications. Third, it is only if statements have such implications that they can be put to an empirical test. Given these assumptions, we can deal with the two questions simultaneously. By showing how statements involving the term repress give rise to implications of a sort that make an empirical test possible, we will at the same time be showing what scientific significance the term has over and above any of its pictorial associations—for example, a man firmly clamping a lid down on a pot of molten metal. With this equivalence in mind, the following discussion will be explicitly directed to the second question: How can the theoretical principles of psychoanalysis be empirically tested?
There is a commonly accepted doctrine, largely derived from a consideration of physics, according to which a theory involving unobservables gets empirical significance by virtue of the fact that it, together with subsidiary assumptions, implies various general lawlike hypotheses that can be directly tested empirically. In this way the theory can be assessed in terms of the extent to which it succeeds in explaining and unifying a variety of lower-level laws that have been empirically confirmed and, on the negative side, the extent to which it does not imply lower-level hypotheses that have been empirically disconfirmed. The Bohr theory of atomic structure, which represents an atom as a sort of miniature solar system with electrons revolving in orbits around the nucleus, cannot be tested directly, for an individual atom cannot be observed. However, from the theory we can derive a variety of testable hypotheses—for instance, those concerning the constitution of the spectrum of the light emitted from a given element.
Deriving Treatable Hypotheses
One might well expect to have difficulty deriving testable hypotheses from psychoanalytic theory. The theory represents the postulated unconscious processes mediating between events that are accessible to either introspection or observation, just as do unobservable processes within the atom in the Bohr theory.
In a typical sequence we start with conscious Oedipal desires in a child. Tentative attempts at satisfaction of the desires are met with violent opposition, and as a result the child builds up strong fear and/or horror of the realization of the Oedipal desires. Thus far, everything is, in principle, directly accessible to one or more observers. Then, according to the theory, the complex of desires, fears, and guilt is repressed, whereupon it undergoes various transformations, the exact nature of which is influenced by things that happen to the person, these things again being directly observable. In particular, the associations formed in the unconscious are largely determined by conscious experiences of the person. Finally, the unconscious complex is manifested in various ways—dreams, memory failures, slips of the tongue, compulsions, obsessions, psychosomatic illnesses—all of which are again accessible to experience. This being the case, one would suppose that the theory would yield general hypotheses to the effect that whenever strong desires of a certain kind are met with strong internal and/or external opposition, then (perhaps with the further assumption of certain kinds of intervening experiences) abnormal symptoms of certain kinds will be forthcoming. In other words, since unconscious psychic processes are supposed to provide connecting links between observables, a theory about them should imply that certain antecedent observables would lead to certain consequent observables.
In fact, however, we find little of this. Some attempts have been made to derive hypotheses about statistical distributions from parts of the theory. For example, the theory of dreams holds that dreams partially satisfy repressed desires by representing them as satisfied. It would follow from this that if a group of people were prevented from dreaming for several nights, they would then show a higher average level of tension than a control group. This hypothesis has been tested, using eyeball movement as a criterion of the occurrence of dreams. Most efforts of this sort have stemmed from relatively peripheral components of the theory; in particular, virtually nothing has been done to derive testable hypotheses specifying sufficient conditions for the occurrence of abnormal symptoms. It is only if this were done that the theory could be used for the prediction of such phenomena. Perhaps this is because of the psychoanalyst's preoccupation with the treatment of particular cases rather than with controlled testing of general hypotheses.
There are other features of the situation that also make the formulation of testable hypotheses extraordinarily difficult. Psychoanalytic theory has not been developed to the point where one can give sufficient conditions for one outcome rather than another even on the theoretical level of unconscious processes. Repression is said to occur when a desire arouses great anxiety, but just how much anxiety is required? Obviously, the amount is crucial, but the measurement problem has yet to be solved. Again, given a certain level of anxiety aroused by Oedipal desires, repression is not the only possible outcome. There might, instead, be a regression to the oral or anal phase, or the libido might be redirected into homosexual channels. There are some suggestions about what makes the difference—for example, if one never fully outgrew an earlier stage, this makes regression more likely. But at present this is all rather loose.
Moreover, once repression has occurred, the repressed material may develop in a great many different ways. The fear of sexual contact with the mother may or may not generalize, and if it does, it may generalize along various dimensions. Thus, the person may develop a dread of sexual contact with anyone or only with anyone who is like his mother in some respect. A part of the complex may come to be associated with things that have little or no intrinsic connection with it, as the girl in the example cited above formed an association between sexual intercourse and the breaking of a vase. It may well seem impossible to develop principles that would take into account all the determinants of unconscious trains of thought in a way that makes possible, in principle, the prediction of such associations. This impression is reinforced by the fact that these associations are often powerfully influenced by the person's external experiences, which could not be predicted on the basis of psychological facts about him. Thus, in the above example the girl had once broken a vase and cut her finger, which had bled profusely, an incident that then was associated in her mind with the bleeding accompanying defloration.
But even if connections were strong on the level of unconscious processes, there would still remain the job of formulating sufficient conditions for the occurrence of the ultimate facts to be explained. One and the same unconscious complex, given our present powers of discrimination, may issue in a phobia, hysterical paralysis or anesthesia, obsessive concern over bodily symptoms, or a generalized feeling of unworthiness, to mention only a few possibilities. No doubt the choice of symptom is due to other factors, but the problem has not been investigated sufficiently to yield even promising general hypotheses.
Background for Clinical Interpretation
In view of the extreme difficulty of empirically verifying psychoanalytic theory, one might ask why it should be regarded as anything other than an imaginatively satisfying fantasy. Why does it seem to have an empirical foundation? The answer is that it has significant connections with empirical facts but not connections of the sort insisted on by philosophers of science who take their models from physical theory. Psychoanalytic theory has grown out of the clinical treatment of neurotics, and in that context it has the function of providing suggestions for the interpretation of particular cases. Thus, if we are dealing with a compulsion neurosis, the theory tells us that compulsive behavior simultaneously provides substitute satisfactions for repressed desires (through the realization of states of affairs unconsciously associated with the realization of the desires) and guards against the arousal and/or satisfaction of the desire. (See the clinical case described above.)
Furthermore, the theory tells us what kinds of desires are most often repressed—incestuous, homosexual, aggressive. Also, psychoanalytic theory is associated with certain techniques—the analysis of dreams, of free associations, and of reactions to the analyst—for ferreting out repressed material in particular cases. Thus, the theory provides leads for the analyst. Insofar as it has this function rather than that of explaining and unifying testable hypotheses about the conditions under which, in general, we will get one outcome rather than another, it is no defect that it is largely made up of rather loose statements about what can happen, given certain conditions, and what can be responsible for a given symptom. In explaining an event, E, that has already occurred, our needs are simpler than when we are engaged in predicting or establishing general principles. In retrospective explanation we can take advantage of our knowledge that E has already occurred; we are reasoning backward to its sources. Therefore, provided we have a list of possible causes and some way of telling which of these are present, we have something to go on, even if each statement of possible cause is only to the effect that C can result in E. If we were setting out to predict, however, we would need a further specification of the conditions under which C will in fact lead to E. The knowledge that an unconscious desire for and fear of intercourse with the father, plus an association between intercourse and breaking a vase, can lead to a compulsive tendency to arrange vases so as to minimize chances of breakage is general knowledge of a sort, but not of the sort exemplified by the Newtonian theory of gravitation, in which the general principles enable one to predict one state of the system from any other state of the system.
Thus, one can say that psychoanalytic theory, given the way it has developed up to now, makes contact with empirical reality through being used as a basis for explanations of certain kinds of observable occurrences and that the theory receives empirical support to the extent that such explanations are adequate. To many methodologists this situation is profoundly unsatisfying. If a theory yields predictively confirmed hypotheses, we have a strong indication that contact with something real has been made, for by thinking in these terms, we have succeeded in anticipating the course of nature. But if the theory can provide only suggestions for retrospective explanations, it is not so clear what this shows. More specifically, many have suspected that the success of psychoanalysts in devising explanations of their patients' symptoms is more a function of the analysts' ingenuity than of the soundness of their theory. It is easy to get the impression that a plausible explanation in psychoanalytic terms could be framed for any behavior, no matter what the facts. If it is not a reaction formation from overattachment to mother, then it is a projection of a self-directed death wish, and so on.
Adequacy of Clinical Interpretations
Clearly, what is needed is a set of objective criteria for the adequacy of an explanation in terms of unconscious psychic factors, criteria that would permit us to assess a proposed explanation on some grounds other than the way it seems to make sense of the phenomena. If and only if such criteria can be formulated can explanations of particular cases provide any empirical basis for the theory.
Within the limits of this article, we can only touch briefly on the problems involved in formulating and defending such criteria. The problems fall into three groups.
status of the data
Questions have often been raised about the status of the ultimate data to which the psychoanalyst appeals in justifying an interpretation. These consist of the behavior of the patient, verbal and otherwise, in therapeutic sessions. Criticisms have been of three sorts.
First, the data actually presented are a small sample of all the behavior engaged in by the patient in the presence of the analyst. We are almost never given any reason for supposing that this is a representative sample, that the analyst has not, perhaps unconsciously, selected those items that best support his hypothesis.
Second, a given patient is rarely, if ever, compared with controls who do not have his difficulties. Without this we cannot show that the data cited have bearing on the abnormalities to be explained. For example, if almost anyone would get annoyed when the analyst acts bored with the session, then the fact that patient A does so is not likely to reveal anything that is responsible for any idiosyncrasy of his.
Third, the analyst may often be guilty of contaminating the data through, perhaps unconsciously, tipping the patient off about his interpretation, thus implicitly inviting the patient to produce associations that will support that interpretation.
These are serious problems in data collection and assessment, and they will have to be solved if psychoanalysis is to become more respectable scientifically. But since it seems in principle possible to overcome them, they are less crucial for the logical status of the theory than problems in the other groups.
An explanation of E in terms of C is not warranted unless C actually exists. What objective tests are there for the actual existence of the unconscious psychic factors appealed to by the analyst? Analysts regularly use a number of detection procedures.
Among the things they consider significant are the following: (1) Patterns of behavior that are as they would be if A had a desire of which he is not conscious. For example, a seventeen-year-old girl devotes a great deal of time and energy to the small children of a youngish widower friend of the family, though she is not aware of being in love with him. (2) Patterns of feeling that have the same status. In the same example, the girl gets very depressed when the widower does not send her a birthday present. (3) Analysis of dreams and of free associations. Such analysis proceeds in a rather devious fashion and cannot be illustrated briefly. It is based on the principle that unconscious complexes influence conscious thought and fantasy, including dreaming, by producing relatively safe conscious derivatives of these complexes. (4) Final realization by the patient, after treatment, that he had the desire in question all along.
The inferences involved in the use of these procedures are extremely complex, and it is difficult to say just how conclusively anyone has ever demonstrated the existence of certain unconscious material in a given case. It is worth noting that the use of (3) and (4), unlike (1) and (2), requires the assumption of certain parts of the theory. Thus, for example, we cannot take dreams to reveal unconscious desires in the way analysts do unless we assume that dreams are formed in the manner postulated by the theory. This means that insofar as explanations that are supported in part by dream interpretation are adduced in support of the theory, we are going round in a circle.
unconscious complexes and symptoms
The most difficult problem is that of showing that a given unconscious complex is responsible for certain symptoms. Granted that the girl does have a repressed desire for and dread of sexual intercourse with her father, why should we suppose that this is what led her to develop a compulsive tendency to arrange the vases in her room in a certain way before retiring? In order to answer this question, we shall have to decide what kind of explanation this is supposed to be. Freud often gives the impression that it has the ordinary pattern of an "in-order-to" explanation ("I went into the kitchen in order to get a bottle of beer" or "I went into the kitchen because I wanted a bottle of beer"), except that here the want is unconscious. But the ordinary "in-order-to" explanation carries the assumption that the agent believes that the action in question is, or may be, instrumental in the satisfaction of the want in question. Can we say that the girl unconsciously believed that preventing the vases from breaking would be instrumental in preventing intercourse with her father? A strange belief, but Freud did say that the unconscious is quite illogical. Or should we say, rather, that no belief is involved here but only an association between breaking a vase and intercourse? However this issue is resolved, this assimilation will not help us to justify the explanation, for the fundamental method of justifying an ordinary "in-order-to" explanation—getting a sincere report by the agent of why he did what he did—is not available here.
Freud might claim that an analogue is available—the realization by the patient, after treatment, that that was why she had to arrange the vases as she did. However, if one rests the adequacy of the explanation on the patient's posttherapeutic insight, he leaves himself open to the charge of undue influence on the source of data. Moreover, circularity comes up again, for if the patient came to have this conviction as a result of being presented with this explanation under hypnosis, this would not count in favor of the explanation. Only insight that comes after certain kinds of therapeutic interactions is relevant, and the claim that insight produced in that way is valid depends on the psychoanalytic theory about the effects that can be expected from psychoanalytic therapy. Thus, there are difficulties in construing the explanation on the model of "I went to the kitchen because I wanted a bottle of beer." On the other hand, if we take as our model an everyday explanation in terms of physical causation, like "The window broke because a baseball hit it," we will have to support it by reference to general principles to the effect that factors of the sort cited have results of the kind we are seeking to explain. And the absence of such tested generalizations in psychoanalysis has already been noted.
Thus, it would seem that before psychoanalytic theory can enjoy a firm empirical foundation, its practitioners must either develop explicit and workable objective criteria for the adequacy of interpretations of clinical phenomena in terms of unconscious factors, or do more to derive testable general hypotheses from the theory, or do both.
The basic expositions of psychoanalytic theory are to be found in the works of Sigmund Freud. See especially his General Introduction to Psychoanalysis (New York: Garden City, 1949), New Introductory Lectures on Psychoanalysis (New York: Norton, 1933), The Interpretation of Dreams (New York: Macmillan, 1937), The Ego and the Id (London: Hogarth Press, 1927), Inhibitions, Symptoms, and Anxiety (London, 1961), and the theoretical essays in Vol. IV of the Collected Papers (London: Hogarth Press, 1948).
Important later additions include Anna Freud, The Ego and the Mechanisms of Defence (London: L. and Virginia Woolf at the Hogarth Press, and the Institute of Psycho-analysis, 1937), and the new developments in ego psychology to be found in two articles by Heinz Hartmann, "Comments on the Psychoanalytic Theory of the Ego," in The Psychoanalytic Study of the Child, Vol. V (New Haven, CT: Yale University Press, 1950), and "The Mutual Influences in the Development of the Ego and Id," ibid., Vol. VII (New Haven, CT: Yale University Press, 1952).
Useful expositions are to be found in R. L. Munroe, Schools of Psychoanalytic Thought (New York: Dryden, 1955); Otto Fenichel, The Psychoanalytic Theory of Neurosis (New York: Norton, 1945); C. S. Hall, A Primer of Freudian Psychology (Cleveland, OH: World Publishing, 1954); and David Rapaport, "The Structure of Psychoanalytic Theory," in Psychology: A Study of a Science, edited by Sigmund Koch, Vol. III (New York: McGraw-Hill, 1959).
Methodological problems are explored in Psychoanalysis, Scientific Method, and Philosophy, edited by Sidney Hook (New York: New York University Press, 1959); B. F. Skinner, "Critique of Psychoanalytic Concepts and Theories," in Minnesota Studies in the Philosophy of Science, Vol. I (Minneapolis: University of Minnesota Press, 1956); Antony Flew, "Motives and the Unconscious," ibid.; Gustav Bergmann, "Psychoanalysis and Experimental Psychology," in Psychological Theory, edited by M. H. Marx (New York: Macmillan, 1951); Ch. 3 of R. S. Peters, The Concept of Motivation (London: Routledge and K. Paul, 1958); a series of essays reprinted in Philosophy and Analysis, edited by Margaret Macdonald (Oxford, 1954)—Stephen Toulmin's "The Logical Status of Psycho-analysis," Antony Flew's "Psycho-analytic Explanations," and R. S. Peters's "Cause, Cure, and Motive"; and B. A. Farrell's articles "The Criteria for a Psychoanalytic Interpretation," in PAS, supp., 36 (1962)—reprinted in Essays in Philosophical Psychology, edited by D. F. Gustafson (Garden City, NY: Anchor, 1964), pp. 299–323—and "Can Psychoanalysis Be Refuted?," in Inquiry 4 (1961): 16–36.
Attempts to derive testable general hypotheses from psychoanalytic theory are canvassed in R. R. Sears, Survey of Objective Studies of Psychoanalytic Concepts (New York: Social Science Research Council, 1943).
William P. Alston (1967)