Reply to Holland

Is psychoanalysis a science? The Spring/Summer 2005 issue of The Scientific Review of Alternative Medicine (vol. 9, no. 1) will contain a debate on the scientific merits of psychoanalysis. The exchange will include a 2000-word summary by the literary critic Norman N. Holland of his essay “Psychoanalysis as Science”; a 1000-word critique by Frederick Crews; a reply from Holland to that critique; and a commentary on both submissions by the psychiatrist Peter Barglow. Holland’s full essay can already be found on the Web here. In anticipation of the SRAM publication, concerned readers may be interested in an early view both of Holland’s summary version and of Crews’s response to the longer piece. The editor of SRAM has granted permission for these postings.

Although Norman Holland’s synopsis conveys the gist of “Psychoanalysis as Science,” the devil is in the details. Necessarily, I will be referring here to the paper itself, which displays shortcomings of coverage and logic that are less discernible in the synopsis.

Holland maintains that important parts of psychoanalytic theory have been experimentally confirmed and that analysts in their daily practice employ a methodologically sound means of gathering knowledge. As he recognizes, this judgment stands at odds with the tacit, all but unanimous verdict of North American psychology faculties. Where psychoanalysis appears at all in the catalogs of well-regarded university departments of psychology, it usually figures as a prescientific historical curiosity, not as a viable body of theory. And a study of citations in the flagship psychology journals concludes that “psychoanalytic research has been virtually ignored by mainstream scientific psychology over the past several decades.” [1, p. 117]

Holland asserts that this snub bespeaks not a considered scientific assessment but rather “a deep-seated prejudice against psychoanalysis” on the part of psychology professors and textbook authors. The academic establishment, he holds, has turned its back on a mountain of studies validating key portions of psychoanalytic doctrine while disallowing some others. Indeed, according to Holland even the most adamant critics of psychoanalysis are unaware of that literature. The main task that he sets for himself in “Psychoanalysis as Science” is therefore an easy one: he will correct an unfair negative impression by calling attention to the somewhat positive results of hitherto overlooked experimental trials.

But how can Holland be sure that those results have been overlooked? One could not tell from his paper that he has read a single page of the revisionist scholarship and reasoning that have revolutionized our perception of the psychoanalytic movement and its claims of scientific validation. His 64 references include no dissenters’ texts; and only one dissenter’s name, my own, is briefly mentioned. Moreover, Holland’s characterization of my position, that I find all of psychoanalytic theory untestable and therefore merely “literary” in nature, is off the mark. I regard psychoanalytic doctrine not as literature but as partly unfalsifiable, partly falsified pseudoscience which, when it was widely believed, caused harm to people whom it demeaned, stigmatized, and misdiagnosed. [2, 3, 4; see also 5]

Unfortunately, the facts and arguments that Holland ignores bear crucially on the question he proposes to answer: whether psychoanalysis deserves to be called a science. He could have learned much, for example, from the work of two major Freud scholars, Frank Cioffi [6] and Malcolm Macmillan [7], who have extensively traced Freud’s initial confusions and misrepresentations, the many unclarities and cross-purposes that have continued to plague psychoanalytic doctrine, and the chronic flight from exposure to potential disconfirmation that has typified the entire record from Freud’s day through our own. Science is as science does. If neither Freud nor his successors have shown a due regard for objections to their pet ideas, psychoanalysis is ipso facto not a science.

Condensing the findings of Cioffi, Macmillan, and other knowledgeable philosophers of science and historians such as Adolf Grünbaum [8], Edward Erwin [9], and Allen Esterson [10], I have elsewhere put into one long sentence the anti-empirical features of the psychoanalytic movement [3, pp. 61n-62n]:

They include its cult of the founder’s personality; its casually anecdotal approach to corroboration; its cavalier dismissal of its most besetting epistemic problem, that of suggestion; its habitual confusion of speculation with fact; its penchant for generalizing from a small number of imperfectly examined instances; its proliferation of theoretical entities bearing no testable referents; its lack of vigilance against self-contradiction; its selective reporting of raw data to fit the latest theoretical enthusiasm; its ambiguities and exit clauses, allowing negative results to be counted as positive ones; its indifference to rival explanations and to mainstream science; its absence of any specified means for preferring one interpretation to another; its insistence that only the initiated are entitled to criticize; its stigmatizing of disagreement as “resistance,” along with the corollary that, as Freud put it, all such resistance constitutes “actual evidence in favour of the correctness” of the theory (SE, 13:180); and its narcissistic faith that, again in Freud’s words, “applications of analysis are always confirmations of it as well” (SE, 22:146).

This indictment is sometimes dismissed by Freudians as the raving of an unhinged mind. The justice of every item, however, has been conceded piecemeal by a number of psychoanalysts who are still unready to take in the total picture. And other previously sanguine pro-psychoanalytic commentators now grant that the Freudian community has shown none of the traits we associate with serious investigators.

Robert F. Bornstein, for example, whom Holland repeatedly cites as a compiler of positive experimental evidence, recently published an article, significantly entitled “The Impending Death of Psychoanalysis,” in which he charged analysts with “the seven deadly sins” of “insularity, inaccuracy, indifference, irrelevance, inefficiency, indeterminacy [that is, conceptual vagueness], and insolence.” [11] Bornstein portrays a self-isolated sect that is not just out of step with the march of knowledge but incapable of understanding where it went wrong. In order for Bornstein to bring his revised view into full alignment with that of the revisionist critics (whom Holland is pleased to malign en masse as “the bashers”), he need only grasp that the dysfunctional attitudes he has listed are traceable to Freud’s own arbitrary system building, to his dismissal of the need to reconcile psychoanalytic theory with mainstream science, to his heaping of scorn on all who questioned his authority, and to his declarations that backsliders from his movement had fallen into psychosis.

In the estimation of Bornstein and some other would-be reformers, psychoanalysis must now rapidly embrace commonly held scientific standards or vanish altogether from the scene. But what would become of the remaining shards of Freudian theory if their proponents took Bornstein’s ultimatum to heart? More than a century has passed since analysts, on no examinable grounds, began launching fanciful propositions about the deep structure of the mind, the stages of psychosexual development, and the unconscious symbolic thought processes in early childhood that supposedly issue in adult mental illness. Medical science has moved decisively away from that approach to explanation, which, as Freud privately observed in acknowledgment of kindred thinkers, harkened back to the “spirit possession” lore drawn upon by the judges in witchcraft trials. [12]

It would be surprising if such an undisciplined and retrograde movement had received support from well-designed experiments, and it would be no less surprising if the critics of psychoanalysis had failed to address the experimental literature. In fact, Holland’s claims on both counts are false. An extensive body of penetrating and disillusioning commentary about pro-Freudian experimentation can be found, beginning with Eysenck and Wilson’s small masterpiece of 1973 [13] and running through Edward Erwin’s meticulous study of 1996. [9] It is apparent that Holland, who innocently equates the terms “experimental” and “empirical,” hasn’t pondered these widely discussed and important works. Yet if he had attended to no other writings than my own, he would have found me engaged in pertinent debate with several of the psychodynamically committed experimental authorities on whom he relies: Seymour Fisher, Roger P. Greenberg, Lester Luborsky, and Matthew H. Erdelyi. [2, 3, 4]

As its scientific critics have shown, most of the research admired by Holland suffers from grave and obvious flaws. These studies, having been conducted by people holding a prior affinity for psychoanalysis, are riddled with confirmation bias and demand characteristics:

  • Instead of testing psychoanalytic hypotheses against rival ones that might have fared better under Ockham’s razor, the experimenters have used Freudian theory as their starting point and have looked for confirming instances, which have been located with the same facility with which Holland once found oral and anal images suffusing the world’s literature.
  • Terms have been construed with suspect broadness; strong causal claims have been reinterpreted as weak descriptive ones; and generous psychoanalytic rules of interpretation have helped to shape positive results.
  • Freudian propositions have been assessed through the application of such questionable instruments as the psychoanalytically tendentious Blacky pictures and the Rorschach test, which already lacked validity before believers in Freudian projection twisted it to their own purposes. [14] (Holland himself twice appeals to psychoanalytic Rorschach findings as sound evidence.)
  • Signs of unconscious cognitive operations have been misidentified as evidence of the very different Freudian unconscious at work. [15] (Holland’s paper indulges in the same confusion.)
  • Replication of tentative outcomes by independent investigators-an essential requirement of experimentation in any field-has not been achieved or even sought.

It is Holland’s countenancing of these lax and biased practices that allows him to proclaim that research “supports an oedipal stage,” that “the penis=baby equation” has been vindicated, that “links between depression and oral fixation” have been found, and that “Freud’s account of paranoia gets confirmation.” Such “confirmation” is a strictly parochial affair, and that is why it has been left out of account by scientifically responsible textbook authors.

Critics of psychoanalysis hold that no distinctively psychoanalytic hypotheses, such as those just mentioned, have earned significant evidential backing. Freudians, however, typically credit psychoanalysis with having introduced broader notions that were, in fact, already commonplace in the middle of the nineteenth century. As the great historian of psychiatry Henri F. Ellenberger observed in 1970, “The current legend attributes to Freud much of what belongs, notably, to Herbart, Fechner, Nietzsche, Meynert, Benedikt, and Janet, and overlooks the work of previous explorers of the unconscious, dreams, and sexual pathology. Much of what is credited to Freud was diffuse current lore, and his role was to crystallize these ideas and give them an original shape.” [16, p. 548]

It is only Freud’s novelties and unique adaptations, along with those of his most emulated revisers, that ought to concern us here. Self-evidently, support for ideas that originated elsewhere, much less those that express the traditional wisdom of the ages, cannot be counted as favoring psychoanalysis. Apparently, however, Holland does not consider himself bound by this axiom.

Holland reports, for example, that research has validated such assertedly psychoanalytic propositions as that “much mental life . . . is unconscious,” that “stable personality patterns form in childhood and shape later relationships,” that “mental representations of the self, others, and relationships guide interactions with others . . . ,” and that “personality development is . . . moving from immature dependency to mature interdependency.” Insofar as these vapid truisms constitute the ground to which psychoanalysis has now fled in its retreat from Freud’s heedless guesswork, they illustrate the bankruptcy, not the scientific vindication, of his movement.

In the second half of his argument, Holland seeks to confer respectability on psychoanalysis by assimilating it to sciences that enjoy unchallenged recognition as such. His reasoning here is notably fallacious. By progressing from single inductions to themes and patterns that are then checked for adequacy, he writes, psychoanalysts employ the same “holistic” method as social scientists and some physical scientists as well; and since neither psychoanalysis nor geology nor astronomy attempts to predict the future, “psychoanalysis is not that far removed from geology or astronomy.” (Nor, in that one respect, is phrenology or the channeling of ancestors.) Needless to say, a perceived or imagined resemblance between the data gathering in one field and that in another tells us nothing about whether their eventual hypotheses are comparably parsimonious and well supported.

Holland labors to portray the psychoanalytic clinician as a scientist in his own right who cautiously moves from a theory-free study of word associations to hypotheses that make full sense of the resultant inferences. Yet he approvingly quotes a pair of experts who point out that the analyst “listens for noises that signify in psychoanalytic terms” (emphasis added); he further admits that “Freudians will see Freudian patterns” everywhere; and he adds that “Freudian patients have Freudian dreams and make Freudian statements and focus on Freudian issues”-thus providing the analyst, we may be sure, with more Freudian evidence for the confirmation of his Freudian hunches. Perversely, however, Holland still clings to his ideal vision of the tabula rasa clinician-scientist.

Freud, Holland maintains, arrived at his theory in just this inductive manner, building hypotheses from sheer attentive listening in the consulting room. We now know, however, that this hoary legend, propagated by Freud himself and his inner circle, is utterly untrue. Far from suspending judgment as a clinician, Freud typically demanded that his patients agree with his theory-driven accusations of incestuous desires, homosexual leanings, and early masturbation.

As a theorist Freud was a rashly deductive bioenergetic speculator who routinely invented “clinical evidence” to fit his predetermined ideas and who altered the facts again when a new speculation required adornment. Contemporaries accused him with good reason of having plagiarized some of his most basic notions, including repression, infantile sexuality, and “universal bisexuality.” When it proved impossible for him to deny such unacknowledged borrowings, he brazenly ascribed them to psychodynamically induced “amnesia.” [16, 17, 18]

Holland’s illustrations of Freud’s supposed method show that he has not fathomed the cardinal difference between the first psychoanalyst’s actual means of reaching conclusions and his seductive rhetorical reconstructions, which offered the trusting reader sequences of ingeniously solved little puzzles that may or may not have preceded his theorizing. Freud’s subtle diagnostic skill as manifested in the Wolf Man case history, for example, earns Holland’s praise; no one has told him about the cunning fibs in that story that were uncovered by the psychoanalyst Patrick Mahony 20 years ago. [19] And in reading Freud’s famous “aliquis slip” narrative in The Psychopathology of Everyday Life, Holland takes at face value a (probably fictitious) young man’s narrated “associations” of liquefied blood and calendar saints. Alas, it has been ascertained that Freud lifted those and other references from a current newspaper article and worked them into a self-flattering and mendacious yarn about Sherlock Holmes-like psychic detection on his part. [20]

Of course, the fact that Freud himself didn’t faithfully employ “the psychoanalytic method” doesn’t impugn that method in other hands. Yet it is impugned, as Adolf Grünbaum in particular has shown, by the circular procedures that Holland now dimly perceives to be a problem. Whereas Holland would like to believe that a clinician need only exercise “integrity” to avoid imposing his presuppositions on the patient, Grünbaum makes it clear that question begging in the therapeutic interchange is structurally unavoidable. [8, 21]

Grünbaum’s demonstration is devastating to the claim, still advanced by Holland, that modern psychoanalysis rests on a secure knowledge base. “Psychoanalytic method”-the analysis of (allegedly) free associations, of dreams and slips, and of the “transference”-is much the same as it was a hundred years ago, and it is helpless against the contaminating effect of suggestion. That is why we see so many warring psychoanalytic schools, each boasting “clinical validation” of its tenets.

Holland’s final misstep is to bracket psychoanalysis with plate tectonics and natural selection, which met with resistance until they were eventually vindicated by consilient findings. The fate of psychoanalysis has been exactly the reverse; it quickly won popular acclaim through its emphasis on taboo breaking but then gradually lost favor as its overweening claims met with no scientific consilience at all. It is that absence of corroboration, not “deep-seated prejudice” or the efforts of debunkers such as myself, that chiefly accounts for the moribund state of psychoanalysis today.


1. Robins RW and others. ‘An empirical analysis of trends in psychology’. American Psychologist 54:117-128, 1999.

2. Crews F. Skeptical Engagements. New York, NY: Oxford University Press, 1986.

3. Crews F and others. The Memory Wars: Freud’s Legacy in Dispute. New York, NY: New York Review Books, 1995.

4. Crews F, editor. Unauthorized Freud: Doubters Confront a Legend. New York, NY: Viking, 1998.

5. Dolnick E. Madness on the Couch: Blaming the Victim in the Heyday of Psychoanalysis. New York, NY: Simon & Schuster, 1998.

6. Cioffi F. Freud and the Question of Pseudoscience. Chicago, IL: Open Court, 1998.

7. Macmillan M. Freud Evaluated: The Completed Arc. Cambridge, MA: MIT Press, 1997.

8. Grünbaum A. The Foundations of Psychoanalysis: A Philosophical Critique. Berkeley, Los Angeles, London: University of California Press, 1984.

9. Erwin E. A Final Accounting: Philosophical and Empirical Issues in Freudian Psychology. Cambridge, MA: MIT Press, 1996.

10. Esterson A. Seductive Mirage: An Exploration of the Work of Sigmund Freud. Chicago, IL: Open Court, 1993.

11. Bornstein RF. ‘The impending death of psychoanalysis.’ Psychoanalytic Psychology 18:3-20, 2001.

12. Crews F. ‘The legacy of Salem: Demonology for an age of science.’ Skeptic 5:36-44, 1997.

13. Eysenck HJ, Wilson GD. The Experimental Study of Freudian Theories. London: Methuen, 1973.

14. Wood JM and others. What’s Wrong with the Rorschach: Science Confronts the Controversial Inkblot Test. San Francisco CA: Jossey-Bass, 2003.

15. Kihlstrom JF. ‘The cognitive unconscious.’ Science 237:1445-1452, 1987.

16. Ellenberger HF. The Discovery of the Unconscious: The History and Evolution of Dynamic Psychiatry. New York, NY: Basic Books, 1970.

17. Sulloway FJ. Freud, Biologist of the Mind: Beyond the Psychoanalytic Legend. Cambridge, MA, 1992.

18. Borch-Jacobsen M. Remembering Anna O.: A Century of Mystification. New York, NY: Routledge, 1996.

19. Mahony P. Cries of the Wolf Man. New York, NY: International Universities Press, 1984.

20. Skues R. ‘On the dating of Freud’s aliquis slip’. International Journal of Psychoanalysis 82:1185-1204, 2001.

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

Comments are closed.