Psychoanalysis as Science

by Norman N. Holland

May 22, 2004


Current objections to psychoanalysis as untestable and unscientific ignore two facts. First, a large body of experimental evidence has tested psychoanalytic ideas, confirming some and not others. Second, psychoanalysis itself, while it does not usually use experimentation, does use holistic method. This is a procedure widely deployed in both the social sciences and the "hard" sciences. Recognizing the holistic nature of psychoanalytic ideas and therapy suggests some kinds of interpretation are more valid than others. It also shows that the debate whether psychoanalysis is a science or a hermeneutic, merely "literary," states the issue falsely. Recognizing the holistic nature of psychoanalytic ideas and therapy suggests some kinds of interpretation are more valid than others. It also shows that the debate whether psychoanalysis is a science or merely "literary," rests on a false dichotomy.


I believe that today's low estimate of the validity of psychoanalysis stems from two mistaken ideas. First, critics of psychoanalysis like Frederick Crews often assert as proven facts that, one, psychoanalysis lacks any experimental confirmation and two, it is in fact untestable. It is merely "literary." "There is not a shred of scientific evidence for psychoanalysis" has become a kind of ceremonial utterance to be repeated in any psychological paper. Second, I believe that, by and large, philosophers and academic psychologists and, indeed, psychoanalysts fail to understand the situation of psychoanalytic method among the sciences. It is this second point that I want to address, but I am unwilling to ignore the first.

Psychology vs. psychoanalysis

About the time Freud began developing psychoanalysis, academic psychology began making itself more and more rigorous as a science. Freud, however, minimized the influence of academic, non-psychoanalytic psychology. Psychologists in turn regarded psychoanalysis as infra dig so far as their scientific agenda was concerned. Thus began a Hundred Years' War, not over yet. Psychologists see psychoanalysis as unscientific. Psychoanalysts see the psychologists' careful experiments as irrelevant to their clinical concerns.1 To be sure, in the forties and fifties, a number of leading psychologists, Gordon Allport, Henry A. Murray, John Dollard, Neal Miller, Robert Sears, and others tried to bring about a rapprochement. But their books and articles tended to show that psychoanalysis could be scientific enough for psychologists, not that it was. So does the recent and encyclopedic Interface of Psychoanalysis and Psychology.2

Thus the question of psychoanalysis' scientific status has nagged the discipline since the beginning. In fact, however, there is, considerable experimental or "empirical" confirmation of psychoanalysis. Unfortunately, these experiments testing psychoanalysis are mentioned rarely if at all in the Freud wars. The bashers or those who believe them simply repeat like a mantra, There is no scientific evidence for psychoanalysis, not could there be. Neither part of that statement holds water. Experimental psychologists have shown that psychoanalysis (at least early psychoanalysis) is not only testable and falsifiable, but it has been tested and to some extent confirmed, to some extent falsified.

Experimental evidence for psychoanalysis

Two impressive books by Seymour Fisher and Roger Greenberg and summarize some 2500 of these studies. Fisher and Greenberg speak of "the incredible amount of effort that has been invested in testing Freud's ideas fairly." In quality, they say, these studies "probably compare well with the accumulated studies relevant to any other major area of psychology."3 And, of course, Fisher and Greenberg are not the only ones looking at tests of Freud's ideas.

There are several earlier surveys of the experimental literature: Robert R. Sears' in 1943; Paul Kline's in 1972; Hartwig Dahl et al.'s studies of verbal protocols; Weiss and Sampson's study of alternative modes of therapy; and, recently and importantly, the annual series Empirical Studies of Psychoanalytic Theories, edited by Joseph M. Masling, ten volumes so far. Masling's first eight volumes contain 327 pages of references with an estimated 6300 citations, most of which deal with experimental testing of psychoanalytic ideas.4

Since this branch of writing on psychoanalysis is so widely ignored, it is perhaps worth a digression to summarize some of these writers' conclusions. Fisher and Greenberg define the range of their survey as the empirical testing of Freud's assertions. That is a considerable limitation, obviously, since psychoanalysis has changed so much since Freud's death. Within that range, however, they draw a number of conclusions. Dreams are transparent, just another form of thinking, not necessarily wish-fulfillments. But Freud's very general idea that the dream has adaptive functions does survive. Experimenters confirm the clusters of traits associated with orality and anality. The literature supports an oedipal stage, but finds that a good superego is likely to come from a loving relationship with the father rather than castration anxiety. There is no support for any of Freud's ideas on female development except the penis=baby equation. Homosexuals of both sexes tended to have negative father images. The literature confirms Freud's link of depression to parents who were disapproving and unnurturing, but not to early experiences of loss, although early loss gives recent losses more effect. There was strong confirmation of links between depression and oral fixation and between depression and tendencies to be self-critical and self-attacking. Freud's account of paranoia gets confirmation.

Importantly, the empirical literature does not confirm a correlation between insight, interpretations, and outcome. Insight does not equal cure. The literature says that psychotherapy is more complex than simply understanding one's problems or where they came from. Although psychoanalytic therapy resulted in more improvement than psychoanalysis, it did not show more effectiveness than other kinds of therapy. Neither psychoanalysis nor any other therapeutic technique shows the superiority Freud claimed for psychoanalysis, if one defines "improved" as symptom removal.

Like most in the psychoanalytic community, I am well aware that all forms of psychotherapy, the various "talking cures," can claim about the same percentages of improved outcomes and not a very high percentage at that. The claim that psychoanalysis provides a special cure, that symptoms will recur without (psychoanalytic) insight into their cause, has not held up. Therapies like behavior therapy, contingency management, or cognitive-behavior therapy all work.5

One line of research, however, has empirically compared two psychoanalytic theories of mental functioning and therapy. Joseph Weiss and Harold Sampson and others working at the Mount Zion Psychotherapy Group in San Francisco have compared two such theories, one close to Freud's early thinking, the other derived from later Freud and ego-psychology. The early "dynamic hypothesis" says that indications of pleasure and pain regulate the patients' defenses automatically, leaving the patients no control over, say, their repressions. The later "unconscious control hypothesis" states that a patient has some control over repressions (through unconscious ego). Weiss and Sampson tested these different hypotheses by transcribing the first 100 hours of one patient's treatment and using scorers and coders to categorize and count the moves and meanings the patient and the analyst were making. They confirmed the unconscious control hypothesis against the dynamic hypothesis.6 They thus showed that it is possible to test hypotheses about psychoanalytic therapy in a natural, unconstrained setting, in ways that are more subtle and particular than simply counting how many treatments achieved a measure of success and how many did not.

But I am not concerned here with the value or validity of psychoanalytic therapy except to the degree it tends to confirm psychoanalysis as a theory of mind. Can we consider today's psychoanalytic theory of mind "scientific"?

On this issue Westen's 1998 survey becomes cogent. What needs to be tested, Westen notes, is psychoanalysis as it is now, not as it was in Freud's day. In particular, Westen points to the advent of object-relations theory and the idea that the basic drives are not just sex and aggression but include a drive to relate to other persons.7

    Westen's survey sets out five general propositions that, he says, today's practitioners would agree are the foundational ideas of modern psychoanalysis. Then Westen then adduces the evidence (in some 350 references) that support these ideas.

    1. Much mental life, including thoughts, feelings, and motives is unconscious. Neurology supplies massive evidence of unconscious processes of cognition. Both neurologists and psychologists abundantly demonstrate unconscious emotional learning and defense mechanisms evoked by emotions, in particular, "avoidant" mechanisms like repression and denial against emotions that are unpleasant. As for motives, both these defenses and priming experients show unconscious motivations.

    2. Mental processes operate in parallel and, often, in conflict. Westen points to the similarity between psychoanalytic accounts of conflict like multiple functioning" and the PDP (parallel distributed processing) theories now gaining wide acceptance among neurologists. PDP uses Hebb's well-known hypothesis: neurons that fire together wire together. A PDP network models how competing inputs gradually change the weighting of various paths through the network until it arrives at a compromise output. Other experiments demonstrate that people act so as to compromise impulses of which they are unconscious.

    3. Stable personality patterns form in childhood and shape later relationships. Observational and longitudinal studies read forward to show adult personality traits beginning to crystallize in childhood and read backward to show that childhood experiences were formative for later personality and social functioning.

    4. Mental representations of the self, others, and relationships guide interactions with others and shape symptoms. Westen points to considerable research on attachment behaviors both in animals and humans; on people's transferring their feelings from significant others to unknown or fictional people; and on links between maternal separation or poor mothering or social support and depression or health in general.

    5. Personality development is not just learning regulation of sex and aggression (Freud's theory) but also moving from immature dependency to mature interdependency. Most evidence for Freud's theory comes from clinical experience, but laboratory studies do support an "oral" style, an "anal" style, and the Oedipus complex, although evidence linking these to childhood stages is much more tenuous. [As in Fisher and Greenberg's survey--nnh.]8

Another key figure in the experimental studies of psychoanalysis is Joseph Masling. He too has surveyed the huge literature testing psychoanalytic ideas.9 What in his view have not held up are Freud's ideas about women; claims about the psychosomatic origins of asthma, ulcers, allergies, and even myopia and astigmatism; autism as a disease caused by mothers; and the recurrence or substitution of symptoms without the supposed efficacy of insight. Against Freud's first idea of "the unconscious" as chaotic and disorganized, many, many experiments in cognitive science and infant development show that the early or unconscious mind has a lot of skills. Interestingly, Freud' s second, ego-psychological concept of unconscious processes, when "unconscious" ceased being a noun and became an adjective, posited unconscious ego skills that coincide with the demonstrations of cognitive science.

Masling points to other psychoanalytic ideas that have held up: unconscious processes (demonstrated by subliminal stimuli); the personality types; and reaction formation. His work showed, for example, that obese and alcoholic and stuttering patients all showed high oral responses on the Rorschach test, as did subjects who yielded to the majority in an Asch conformity experiment or showed greater arousal when left alone or with an unfriendly companion in a soundproof room. Further work showed that those who report many oral dependent images on the Rorschach test (based on explicit rules for scoring) were generally more sensitive to interpersonal cues.10 Masling has gone on to direct a number of dissertations and graduate student projects that lead to similar confirmations. For example, a long series of experiments showed that psychoanalytically charged messages such as NO ONE LOVES ME or MAMA IS LEAVING ME could lead to depressive responses and changes in behavior even though they were only presented subliminally.11

Other, unpublished experiments demonstrated reaction-formation. Specifically, animal rights advocates showed markedly more hostile responses when asked to apply to various criminals jail sentences, the death penalty, or monetary fines. Homophobics showed greater interest in pictures of male nudes and male-male sexual activity than men neutral on the subject.12

Masling and his co-editors have contributed most importantly, however, by starting, editing, and supervising the book series, Empirical Studies of Psychoanalytic Theory. These ten volumes have put in one place experimental tests and modifications of psychoanalytic theory by many distinguished psychologists and psychiatrists: explorations through subliminal studies, by analyses of transcripts of psychoanalytic sessions, explorations of the psychodynamics of gender and gender role; of psychological factors in health and illness; unconscious feelings and motives and thought processes; of development in infancy and across the lifespan; of object-relations theory; of psychopathology (bulimia, depression, schizophrenia, etc.); of dreaming; of wishes; of the therapeutic alliance, and on and on. These volumes offer a wealth of empirical support and modification for an extraordinary range of particular psychoanalytic hypotheses.13

In short, Westen plus Masling plus Fisher and Greenberg plus the earlier summarizers of experimental work plus the thousands of experimental studies themselves constitute a very large body of empirical evidence. Yet, as the squibs and squillets fly in the Freud wars, neither psychoanalysis' detractors nor its supporters often mention these experimental results. Psychoanalysts do not find experiments relevant to their practice. Non-analytic psychologists pay little heed, because this kind of evidence runs counter to the deep-seated prejudice against psychoanalysis among academic psychologists, a prejudice embedded in textbooks and indoctrinated in beginning psychology courses. And psychologists rightly point to the unreliability of clinical evidence.

Clinical evidence

The clinical setting puts up two insuperable obstacles toward public, repeatable data. First, the demands of confidentiality require that even if analysts publish data, they have to alter them so as to make the patient unrecognizable--even to the patient. Worse, as Donald Spence points out, the custom has emerged among psychoanalytic authors of simply alluding to facts or the patients' words, making it impossible to base reasoned argument on such papers.14 Someone reading this evidence cannot tell what has been changed or how the changes might alter the interpretation offered. You have to trust the analyst-scientist. And, with psychoanalysis, the reporter's vagaries of memory, repression, and motive all affect case history material.

To be sure, this is true in all the sciences, where published conclusions always depend on the honesty and reliability of the scientist who reports them. Bias is harder to detect in the experimental situation, because of the seeming objectivity of the various procedures. It is, however, no less a problem, as the long line of Rosenthal experimentation made clear.15 Experimenters' prejudgments affect outcomes. But in public sciences, subsequent scientists can test the innovating scientist's claims. Psychoanalysis is different, and that leads to the second problem with psychoanalysis' claims to validity based on clinical evidence.

Each patient is unique and therefore each analysis or therapy is, too. The detractors make an important point here, but one needs to qualify it by recognizing that other branches of social and "hard" sciences have somewhat the same problem. Other sciences cannot conduct repeatable experiments. Sciences like geology, astronomy, oceanology, meteorology, ecology, large parts of biology, and even areas within physics do not lend themselves to repeatable experiments. But no one doubts that these sciences are sciences. In judging psychoanalysis as a science, then, one needs to consider the range of scientists' methodologies, initially in the social sciences. One has to get over the idea that the only "science" with a claim to validity is experimental.

The variety of social science methods

In thinking about psychoanalysis' claims, the most helpful books I have found are Paul Diesing's Patterns of Discovery in the Social Sciences (1971) and How Does Social Science Work? (1991).16 Diesing is a down-to-earth philosopher. He does not fall into the mistake of starting with an abstract or a priori definition of "science" and question whether this or that activity fits into it. Diesing, by contrast, read papers and visited laboratories to see what social scientists actually were doing. From this observation of social scientists at work, he singled out four methods in common use in the social sciences.


Statistical survey research

Formal methods

Participant-observer or clinical or holistic methods

Very briefly, experimentation works with variables (natural occurrences that offer measurable variations). The experimenter seeks numerical correlations among independent, dependent, and controlled variables. We see these techniques commonly among academic psychologists, for example, those testing psychoanalytic theories. We generally regard experimentation as the most rigorous and "scientific" of the methods in the social sciences, but it has problems.

For one thing, because each experimenter defines variables and methods very precisely, experimenters have difficulty in generalizing results beyond the particular experimental method used.17 Also, most psychological experiments treat some stimulus as the independent variable and the response as the dependent variable. The individuals in between the stimulus and the response get averaged out by various statistical techniques. Thus, the model of one variable depending on another more or less locks the experimenter into a stimulus-response model of the human being. That commitment tends to model us human beings as though we were born blank slates on which the environment writes its influences. Thinkers about scientific method like Chomsky and Pinker have raised serious doubts about that "Standard Social Science Model."

Matthew Erdelyi contrasts clinical and experimental evidence. The clinical method can reveal complex processes--that is its strength. Its weakness is its looseness of method. It rests on interpretation, which is a "subjective" matter. The therapist's influence affects what happens. The therapist's reports are also necessarily personal and subjective. By contrast, the experimental method is both rigorous and public. The many loose influences in the clinical approach get eliminated or controlled and quantified. But the method has a weakness in its inability to deal with complex processes--like the origins in childhood of a repression. In the debate about psychoanalysis, experimental psychologists sometimes delcare they cannot confirm this or that complex psychoanalytic concept, say, repression. What is striking, Erdelyi points out, is that they then dismiss the concept instead of holding the method accountable for the failure.18

Diesing's second category, survey method, serves to study larger populations than the experimenter can, substituting statistical for experimental controls. Using survey methods, social scientists collect data and formulate general if-then principles for organizing the data. One finds survey method used by sociologists, political scientists, marketers, or public health officials, for example.19 One even finds it with particle physicists. For instance, the immense Super-Kamiokande neutrino detector in Japan had 11,000 big photodetectors simply waiting and watching and counting any neutrino interactions generated in the sun and by cosmic rays. That is survey research.20

A social scientist using formal methods will develop a model--nowdays it is often a computer model--and compare its behavior to the real world's. Among the social sciences, one sees this method most clearly in linguistics and economics, but also in international politics and sociology.21

Holistic method is another mode that social scientists commonly use, and a great many philosophers and some scientists have written about holistic methods: Susan James, D. C. Philips, Christopher Peacocke, Karl Pribram Ulrich Gahde, and many others. Much of this discussion, however, has addressed abstract questions such as whether the whole can be greater than the sum of its parts or what is the relation between subjective and objective. Diesing's focus on just the method the social scientists use has proved the most useful for my purposes.

The holistic researcher gathers data into a coherent mass, using themes to explain interrelations within the data. This method serves best to study unique systems that cannot readily be multiplied for experimental or survey manipulations: a patient, a community, a family, or a corporation, for example. Holistic method is used by archaeologists and anthropologists, clinical psychologists, but above all by psychoanalysts.22

Anthropologist Claude Lévi-Strauss, for example, offers an elaborate discussion of various customs of the Bororo Indians, the layout of the village, dances, theories of the dead, myths, clothing, and so on. He then pulls them all together into "one regulation [that] takes precedence over all others."23 As for archaeologists, Freud's description of their method is quite accurate and clearly shows the holistic method. Finding ruins, the explorers see what lies exposed, question the inhabitants, perhaps clear away rubbish, and uncover what is buried. Finding inscriptions, they decipher them. If they are successful, "The discoveries are self-explanatory."24

Holistic method

Holistic method tries to interrelate all aspects of a unique system into a whole that will "make sense." Often, the researcher hypothesizes an organic unity to the system, but that is not necessary. Holistic research typically proceeds through four stages.25

First, the researcher gathers data. These data consist of raw facts, as free as possible of confining hypotheses (unlike the experimenter for whom hypothesis is essential). The holistic researcher focuses on the subject-matter, which can be virtually anything. In, say, the case of anthropology, it could be taboos, stories, customs, housing arrangements, and so on. If the unique system is a human being, the data will consist of the "observable behaviors" prized by experimental psychologists.

Where an experimenter is likely to begin with a carefully worked out procedure, for the holist, the data are the important thing. Thus, to an experimenter, the holist's methods of gathering data can seem haphazard or slapdash: interviews, informants, everyday observations, writings, anything that might be part of a total picture. But for the holist, the issue at this stage is not how the data were gathered, but how reliable are the data and how reliable their source.26

Typically, holistic researchers are themselves involved with the data. They become participant-observers, like the anthropologist by the Navajos' campfire or the psychoanalyst behind the couch. Holistic researchers acknowledge, indeed use, their involvement in the unique situation they are studying. In fact, the researcher's reactions often become part of the data, like the countertransference in an analysis. Because of the holistic worker's presence in the material and methods, the problem of bias stands out, even at the data-gathering stage.

In the second stage, the researcher groups the data into themes, each theme representing a certain uniformity in the material. A theme is the lowest level of interpretive statement in a case study. By formulating a theme, the experimenter asserts simply that a certain uniformity exists in the data, some sort of clustering or syndrome or repetition or contrast. A theme asserts that the uniformity will continue to appear in new data. One tests a theme by seeing whether further instances do appear. If only a few turn up, one stops paying attention to that theme. If more than one or two negative instances appear, one discards it. A theme, says Diesing, is like a pawn, easily gotten and easily discarded.27

Holistic method, it seems to me (as to Diesing28), corresponds precisely with the methods of psychoanalysis. First, one collects data in the form of the words of free association. Second, one formulates particular wordings into themes.

Consider a particularly clear example, Freud's analysis of a young academic's forgetting the world aliquis in a quotation from Virgil.29 (The allegation that this slip is Freud's own is not relevant here, since I am addressing only the method of interpretation.) Freud listened to the associations (the young man's or his own) and grouped them into themes (conveniently italicized in the Standard Edition): relics, liquefying, fluidity, fluid, saints, calendar saints, "the blood that starts to flow on a particular day, the disturbance when the event fails to take place, the open threats that the miracle must be vouchsafed, or else." He concluded that the young man was worried because his mistress had begun missing her periods. He had arrived holistically at an overall theme that unified all the various associations he had previously grouped into themes. That overall theme constituted a narrative explanation of the original slip, a holistic explanation that brought all the details together around that center.

Freud's early dream analyses also build on holistic reasoning. In Freud's first published dream analysis, for example, the Irma dream, he began by spelling out his associations (his data). In doing so, he indicated a variety of recurring themes: people recalcitrant to treatment; use of cocaine; lack of medical conscientiousness; organic illnesses--dysentery, diphtheria--as opposed to hysteria; smells of organic chemicals; trimethylamin and sexuality; "Otto's" thoughtlessness; Freud's own conscientiousness. Finally, he concluded: "They could all be collected into a single group of ideas and labelled, as it were, concern about my own and other people's health--professional conscientiousness."30 This purely exemplifies holistic reasoning.

In interpreting the table d'hôte dream, Freud says, "By following the associations which arose from the separate elements of the dream divorced from their context, I arrived at a number of thoughts and recollections." These associations, thoughts, and recollections make up data for a holistic interpretation into themes: getting too little or the worst of the bargain; getting something without paying for it; being in debt; being guilty. These in turn he focuses into a single unifying idea--he calls it "a single nodal point": a wish for a love that would call for no expenditure.31

These first stages, listening to the associations and interpreting them for themes and then a single unifying theme, constitute what Robert Waelder called "clinical interpretation." They are, he wrote, "entirely indispensable, not only for the practice of psychoanalysis but for any degree of understanding of it."32 Interestingly, Levine and Luborsky were able in experiments to achieve inter-judge reliability for such interpretations.33 Similarly, Donald Spence achieved inter-judge agreement on a low-level interpretation of a dream.34

This holistic procedure is the rock-bottom solid core on which psychoanalysis builds. First, the researcher-analyst takes as data the patient's free associations. Second, analysts sense themes for the data. Next, they group themes into an overall configuration or model or narrative, some systematic description of the case that will constitute an explanation of the themes and data arrived at so far. The tests at this third stage are: How many themes are included in the configuration and how many are left out? Then, how coherent or well-organized does the model make the themes.35

In this third phase, the analyst proceeds from the themes into a description of the dream, parapraxis, symptom, or even a whole character or identity as a coherent system. In the Wolf Man case, for example, Freud described the analyst's experience when he has arrived at that central theme or issue for a patient, as the feeling of "how, after a certain phase of the treatment, everything seemed to converge upon it, and how later, in the synthesis, the most various and remarkable results radiated out from it; how not only the large problems but the smallest peculiarities in the history of the case were cleared up by this single assumption."36 As he made clear, "everything" was to include every word and every detail, no matter how small.37

To be sure, in his later writings, as Freud became more at ease with the method, he tended to move abruptly to his conclusion without detailing the separate steps of associations, themes, and unification. He did this more and more in symbolic interpretations and his general theory or metapsychology.38 Nevertheless, he had at the outset established what seems to me the basic method of psychoanalysis.

You can see it, for example, in the many quite personal associations recorded by Theodore Reik in his remarkable collection, Listening with the Third Ear. His daughter goes out on a date, and his thoughts ramble from the suitor he fearfully imagines, an alcoholic playboy, to an alcoholic patient of that kind, to an anti-semitic remark the patient made, from there to speculations on Shylock and his daughter Jessica in The Merchant of Venice, to a theory about the play, to a succession of names, Jones, Jericho, Jephthah, Jehovah, Jesus . . . All come together around "jealousy of my daughter, also possessiveness, fury against the unknown young man who will take her away from me," a fury imaged in the trial scene of Merchant.39

A contemporary analyst, Christopher Bollas, tells how he listens to a patient's associations "until we reach a revelation -- a point when suddenly we are struck by a pattern of thought." "Patients have these organised inner compositions which, like magnets, attract further impressions, and serve as the core of the self's creative articulation of the inner compositions themselves."40 Bollas goes on in this and other writings to tell of taking the third step in a holistic analysis. He assembles the various themes into a sense of the style of a certain person, the "idiom" as he calls it, very much as one identifies the style of a poet. "Each of us at birth is equipped with a unique idiom of psychic organization that constitutes the core of our self, and then in the subsequent first years of our lives we become our parents' child, instructed . . . in the family's way of being; we become a complex theory for being a self that the toddler does not think about but acquires operationally."41 This unthought theory becomes embodied in us, and therefore another, an analyst perhaps or simply a wise friend, can read it from outside us. Bollas is describing as "idiom" what I, following Lichtenstein,42 call an "identity theme" or "style."43

In short, both early and contemporary psychoanalysts work by holistic method. First, one collects data in the form of the words of free association by listening in a special way, as described by Makari and Shapiro.44 Analytic listening, they say, involves reversing the usual relation of signal and noise. The analyst, they say, "listens for noises that signify in psychoanalytic terms; slips, metaphors, similes, archaic or idiosyncratic usages, repeatedly used words, words used in the wrong context; he [or she] listens for odd syntax, passive voice, obsessively formed sentences; in short, he [or she] listens for the poetics that structure another's signifying repertoire."45 This is Bollas' "idiom." For example, they give the example of a patient, "who referred to his nose and his penis as `the' nose and `the' penis, and was thereby enacting the isolation from urgency he felt, and the objectification of his body as things, because he feared confronting his dangerous sexual thoughts and feelings."46 Similarly, Donald Spence successfully used the co-occurrence of pronouns referring to the analyst and pronouns referring to the patient as a measure of the likelihood the analyst would intervene.47

Listening to the linguistic surface is precisely Freud's way in those early dreams and cases and later. As the poet H.D. wrote of her 1933 analysis with Freud, "I told Freud a long tale yesterday, full of important details. When I had used up the hour he said, `I can tell from the way you speak, that you are hiding things. So I did not have to listen. You ran your articles together. You did not speak clearly.' What a sell! Was I mad?????????????????"48 In this mode, one does not read through the patient's words to a meaning so much as one listens to the choice of words--diction. The linguistic surface is the data, because style is character, as in the many psychoanalytic readings of writers' styles.49

Second, one formulates those details of speech into themes. Third, one proceeds to combine those themes into a pattern of repetitions and contrasts that applies to all one's data. This is a "pattern explanation" of the dream, parapraxis, symptom, or even a whole character or identity as a coherent system. In effect, the holist is recounting repetitions, hence a pattern. By describing a pattern, the holist brings disparate behaviors together into a single coherent picture. The model or configuration or narrative that unites the themes constitutes a pattern explanation. Fourth, one checks the explanation against new data, refining and correcting it.50

Obviously, this procedure risks becoming a self-fulfilling prophecy. Freudians will see Freudian patterns, Jungians Jungian, Reichians Reichian, and so on. Theory will drive interpretation (although it is well to remember that theory drives experiment, too). Here, as in other sciences, good results rest on the integrity of the scientist. Holistic researchers need to keep their minds as free of theory as possible when gathering data. They need simply to be open to the facts. And in all the subsequent stages, holists need to let that data talk to them rather than impose a theory on it. This is not easy, particularly in a clinical setting, but it is possible. If one does holistic research well, the final thematic interpretation will add to existing theories rather than merely parrot them.

As W. W. Meissner puts it (with philosophical exactitude), "Analysis is not concerned with the repeatability of data from case to case, but rather with the inner consistency and pattern of meaning that obtains within each case."51 "Meaning is the central fact of human existence with which psychoanalysis has to deal. The whole direction of the therapist's effort is toward elaboration of the full context of meaning in which the whole range of data that . . . gathered about the patient falls into a consistent, coherent, and intelligible pattern." "This pattern of meaning at its highest level of generalization encompasses the entire life experience of the patient."52

My first quotation from Meissner sounds as though he had given up all claims to scientific generality. But he is right to say that one cannot expect that the data--the surface of the patient's language--will repeat from case to case. Each case, each set of free associations, will be unique. In this respect, Donald Spence draws an extensive analogy between psychoanalytic principles and rules of law. Both are derived from a series of cases. In legal cases, there is, as in psychoanalysis, much flexibility as to whether the facts of the first case fall within the rule, whether an exception should be made, whether the rule should be modified or limited, and so on. "Psychoanalyis (like the law) is rule-governed but not rule-bound."53

Then, is it impossible to generalize from psychoanalytic cases? If psychoanalysis does not admit generalization, it can hardly qualify as a science. How could it live up to the usual demand of psychologists that the principles they discover predict and control behavior?

Generalizing and predicting

In a very limited way, a pattern explanation does enable one to predict. One would expect behaviors not yet brought into the picture for a given person to fit the pattern as the other data do. If they do not, then the pattern explanation needs to be modified to take the new data into account. This kind of testing marks a significant difference from the experimental method. In holistic work, experiments do not test the explanation, new data do.

Since each case is unique, though, how can one generalize from case to case? In his 1971 book, Diesing suggests that one can generalize by means of typologies. This is the method Jack Block and his team used in his well-known longitudinal studies of character.54

I would suggest another approach. A researcher could go from one unique case to another by learning questions to apply to this kind of unique case. These will not be questions that presuppose underlying laws, but questions that point to aspects of the unique case that one found important in another unique case and that otherwise one might not notice.

Maps, for example--each map is unique; how would you generalize about maps? You could move from map to map by a series of questions that tell you what to look for in the next map. Does this map preserve actual distances between features? Does it preserve areas? Compass directions? Can you use it to drive roads? To navigate the seas? Does it give altitudes? Does it give you non-geographical information, like population density or voter history? And so on. Such questioning would lead to an idea (or an ideal) of research as a conversation between the researcher and the data and, indeed, among researchers. Surely, in their early stages, scientific theories proceed by a such conversation in "the literature," as one can see in the earliest issues of the Proceedings of the Royal Society and in the current literature of a developing science like neuropsychology.

Freud certainly proceeded this way. Consider this wording from the famous letter of October 15, 1897 in which he announced his discovery of the oedipus complex. "A single idea of general value dawned on me. I have found, in my own case too, being in love with my mother and jealous of my father, and I now consider it a universal event in early childhood."55 Evidently Freud had been looking at his cases, asking himself, Do I find love of the mother and jealousy of the father?; and he looked at himself in his self-analysis with the same question. Having answered yes a sufficient number of times, he felt entitled to say he had found an idea of general value. And his presenting his idea to Fliess constituted the mini-conversation among scientists that he needed.

Having answered a holistic query the same way in a number of cases, one has a generalization that can be tested experimentally. It can be confirmed or disconfirmed as a universal principle, as a probability, or as a non-starter. And, in the first part of this paper, we have seen an abundance of experimental literature doing just that with a great many psychoanalytic generalizations. Holistic method, despite its loose-jointed procedures, can lead by stages to scientifically establishable hypotheses. And it has, and not just in psychoanalysis or the social sciences.

Whatever the virtues of experiment, the plain fact of the matter is that many social scientists do use holistic method, and they consider it valid and scientific. I am thinking of anthropologists, archaeologistsl (notably in studying hominids and long-buried cultures), sociologists and psychologists studying special groups, some market researchers, some geographers, and most political scientists and historians. Which are we to weigh more heavily, the praxis of working researchers or the abstract prescriptions of philosophers? Further, holistic method plays a part not only in the social but also in the "hard" sciences.

Holistic method in "hard" science

One major example of holistic reasoning in the "hard" sciences, one which revolutionized its science, is Wegener's theory of continental drift. In 1912 the German meteorologist Alfred Wegener proposed a theory that about 200 million years ago there was one supercontinent, Pangaea, which split into two vast land masses, Laurasia and Gondwanaland. He pointed to the jigsaw fit of the opposing Atlantic coasts and geologic and paleontologic correlations on both sides of the Atlantic. Until 1954, his theory met only controversy, but then more kinds of evidence accumulated. For example, British geophysicists conducted studies of magnetism in rocks which indicated that in the geological past the earth's magnetic field has often reversed its polarity. Therefore rock movement relative to the magnetic poles must have occurred.

The modern theory of plate tectonics evolved from and replaced Wegener's original thesis, and it has revolutionized geologists' understanding of the earth's history. The modern theory holds that the earth's crust is divided into contiguous, moving plates in which the continents are embedded. As the plates move, the continents "drift," as Wegener had suggested. At the plate boundaries, earthquakes and volcanoes occur. There is one kind of boundary at the mid-ocean ridges. There tensions open rifts, allowing new material to well up from the earth's mantle. At another boundary type, plates shear past each other along great faults, causing earthquakes (like San Francisco earthquake over the San Andreas fault in California or the Afraz fault that destroyed the city of Bam in Iran). Mountain ranges form where two plates carrying continents collide (e.g., the Himalayas), or where ocean crust is subducted along a continental margin (e.g., the Andes).

In short, the theory begins with a hodgepodge of data and ends by offering a narrative account of the earth's history to explain a great variety of phenomena: the matching shapes of the opposing Atlantic coasts; movements in geological time of the magnetic poles; the positions of earthquakes, ridges on the ocean floor, and volcanoes; the relative ages and positions of different mountain ranges; even the presence of marsupial mammals in Australasia and their absence elsewhere. This is, quite simply, holistic reasoning leading to a pattern or, in this case, narrative explanation. It is as "scientific" as scientific gets. We teach it to our schoolchildren.56

The major example of holistic reasoning and pattern explanation in the hard sciences is, of course, the theory of evolution. It, too, revolutionized its discipline. Darwin and Wallace gathered data from comparing closely related animals, the geographic distribution of those animals, variations in environment, evidence from fossils, embryonic development, and so on. "Themes" could explain variations in, say, Galapagos finches, and theme after theme formed a total narrative that explained the variations in species. What was missing was an explanation for initial changes, which had to wait for twentieth-century genetics. Now, however, we have a coherent narrative that largely explains the origin of species.

The examples of plate tectonics and evolutionary theory show that there is, really, no contradiction between an interpretive or "hermeneutic" explanation of data and a "scientific" explanation of data. In this sense, I think it is perfectly legitimate to claim that psychoanalysis is "scientific."

Notice that evolution itself would not meet the demands of positivist philosophers that "science" predict and control. The world of living organisms is unpredictable because the base from which we understand it, evolution, is unpredictable. Natural selection involves two intrinsically random and unpredictable processes: mutation and survival. Indeed, much of the scientific (as contrasted to religious) resistance to Darwin's theory in the nineteenth century came from the demand that a theory, to be scientific, yield predictable results. As Ernst Mayr has pointed out, much of twentieth-century biology, like much of our physics, deals with the random and the probabilistic, making prediction impossible.57 As it is in literary studies. As it is in psychoanalysis. A demand for prediction demands too much.

In short, some of the social sciences and some of the natural sciences depend on holistic method. And this is the crux of my argument: psychoanalysis also rests on holistic method, which can lead, ultimately, to principles that one can test experimentally. Psychoanalysis can claim to be a science, a holistic science.

Clinical issues

While this discussion of psychoanalysis' holism and its scientific status waxes philosophical and abstract, it seems to me, even as a non-clinician, to bear on clinical practice. It is one thing to say to a patient, When you mention your lover, you sound aggressive. It is another to say, When you mentioned your lover, you spoke of struggling, battling, retreating, fighting off . . . The first phrasing uses the kind of category thinking that fits experimental method. The second phrasing points to a holistic theme in the patient's speech. I hope I am not going beyond my brief if I say that the second surely speaks more directly to the patient than the first, because it uses the patient's own words.

Character will reveal itself no matter what a person is saying.58 Anything a person says, in or out of a therapeutic relationship, expresses unconscious concerns and ultimately that person's deepest themes, their "idiom." We can, after all, discern personal styles in the most abstruse of human activities: writing philosophy or mathematics, playing chess, programming computers. We can discern styles in the most bodily of acts, playing tennis, boxing, walking, making love, handwriting. I am simply repeating the old psychoanalytic metaphor of the iceberg. Our conscious activities show just the tip. Underneath, 90% of our mental life is unconscious, and unconscious determinants show in everything we do.

To be sure, what the patient says will reflect what the analyst thinks important. As is well known, Freudian patients have Freudian dreams and make Freudian statements and focus on Freudian issues. Jungians or Kleinians or Horneyans do the same for their theories.59 If the therapist interprets in terms of a theory (that is, as if the treatment were an experiment), then what the patient says offers no confirmation of the theory. If the therapist interprets using the patient's own words (that is, in holistic mode), the patient's words reveal the patient's "idiom" regardless of the theory. The therapist's theory will not change the patient's style. If we regard psychoanalytic free association as a method of self-understanding (whatever its therapeutic consequences), then the themes the free associator discovers will be his or her themes regardless of the school of therapy in play.

At a theoretical level, it is one thing to say that countertransference is important, that an analysis is a two-person affair, that the analysis is a joint product of analyst and analysand, that the analysis will be different for each analytic pair, that analysis is interpersonal. It is another to substitute abstract terms like "depressed," "aggressive," "envious," for the patient's own words. Such terms risk turning psychoanalysis into just another form of counseling.

Is psychoanalysis "merely literary"?

Prompted by these clinical questions, the psychoanalytic community debates, Is psychoanalysis a hermeneutic, a method of interpretation, and therefore merely "literary"? Or is it a science that can make the claims on our belief that a science? is entitled to? If we recognize the holistic and therefore (I claim) scientific basis for psychoanalysis, we can see that this is a false dichotomy. Psychoanalysis entails the gathering of data and working it into a generalizable explanation, just as other sciences do. No hard-and-fast line marks off A holistic science from "real" (i.e., experimental) science.

The statement that psychoanalysis is "not scientific" seems to me altogether too simple. While it is hardly physics or chemistry, psychoanalysis is not that far removed from geology or astronomy. It certainly falls quite naturally within medicine, where diagnosis is a holistic skill. And it is, if you will, also and in addition and not at all contradictorily, "literary."

Works Cited

    Adams, H. E., Wright, L. W., and Lohr, B. A. (1996, August). Is Homophobia Associated with Homosexual Arousal? Journal of Abnormal Psychology, 105(3), 440-445.

    Barron, J. W., Eagle, M. N., and Wolitzky, D. L. (Eds.). (1992). Interface of Psychoanalysis and Psychology. Washington, DC: American Psychological Association.

    Barron, J. W., Eagle, M. N., and Wolitzky, D. L. (1992). Preface. In J. W. Barron, M. N. Eagle and D. L. Wolitzky (eds), Interface of Psychoanalysis and psychology (pp. xi-xiv). Washington, DC: American Psychological Association.

    Block, J. (1971). Lives Through Time (N. Haan, Collaborator). Berkeley CA: Bancroft Books.

    Bollas, C. (1992). Being a Character Psychoanalysis and Self Experience. New York: Hill and Wang.

    Bollas, C. (2002). Free Association. Ideas in Psychoanalysis. London: Icon Books.

    Bornstein, R. F., and Masling, J. M. (Eds.). (1998a). Empirical Perspectives on the Psychoanalytic Unconscious. Empirical studies of psychoanalytic theories. Washington, D.C.: American Psychological Association.

    Bornstein, R. F., and Masling, J. M. (Eds.). (1998b). Empirical Studies of the Therapeutic Hour. Empirical Studies of Psychoanalytic Theories. Washington, D.C.: American Psychological Association.

    Bornstein, R. F., and Masling, J. M. (Eds.). (2002). The Psychodynamics of Gender and Gender Role. Empirical Studies in Psychoanalytic Theories, vol. 10. Washington, DC: American Psychological Association.

    Bornstein, R. F. (2001). The Impending Death of Psychoanalysis. Psychoanalytic Psychology, 18(1), 3-20.

    Collins, G. P. (2002, February). Setback for Super-K. Scientific American, p. 26.

    Dahl, H., K”chele, T., and Thom”, H. (Eds.). (1988). Psychoanalytic Process Research Strategies. Berlin and New York: Springer-Verlag.

    Diesing, P. (1971). Patterns of Discovery in the Social Sciences.. Chicago: Aldine-Atherton.

    Diesing, P. (1985). Comments on 'Why Freud's Research Method was Unscientific' by Von Eckhardt. Psychoanalysis and Contemporary Thought, 8, 551-66.

    Diesing, P. (1986). Review of Edwin Wallace, Historiography and Causation in Psychoanalysis. Contemporary Psychology, 31, 417-18.

    Diesing, P. (1991). How Does Social Science Work? Reflections on Practice. Pittsburgh: University of Pittsburgh Press.

    Duberstein, P. R., and Masling, J. M. (Eds.). (2000). Psychodynamic Perspectives on Sickness and Health. Empirical studies of psychoanalytic theories. Washington, DC: American Psychological Association.

    Erdelyi, M. H. (1985). Psychoanalysis: Freud's Cognitive Psychology. Series: Books in Psychology. New York: W.H. Freeman.

    Fisher, S., and Greenberg, R. P. (1977). The Scientific Credibility of Freud's Theories and Therapy. New York: Basic Books.

    Fisher, S., and Greenberg, R. P. (1996). Freud Scientifically Reappraised: Testing the Theories and Therapy. New York: John Wiley.

    Freud, S. (1896c). The Aetiology of Hysteria. Std. Edn. 3: 187-221.

    Freud, S. (1900a). The Interpretation of dreams. Std. Edn. 4-5: 1-627.

    Freud, S. (1901a). On Dreams. Std. Edn. 5: 629-685.

    Freud, S. (1901b). The Psychopathology of Everyday Life. Std. Edn. 6: 1-279.

    Freud, S. (1918b). From the History of an Infantile Neurosis. Std. Edn. 17: 7-122.

    Freud, S. (1937d). Constructions in Analysis. Std. Edn. 23: 129-221.

    Freud, S. (1985). The Complete Letters of Sigmund Freud to Wilhelm Fliess, 1887-1904 (J. M. Masson, Trans. and Ed.). Cambridge and London: Belknap Press of Harvard University Press.

    Gahde, Ulrich. "Holism, Underdetermination, and the Dynamics of Empirical Theories." Synthese: An International Journal for Epistemology, Methodology and Philosophy of Science 130.1 (Jan 2002): 69-90.

    Holland, N. N. (1985). The I. New Haven and London: Yale University Press. ( Accessed Jun 10, 2002.)

    Holland, N. N. (2001, March). 'The barge she sat in': Psychoanalysis and Diction. Psychoanalytic Studies, 3(1): 79-94.

    Holland, N. N. (2002). H.D.'s Analysis with Freud. PsyArt: An Online Journal for the Psychology of the Arts. ( Accessed June 10, 2002.)

    James, Susan. The Content of Social Explanation. Cambridge: Cambridge UP, 1984.

    Kline, P. (1981). Fact and Fantasy in freudian Theory (2nd ed.). London and New York: Methuen.

    Levine, F. J., and Luborsky, L. (1981). The Core Conflictual Relationship Theme; A Demonstration of Reliable Clinical Inference by the Method of Mismatched Cases. In S. Tuttman, C. Kaye and M. Zimmerman (Eds.), Object and Self: A Developmental Approach: Essays in Honor of Edith Jacobson (pp. 501-526). New York: International Universities Press.

    LÈvi-Strauss, C. (1964). Tristes Tropiques. (J. Russell, Trans.). New York: Atheneum.

    Lichtenstein, H. (1961). Identity and Sexuality: A Study of Their Interrelationship in Man. J. Amer. Psychoanal. Assn., 9, 179-260.

    Makari, G., and Shapiro, T. (1993). On Psychoanalytic Listening: Language and Unconscious Communication. J. Amer. Psychoanal. Assn., 41(4), 991-1020.

    Masling, J. M. (2000, Winter). Empirical Evidence and the Health of Psychoanalysis. J. Amer. Acad. of Psychoanalysis, 28(4): 665-685.

    Masling, J. M. (2002). Speak, Memory, or Goodbye, Columbus. Journal of Personality Assessment, 78(1), 4-30.

    Masling, J. M., and Cohen, I. S. (1987). Psychotherapy, Clinical Evidence, and the Self-Fulfilling Prophecy. Psychoanalytic Psychology, 4(1), 65-79.

    Masling, J. M., and Bornstein, R. F. (Eds.). (1993). Psychoanalytic Perspectives on Psychopathology. Empirical Studies of Psychoanalytic Theories. Washington, DC: American Psychological Association.

    Masling, J. M., and Bornstein, R. F. (Eds.). (1994). Empirical Perspectives on Object Relations Theory. Washington, DC: American Psychological Association.

    Masling, J. M., and Bornstein, R. F. (Eds.). (1996). Psychoanalytic Perspectives on Developmental Psychology. Empirical Studies of Psychoanalytic Theories. Washington, DC: American Psychological Association.

    Masling, J. M. (Ed.). (1983). Empirical Studies of Psychoanalytical Theories, vol. 1. Hillsdale, N.J.: Analytic Press.

    Masling, J. M. (Ed.). (1986). Empirical Studies of Psychoanalytic Theories, vol. 2. Hillsdale, N.J.: Analytic Press.

    Masling, J. M. (Ed.). (1990). Empirical Studies of Psychoanalytic Theories, vol. 3. Hillsdale, N.J.: Analytic Press.

    Mayr, E. (1978, September). Evolution. Scientific American, pp. 47-55.

    Meissner, W. W. (1966). The Operational Principle and Meaning in Psychoanalysis. Psychoanal. Q., 35, 233-255.

    Meissner, W. W. (1971). Freud's Methodology. J. Amer. Psychoanal. Assn., 19: 265-309.

    Peacocke, Christopher. Holistic Explanation: Action, Space, Interpretation. Oxford: Clarendon Press, 1979.

    Phillips, D. C. Holistic Thought in Social Science. Stanford: Stanford UP, 1976.

    Plate tectonics. In EncyclopÊdia Britannica. (<> Accessed June 10, 2002.)

    Pribram, Karl H. "Behaviorism, Phenomenology, and Holism in Psychology." The Metaphors of Consciousness. Ed. Rolf von Eckartsburg. New York: Plenum Press, 1981. 141-51.

    Reik, T. (1948). Listening With the Third Ear: The Inner Experience of a Psychoanalyst. New York: Farrar, Straus.

    Rosenthal, R. (1976). Experimenter Effects in Behavioral Research [Enl. edn.]. New York: Irvington Publishers.

    Sears, R. R. (1943). Survey of Objective Studies of Psychoanalytic Concepts. Social Science Research Council Bulletin 51. New York: Social Science Research Council.

    Spence, D. P. (1987). The Freudian Metaphor. New York: Norton.

    Spence, D. P. (1992). Interpretation: A Critical Perspective. In J. W. Barron, M. N. Eagle and D. L. Wolitzky (eds), Interface of Psychoanalysis and psychology (pp. 558-572). Washington, DC: American Psychological Association.

    Spence, D. P. (1994). The Rhetorical Voice of Psychoanalysis: Displacement of Evidence by Theory. Cambridge MA: Harvard University Press.

    Spence, D. P. (1998). Pronoun Co-Occurrence as a Measure of Shared Understanding. In R. F. Bornstein and J. M. Masling (Eds.), Empirical Studies of the Therapeutic Hour. Empirical studies of psychoanalytic theories, vol.  8. Washington, D.C.: American Psychological Association.

    Waelder, R. (1962). Psychoanalysis, Scientific Method, and Philosophy. J. Amer. Psychoanal. Assn., 10, 617-637.

    Waelder, R. (1976). Psychoanalysis, Scientific Method, and Philosophy. In S. Guttman (Ed.), Psychoanalysis: Observation, Theory, Application: Selected Papers of Robert Waelder (pp. 248-274). New York: International Universities Press. (Original work published 1962)

    Weiss, J., Sampson, H., and the Mount Zion Psychotherapy Research Group. (1986). The Psychoanalytic Process : Theory, Clinical Observation, and Empirical Research. New York: Guilford Press.

    Westen, D. (1998). The Scientific Legacy of Sigmund Freud: Toward a Psychodynamically Informed Psychological Science. Psychological Bulletin, 124(3): 333-371.


    2 Barron, Eagle, and Wolitzky, Interface 1992; Barron, Eagle, and Wolitzky, "Interface" 1992. (Back To Main Text)

    5 Masling and Cohen 1987. Masling, "Empirical evidence," 2000,  670. (Back To Main Text)

    6 Weiss, Sampson and the Mount Zion Psychotherapy Research Group 1986. (Back To Main Text)

    11 Masling, "Speak," 2002; Masling, J., "Empirical evidence," 2000. (Back To Main Text)

    12 Masling, "Speak," 2002, a result confirmed by others (Adams, Wright and Lohr 1996). (Back To Main Text)

    13 Masling 1983; 1986; 1990; Masling and Bornstein 1993; 1994; 1996; 1998b; Bornstein and Masling 1998a; Duberstein and Masling 2000; Bornstein and Masling 2002. (Back To Main Text)

To cite this article, use this bibliographical entry: Norman N. Holland "Psychoanalysis as Science". PSYART: A Hyperlink Journal for the Psychological Study of the Arts. December 15, 2009. Available May 22, 2004 [or whatever date you accessed the article].
Received: May 10, 2004, Published: May 22, 2004. Copyright © 2004 Norman N. Holland