3. Adolf Grünbaum: The Philosophical Critique of Freud

Frank Sulloway and Jeffrey Masson address themselves to questions of Freud’s intellectual biography. The target of their criticism is the received account of how Freud’s psychoanalytic ideas originated in the crucial decade of the 1890s. Adolf Grünbaum’s critique of psychoanalysis takes a very different form. It is above all an evaluation of the evidence and arguments that Freud used to justify his ideas. Grünbaum asks, How does Freud know that human behavior is significantly influenced by unconscious thoughts? Grünbaum poses this question most systematically in The Foundations of Psychoanalysis: A Philosophical Critique, published in 1984, although the book’s main contentions had been aired in a series of papers that began appearing in philosophical journals during the late 1970s. By that time, Grünbaum, who was born in 1923, already had behind him a distinguished career as a philosopher of science. He was especially admired for his “magisterial” studies in the philosophy of time and space.^[1] In contrast to Sulloway and Masson, then, Grünbaum came to his engagement with Freud relatively late in life and with an impressive record of accomplishment in another field of inquiry.

The Foundations of Psychoanalysis has been widely hailed as the most substantial philosophical critique of Freud ever written. One might suspect the enthusiasm of so resolute an anti-Freudian as Frederick Crews, who greeted Grünbaum’s book, in The New Republic, as “monumental” and “epoch-making.” “After Grünbaum,” wrote Crews, “the wholesale debunking of Freudian claims, both therapeutic and theoretic, will be not just thinkable but inescapable.”^[2] Yet even a critic of the book like the philosopher David Sachs acknowledged it as “an ‘event’ in philosophical criticism of psychoanalysis.”^[3] When, in 1986, the book was made the subject of “Open Peer Commentary” in the journal The Behavioral and Brain Sciences, forty-one of Grünbaum’s colleagues paid tribute to his achievement. Robert R. Holt, professor of psychology at New York University, wrote that “the power and subtlety of the analysis and arguments Adolf Grünbaum presents in this book far surpass those of any previous philosophical evaluation of psychoanalysis,” and Irwin Savodnik of UCLA called it “the most exhaustive and powerful critique of psychoanalysis to date.”^[4] Psychoanalysts were hardly less admiring—in sharp contrast to their response to Sulloway and especially Masson. The analysts Robert Wallerstein and Judd Marmor praised Grünbaum’s incisiveness and his mastery of Freudian theory, while Marshall Edelson paid him the high compliment of writing an entire book to answer his criticisms.^[5] Of all the attacks to which Freud has been subject in the past decade, Grünbaum’s is undoubtedly the weightiest.

Although philosophical critiques of psychoanalysis have existed almost since the doctrine first made its appearance at the turn of the century, Grünbaum’s analysis is distinguished from these earlier efforts by several features. Most notable are his extraordinary rigor and precision. Grünbaum is manifestly both very smart and very sophisticated, and his critique maintains an unprecedented level of dialectical intensity. At least for the philosophically untutored (to borrow one of Grünbaum’s own favorite words), virtually every sentence must be carefully unpacked, so thick and unforgiving (although never obscure) is his habit of thought. At the same time, Grünbaum surpasses all previous philosophical critics of psychoanalysis in the breadth and suppleness of his knowledge of Freud’s writings. This intimate relationship with the object of his criticism is closely linked to yet another characteristic of Grünbaum’s work: its highly interesting ambivalence. Strange as it may seem, much of Grünbaum’s energy goes into defending Freud’s philosophical astuteness. In part, this effort at building Freud up is a dialectical ploy—a disingenuous show of admiration that renders Grünbaum’s coup de grâce all the more stunning when it is finally delivered. Yet the fact remains that Grünbaum reserves his most withering criticisms not for Freud himself but for those of Freud’s followers who fail to measure up to the master’s methodological standards, and, even more so, for Grünbaum’s fellow philosophers: whether friendly to analysis (Jürgen Habermas, Paul Ricoeur) or hostile to it (Karl Popper, Frank Cioffi), they are invariably described as slovenly readers of Freud as well as poor logicians. This ambivalence renders Grünbaum’s critique all the more compelling.

Yet despite his superior critical muscle, Grünbaum has not achieved the prominence of Sulloway or Masson, and I doubt that he will have as marked an influence on Freud’s reputation as either of them. The explanation for Grünbaum’s relative obscurity lies in the exceptional difficulty of his writing: The Foundations of Psychoanalysis is profoundly inaccessible. Grünbaum has only the most primitive sense of how to compose a book, taking up topics in no discernible order and seldom resisting the temptation to digress. A chapter of fifty pages is followed by one of three pages; internal subdivisions and titles provide little useful guidance to the intellectual proceedings. But the idiosyncrasies of the book’s organization would be tolerable were its language less forbidding. Most bothersome is Grünbaum’s addiction to a highly technical philosophical vocabulary. To a certain extent, of course, the use of this terminology reflects his commitment to intellectual precision. But a devotion to precision will not explain the ornate and pedantic rhetorical structures into which Grünbaum embeds his technical terms. Again and again, the reader must struggle to disentangle fearsome syntactical complexities. Here is a representative sample: “Duhem showed, before Habermas was born, that the presence of auxiliary hypotheses in the experimental testing of a major hypothesis in physics precludes a deductively conclusive refutation of the latter, despite the deductive validity of a modus tollens inference.”^[6] This sort of writing serves principally to inhibit comprehension. Reading The Foundations of Psychoanalysis, one can easily doubt that anything so disorganized and impenetrable could ever pose a serious threat to Freud.

Quite apart from its sheer inscrutability, Grünbaum’s writing occasionally succumbs to disconcerting lapses in tone. With no warning he will plunge from the philosophic empyrean to the journalistic lowlands. Admittedly, this habit lends an intriguingly human touch to the book, although it also makes one wonder about Grünbaum’s judgment. For example, consider his discussion of Freud’s theory that paranoia results from repressed homosexuality. Grünbaum’s general philosophical point is this: because Freud’s theory invites the (clearly falsifiable) prediction that a decline in the repression of homosexuality ought to result in a corresponding decline in paranoia, it disproves Karl Popper’s claim about the unfalsifiability of psychoanalytic propositions. But Grünbaum chooses to make his point in a most curious fashion:

The recently revealed likelihood that, in 1893, Tchaikovsky was blackmailed into suicide—under threat of exposure of a homosexual liaison—is a measure of the lethal power possessed by the ban on homosexuality in the Christian world less than a century ago. This suicide occurred at the pinnacle of his career, less than a week after the Saint Petersburg premiere of his celebrated Pathétique symphony. Yet, for nine decades the standard biographies of him attributed his death at the age of fifty-three to cholera, probably yet another manifestation of the prevailing taboo on homosexual behavior and on suicide as well. Since 1893, which was also the year of Breuer and Freud’s momentous “Preliminary Communication,” even prominent members of both sexes have publicly identified themselves as homosexuals despite the harassing agitation of Anita Bryant. Perhaps it is therefore not too early now to begin garnering appropriate statistics on the incidence of paranoia with a view to ascertaining in due course whether these epidemiologic data bear out the psychoanalytically expected decline. (38–39)

How did we get to Tchaikovsky and Anita Bryant? Such apparently gratuitous observations erupt into Grünbaum’s otherwise relentlessly abstract discourse rather like neurotic symptoms, disturbing the philosophical peace. At the very least they bespeak a remarkable—albeit not unattractive—indifference to appearances. One begins to suspect that Adolf Grünbaum is a deeply eccentric man.

The sensibility on display in these pages is light-years removed from Freud’s. Even his harshest critics agree that Freud was among the most lucid and least pedantic of writers. So far as I know, he was never guilty of the sorts of lapses in tone that give Grünbaum’s book a sometimes surreal effect. The discrepancy might well be irrelevant. But so profound a difference in literary manner carries certain intellectual implications. One begins to worry that, for all his philosophical acuity and his knowledge of the psychoanalytic literature, Grünbaum cannot actually engage Freud. One fears, in other words, that because Freud and Grünbaum inhabit such manifestly alien rhetorical universes, the encounter between the two risks becoming an intellectual nonevent.

The Hermeneutic Freud

The first third of The Foundations of Psychoanalysis is devoted to the hermeneutic interpretation of Freud advanced by Jürgen Habermas and Paul Ricoeur. Grünbaum’s massively detailed attack on these two philosophers, coming right at the start of his book, is rather disorienting for the unwary reader. Why, one is apt to wonder, does a book purporting to deal with the foundations of psychoanalysis begin so arbitrarily with this obscure and difficult philosophical disagreement? And why does its author inveigh so passionately and at such length against the interpretations offered by these two thinkers?

Typically, Grünbaum does not explain himself. But his decision to begin with Habermas and Ricoeur in fact makes perfect sense. Grünbaum plans to argue that psychoanalysis does not meet the standards of proof expected in the natural sciences—an enterprise Habermas and Ricoeur threaten to undermine from the start by getting Freud off the evidential hook. They have been leading voices in the campaign to loose psychoanalysis from its scientific roots and replant it, so to speak, in the more hospitable soil of the humanities. Their hermeneutic reinterpretation turns on the idea that psychoanalysis is not a science, even though Freud made the mistake of claiming it was. Rather, it is an interpretive discipline, more akin in its intellectual methods and goals to philosophy, history, and literature than to physics, chemistry, or biology. And because psychoanalysis is not a science, it obviously cannot be held to the standards of proof that obtain in science. To complain that Freud’s ideas fail to measure up to the canons of inductivism demanded in the sciences (as opposed to the ways of knowing that prevail in the humanities) is to misunderstand what sort of knowledge psychoanalysis imparts.

Grünbaum views the hermeneutic approach as a fiendishly clever stratagem designed to save psychoanalysis from intellectual oblivion. He distrusts its easy appeal to a profession that has long found itself on the epistemological defensive, alluding darkly to the “desire to safeguard a lifetime professional investment in the practice of psychoanalytic treatment”:^[7]

Faced with the bleak import of skeptical indictments of their legacy, they [psychoanalysts] are intent on salvaging it in some form. Hence, some of them will be understandably receptive to a rationale that promises them absolution from their failure to validate the cardinal hypotheses of their clinical theory, a failure I demonstrate in depth…below. Be of stout heart, they are told, and take the radical hermeneutic turn. Freud, they learn, brought the incubus of validation on himself by his scientistic pretensions. Abjure his program of causal explanation, the more drastic hermeneuticians beckon them, and you will no longer be saddled with the harassing demand to justify Freud’s causal hypotheses. (57)

Grünbaum regrets that a number of prominent analysts—such as George Klein, Roy Schafer, and Donald Spence—have thrown in their lot with the hermeneuticians. They argue, in Donald Spence’s words, that psychoanalysis seeks not “historical truth” but “narrative truth,” and they urge analysts to abandon the claim to offer objective explanations of human behavior in favor of more modest interpretive goals, such as providing the analysand with a coherent (though not necessarily accurate) account of his experience. Understandably, Grünbaum is eager to head off this intellectual retreat, because, if successful, it would render his inductivist critique of psychoanalysis otiose. Hence the savage attack on Habermas and Ricoeur with which he begins.

Grünbaum complains that the hermeneutic interpreters radically misrepresent Freud’s views. In particular, Grünbaum objects to their effort to portray Freud as a helpless victim of nineteenth-century positivism, caught in the embrace of unreconstructed philosophical materialism. The hermeneutic interpreters err, Grünbaum insists, in thinking that Freud’s claim to be a scientist rested on his continued faith in an outdated materialist ontology of mind—classically embodied in the Helmholtzian Project for a Scientific Psychology—whose residues, they argue, infected even his mature metapsychological conceptions. Rather, asserts Grünbaum, once Freud had abandoned the 1895 Project, he held to a strictly methodological or “epistemic” (6) notion of the scientific credentials of psychoanalysis: psychoanalysis was scientific because it arrived at general truths by way of induction, not because it had identified the fundamental material entities of the human mind. Accordingly, for Freud the most scientific part of psychoanalysis was its clinical theory—its explanations for various mental illnesses and its ideas about dreams and slips, all of which, Freud maintained, rested first and foremost on a “wealth of dependable observations” gathered from the couch.^[8] In keeping with the inductive methods of science, the metapsychological constructs were a secondary development, intended to account for those primary clinical findings. The metapsychology was frankly speculative, and, as Grünbaum shows, Freud was always ready to give it up if it came into conflict with clinical material. Thus, writing in 1925 about his metapsychological description of the mind in terms of “agencies” or “systems” (like the unconscious and the preconscious), Freud noted: “Such ideas as these are part of a speculative superstructure of psycho-analysis, any portion of which can be abandoned or changed without loss or regret the moment its inadequacy has been proved.”^[9] The hermeneutic revisionists, Grünbaum complains, erroneously treat the metapsychology not as “a speculative superstructure,” which Freud truly deemed it to be, but as the discreditable philosophical base of his thought. They thereby fundamentally misconstrue the nature of Freud’s commitment to science.

Jürgen Habermas is the hermeneutic interpreter whose version of psychoanalysis most offends Grünbaum. Habermas’s principal discussion of psychoanalysis occurs in Knowledge and Human Interests (1968), where he argues that Freud was guilty of “scientistic self-misunderstanding” in thinking his discoveries were a contribution to science.^[10] Habermas’s case depends on showing that the essential features of psychoanalytic theory deviate from the intellectual procedures routinely found in the sciences. Grünbaum criticizes both ends of Habermas’s dichotomy: psychoanalysis, Grünbaum insists, does not depart from the scientific norm (at least in its aspirations) as Habermas says it does, and science, for its part, turns out to be a rather different intellectual animal from the caricature conjured up by Habermas. Here Grünbaum draws on his profound knowledge of modern physics to suggest that Habermas is something of a scientific illiterate, whose pronouncements about what is and what is not scientific are based on massive ignorance of the intellectual practices actually employed by scientists.

According to Habermas, psychoanalysis differs most importantly from science in that it does not aspire to causal knowledge. Rather than trying to explain human behavior in terms of general causal laws, it aims to dissolve the causal nexus of the natural world. “In technical control over nature,” Habermas writes,

we get nature to work for us through our knowledge of causal connections. Analytic insight, however, affects the causality of the unconscious as such. Psychoanalytic therapy is not based, like somatic medicine, which is “causal” in the narrower sense, on making use of known causal connections. Rather, it owes its efficacy to overcoming causal connections themselves.^[11]

Habermas’s thinking here is not easy to grasp. He apparently believes that an analytic cure actually destroys the causal tie between a repression and its neurotic symptom. Analysis, one might say, rescues the patient from the causal regime of nature. It lifts the patient out of the material world of causality into a more purely intellectual or spiritual realm where mundane causality is transcended. Psychoanalysis is, in effect, a doctrine of liberation.

Grünbaum is eager to dispatch this conception, because his own critique of Freud will focus above all on Freud’s failure to provide adequate evidence for the causal propositions at the heart of psychoanalytic theory. If Habermas is correct, Freud’s theory does not advance causal propositions at all; on the contrary, the theory radically undermines the role of causality in human behavior. Grünbaum has little trouble showing that this notion is repeatedly contradicted in Freud’s writings, which are saturated with causal claims. Indeed, they exhibit an almost compulsive search for causes—whether of neurotic symptoms, dreams, or slips. Grünbaum also argues that Habermas’s claim is philosophically “incoherent”: “Habermas slides from the therapeutic conquest of effects by the removal of their cause into the dissolution of the causal linkage between the pathogen and the neurosis. Overcoming an effect by undercutting its cause is hardly tantamount to dissolving the causal connection that links them” (11–12). In other words, psychoanalytic therapy does not abolish causal connections but instead makes use of them: the patient escapes his symptom by bringing to consciousness the repressed experience that gave rise to the symptom. Far from attempting to effect some mysterious liberation of human experience from the causal order, psychoanalysis traffics precisely in causal propositions—and in this regard it is indistinguishable from any natural science.

Another difference between psychoanalysis and science, in Habermas’s account, is that the analyst does not enjoy the sovereign intellectual authority over his subject exercised by the natural scientist. The order of nature, which the scientist investigates, exists only as an object, and the scientist alone is the source of knowledge about its functioning. In analysis a very different situation obtains. Here the “object” of investigation is himself or herself a subject. Indeed, in Habermas’s view, the patient stands in a privileged relation to the knowledge generated under analysis: not only is the patient the active source of all the material out of which interpretations are fashioned, but the truth of those interpretations ultimately depends on the patient’s embrace of them. “Analytic insights,” Habermas writes, “possess validity for the analyst only after they have been accepted as knowledge by the analysand himself.”^[12] By insisting on the privileged epistemological position of the analysand, Habermas undermines any notion of the psychoanalyst as functioning like a physicist or chemist, who confidently determines the laws to which the objects of his study must submit. Instead, the analyst’s constructions have the tentative or heuristic quality of interpretations in the humanities; as such they must be confirmed by the patient before they can be accepted as true. Once again, Habermas seeks to distance psychoanalysis from science by showing that it assumes a fundamentally different way of knowing, and hence a different method of validation.

Habermas’s argument for the patient’s epistemological sovereignty is perhaps less threatening to Grünbaum than his outright denial that psychoanalysis deals in causes. But it stands in opposition to one of Grünbaum’s central complaints about analysis, namely, that its findings are subject to evidential contamination. In Grünbaum’s view, the patient is a source not of truth but of error, a highly dubious authority, whose memory cannot be trusted and whose reliability is compromised by a desire to fulfill the analyst’s expectations. Indeed, the “suggestible” patient is the weak link—the Achilles’ heel—in Freud’s intellectual system. Grünbaum consequently cannot tolerate Habermas’s exaltation of that patient into the arbiter of analytic truth. He is thus eager to discredit the notion of the patient’s unique intellectual authority.

Grünbaum can easily show that Freud awarded the analysand no such veto power over analytic interpretations. In fact Freud’s prejudices were just the opposite: in the best medical tradition, he believed in the absolute intellectual superiority of the doctor to the patient. In several of the case histories Freud insists that his interpretations are correct even when the patient expressly rejects them—indeed, precisely because the patient rejects them. An especially clear-cut example is provided by the young lesbian subject of “The Psychogenesis of a Case of Homosexuality in a Woman” (1920), whose refusal of his analytic insights Freud attributes to a desire to punish him, as a surrogate for her father. The same prejudice is famously on display in the Dora case, where Freud shows an overweening contempt for the objections of his patient. More generally, the doctrine of the unconscious—the very heart of psychoanalysis—contravenes Habermas’s notion that the patient enjoys unique intellectual authority. The entire logic of psychoanalysis, one might say, undermines the traditional pretensions of the individual to self-knowledge: the psychoanalytic self is largely ignorant of its desires and deluded about its intentions. Habermas’s effort to restore to this self its traditional authority reveals a quaint loyalty to the ideals of the Enlightenment, but it is entirely foreign to the teachings of analysis.

Grünbaum goes on to argue that evidence from cognitive psychology supports Freud’s skepticism about the patient’s self-knowledge. Recent studies cast doubt on the ability of individuals to achieve genuine understanding of even their conscious motives. “When a subject attributes a causal relation to some of his own mental states, he does so—just like outside observers—by invoking theory-based causal schemata endorsed by the prevailing belief-system” (30). In other words, what passes for self-knowledge is often only the projection of one’s current intellectual prejudices onto one’s past. Habermas’s insistence on the analysand’s intellectual authority is thus contradicted not only by Freud’s practice but also by the latest experimental findings, which show that the individual is a poor judge of what has caused him to behave or think the way he does. It follows that the relationship between the analyst and his subject (a person) is not categorically different from that between the scientist and his subject (the world of nature). Once again, Habermas’s effort to create a conceptual opposition between psychoanalysis and science turns out to be misguided.

Finally, Habermas contends that analytic knowledge also differs from scientific knowledge in its radically historical character. The truths of psychoanalysis are time-bound and contingent, Habermas argues, while those of science are timeless and absolute. Grünbaum rejects this dichotomy just as vigorously as the others Habermas adduces. In this instance, however, Habermas’s error stems not from a misunderstanding of Freud but from his ignorance of science. Many of the propositions of the natural sciences, Grünbaum maintains, are just as historical—as “contextual”—as the propositions of analysis. Rather boldly, Grünbaum chooses his examples not from biology or geology—which on first blush might seem the most historical of the sciences—but from physics. Thus he argues that the physical theory of classical electrodynamics is every bit as historical as is psychoanalytic theory: “At any one instant, the electric and magnetic fields produced throughout infinite space by a charge moving with arbitrary acceleration depend on its own particular entire infinite past kinematic history!” (17). Grünbaum presents several other examples of this sort, which unfortunately convey the impression that he drastically misunderstands what Habermas means by “historical.” Grünbaum seems to confuse “history”—the cumulative and meaningful realm of human experience—with “time”—the hermeneutically neutral movement of particles through space. Put another way, the “infinite past history” of an electric charge is hardly the same as the finite past of a human life. As Grünbaum himself observes, almost wistfully, “Some hermeneuticians may retort that these physical cases do not capture the relevant sense of ‘history’ ” (19).

While one might sympathize with the objection that Habermas trades on “stone age physics” (20), making a virtue of scientific ignorance, there remains something wooden about Grünbaum’s uninflected insistence that the human and natural realms are entirely comparable—that knowledge of human behavior can assume exactly the same form as knowledge of the behavior of magnetic particles. Habermas may exaggerate the difference between science and psychoanalysis, and he certainly misrepresents Freud’s view of the matter, but Grünbaum’s hard-hat examples remind one that the hermeneutic construction of psychoanalysis has a certain credibility. Put another way, if the hermeneuticians err in failing to see that psychoanalysis differs from literature, Grünbaum errs in failing to see that it also differs from physics. Psychoanalysis aims to be a science of the self, but ideas about the self can never achieve the rigor of ideas about nature. Analysis thus necessarily exists on the border between science and the humanities. As much as Freud himself wanted his theory to conform to the model of physics, chemistry, and biology, his human subjects forced him to make a number of intellectual compromises, all of which bring his ideas legitimately into the hermeneutic orbit.

Paul Ricoeur first presented his hermeneutic interpretation of psychoanalysis in Freud and Philosophy (1970), followed in due course by Hermeneutics and the Human Sciences (1981). Ricoeur’s master theme is that psychoanalysis is not a science but a language—or, as Ricoeur puts it, psychoanalysis is a “semantics of desire.”^[13] Even more so than Habermas, Ricoeur seeks to bring Freud’s ideas into conformity with the linguistic turn of recent intellectual history—the effort to understand virtually all aspects of human behavior in terms of language. As a result of this linguistic perspective, Ricoeur is led to limit the proper subject of psychoanalysis to the verbal communications of the patient in the analytic situation. Psychoanalysis, he writes, is “a work of speech with the patient.”^[14] All the facts of psychoanalysis are linguistic facts, to wit, the words actually uttered by the patient and the interpretations those words inspire. As Jacques Derrida has it, “There is nothing outside the text”^[15]—in this case, nothing outside the text produced by the patient under analysis. Unlike the natural scientist, therefore, the psychoanalyst does not have access to a supposedly objective realm beyond the patient’s story. “There are no ‘facts’ nor any observation of ‘facts’ in psychoanalysis,” writes Ricoeur, “but rather the interpretation of a narrated history.”^[16] Ricoeur finds evidence for his restriction of psychoanalysis to the patient’s “narrated history” in Freud’s abandonment of the seduction theory: Freud rightly gave up any claim to identify an objective external experience as the source of the patient’s illness, confining his attention instead to the patient’s fantasies—a narrative rather than a historical reality—as revealed in the analytic session.

Ricoeur’s limitation of the proper domain of analysis to the patient’s utterances sends Grünbaum into paroxysms of philosophical castration anxiety. He protests against “this ideological surgery on the psychoanalytic corpus” (47), which he also describes as a “mutilation” (43), an “ontological amputation” (44), and an “emasculation” (60) of Freud’s views. Grünbaum shows that Freud himself entertained no such “truncated” (43) conception of the analytic domain. In fact, Freud often turned his attention to the patient’s nonverbal behavior (our fingers betray us, he said, even when our mouths don’t), and he speculated freely about the psychological meaning of mute artifacts like statues and paintings. Similarly, Freud’s substitution of fantasied for actual seductions in no way limits the scope of analysis to the patient’s narration, any more than it relieves Freud of the burden, which he shares with any scientist who makes a causal assertion, of proving that imaginary seductions in fact have the pathological effects attributed to them. Most important of all, in Grünbaum’s view, Freud clearly believed that his discoveries held true for individuals who had never been analyzed and thus never had occasion to produce a narrative account of their symptoms. In effect, Freud was convinced he had created a general psychology, one quite capable of universally valid statements about human behavior and motivation. He was hardly so modest as to limit his intellectual claims to whatever emerged directly from his patients’ stories. In keeping with the linguistic perspective on analysis, Ricoeur regards all analytic productions, including slips, dream symbols, and neurotic symptoms, as semantic structures. They are elements of a language, and their purpose, like that of ordinary language, is to convey meaning. Against this interpretation Grünbaum argues that Freud did not conceive of the “meaning” of symptoms or slips according to the communicative model of a language. Rather, the “meaning” of symptoms and slips refers to their “definite causal origin” (63). They resemble not linguistic forms but causal traces, like footprints in the sand: “The footprint is not, as such, a vehicle of communication; it is not a linguistic sign or symbol; it does not semantically stand for, denote, designate, or refer to the past pedal incursion” (64). Grünbaum illustrates this difference by way of the psychoanalytic theory of paranoia, which (as I have already indicated) especially fascinates him. When the paranoid expresses his repressed sexual desires through feelings of persecution, he is not—as the semantic theory would seem to imply—seeking to communicate the fact that he is a homosexual:

Paranoid behavior may well be a vicarious outlet for repressed homosexuality, but in no case is it a verbal label for it! Thus, as we saw, etiologically that behavior is the afflicted person’s attempt to cope with the anxieties generated by his unconscious sexual urges, not his/her attempt to communicate these yearnings by means of persecutory delusions and behavior. (66)

Here as elsewhere, Grünbaum argues, Ricoeur’s effort to treat psychoanalysis as a language, rather than a causal theory, stands at odds with Freud’s own conception. It is another misguided attempt to rescue psychoanalysis from the clutches of science.

Grünbaum ultimately judges the hermeneutic interpretation of Freud as unworthy of its subject. Habermas and Ricoeur represent “a nihilistic, if not frivolous, trivialization of Freud’s entire clinical theory” (58). Moreover, Grünbaum predicts, the hermeneutic rescue operation will in fact lead to the ruin of psychoanalysis. “Far from serving as a new citadel for psychoanalytic apologetics,” he warns, “the embrace of such hermeneuticians is, I submit, the kiss of death for the legacy that was to be saved” (58). Convinced that he has disposed of analysis’s hermeneuticist friends, Grünbaum turns next to its positivist enemies, in particular the philosopher Karl Popper. Significantly, he finds Popper’s criticism of Freud no more cogent than Habermas’s or Ricoeur’s defense of him.

Karl Popper and the Question of Falsifiability

Grünbaum was a critic of Karl Popper before he became a critic of Freud. Indeed, it was through Popper that Grünbaum was drawn to Freud. “The first impetus for my inquiry into the intellectual merits of the psychoanalytic enterprise,” he writes, “came from my doubts concerning Karl Popper’s philosophy of science” (xii). The philosophy of science in question was the theory of falsifiability, which Popper proposed in place of inductivism as the essential measure of scientific knowledge. Popper saw the need for this supposedly more rigorous standard, he said, when he encountered intellectual systems like Marxism and psychoanalysis that claim to derive from observation—that is, from some form of inductive reasoning—yet obviously fall far short of true science. “My problem perhaps took the simple form, ‘What is wrong with Marxism, psychoanalysis, and individual [Adlerian] psychology? Why are they so different from physical theories, from Newton’s theory, and especially from the theory of relativity?’ ”^[17] Popper’s solution was the notion of falsifiability: these pseudosciences may base their ideas on observation, but, unlike true science, they advance propositions that are not open to the possibility of disproof. A scientific theory is a high-risk affair; it asserts things that have a real chance of being contradicted by as yet undisclosed facts. Indeed, science conducts its business so as to encourage the discovery of precisely such disconfirming facts. A pseudoscience, by contrast, is never in danger of this embarrassment. Its propositions are so designed as to be immune to contradictory evidence, because every imaginable state of affairs can somehow be reconciled with them. In Popper’s view, the theories of Freud and Adler offer especially clear-cut examples of such nonfalsifiable (and hence unscientific) intellectual systems:

Neither Freud nor Adler excludes any particular person’s acting in any particular way, whatever the outward circumstances. Whether a man sacrificed his life to rescue a drowning child (a case of sublimation) or whether he murdered the child by drowning him (a case of repression) could not possibly be predicted or excluded by Freud’s theory; the theory was compatible with everything that could happen.^[18]

Psychoanalysis was “simply non-testable, irrefutable. There was no conceivable human behaviour which would contradict” it.^[19]

Grünbaum views this charge as astonishingly ignorant. In fact, Grünbaum professes amazement at Popper’s “obliviousness to Freud’s actual writings” (124). Instead of citing real instances of intellectual malfeasance, Popper rests his entire case on a hypothetical example (known as the drowning baby), which Grünbaum dismisses as “grossly contrived” (114). (It is more than contrived: the notion that intentionally drowning a baby might count as an instance of “repression” makes little psychoanalytic sense.) Grünbaum then proceeds to offer evidence, from the Freudian corpus, that controverts Popper’s strictures. In both theory and practice, Grünbaum insists, Freud honored the criterion of falsifiability.

An early example is provided by “A Reply to Criticism of My Paper on Anxiety Neurosis” (1895), in which Freud defends his hypothesis that anxiety neuroses are caused by a contemporary disturbance in sexual life, such as masturbation or coitus interruptus. In the paper Freud admits that his theory would be discredited if his critics could produce a case of anxiety neurosis in the absence of sexual anomalies. “My theory can only be refuted,” he writes, “when I have been shown phobias where sexual life is normal.”^[20] A Popperian might complain that Freud does not make this concession in the spirit of falsifiability; his “only” suggests a less than open mind about the likelihood of contrary instances. But the passage satisfies Grünbaum that Freud was prepared to play the scientific game by the rules—that he advanced empirically “risky” propositions. Freud’s statement, Grünbaum writes, “would do any falsificationist proud” (120).

Grünbaum is even more impressed by the 1915 paper “A Case of Paranoia Running Counter to the Psycho-Analytic Theory of the Disease,” whose very title seems to have a falsificationist ring about it. Here Freud discusses a seemingly paranoid woman in whom he finds no evidence of homosexuality. In other words, the case contradicts the psychoanalytic theory that paranoia is caused by repressed homosexual desire. Grünbaum admires the paper because in it Freud expressly recognizes the intellectual consequences of his negative finding: either he must give up his theory of paranoia, or he must contemplate the possibility that the woman is not in fact suffering from the disease. “Freud explicitly allowed that if the young woman was paranoid, then her case was a refuting instance of the etiology he had postulated for that disorder. Alternatively, he reckoned with the possibility that she was not paranoid” (109). Like Freud’s much earlier “Reply to Criticisms of My Paper on Anxiety Neurosis,” “A Case of Paranoia Running Counter to the Psycho-Analytic Theory of the Disease,” Grünbaum argues, is admirably sensitive to the criterion of falsifiability.

Unfortunately for Grünbaum’s latter illustration, Freud did not in fact abandon his hypothesis in the face of recalcitrant facts. Instead, he managed to wriggle off the methodological hook, and in a manner a Popperian would surely find incriminating. A second session with his patient turned up the gratifying evidence that she did indeed have an unconscious homosexual attachment to a woman under whom she worked: lo and behold, her paranoia had an orthodox analytic source after all. Thus, while the case may suggest a commendably rigorous theoretical commitment to the principle of falsifiability, in practice it shows Freud doing just what his Popperian critics find most objectionable about psychoanalysis: he gen erates new facts to make his patient’s circumstances fit the theory.

According to Grünbaum, then, the papers of 1895 and 1915 demonstrate that Freud understood and accepted the logic of falsification. But Grünbaum also insists that Freud practiced what he preached: on several occasions Freud actually gave up ideas because they proved empirically insupportable. That is, he operated just as Popperian theory says a scientist ought to, abandoning theoretical positions when they were contradicted by the facts:

Freud’s successive modifications of many of his hypotheses throughout most of his life were hardly empirically unmotivated, capricious, or idiosyncratic. What reconstruction, I ask, would or could Popper give us of Freud’s rationale for these repeated theory changes, and still cling to his charge of nonfalsifiability and/or to his charge that Freud was inhospitable to adverse evidence? (117)

The classic example of Freud’s changing his mind when the facts so obliged him is, of course, his decision to abandon the seduction theory. Grünbaum frequently draws attention to the famous renunciation letter of September 21, 1897, in which, Grünbaum says, Freud confesses “how adverse evidence that he himself had uncovered drove him to repudiate his previously cherished seduction etiology of hysteria” (117). Given all the controversy over Freud’s decision, this amounts to a rather simplistic account of the episode. Grünbaum leaves the impression that Freud had come upon new information that contradicted the seduction hypothesis, forcing him to renounce it. As we have seen, however, Freud’s letter in fact mentions no new evidence. Rather, it draws attention to a consequence of the seduction theory that Freud now finds empirically doubtful, namely, that the real sexual abuse of children was as common as the incidence of hysteria would require it to be (“surely such widespread perversions against children are not very probable,” he writes).^[21] Had Freud actually cited some sort of new evidence—that, for example, his hysterics’ stories of seduction had, in concrete instances, been contradicted by the testimony of adults—Jeffrey Masson would never have been able to mount his argument that Freud betrayed a discovery he knew to be true. The import of the seduction theory for Freud’s reputation as a scientific methodologist is, I’m afraid, less clear than Grünbaum would like to think.

Whether or not Freud himself believed in or practiced science according to Popperian criteria, his writings, Grünbaum insists, are full of assertions that can in fact be shown to be falsifiable. Psychoanalytic theory repeatedly predicts circumstances whose failure to materialize must result in its disconfirmation. As we’ve noted, Grünbaum’s favorite example of such an obviously falsifiable idea is the Freudian etiology of paranoia, which sets up a decidedly risky epidemiological prediction: “If repressed homosexuality is indeed the specific etiologic factor in paranoia, then the decline of the taboo on homosexuality in our society should be accompanied by a decreased incidence of male paranoia” (111). The theory also invites disconfirmation through historical or anthropological research, because it implies that paranoia will be relatively uncommon in societies, such as ancient Greece, less hostile to homosexuality.

Finally, Grünbaum’s strongest evidence against Popper is supplied by psychoanalytic ideas that have actually been proven false—the ultimate scientific compliment, one might say, in a Popperian universe. Here Grünbaum cites Seymour Fisher and Roger Greenberg’s “monumental” book The Scientific Credibility of Freud’s Theories and Therapy (1977), which summarizes a vast number of empirical studies and argues that, among other psychoanalytic propositions, Freud’s theory of dreams “has been contradicted by many scientific observations.”^[22] Freud’s ideas, in other words, both invite and receive empirical refutation. They are eminently falsifiable.

Grünbaum’s vigorous defense of Freud against Popper begs for explanation. As with his polemic against Habermas and Ricoeur, the reader wonders why Grünbaum has allowed himself to get so exercised over a philosophical critic whose animadversions on Freud seem to have been both brief and superficial. Part of the reason has nothing to do with psychoanalysis per se. As a philosopher of science, Grünbaum’s stock in trade has been the defense of classic Baconian inductivism, which he regards as the epistemological foundation of modern science. Indeed, his critique of psychoanalysis is framed as a defense of inductivism. Grünbaum agrees with Popper that psychoanalysis “does not come up to scientific standards” (106). But the fact that it nonetheless passes Popper’s falsifiability test proves that the test itself is a poor means of distinguishing science from pseudoscience. The bulk of The Foundations of Psychoanalysis is given over to arguing that Freud’s theory, despite its genuinely scientific aspirations, does not meet the traditional Baconian requirements. “Popper’s application of his falsifiability criterion is too insensitive to exhibit the most egregious of the epistemic defects bedeviling” psychoanalysis, while the “time-honored inductive canons for the validation of causal claims have precisely that capability” (124–25). The Freudian case thus serves to vindicate the familiar inductivist understanding of science and discredit the Popperian challenger.

But Grünbaum’s championing of Freud against Popper cannot be explained solely in terms of a long-standing philosophical commitment to inductivism: his heavily inflected language suggests that something else is going on. My own view is that Grünbaum is profoundly divided in his attitude toward Freud. Without doubt, his main purpose in The Foundations of Psychoanalysis is to prove that Freud’s intellectual brainchild is, as Grünbaum says more than once, “fundamentally flawed” (xii, 94, 124, 128). Put bluntly, he is out to get psychoanalysis. In the course of prosecuting his case, however, Grünbaum seems to have grown remarkably fond of the object of his abuse. Thus his later writings, like Foundations, are much warmer in their enthusiasm than are his early papers. It is difficult to avoid psychoanalytic language in describing this phenomenon. There seems to be an element of deferred allegiance to the repudiated father, a perhaps guilty attempt to atone for intellectual aggressions. Grünbaum’s ambivalence is the exact inverse of Frank Sulloway’s. Where Sulloway pretends to reveal Freud’s scientific originality while secretly holding him in contempt, Grünbaum sets out to dispute the scientific credentials of psychoanalysis but is unable to suppress a growing admiration for its creator. He seems truly angered by Karl Popper’s disrespectful attitude toward Freud, who emerges from Grünbaum’s pages as an intellectual giant, albeit a blemished one, and a scientific methodologist of the first order.

Grünbaum’s writings are in fact studded with tributes to Freud. Sometimes this praise offers Freud no more than his historical due as a force in modern intellectual life: his thought is “momentous” (39), “pioneering” (148), “epoch-making” (10). But Grünbaum also insists on the power and profundity of Freud’s ideas. No adjective comes more readily to Grünbaum’s mind in connection with Freud than “brilliant” (13, 93, 135). Freud displays “a soaring mind” (189) and a “brilliant theoretical imagination” (278). The caliber of his arguments is “astronomically higher, and their often brilliant content incomparably more instructive” (93) than are the glosses of his hermeneutic critics. He is the author of a “monumental clinical theory of personality” (94). His thinking is “admirably rich and lucid” (168). Even more striking than these generic celebrations of Freud’s intelligence is Grünbaum’s repeated insistence on Freud’s methodological acuity—in other words, his distinction in precisely that department of intellectual affairs where Popperians find him most wanting. Freud’s reflections on the philosophy of science are “pregnant” (42), and his concerns with the contaminating effects of suggestion “always unflagging” (129). He displays a “keen appreciation of methodological pitfalls that are commonly laid at his door by critics” (168). Indeed, he must be acknowledged “a sophisticated scientific methodologist, far superior than is allowed by the appraisals of friendly critics like Fisher and Greenberg or Glymour, let alone by very severe critics like Eysenck” (128, and see 172). In other words, Freud is to be admired not simply as a daring theorist but also as a “meticulous” (135) and “careful” (169) practitioner. When Freud falls short of his own high methodological standards (as he inevitably does), Grünbaum records such failures more in sorrow than in anger. He seems disappointed that Freud’s arguments are sometimes “unworthy” (141) of so incisive a scientific mind.

At the same time, Grünbaum’s animus toward Popper is not just a matter of an inadequately repressed admiration for Freud. It also reflects Grünbaum’s fear that the dismissal of psychoanalysis as “non-testable, irrefutable” will divert attention from Freud’s real shortcomings. Popper’s “indictment of the Freudian corpus as inherently untestable,” Grünbaum writes, “fundamentally misdiagnosed its very genuine epistemic defects” (xii). In this respect, Grünbaum’s hostility to Popper ultimately stems from the same source—or has the same rationale—as his hostility to Habermas and Ricoeur. Although Popper and the hermeneuticians hold diametrically opposed attitudes toward science—for Popper science is the greatest intellectual achievement in history, whereas for Habermas and Ricoeur science is an imperialistic threat to the interpretive understanding of human affairs—they agree, ironically, that psychoanalysis falls beyond the scientific pale. For the hermeneuticians, of course, this is a blessing, while for Popper it is a curse. But from Grünbaum’s point of view, both parties threaten to excuse Freud’s gravest defect as a thinker: his failure to supply adequate evidence for his ideas. Habermas and Ricoeur excuse this failure by arguing that the sort of hermeneutic knowledge psychoanalysis supplies does not require the same inductive support as do the truths of science. Popper, for his part, excuses it by arguing that the question of empirical evidence is beside the point when every conceivable state of affairs is compatible with analytic theory. Against both the apologetic left and the critical right, Grünbaum maintains that psychoanalysis is a science in aspiration, if not in fact. He wants to disprove Popper’s claim that psychoanalysis is unfalsifiable precisely because he is eager to falsify it.

Grünbaum has little difficulty showing that Popper’s case against Freud is slight and ill-informed. But Grünbaum’s dismissal of the charge of nonfalsifiability overlooks a genuine intellectual weakness of psychoanalysis, one that Popper obviously sensed and gestured toward—however crudely—with his example of the drowning baby. I have in mind the “heads-I-win-tails-you-lose” style of argument that pervades psychoanalytic reasoning. (Freud himself acknowledged the problem in his late essay “Constructions in Analysis,” where he responded to the charge that analysts construe the patient’s “no” to mean “yes” whenever it serves their purpose.) Psychoanalytic theory provides its adepts with too many interpretative alternatives—too many choices—which often seem to function as intellectual escape routes when the evidence is unaccommodating. In particular, concepts like resistance, ambivalence, overdetermination, and reaction formation let the analyst have it both ways—or, as Popper would insist, have it any way whatsoever. Thus, when one of Freud’s patients reported dreams that apparently revealed no hidden wish, Freud notoriously interpreted them as revealing the wish to disprove his dream theory! Clearly Popper was onto something when he charged that analysis is closed to the possibility of contradiction.

The answer to this line of criticism, put very generally, is that human beings are complex and the contingencies of life innumerable. A given action does not always have the same meaning, not even for a single individual. In terms of Popper’s hypothetical example of the drowning baby, a person may repress an impulse in one instance while sublimating it in another, and the desire to save (or drown) a baby might be variously motivated. A psychological system must leave room to maneuver if it wishes to make sense of the vagaries of human experience. By contrast, a theory that always gives unambiguous answers—that provides the binary predictions Popper seems to want—will inevitably flatten out the human reality it seeks to explain. Of course, the very flexibility of psychoanalytic theory means that the analyst must be exceptionally disciplined in its application, because the temptation to intellectual abuse is so great. In lesser hands, Freud’s ideas too often invite the sort of bad intellectual manners that Popper complains of.

The Tally Argument

Having disposed of Habermas, Ricoeur, and Popper, Grünbaum arrives at the centerpiece of his critique: his claim to have discovered in Freud a hitherto ignored philosophical defense of psychoanalysis, which he christens the Tally Argument. Grünbaum divides his energy about equally between celebrating the Argument’s virtues and exposing its weaknesses. His method is dialectical, or at least dramatic: he builds the Argument up so that its ultimate collapse will seem all the more ruinous. In effect, Grünbaum establishes Freud’s methodological sophistication in order to use it against him, suggesting that Freud neglected standards of proof whose legitimacy he fully understood. Freud’s shortcomings thereby seem willful rather than naive.

Grünbaum contends that the Tally Argument provides the philosophical justification for virtually all of Freud’s psychoanalytic concepts. But in fact the Argument bears mainly on the idea of the unconscious. Its implications for that other great pillar of psychoanalysis, the theory of infantile sexuality, are at best indirect. Significantly, Freud’s ideas about the beginnings of sexual life figure only marginally in Grünbaum’s analysis. His focus on the unconscious to the neglect of infantile sexuality makes his treatment of Freud very unlike the critiques of Frank Sulloway and Jeffrey Masson, in which Freud’s notions about the sexual lives of children are always the center of attention, while the unconscious is largely ignored.

As Grünbaum recognizes, Freud’s theory of the unconscious is distinguished by two features. First, it holds that ideas we are unaware of exert a significant influence over our behavior. Neurotic symptoms provide the prime example of such influence. Second, the mechanism through which ideas become unconscious is repression. Originally conscious, these ideas are forced into the unconscious when they prove incompatible with conscious convictions. “The theory of repression,” Freud says, “is the cornerstone on which the whole structure of psychoanalysis rests.”^[23] Thus the archetypal Freudian claim, whose credentials Grünbaum sets out to evaluate, asserts that a particular piece of behavior—typically a neurotic symptom—is caused by an unconscious idea, whose repression usually occurred some distance in the past.

If one asks Freud and other analysts what persuades them that repressed ideas actually have such consequences, they point over and over to the evidence gathered from clinical practice: their conviction, they say, derives above all from the observation of patients in analysis. “Most advocates” of psychoanalysis, Grünbaum writes, “regard the analyst’s many observations of the patient’s interactions with him in the treatment session as the source of findings that are simply peerless, not only heuristically but also probatively” (99–100). The clinical situation provides unique insights, in this view, both because a typical analysis lasts for several years and because the analyst follows the patient’s free associations wherever they may lead. Analysis thus allows for the accumulation of evidence about the patient’s life that is unparalleled in its quantity, detail, and nuance.

Not only do analysts insist that evidence from the couch supplies a firm empirical base for psychoanalytic ideas, but they are deeply skeptical about the value of any other kind of evidence, in particular the experimental and epidemiological data so beloved of academic psychology. Such statistical information, derived from a transitory laboratory setting, is, in their estimate, contrived and superficial and hence incapable of yielding the deep insights obtained in analysis. Freud himself set the pattern for the analytic tendency to belittle experimental evidence. When the psychologist Saul Rosenzweig sent Freud an account of experimental results that purportedly confirmed Freud’s ideas, Freud responded: “I have examined your experimental studies for the verification of the psychoanalytic assertions with interest. I cannot put much value on these confirmations because the wealth of reliable observations on which these assertions rest make them independent of experimental verification. Still, it can do no harm.”^[24]

The clinical defense of psychoanalysis suffers one great philosophical weakness: the possibility that information gathered from patients under analysis cannot be trusted. For some critics that information is unreliable because the sample on which it rests—persons who seek analysis—is unrepresentative. But a far weightier objection—and the one to which virtually all of Grünbaum’s attention is devoted—is that analytic patients are victims of suggestion. The interpretations that emerge in analysis, critics charge, are compromised by the analyst’s theoretical expectations. Far too often, the patient simply tells the analyst what the analyst wants to hear. Because information from the couch is so hopelessly tainted, it cannot be considered “probative.”

Grünbaum argues that, contrary to general belief, Freud was exquisitely sensitive to this criticism and went to great trouble to refute it. The principal evidence lies in two installments from the Introductory Lectures on Psychoanalysis, which Freud delivered to medical students at the University of Vienna during World War I. The first of these, on “Transference,” frankly acknowledges that suggestibility poses a more serious threat to psychoanalysis than it does to any other therapeutic procedure for the treatment of mental illness. Analysis makes the emotional tie to the doctor, the transference, absolutely central to the resolution of the patient’s neurosis. The transference stems from the recreation, in an analytic setting, of significant emotional relationships of childhood, with the doctor typically assuming the role of parent. Analysis thus invests the doctor—already an authoritative figure in any therapeutic situation—with the added authority of a parental surrogate. As a consequence, the patient is rendered even more vulnerable to the doctor’s intellectual influence. The transference, writes Freud, “clothes the doctor with authority and is transformed into belief in his communications and explanations.”^[25] Thus, as “Freud knew all too well,” the notion of the transference virtually invites the criticism that clinical findings reflect not “true insightful self-discovery” (130) but the patient’s compliance with the analyst’s suggestions.

If Freud’s lecture on “Transference” candidly acknowledges the extent of the problem, the next lecture, on “Analytic Therapy,” provides what is for Grünbaum the most considered methodological defense of psychoanalysis ever written. This is the so-called Tally Argument, in which Freud “brilliantly, albeit unsuccessfully, came to grips with the full dimensions of the mortal challenge of suggestibility” (135). Grünbaum returns to the crucial passage over and over in his writings, and one can fairly say that his entire philosophical critique of psychoanalysis ultimately depends on his reading of it. The passage exhibits the sweet reasonableness so characteristic of Freud’s expository works, in which he shows a masterly skill at anticipating his listeners’ objections:

But you will now tell me that, no matter whether we call the motive force of our analysis transference or suggestion, there is a risk that the influencing of our patient may make the objective certainty of our findings doubtful. What is advantageous to our therapy is damaging to our researches. This is the objection that is most often raised against psycho-analysis, and it must be admitted that, though it is groundless, it cannot be rejected as unreasonable. If it were justified, psycho-analysis would be nothing more than a particularly well-disguised and particularly effective form of suggestive treatment and we should have to attach little weight to all that it tells us about what influences our lives, the dynamics of the mind or the unconscious. That is what our opponents believe; and in especial they think that we have “talked” the patients into everything relating to the importance of sexual experiences—or even into those experiences themselves—after such notions have grown up in our own depraved imagination. These accusations are contradicted more easily by an appeal to experience than by the help of theory. Anyone who has himself carried out psycho-analyses will have been able to convince himself on countless occasions that it is impossible to make suggestions to a patient in that way. The doctor has no difficulty, of course, in making him a supporter of some particular theory and in thus making him share some possible error of his own. In this respect the patient is behaving like anyone else—like a pupil—but this only affects his intelligence, not his illness. After all, his conflicts will only be successfully solved and his resistances overcome if the anticipatory ideas he is given tally with what is real in him. Whatever in the doctor’s conjectures is inaccurate drops out in the course of analysis; it has to be withdrawn and replaced by something more correct.^[26]

Grünbaum of course dubs this the Tally Argument after the crucial verb in the penultimate sentence: the patient’s difficulties will be solved—his neurosis cured—only if the analyst’s interpretations “tally with what is real in him.” The passage, Grünbaum writes, contains Freud’s “cardinal epistemological defense of the psychoanalytic method of clinical investigation and testing, a pivotal vindication whose import had gone completely unnoticed in the literature, as far as I know, until I called attention to its significance in two recent papers” (135).

Grünbaum proceeds to “tease out” and give more precise philosophical expression to the assumptions of the Tally Argument.^[27] In essence, Grünbaum suggests, the Argument involves two propositions, whose “conjunction” (139) he calls the Necessary Condition Thesis, or NCT. The first proposition is that psychoanalysis alone provides insight into the unconscious causes of the patient’s illness: “Only the psychoanalytic method of interpretation and treatment can yield or mediate to the patient correct insight into the unconscious pathogens of his psychoneurosis” (139). The second proposition is that such insight is essential to the patient’s cure: “The analysand’s correct insight into the etiology of his affliction and into the unconscious dynamics of his character is, in turn, causally necessary for the therapeutic conquest of his neurosis” (139–40). Simply put, the truth of Freud’s ideas is guaranteed by the success of his therapy: his theories are validated by the fact that patients are cured. (Freud does not, be it noted, claim that analysis always results in cures; more modestly, according to Grünbaum, Freud says that analytic insight is a necessary but not a sufficient cause of therapeutic success.) The Tally Argument protects analytic interpretations from the charge of suggestion because only if those interpretations are true, the Argument asserts, will the patient get well. Interpretations that do not reflect the patient’s reality will not result in cures and, Freud asserts optimistically, will in fact wither away as the analysis proceeds. Grünbaum adds that just as individual cures assure Freud of the correctness of particular interpretations, so the cumulative therapeutic successes of analysis guarantee its general ideas: “Collectively, the successful outcomes of analyses…constitute cogent evidence for all that general psychoanalytic theory tells us about the influences of the unconscious dynamics of the mind on our lives” (140).

The Tally Argument, in Grünbaum’s construction, has two further implications, although Freud expressly mentions neither. Both involve empirical matters, and, as one might expect, they contain the seeds of the Argument’s downfall. The first is that the Argument implicitly rules out the possibility of spontaneous remissions—cures that happen without any kind of professional intervention. This conclusion follows logically from the Necessary Condition Thesis, which asserts that only analysis can provide the insights needed to effect a cure: spontaneous remissions, whatever their cause, are not produced by the insights of analysts. (Freud’s self-analysis might qualify as an exception, although “spontaneous remission” usually designates a return to health that results from nothing more strenuous than the ordinary business of living.) By the same logic, the Tally Argument commits Freud to the belief that analysis is therapeutically superior to all rival psychiatric methods, none of which, in Freud’s view, delivers insight into the repressed causes of neurosis—the sine qua non of therapeutic success, according to the Necessary Condition Thesis. Thus the twin spectres of spontaneous remission and rival cures hang like threatening empirical clouds over the Tally defense. If neurotics get well without psychiatric help, or if they get well through the ministrations of a non-Freudian therapist, then the Tally defense collapses. As we shall see, the philosopher David Sachs accuses Grünbaum of overburdening the Tally passage when he makes it an argument not just against suggestibility but against spontaneous remission and rival cures as well.

Grünbaum believes that the Tally Argument constituted the deepest source of Freud’s confidence in the truth of his ideas. It was “a veritable pillar” (163) of his doctrine and the ground for his “sovereign patronizing serenity” (170) in the face of charges that analytic “insights” were but spurious products of suggestion. It assured Freud, in a profound psychological way, that his method of clinical investigation was intellectually sound. Above all, the Argument gave Freud the unshakable conviction that clinical evidence was sufficient to validate his claims about the role of unconscious ideas in mental illness—that there was no need for recourse to statistical comparisons with untreated control groups. As Grünbaum portrays him, Freud always had the Tally defense hovering in the back of his mind as a kind of philosophical security blanket.

Grünbaum ignores certain obvious objections to this promotion of the Tally Argument. More than once he expresses surprise that no one before him seems to have recognized the Argument’s significance. But, of course, this neglect could simply mean that the Argument possesses neither the cogency nor, more important, the centrality in Freud’s thinking that Grünbaum claims for it. After all, Grünbaum’s entire case comes down to his reading of a single sentence in the vast Freudian corpus, and that sentence occurs in what Freud himself regarded as a piece of popular writing—a kind of haute vulgarisation—in which he presented his ideas to a nonanalytic audience. If the Argument were as fundamental to Freud’s thinking as Grünbaum says, Freud might have been expected to follow his normal practice of making it the subject of a technical paper or monograph. At the very least, he presumably would have offered a more systematic and extended discussion of its logic, rather than contenting himself with a single, terse sentence embedded in the middle of a university lecture (and introduced almost offhandedly with “after all”). One cannot escape the impression that Grünbaum has seized on a relatively casual remark and blown it up into a major intellectual event—making a philosophical mountain out of an expository molehill.

Grünbaum nowhere suggests, however, that this objection has crossed his mind. I don’t know whether his silence is a sign of cunning or of simple insensitivity to matters of tone and proportion. Nonetheless, Grünbaum apparently feels the need to lend a broader textual resonance to his claims for the Tally defense. Accordingly, he combs Freud’s writings for further passages that might be construed to reflect a similar line of reasoning. Perhaps the strongest support comes from the Little Hans case, in which Freud again defends his practice of providing the patient with “anticipatory ideas” (“Erwartungsvorstellungen”) and, in contrast to the Tally passage, explicitly casts doubt on spontaneous remissions. “Slight disorders,” Freud writes, “may perhaps be brought to an end by the subject’s unaided efforts, but never a neurosis.” To overcome a neurosis, “another person [i.e., the analyst] must be brought in, and in so far as that other person can be of assistance the neurosis will be curable.”^[28] Unfortunately for Grünbaum, the passage does not explicitly associate the analyst’s “anticipatory ideas” with the charge of suggestibility: it contains part of the Tally defense, but not all of it. It thus fails to dispel the suspicion that the Tally Argument is Grünbaum’s own artful concoction, which he has forced on Freud’s innocent philosophical imagination.

Grünbaum hears echoes of the Tally Argument whenever Freud makes a remark (however general) that associates psychoanalytic theory with its therapeutic application. Thus the Argument lies behind Freud’s assertion that in psychoanalysis “scientific research and therapeutic effort coincide” as well as his later claim that “in psycho-analysis there has existed from the very beginning an inseparable bond between cure and research.”^[29] Grünbaum professes to be “dumbfounded” (146) by contemporary analysts who would separate the theory from the therapy, as does Judd Marmor when he writes: “I suspect that it was largely the historical accident that Freud was attempting to earn a living as a psychiatric practitioner that drove him to utilize his investigative tool simultaneously as a therapeutic instrument.”^[30] Grünbaum warns that such a separation of theory from therapy invites disaster for psychoanalysis, since it ignores Freud’s profound understanding, as recorded in the Tally Argument, that the intellectual fortunes of analysis are inextricably linked to achieving cures.

Ironically, Grünbaum finds one of the earliest invocations of the Tally defense in Freud’s 1896 paper on “The Aetiology of Hysteria”—the very paper that Jeffrey Masson celebrates as “Freud’s most brilliant” because it contains his boldest assertion of the seduction theory.^[31] The irony, of course, is that therapeutic success is here made to testify on behalf of an idea Freud would repudiate a year and a half later in the most controversial intellectual about-face of his career. Nonetheless, Grünbaum detects the Tally Argument at work in Freud’s claim, in “The Aetiology of Hysteria,” that the seduction hypothesis had been “confirmed” therapeutically:

If you submit my assertion that the aetiology of hysteria lies in sexual life to the strictest examination, you will find that it is supported by the fact that in some eighteen cases of hysteria I have been able to discover this connection in every single symptom, and, where the circumstances allowed, to confirm it by therapeutic success.^[32]

Grünbaum also maintains that the collapse of the seduction theory did not lessen Freud’s confidence in the Tally Argument. Nor, in Grünbaum’s opinion, should it have. To be sure, if Freud’s hysterical patients had actually been cured by being given false insights into childhood events (seductions) that never occurred, the Necessary Condition Thesis would have been “strongly disconfirmed” (159). But Grünbaum infers—rather generously—that Freud must have come to regard at least some of those cures as bogus, perhaps because the patients in question suffered relapses. In support of this inference Grünbaum cites the famous renunciation letter of September 21, 1897, in which Freud points to therapeutic disappointment as a major reason for his loss of confidence in the seduction theory. (Freud speaks of “the absence of the complete successes on which I had counted.”)^[33] Thus, far from discrediting the Tally defense, the abandonment of the seduction hypothesis implies that Freud continued to rely on the assumption that cures are the guarantor of truth: he gave up the hypothesis precisely because of therapeutic failures. The seduction debacle, Grünbaum concludes, “provides no basis for judging Freud to have been intellectually dishonest when he explicitly enunciated NCT in 1909 [the Little Hans case] and 1917 [the “Analytic Therapy” lecture]” (159). But, one could object, while Freud may not have been dishonest, he was surely imprudent. Having confidently asserted in 1896 that therapeutic success confirmed his seduction hypothesis only to conclude the following year that at least some of those successes were bogus, Freud might sensibly have decided not to place so much trust in the evidence of cures. Certainly the experience ought to have made him leery about invoking cures as testimony to the correctness of his views.

Grünbaum’s master theme in The Foundations of Psychoanalysis is that the Tally Argument constituted a powerful defense of analytic ideas and that Freud rightly derived great comfort from it. Nonetheless, Grünbaum also insists that, for all its strengths, the Argument ultimately fails. Its failure is in fact his second major preoccupation, creating the strong suspicion that his original enthusiasm for the Argument was something of a setup. When Freud does not acknowledge that psychoanalysis is defenseless without the Tally Argument, his otherwise admirable methodological sophistication becomes grounds for questioning his integrity.

Grünbaum is surprisingly elusive about the exact reasons for the Argument’s failure. He gives two versions of the story, without saying which he considers the more important. And should we find either of these accounts unconvincing or think that the damage is not irreparable, Grünbaum holds yet another version—the definitive one, I suspect—in reserve.

In the first version, the Tally Argument collapses not for logical or empirical reasons but because Freud himself abandoned it. Like Marx’s capitalist, Freud in effect becomes his own gravedigger. This focus on Freud’s attitude toward the Argument, rather than on its inherent intellectual weaknesses, confirms the impression that Grünbaum’s perspective on Freud is as much psychological as philosophical. He is interested in the character of Freud’s conviction—the source of his persuasion—not just in whether that conviction was justified.

The key to the Argument’s fate, in Grünbaum’s first version of its downfall, lies in the evolution of Freud’s therapeutic views. All historians of psychoanalysis agree that Freud grew more pessimistic about analytic therapy during the final decades of his life—a pessimism fully in keeping with the darker atmosphere of his later thought, as reflected in such speculative works as Beyond the Pleasure Principle (1920) and Civilization and Its Discontents (1930). It was given its most melancholy expression in the 1937 essay “Analysis Terminable and Interminable,” in which Freud reveals grave doubts about the thoroughness and durability of analytic cures. Analysis, the essay concedes, cannot guarantee that the patient won’t suffer a recurrence of his affliction, any more than it can provide immunization against the outbreak of a different neurosis. Mental illness now appears to Freud more elusive and intractable than ever before. Analysis, accordingly, becomes “an interminable task.”^[34]

Grünbaum reads Freud’s late therapeutic pessimism as an implicit disavowal of the Tally Argument. The Argument posits a radical dependence of analytic ideas on therapeutic success, but Freud’s growing doubts about his ability to achieve genuine and lasting cures effectively stripped the Argument of its essential premise. “The import of this therapeutic pessimism is shattering” (160), Grünbaum writes. If analysis cannot produce cures, it forfeits its sole guarantee against the crippling charge of suggestibility. Freud thus gave up his only defense when he lost faith in analysis’s healing powers.

The gloomy outlook of “Analysis Terminable and Interminable” represents, according to Grünbaum, the culmination of a therapeutic retreat that began over a decade earlier. Although Grünbaum generally speaks of this retreat as a gradual process, he is inclined to regard the publication of Inhibitions, Symptoms and Anxiety in 1926 as a watershed: in it Freud expressly abandons one of the supposed pillars of the Tally defense, the denial of spontaneous remissions. “As a rule,” Freud there writes, “our therapy must be content with bringing about more quickly, more reliably and with less expenditure of energy than would otherwise be the case the good result which in favourable circumstances would have occurred of itself.”^[35] Once he demoted analysis in this fashion—making it a mere expediter rather than the indispensable cause of the patient’s recovery—Freud effectively gave up the Tally defense. “Unless analytic treatment is the paragon of the therapies as claimed in the Tally Argument,” Grünbaum concludes, “Freud himself has acknowledged that he cannot be assured of the inherent scientific value of psychoanalysis” (172).

Without ever saying so forthrightly, Grünbaum constructs a history of Freud’s methodological opinions, according to which Freud adopted the Tally Argument at the end of the nineteenth century (in his paper on “The Aetiology of Hysteria”), held to it for thirty years, and then abandoned it in 1926. Admittedly, the Argument received explicit formulation only in the Little Hans case of 1909 and the “Analytic Therapy” lecture of 1917. But Grünbaum implies that Freud nonetheless depended on it from 1896 to 1926—in other words, throughout the central three decades of his creative life.

This construction is superficially attractive, because many of Freud’s disparaging remarks about therapy do in fact date from his later years. For instance, in 1926—the very year in which he admitted the reality of spontaneous remissions—Freud wrote: “The future will probably attribute far greater importance to psycho-analysis as the science of the unconscious than as a therapeutic procedure.”^[36] That same year he also called himself “a supporter of the inherent value of psycho-analysis and its independence of its application to medicine.”^[37] And in the New Introductory Lectures of 1933 he remarked, only half-jokingly: “I do not think our cures can compete with those of Lourdes. There are so many more people who believe in the miracles of the Blessed Virgin than in the existence of the unconscious.”^[38] Clearly, as he reached the end of his life, Freud took a sober view of the therapeutic situation.

Nonetheless, Grünbaum’s historical construction is highly dubious. In fact, it is a phony history. One can assemble a substantial body of evidence to show that doubts about analytic therapy are by no means confined to Freud’s later thinking. “I have never been a therapeutic enthusiast,” he rightly said of himself.^[39] Nor are assertions about the independence of analytic ideas from their practical application unique to his final years. One of the most emphatic of such assertions occurs in the locus classicus of the Tally Argument, The Introductory Lectures of 1916–17, where Freud writes: “Even if psycho-analysis showed itself as unsuccessful in every other form of nervous and psychical disease as it does in delusions, it would still remain completely justified as an irreplaceable instrument of scientific research.”^[40] Grünbaum dismisses this embarrassing denial of the link between science and therapy as “a gratuitous piece of salesmanship, unworthy of the Freud who gave us the Tally Argument” (141). But in fact, during the very years when he was supposedly devoted to the Tally defense, Freud displayed a remarkable willingness to associate psychoanalysis with therapeutic failure—most spectacularly in the famous case histories. Dora (1901) and the Wolf Man (1914) were therapeutic fiascos, yet Freud insisted they yielded valuable insights. Because analytic failure prolonged treatment, it had the ironic effect of aiding discovery. Even in what he regarded as his one unqualified success, the Rat Man case (1909), Freud wrote that “the scientific results of psycho-analysis are at present only a by-product of its therapeutic aims, and for that reason it is often just in those cases where treatment fails that most discoveries are made.”^[41]

Judged in its entirety, then, the evidence does not support the historical pattern Grünbaum claims to detect. Rather, it suggests that throughout his career Freud’s view of the relation between therapy and science remained inconsistent: he seems to have been in the grips of a permanent ambivalence. Sometimes, allowing his hopes to get the better of him, Freud overstated both the therapeutic prospects and their significance for the truth of his ideas. At other times, his inherent conservatism or his annoyance with the excesses of enthusiasts such as Reich or Ferenczi led him to adopt a more circumspect view—to speak of the value of his ideas as independent of their practical consequences. Despite the undeniable dimming of his hopes in later years, Freud’s career cannot be neatly divided into three decades of therapeutic optimism (thanks to the Tally Argument) followed by thirteen years of despair—when, as Freud should have recognized, the Tally defense was discredited and analysis left in methodological shambles.

Grünbaum’s abortive attempt to link Freud’s notions about therapy and truth in an unambiguous historical pattern suggests that the Tally Argument may not have been the deep source of Freud’s confidence in psychoanalysis after all. His undiminished belief in his ideas after 1926 seems to Grünbaum incomprehensible—an act of sheer willfulness in the face of the unambiguously negative implications of his own therapeutic doubts. But Freud’s failure to draw these allegedly inescapable conclusions implies that in Freud’s mind the connection between therapy and science—between cures and truth—was much looser than Grünbaum would have us believe. This in turn implies that the real source of Freud’s conviction lay elsewhere and hence was undisturbed by the decline of his therapeutic hopes. Such is the conclusion reached by two of Grünbaum’s severest philosophical critics, David Sachs and Frank Cioffi—the former a friend of analysis, the latter an inveterate enemy—who, as we shall see, argue that Freud’s sublime confidence in his ideas had entirely different foundations. Not only for Freud himself but for many others who have found his ideas compelling, the intellectual attractions of psychoanalysis seem quite independent of its therapeutic claims.

Much of Grünbaum’s effort, I have indicated, goes into contending that the Tally Argument was abandoned by its own maker. This dialectical critique—his first version of the Argument’s demise—clearly holds great appeal for him. Because Grünbaum has labored so heroically to prove that the Tally Argument was the most powerful defense of psychoanalysis ever conceived, its dismantlement by Freud’s own hand seems all the more devastating. It conjures up Wagnerian images of self-immolation, a kind of psychoanalytic Götterdämmerung.

But perhaps sensing that his portrayal of Freud’s evolving therapeutic opinions might be open to challenge—and knowing full well that, in purely logical terms, Freud’s view of the Tally Argument is irrelevant to its validity—Grünbaum offers a second version of the Argument’s collapse. In this account the Argument comes to grief not because Freud betrayed it but because recent experimental studies have discredited one of its chief empirical supports, namely, the claim that psychoanalysis produces better results than its therapeutic rivals—of which, Grünbaum says, there are “at least well over 125” (161). In The Foundations of Psychoanalysis Grünbaum touches only briefly on these studies, but in his earlier papers they receive substantial attention and weigh heavily in his case against analysis.

The studies are remarkably similar. All were produced by teams of experimental psychologists, whose names litter Grünbaum’s texts like so many international law firms: Meltzoff and Kornreich; Fisher and Greenberg; Bergin and Lambert; Rachman and Wilson; Smith, Glass, and Miller; Strupp, Hadley, and Gomes-Schwartz. They write books with titles like Research in Psychotherapy, The Effects of Psychological Therapy, The Benefits of Psychotherapy, and Psychotherapy for Better or Worse, which, appropriately, come to more or less the same conclusions. First, the studies suggest that any form of psychotherapy is preferable to no therapy at all, because the recovery rate for persons who seek therapy is higher (though not dramatically so) than the incidence of spontaneous remissions. Second, they find that, among psychotherapies, analysis works no better than its competitors, several of the studies even judging it inferior to behavioral therapy. Finally, they draw the inference that, because the results obtained by different therapies are indistinguishable, the benefits of psychotherapy most likely derive from some feature common to all therapies. Grünbaum likes to refer to this as the placebo effect: therapies work not for the reasons given by their proponents but because they share some inadvertent factor. In all likelihood, the operative common denominator is nothing more mysterious than the therapist’s sympathetic ear.

The upshot of these findings, for Grünbaum, is that Freud’s implicit claim for the superiority of psychoanalytic treatment turns out to be groundless; therefore, the Tally Argument is scuttled. In particular, if psychoanalytic cures are “placebogenic” (161), then Freud was simply wrong when he argued that neurotics get well only through analytic insight into their unconscious motives. This conclusion inspires Grünbaum to a rare exercise of philosophical wit: “The therapeutic achievements of psychoanalysis are not wrought after all by the patient’s acquisition of self-knowledge, much to Socrates’ sorrow” (161). As Grünbaum is careful to note, the recent experimental studies do not actually prove that psychoanalysis works by placebo effect, although the implication is very strong. The Necessary Condition Thesis of the Tally Argument, he concludes, has suffered an “empirical demise” (171), and the principal defense of Freud’s ideas thus lies in ruins.

Grünbaum’s emphasis on recent experimental studies, like his emphasis on Freud’s late therapeutic pessimism, gives the impression that the Tally Argument may once have been valid but that it has since been twice discredited by historical developments. First Freud himself did it in when he lost confidence in analytic cures, and then, for good measure, it was done in again by empirical findings, which showed that analysis is no more effective than other psychotherapies. But, in truth, Grünbaum does not really believe in his own historical scenario. Ultimately, neither Freud’s change of heart nor the recent findings of experimental psychology invalidated the Tally Argument. Rather, the Argument was invalid from the start. Grünbaum’s painstaking account of its historical rise and fall is disingenuous.

The Argument’s fundamental defect was plain from the beginning. It makes an empirical assertion that must be backed up with evidence—but the evidence is not forthcoming, and even the need for it goes unacknowledged. The empirical assertion is that neurotics will get well only if they obtain insight into the unconscious causes of their neuroses: in Freud’s own words, the patient’s “conflicts will only be successfully solved and his resistances overcome if the anticipatory ideas he is given tally with what is real in him.” Despite his initial pretense to take this reasoning seriously, Grünbaum obviously finds Freud’s claim hollow unless it can be supported by statistical comparisons with neurotics who have not been given insight into the presumed unconscious causes of their neuroses. As evidence, analytic successes—even spectacular ones—by themselves mean nothing, because factors other than insight may be at work. That is, these successes may be placebo effects—attributable not to insight but to the analyst’s solicitousness. One can cogently argue that analytic insights produce cures only by carrying out controlled experiment: such insights must be provided to some neurotics but withheld from others, and the results compared. In Grünbaum’s view, then, Freud’s ideas could never be vindicated by purely clinical evidence, as the Tally Argument seems to promise. They always stood in need of extraclinical confirmation. From the perspective of most psychoanalysts, of course, this demand for experimental controls represents a counsel of perfection; it sets a standard of proof that can never be met when the subject under investigation is as complex and elusive as an individual’s psyche. In the study of human behavior, such certainty can be attained only if one is willing to limit inquiry to the banal and the trivial.

Although he does not say so, Grünbaum clearly thinks that the need for experimental controls should have been obvious to Freud when he advanced the Tally defense. Freud’s later therapeutic pessimism and the recent findings of experimental psychologists are therefore, strictly speaking, philosophically irrelevant. One need not await actual empirical disconfirmation in order to pronounce an argument unsound. The Tally Argument, by Grünbaum’s own inductivist criteria, was always unsound. Moreover, Freud would have been a much less astute methodologist than Grünbaum claims he was if he failed to recognize its unsoundness. Indeed, the Argument’s obvious defects again make one doubt Grünbaum’s entire case for its centrality in Freud’s thinking. It appears a product more of Grünbaum’s fertile imagination than of Freud’s presumed philosophical rigor.

In his later writings on Freud, notably in The Foundations of Psychoanalysis, Grünbaum tends to obscure his own negative judgment of the Argument, presumably in order to promote his discovery of the Tally defense. But earlier he was uncompromising. Thus in 1978 he wrote:

Neither Freud nor other psychoanalysts have controlled for inadvertent placebo effects by means of appropriate research studies. And since the rival hypothesis of placebo effect thus stands unrefuted, such treatment successes as analysts may achieve beyond spontaneous remission cannot warrantedly be adduced as support for the therapeuticity of analysis as such.^[42]

In the same essay Grünbaum also expressly dismissed Freud’s central assertion in the Tally passage: “There is no cogency in Freud’s empirical claim that an analysis can be therapeutically successful only if the analyst’s interpretations ‘tally with what is real’ in the patient.”^[43] Thus Grünbaum’s elaborate demonstration, in The Foundations of Psychoanalysis, that the Tally Argument remained alive and well until it was sabotaged by Freud himself and subsequently by empirical studies is ultimately a charade. The Argument was dead in the philosophical water from the start. The only real mystery is why Grünbaum has expended such energy on it—first coaxing it out of Freud’s reluctant text, then celebrating its supposed philosophical virtues, and finally composing a mythical trajectory for its rise and fall—when he obviously considered it “fatally flawed” from the outset.

Dreams and Slips

Freud arrived at his idea of the unconscious by way of the neuroses, and neurotic symptoms always remained for him the principal evidence for the influence of unconscious thoughts on human behavior. But very early in his psychoanalytic career Freud also sought support for his theory of unconscious motivation in two other phenomena: dreams and slips (or “parapraxes”). Dreams and slips became, respectively, the subjects of two of his most influential books: The Interpretation of Dreams (1900) and The Psychopathology of Everyday Life (1901). Freud returned to these topics over and over again whenever he wanted to argue the case for the unconscious.

Most discussions of Freud’s ideas about dreams and slips fail to recognize that their primary function was always an evidentiary one. Freud himself was responsible for this misunderstanding, because with regard to both dreams and slips he advanced a provocatively universalist hypothesis, thereby drawing attention away from his more modest intellectual goal of providing evidence for the unconscious. Freud insisted that all dreams could be interpreted as the fulfillment of unconscious wishes, just as all slips (at least all significant ones) were caused by unconscious impulses. But for purposes of establishing the reality of the unconscious, these universalist claims were logically unnecessary. All Freud needed to prove was that some dreams and slips can be explained only by the assumption of unconscious motivation.

Much of the critical literature on Freud’s theory of dreams and slips argues that, in many instances, these phenomena can be accounted for more readily in other ways and that Freud’s universalist hypotheses are therefore wrong. Grünbaum cites a good deal of this literature and endorses its general conclusions. He is particularly impressed by Sebastiano Timpanaro’s contention, in The Freudian Slip, that parapraxes allow of a variety of linguistic explanations. He also gives an uncritical account of the views of R. W. McCarley and J. A. Hobson, who maintain that the theory of dreams as wish fulfillments has been discredited by neurophysiological findings. But Grünbaum shrewdly recognizes that disproving Freud’s explicit hypotheses does not get to the heart of the matter. Even if dreams and slips do not always have the specific purpose Freud assigns them, they can still offer evidence for the influence of unconscious motives on human behavior. If Grünbaum is to sustain his case that clinical material alone cannot underwrite psychoanalytic ideas, he must show that dreams and slips fail to supply Freud with cogent reasons for believing in the unconscious.

Substantial portions of Grünbaum’s writings on psychoanalysis are devoted to precisely this endeavor. In essence, Grünbaum contends that dreams and slips do not provide genuinely autonomous evidence for the unconscious. Rather, the argument from dreams and slips is dependent on the historically prior and intellectually more fundamental argument from the neuroses. Or, as Grünbaum likes to put it, the theory of dreams and the theory of slips are “epistemically parasitic” (167) on the theory of neurotic symptoms.

Grümbaum justifies this subordination of dreams and slips to symptoms by an appeal to Freud’s notion that both dreams and slips are analogous to the neuroses. In a sense, Freud suggests, dreams and slips are manifestations, in normal experience, of the same psychic processes that lead to the development of symptoms in neurotics. Like symptoms, they are compromise formations, in which an unconscious impulse finds expression by assuming a strategic disguise. Hence they are appropriately thought of as “mini-neuroses.” Freud returns to this comparison repeatedly. “Dreams,” he writes, “are constructed like a neurotic symptom: they are compromises between the demands of a repressed impulse and the resistance of a censoring force in the ego.”^[44]

Grünbaum holds Freud to the letter of his analogy, pursuing its implications with a vengeance. The dream and slip theories, Grünbaum insists, are in fact merely extrapolations from the theory of the neuroses. They are, so to speak, the theory of the neuroses writ small. The claim that dreams and slips are caused by unconscious ideas therefore ultimately derives its legitimacy from the claim that neurotic symptoms are so caused. This means that, as intellectual parasites on his theory of the neuroses, Freud’s ideas about dreams and slips also must rely on the Tally defense for their methodological validation: they are believable only to the extent that Freud can back up his conception of the neuroses with cures. And because the Tally defense has been discredited, Grünbaum concludes, Freud’s ideas about dreams and slips have been robbed of their essential intellectual support, however indirect that support might seem. The failure to produce cures thus sinks the entire psychoanalytic ship, as the evidence of dreams and slips is washed away along with that for Freud’s neurotic etiologies.

Grünbaum drives this analogical critique to even further lengths. He argues that the repression theories of dreams and slips are actually “misextrapolations” (194) from the repression theory of the neuroses. The analogy is imperfect, he maintains, because it lacks an essential component of the theory of neurotic symptoms: the claim that the “disorder” can be alleviated by bringing its repressed cause to consciousness. In the case of dreams and slips, there is nothing to correspond to the therapeutic payoff that supposedly confirms Freud’s conception of the neuroses, namely, the elimination of the symptom. Grünbaum even suggests what such a payoff might look like if Freud had only seen the analogical necessity for it. In the case of dreams, making the dreamer aware of his unconscious motives ought to have the “therapeutic” effect of causing him to dream less, perhaps ultimately not at all. With slips, a knowledge of their unconscious source should enable the “patient” to overcome the habit of committing such errors. Freud is taken to task for failing to advance these claims, thus depriving the argument that dreams and slips are caused by unconscious motives of even the “prima facie therapeutic support” (194) that lends plausibility to the psychoanalytic theory of the neuroses. Without confirmation by something analogous to cures, the assertion that dreams and slips testify to the influence of the unconscious is, in Grünbaum’s view, entirely empty.

Freud himself clearly never intended his analogy to be taken so literally or pushed to such surrealistic extremes. The analogy with symptoms was meant primarily as a heuristic device, intended to shed light on his ideas about dreams and slips by drawing attention to certain resemblances with his findings about the neuroses. In other words, the analogy was not the central epistemological bulwark that Grünbaum imagines. Freud obviously felt that the evidence for the unconscious provided by dreams and slips was persuasive on its own terms. It spoke directly; it did not depend on some circuitous justification by way of neurotic cures. Grünbaum is of course aware of this conviction, and, accordingly, he is at pains to show that Freud’s way of defending his ideas about dreams and slips does not pass philosophical muster.

When Freud argues that a slip or a dream originates in an unconscious idea, he appeals to two kinds of evidence. The first is thematic affinity: he identifies a substantive likeness between the slip or the dream on the one hand and the unconscious idea on the other. Second, Freud points to free association. If the dreamer or the person who commits the slip allows his mind to travel undisturbed along the chain of associations that the dream or slip gives rise to, he will eventually be led to the motivating idea.

Grünbaum illustrates Freud’s dependence on both thematic affinity and free association—as well as their liabilities—with an example that Freud himself discusses at the beginning of The Psychopathology of Everyday Life, the so-called aliquis slip. A young Jewish man expresses his anger about anti-Semitism by quoting Freud the line from the Aeneid in which Dido invites posterity to wreak vengeance on Aeneas: Exoriare aliquis nostris ex ossibus ultor (“Let someone arise from our bones as an avenger”).^[45] But the young man misremembers the line: he leaves out the word aliquis (“someone”). In order to determine the unconscious meaning of the slip, Freud asks the man to relate whatever comes into his mind in connection with the missing word. The man obliges by producing a series of associations. Aliquis first divides into a and liquis. Liquis then gives rise to Reliquien (“relics”), liquefying, fluidity, and fluid. Reliquien in turn leads to religious associations, notably with Saint Augustine and Saint Januarius (both of whom, Freud observes, have to do with the calendar). Saint Januarius returns the young man’s associations to the earlier notion of liquefying, because the saint’s blood, kept in a phial in a Neapolitan church, is said to liquefy miraculously on a particular holy day. The thought of Saint Januarius’s liquefying blood (in Naples) produces the final association and, according to Freud, the unconscious idea that inspired the original lapse: the young man is reminded of his Italian mistress and in particular of his fear that her periods may have stopped. Freud congratulates his interlocutor: “You’ve made use of the miracle of St. Januarius to manufacture a brilliant allusion to women’s periods.” “And you really mean to say that it was this anxious expectation that made me unable to produce an unimportant word like aliquis?” asks the skeptical young man. “It seems to me undeniable,” Freud responds.^[46]

The stream of the young man’s associations is united by certain thematic affinities, of which the calendar and the blood that flows on a particular day are the most obvious. The original forgetting of aliquis is bound to its unconscious cause by the overarching theme of descent: the young man’s repressed fear that his mistress might be pregnant has interfered with his ability to produce a correct version of his wish for descendants who will avenge the wrongs of the Jews. To be precise, it has blocked out his ability to remember the very word that alludes to the avenging descendant, aliquis. Freud explains the underlying logic of the slip as follows:

The speaker had been deploring the fact that the present generation of his people was deprived of its full rights; a new generation, he prophesied like Dido, would inflict vengeance on the oppressors. He had in this way expressed his wish for descendants. At this moment a contrary thought intruded, “Have you really so keen a wish for descendants? That is not so. How embarrassed you would be if you were to get news just now that you were to expect descendants from the quarter you know of. No: no descendants—however much we need them for vengeance.”^[47]

Readers are apt to respond variously to Freud’s explanation. For some it will seem overwrought: surely nothing so complicated is needed to account for the simple inability to remember a word in a quotation, especially a quotation in a foreign language. The fact that the young man forgot aliquis seems less remarkable—and thus less in need of explanation—than the fact that he remembered the rest of the line. For others, however, the ingenious sequence of associations, the uncanny echo of descendants desired and undesired, and, above all, the frisson of the ultimate revelation, in which high-minded indignation is done in by a dirty erotic secret, will seem undeniably beguiling. Freud, they will conclude, is onto something.

Grünbaum does his best to discourage the latter reaction by showing that it is open to severe philosophical objections. Neither thematic affinity nor free association, Grünbaum argues, carries the evidential weight necessary to justify such a causal assertion. He grants, for the sake of argument, Freud’s assumption that the associations are genuinely free, that is, that they have not been contaminated by the analyst’s proddings or expectations—although Grünbaum thinks there is reason to doubt this. Still, even assuming that the “meandering associations starting out from the restored memory of aliquis” (192) are spontaneous, they simply do not represent cogent grounds for thinking that the young man’s unconscious worry about his mistress’s pregnancy caused him to forget the word aliquis. Put another way, although one might argue that aliquis launched the string of associations that brought forth the man’s repressed anxiety, one cannot reverse the causal sequence and claim that the repressed anxiety caused the original forgetting:

Let it be granted that [the] chain of associations from his corrected parapraxis issued causally in the disclosure of the repressed anxiety afflicting him, and that this unconscious fear of pregnancy had been clamoring for overt expression. How, then, does this assumed motive serve to explain even probabilistically why [he] suffered any memory loss at all, let alone why he forgot aliquis? (197)

The inadequacy of free associations for establishing causes is illustrated for Grünbaum by the problem of how the analyst decides where to end the associative sequence. After all, the associations could be extended indefinitely, with a view to turning up yet deeper unconscious motives. What determines that one association rather than another is the true source of the slip? Grünbaum rejects any appeal to confirmation by the person committing the slip, who may, of course, testify to the reality of a particular fear or desire but who is no more expert than the rest of us concerning “the alleged causal nexus between the given fear and the slip” (208).

Ultimately, Grünbaum insists, to argue that free association can identify causes is to fall victim to the error of post hoc ergo propter hoc: the temporal fact that the revelation of the supposed source occurs after the slip does not prove that the slip was in fact caused by that “source.” Freud’s “epistemic tribute to free associations,” Grünbaum complains, “rests on nothing but a glaring causal fallacy” (186). “To endow the unconscious with cunning, uncanny powers of intrusion upon conscious actions is only to baptize the causal fallacy by giving it an honorific name” (192). Nor can Freud’s defenders fall back on the argument that Freud at least has offered an explanation for slips, whereas such errors have generally been ignored or considered inexplicable—in violation of the scientific principle that nothing in the world is uncaused. On the contrary, Grünbaum responds, Sebastiano Timpanaro and others have provided alternative explanations for slips that beg fewer questions. Moreover, similar psycholinguistic explanations were available to Freud when he wrote The Psychopathology of Everyday Life, and Grünbaum judges Freud’s dismissal of those competing theories particularly weak. “Not even the tortures of the thumbscrew or of the rack,” Grünbaum concludes rather picturesquely, “should persuade a rational being that free associations can certify pathogens or other causes!” (186).

As to thematic affinity, the notion is far too elastic, Grünbaum argues, to provide free association with the sort of discipline that might rescue it from the charge of post hoc ergo propter hoc. With sufficient ingenuity one can generate unlimited thematic echoes, especially if the stream of associations is not prematurely interrupted. “For any emerging repression,” Grünbaum writes, “it will be possible to find some thematic thread, however farfetched, such that there will be some topical kinship with the given lapse” (199). He clearly thinks that the thematic affinity linking aliquis with the feared pregnancy—the play on wanted and unwanted descendants—is an example of just such a farfetched connection. To illustrate the excessive elasticity of thematic affinity—as well as the related problem of where to terminate the sequence of associations—Grünbaum constructs a hypothetical extension of Freud’s interview with the forgetful young man (whom he calls AJ for “Austrian Jew”) and offers his own alternative explanation for the slip:

Suppose that Freud had allowed AJ to continue well past the disclosure of the pregnancy fear. Perhaps it would then have emerged that AJ’s parents had taught him early that the Romans had crucified Jesus, but that Christians had then unfairly blamed the Jews for deicide. It might furthermore have emerged that AJ had repressed his ensuing hatred of the Romans when Virgil, Horace, and other Roman poets were shown great respect in his Austrian educational environment.…Would AJ’s hypothesized repression of his hatred for the Romans not have had greater thematic “suitability as a determinant” of his aliquis slip than his anxiety about the pregnancy, even though the former assumedly emerged only later in the associative chain? After all, Virgil was a Roman, and AJ was citing the line from the Aeneid to express his conscious resentment of Christian anti-Semitism. What a golden opportunity to punish the unconsciously resented Romans simultaneously by spoiling Virgil’s line! Although the repressed hatred for the Romans is, of course, purely hypothetical in the case of AJ, it does lend poignancy to the complaint of selection bias, which is given substance generally by the thematic elasticity of the associations I have emphasized. (209–10)

One suspects that Freud would have responded, “Yes, but of course the man produced no such associations.” Still, the problem of using thematic affinity as a means of exercising intellectual control over free associations is genuine. The difference between Freud’s account and Grünbaum’s is ultimately aesthetic: Freud’s story is a good one, while Grünbaum’s falls flat. Indeed, the example lends support to Habermas’s and Ricoeur’s argument that psychoanalytic interpretation has more in common with literature than with science. In the matter of dreams and slips, Grünbaum would doubtless not only accept this verdict but argue that it makes his point: while thematic affinity may be fine for poets and novelists, it cannot be trusted as a guide to causes.

Because of the intellectual shortcomings of free association and thematic affinity, Freud’s theory of slips and his theory of dreams are thrown back on the theory of the neuroses for their epistemological salvation. But, as we have already seen, Grünbaum believes that this recourse is in vain. The analogy with the neuroses fails because it is imperfectly drawn—it lacks a therapeutic payoff—and in any event the cardinal defense of the theory of the neuroses, the Tally Argument, has itself been discredited. Thus the argument from dreams and slips can neither stand on its own nor find support from its crippled conceptual neighbor.

Grünbaum’s Critics

Grünbaum has elicited at least three significant responses. The analyst Marshall Edelson, who is a professor of psychiatry at Yale, devoted an entire book—Hypothesis and Evidence in Psychoanalysis (1984)—to answering his attack. More recently, two philosophers, David Sachs and Frank Cioffi, have published shorter but nonetheless incisive critiques of Grünbaum’s writings on analysis. Sachs is a defender of analysis, while Cioffi is a staunch opponent, yet, remarkably, they agree about where Grünbaum goes wrong.

Of the three, Marshall Edelson is the most sympathetic to Grünbaum. In fact, Edelson is in many ways an admirer, who thinks that analysts must take Grünbaum’s “formidable argument” very seriously.^[48] Edelson’s main concern is not with Freud but with the future of psychoanalysis. He fears that if Grünbaum’s objections go unanswered, the outlook for the profession is bleak. Young scientists and scholars will not want to pursue careers in the service of a tradition that has been intellectually discredited. Accordingly, the purpose of Edelson’s book is to argue that, despite Grünbaum’s weighty criticisms, analytic ideas can still be defended with clinical evidence.

Edelson accepts Grünbaum’s definition of the problem. He is not particularly interested in the Tally Argument and the supposed historical reasons for its collapse, but he concurs that the fundamental issue in the defense of analysis is therapeutic success. Freud’s ideas ultimately depend on the claim that analytic insights produce cures, although Edelson argues for a more modulated conception of “cure” than the rather mechanical notion of the removal of symptoms. Also like Grünbaum, Edelson rejects the effort to salvage analysis by separating theory from therapy—the ideas from their practical application—just as he rejects the hermeneutic escape route proposed by analysts like Roy Schafer and Donald Spence. For better or worse, psychoanalysis is committed to giving the patient a truthful, not just a narratively coherent, account of his circumstances, and it is distinguished from all other psychotherapies by its belief that such truthful insight alone is the key to psychic health.

Most important, Edelson agrees with Grünbaum that the great weakness of analysis has been its failure to provide its clinical findings with the necessary intellectual controls. Analysts are under the misapprehension that they can establish the truth of their ideas simply by piling up instances of therapeutic success. But Grünbaum has shown that therapeutic successes by themselves prove nothing. One must also eliminate (or at least reduce) the possibility that those successes can be explained otherwise than through analytic insight. That is, one must guard against their being placebo effects—the result of some inadvertent factor in the analytic situation, such as the analysand’s intellectual confidence in the analyst or the peculiar emotional tie between doctor and patient. Edelson thus implicitly endorses my view that the Tally Argument fails not for the historical reasons proffered by Grünbaum—because Freud himself abandoned it or because recent experimental findings contravene it—but because it was inadequate from the start. It always begged the question of how to eliminate alternative explanations.

Edelson nonetheless criticizes Grünbaum’s conclusion that one can eliminate rival explanations only through experimental control studies—that is, statistical comparisons between patients who have undergone analytic treatment and otherwise identical populations who have not been given insight into their unconscious motives. Edelson strongly defends the skepticism that analysts, beginning with Freud himself, have always felt about such studies. The effort to verify analytic ideas through experimental comparisons is bound to fail—not, as Grünbaum would imply, because analytic ideas are insupportable, but because the experiments can never be refined enough to measure the complex and nuanced matters explored under analysis. To demand comparative studies in which all relevant variables are controlled so as to rule out every conceivable alternative explanation is to offer “a counsel of perfection,” which analysts have rightly rejected.^[49]

Fortunately, Edelson argues, there is another way of obtaining the necessary controls, one that avoids the “breathtakingly nontrivial defects” of group studies.^[50] This alternative is to introduce controls within the clinical framework itself. To be exact, Edelson urges substituting time controls for group controls: information gathered from a single subject at different moments during the analysis can take the place of comparative data from an experimental group. The analyst can compare “time slices of the subject”—notably, slices taken before and after a particular interpretation has been given to the patient.^[51] To the extent that other factors have been held constant, the analyst will be able to rule out rival explanations for any changes that have occurred in the patient. In this manner the individual patient can serve as his own “historical” control, and single-subject research can acquire the probative quality of group-comparison research. This proposal for intraclinical testing is the heart of Edelson’s response to Grünbaum: temporal comparisons can lend clinical case studies something of the intellectual texture of group-comparison studies.

Edelson insists that Grünbaum himself gives his blessings to such a “time-series design” or “multiple baseline design” in his discussion of the cases in Studies on Hysteria, especially that of Anna O.^[52] Grünbaum in fact greatly admires the Anna O. case, because it shows that Freud was sensitive to the danger of placebo effect. In answer to the charge that Anna O.’s symptoms were lifted not by recovering their traumatic origins but through suggestion, Freud and Breuer cite the peculiar manner in which her symptoms disappeared: they vanished separately and independently, each one at the very moment its historical source was recalled. If Anna O.’s cure had resulted merely from suggestion—from the more or less constant influence of the doctor’s attentions—no such temporal pattern should have emerged. As Grünbaum concludes, approvingly: “The separate symptom removals are made to carry the vital probative burden of discrediting the threatening rival hypothesis of placebo effect, wrought by mere suggestion” (179).

There is merit, then, in Edelson’s contention that the Anna O. case anticipates the sort of time-controlled inquiry that Edelson hopes will rescue psychoanalysis. But he overestimates Grünbaum’s enthusiasm for the case. True, Grünbaum views it as a move in the right direction: having duly recognized the threat posed by the placebo argument, Freud and Breuer tried to respond to it by emphasizing the separate removal of individual symptoms. But this defense remains for Grünbaum ultimately unsatisfactory. The sequential lifting of symptoms could still be a placebo effect, because Anna O. must have known that Breuer hoped to uncover a thematically appropriate memory whenever he focused her attention on the first appearance of one of her symptoms. In other words, she may still have been obliging the doctor’s strongly felt expectations. Even when symptoms disappear one by one, therefore, the only way to establish that their removal has resulted from insight rather than suggestion is through comparisons with “a suitable control group whose repressions are not lifted” (180). If Anna O. is a prototype for the time-slice method Edelson advocates, the method fails to measure up to Grünbaum’s standards.

Edelson argues that the idea of comparing different moments in a single analysis is more than just a suggestion for the future. Psychoanalytic hypotheses not only can be, but already have been, successfully tested in the clinical setting on at least two occasions. Edelson’s book culminates in his discussion of these two cases, which offer what he considers a “decisive refutation” of Grünbaum.^[53] I fear, however, that the cases make rather a poor climax: Hypothesis and Evidence in Psychoanalysis is long on foreplay but short on action. One of the cases is merely a reconstruction, by Clark Glymour, of Freud’s Rat Man study, in which Glymour argues that Freud might have been seen as using a “bootstrap strategy” to test and reject rival hypotheses.^[54] It proves nothing beyond good intentions. Apparently, then, the sole instance in which the time-slice method has actually been put to the test is Lester Luborsky’s case of “Miss X” (1974). By comparing different moments in her analysis, Luborsky shows that Miss X suffered episodes of forgetting whenever her emotional involvement with the analyst grew especially intense. Luborsky establishes this connection by pairing passages—or, as he prefers to say, “contexts”—from her analytic sessions. As Edelson explains it, “each pair of contexts, differing with respect to whether momentary forgetting has or has not occurred, has been matched…so that the contexts can otherwise in crucial respects be regarded as equivalent.”^[55]

Unfortunately, whatever the merits of this experiment, it is a slender empirical reed on which to rest the argument for a purely clinical defense of psychoanalysis. Grünbaum justly complains that the Miss X case does not even address—let alone substantiate—any of Freud’s significant causal hypotheses. In particular, it has no bearing on the essential Freudian claim that insight into a patient’s unconscious motives is essential to therapeutic success. If anything, it tends to prove just the opposite, namely, that patients respond not to insight but to the emotional influence of the analyst—which Grünbaum, of course, would categorize as a placebo effect. If Luborsky’s Miss X is the best that Edelson has to offer, the prospects for the intraclinical confirmation of analytic ideas must be pronounced exceedingly faint.

In Grünbaum’s view, Edelson’s defense of intraclinical testing is worse than unpersuasive: it is intellectually irresponsible. It holds out a false hope and thus encourages further delaying tactics, when in fact analysts desperately need to recognize the inadequacy of purely clinical evidence and set about devising appropriate experimental control studies. “It is now ninety years since the publication of Studies on Hysteria. The hour is late, and the bell is tolling.”^[56]

Marshall Edelson’s mistake, I would suggest, is to have accepted Grünbaum’s ground rules. Once you embrace the premise that everything hinges on cures, it is difficult to escape the demand for experimental controls. David Sachs and Frank Cioffi both avoid this mistake. They challenge the fundamental assumption that therapeutic success was the source of Freud’s confidence in his ideas. Grünbaum errs, they maintain, in arguing that psychoanalysis lives or dies with its ability to produce cures.

David Sachs attacks Grünbaum along two fronts. First, Sachs demonstrates that the Tally Argument is a product of Grünbaum’s aggressive misreading of the passage in the “Analytic Therapy” lecture from which the argument is extracted. Grünbaum burdens the passage, Sachs insists, with claims it simply does not make. More important, Sachs argues that Freud’s ideas rest on a much broader evidential base than Grünbaum allows. They do not depend exclusively on clinical data, and even less on clinical data illustrating therapeutic success. Grünbaum’s conception of the “foundations” of psychoanalysis is thus shown to be excessively narrow.

According to Sachs, Freud’s real purpose in the Tally passage is quite modest: he is concerned with answering the charge that analysis works entirely by suggestion. “What it says is tantamount to the following: unless the suggestions an analyst makes to his patient correspond to facts about him, an understanding of his conflicts will not be attained, and his resistances will not be defeated.”^[57] The passage contains nothing to justify the two further contentions with which Grünbaum saddles it: that cures never occur spontaneously and that nonanalytic therapies are ineffective. Grünbaum simply puts words in Freud’s mouth when he characterizes the passage as a “bold assertion of the causal indispensability of psychoanalytic insight for the conquest of the patient’s psychoneurosis” (139). In effect, Sachs shows that the Tally Argument, properly speaking, doesn’t really exist. It is a figment of Grünbaum’s overwrought philosophical imagination.

Moreover, Freud would never have argued that analytic insight is indispensable to achieving cures, because, as Sachs documents, Freud always recognized that neurotics may recover either on their own or through the aid of other therapies. In the “Analytic Therapy” lecture itself, Freud recalled the “complete and permanent” cures he had sometimes achieved, in the 1890s, using the pre-analytic technique of hypnotic suggestion.^[58] Freud also speculated that neuroses might one day succumb to purely chemical treatment—which would obviously mean without benefit of analytic insight. On the matter of spontaneous remissions, he acknowledged in 1913 that all the disorders successfully treated by analysis were also “occasionally subject to spontaneous recovery.”^[59] Grünbaum’s insinuation of the “causal indispensability” claim into the Tally passage is thus contradicted by clear evidence that Freud knew all along about both spontaneous remission and rival cures. Furthermore, Freud held that analysis could provide correct views even when it was powerless to relieve the patient’s complaint, notably in the understanding of the psychoses. Dementia praecox and paranoia were immune to analytic therapy, yet Freud confidently insisted that they yielded to analytic explanation. His ideas were reliable, he felt, even when they were therapeutically unavailing.

Sachs’s demonstration that Freud neither formulated nor believed in the Tally Argument points to what is surely the central mystery of Grünbaum’s critique: why he goes to such lengths to create this philosophical mirage, which he even equips with a mythic (and tragic) history. It would be both petty and foolish to suggest that Grünbaum is motivated by base academic instinct, the desire to claim a new textual discovery: he seems too caught up in his philosophical enterprise to be concerned with scoring scholarly points. More likely, as Sachs suggests, Grünbaum has chosen to employ an inflationary and deflationary tactic in his effort to discredit psychoanalysis. First he creates the impression that all Freud’s ideas depend on therapeutic claims (the inflationary opening move); then he shows that those claims are unwarranted, which leads in turn to the inevitable conclusion that the original ideas are without support (the deflationary goal). Because Grünbaum feels so confident he can demolish the therapeutic pretensions of analysis, he is driven to sweep the whole analytic operation into the therapeutic bin. The Tally Argument serves this purpose ideally. It puts the assertion that analytic ideas depend on analytic cures into Freud’s own mouth—where, however, it clearly does not belong.

More telling than Grünbaum’s misconstrual of the Tally passage is, in Sachs’s view, his neglect of the wide variety of nonclinical material on which Freud based his ideas. Many of Freud’s writings, Sachs reminds us, are devoted to showing that analytic ideas find support in such diverse nonclinical phenomena as sexual behavior, jokes, religious ceremonies, mythology, folklore, literature, sculpture, and painting. Not all this extraclinical evidence is equally impressive, but that it did much to inspire Freud’s conviction seems undeniable. The same breadth of intellectual reference also contributes powerfully to Freud’s ability to attract adherents. Grünbaum, however, ignores it, arbitrarily allowing only clinical evidence— indeed only clinical evidence of therapeutic upshot—to bear any weight in Freud’s calculations.

Sachs is particularly offended by Grünbaum’s treatment of the two most important nonclinical sources Freud relies on: dreams and slips. Many of the dreams and at least some of the slips Freud discusses of course originate in a clinical setting: they are produced by his patients, often in association with their neurotic symptoms. But a substantial portion of the dreams and the vast majority of the slips are extraclinical: they come, as Freud’s title indicates, from “everyday life.” Sachs argues that Grünbaum’s attempt to finesse their importance by calling them “misextrapolations” from the theory of the neuroses (and thus clinical, as it were, only by courtesy) bizarrely distorts Freud’s own understanding of them.

Above all, Sachs criticizes Grünbaum’s complete neglect of the particular category of slips that provides the best evidence for the unconscious, so-called compound or accumulated parapraxes, in which two or more errors cooperate in fulfilling the same wish. Compound parapraxes often serve the purpose of what Sachs calls “tendentious forgetting”—forgetting that has a patently self-interested motive.^[60] One of Freud’s examples involves a woman living in Basel, who unconsciously resented a friend’s recent marriage. When the friend, “Selma X. of Berlin,” visited Basel on her honeymoon, the woman managed to forget an afternoon rendezvous with her. Then at the very hour of the rendezvous, the woman was forced into an “unconscious safeguarding” of her first parapraxis through the commission of a second.^[61] Engaged in a conversation about the recent marriage of the famous coloratura soprano Selma Kurz, she ventured some critical remarks about the marriage. But, to her embarrassment, she was unable to think of the singer’s first name, though ordinarily she knew it very well and had heard Kurz sing many times. That evening, with the woman’s friend now safely departed from Basel, the famous singer again came up in conversation, “and without any difficulty the lady produced the name ‘Selma Kurz.’ ‘Oh dear!’ she at once exclaimed, ‘it’s just struck me—I’ve completely forgotten I had an appointment with my friend Selma this afternoon.’ ”^[62] Timpanaro’s linguistic theory sheds no light on this ingenious conspiracy of forgetting, whereas its logic becomes transparent given Freud’s hypothesis of an unconscious wish that pursues its objective first one way and then another. Grünbaum, Sachs complains, nowhere mentions such combined parapraxes, even though Freud believed they offered the most convincing evidence for his theory of unconscious motivation.

Sachs sees an exactly parallel neglect of nonclinical evidence in Grünbaum’s slighting of the psychoanalytic theory of dream symbolism. Why does Grünbaum downplay it so? The answer seems obvious: Freud insists that he learned the unconscious meaning of dream symbols “from fairy tales and myths, from buffoonery and jokes, from folklore (that is, from knowledge about popular manners and customs, sayings and songs) and from poetic and colloquial linguistic usage.”^[63] In other words, all his sources were nonclinical—and, Freud adds, “if we go into these sources in detail, we shall find so many parallels to dream-symbolism that we cannot fail to be convinced of our interpretations.”^[64] Freud’s conviction in this crucial matter of dream symbolism—which figures prominently in the interpretation of neurotic symptoms, slips, and literary works as well—finds its “epistemic basis” entirely outside the clinic.^[65] His patients’ free associations, of which Grünbaum is so contemptuous, play no role whatsoever, and the question of therapeutic success (the sine qua non of analytic conviction, according to the Tally Argument) is even more irrelevant. Once again, Grünbaum refuses to provide a comprehensive and balanced account of the evidence on which Freud relied. Grünbaum is able to transform psychoanalysis into a purely clinical doctrine only by severely distorting Freud’s actual intellectual practices.

Much of Frank Cioffi’s critique of Grünbaum closely resembles David Sachs’s. Cioffi, too, complains that Grünbaum greatly overestimates Freud’s reliance on cures to guarantee his ideas. In particular, Cioffi insists that Freud could never have advanced the Tally Argument as a plausible defense of psychoanalysis. At the heart of the Argument, in Grünbaum’s construction, stands the claim that analytic therapy is superior to all its rivals. But, according to Cioffi, Freud knew perfectly well that he was in no position to make such a claim, because it implied that he had undertaken a comparative study of other therapies and found their results inferior to his own—which of course he hadn’t. Quite illogically, the Tally Argument portrays Freud as deriving comfort from a defense that any minimally rational person would dismiss as patently flimsy.

Freud, then, would have had no grounds to believe that the Argument, as Grünbaum construes it, was true. Even worse, in Cioffi’s view, he would have had good reason to think it false, at least in its central premise that analytic insight is indispensable to cures. Unlike Sachs, Cioffi does not stress the clear evidence that Freud always recognized both spontaneous remission and the success of rival therapies. Instead, Cioffi hones in on Freud’s own earlier claim to have achieved cures even when he was relying on his erroneous seduction theory: Freud himself pretended to cure neurotics before the discovery of the Oedipus complex, and thus before he could possibly provide his patients with “veridical” insight into the unconscious causes of their illness. Hence the contention that analytic insight is indispensable to cures—Grünbaum’s Necessary Condition Thesis—had been contradicted by Freud’s own experience as a therapist.

Cioffi’s case against Grünbaum does not rest solely on an appeal to Freud’s intelligence and his presumed aversion to contradicting himself. Cioffi also shows that Freud’s confidence in his ideas was largely independent of their therapeutic consequences. Much of this evidence is already familiar from David Sachs’s critique: Freud often supported his theories with material drawn from outside the clinic; he believed that his insights were valid even when, as with the psychoses, there was no hope of a cure; in the famous case histories he remained confident of his interpretations despite therapeutic disaster. Cioffi is particularly struck that Freud expressed not the slightest doubts that his analysis of the Wolf Man was correct even after the patient again became deranged. In all this the pattern is unmistakable: Freud’s convictions did not depend on cures.

The real source of Freud’s confidence, Cioffi argues, lay elsewhere. Freud believed in his interpretations, and in the theoretical views on which they rested, above all because of their narrative coherence—their “ability to confer intelligibility on the data.”^[66] Grünbaum refers to this as the argument from “inductive consilience” (275): the harmonious convergence of an interpretive element with other pieces of evidence to form a psychological whole. But Grünbaum insists that Freud resorted to this defense only at the very end of his life, after he had lost faith in the Tally Argument. In Cioffi’s view, by contrast, narrative coherence was always the deep source of Freud’s conviction. The stories Freud constructed to explain his patients’ behavior appealed to him in the manner of a jigsaw puzzle. Indeed, both early and late in his career, Freud invoked precisely this analogy to convey the attraction of his interpretations. In the famous 1896 paper on “The Aetiology of Hysteria,” he defended the scenes of infantile seduction in terms of their perfect fit with “the whole of the rest of the case history”:

It is exactly like putting together a child’s picture-puzzle: after many attempts, we become absolutely certain in the end which piece belongs in the empty gap; for only that one piece fills out the picture and at the same time allows its irregular edges to be fitted into the edges of the other pieces in such a manner as to leave no free space and to entail no overlapping. In the same way, the contents of the infantile scenes turn out to be indispensable supplements to the associative and logical framework of the neurosis, whose insertion makes its course of development for the first time evident, or even, as we might often say, self-evident.^[67]

Over a quarter century later the analogy reappears in “Remarks on the Theory and Practice of Dream Interpretation” to explain the analyst’s faith in his interpretations. “What makes [the analyst] certain in the end,” Freud writes, “is precisely the complication of the problem before him, which is like the solution of a jig-saw puzzle.”^[68] Such avowals, Cioffi asserts, reveal the secret of Freud’s persuasion: they appeal not to cures but to coherence. In effect, very much like the hermeneutic revisionists, Cioffi identifies what might fairly be called the aesthetic power of Freud’s ideas as the source of his confidence. Freud believed in his interpretations because they brought order to material that otherwise remained chaotic.

If one hopes to rout psychoanalysis—as Cioffi clearly does—one must, he argues, attack the reasoning on which it actually relies. And because narrative coherence was Freud’s ultimate court of appeal, the critic of analysis must show that Freud’s narratives do not in fact possess the intelligibility he claimed for them—that they “fail to meet the standards of plausible story telling (vague as these are) current in good historical, biographical, or forensic practice.”^[69] Grünbaum’s exclusive focus on cures is therefore “a great strategic error.”^[70] His heavy philosophical artillery is wasted obliterating a defensive formation of little consequence, while the real enemy—the argument from consilience—escapes unscathed to continue its depredations.

Much of Cioffi’s energy goes to arguing that The Foundations of Psychoanalysis is largely beside the point. But in at least one regard Cioffi considers the book more than innocuous: he profoundly objects to Grünbaum’s effort to rehabilitate Freud’s reputation as a scientific methodologist. Cioffi is especially annoyed by the suggestion that Freud responded hospitably to adverse findings. Against Karl Popper, Grünbaum of course maintains that Freud not only advanced falsifiable propositions but actually withdrew ideas if they were discredited empirically. “As a rule,” writes Grünbaum, Freud’s “repeated modifications of his theories were clearly motivated by evidence.”^[71] The abandonment of the seduction theory is the most famous case in point.

Cioffi will have none of this. Much more typical of Freud’s intellectual practice than any inclination to revise his ideas in the face of contradictory evidence was his tendency to rearrange the data—in a word, to lie. Freud was especially adept at inducing his patients to manufacture information that confirmed his theories. He never let mere facts get in the way of an idea.

Ultimately, then, Cioffi views Grünbaum’s writings on pyschoanalysis as an effort to exonerate Freud—to show that Freud possessed the empirical scruples of a true scientist, even though in practice he often fell short of his own standards. In Cioffi’s view, however, Freud was not a failed empiricist but a sham empiricist. Grünbaum is thus guilty of giving comfort to the enemy and of disarming the unwary. Under the guise of a ruthless philosophical critique, his book in fact restores Freud to scientific respectability. No friend of science, Cioffi concludes, can fail to regret it.

Grünbaum responds to Cioffi by citing again the many passages in which Freud either affirms the therapeutic superiority of analysis or points to analytic cures as confirmation of his ideas. Thus the disagreement between Grünbaum and Cioffi finally comes down to how one weighs these passages against others in which Freud backs off from his therapeutic claims or insists on the validity of his interpretations regardless of their practical consequences. More generally, it is a question of how one measures the relative importance of the argument from cures against the argument from narrative coherence. In my own view, Cioffi is much nearer the truth than is Grünbaum. Freud was perfectly happy to cite the evidence of cures whenever it would lend credibility to his claims. Yet he clearly never felt bound by it. Cures were a kind of bonus. They were certainly welcome, but they were not the cornerstone of his conviction. Freud’s faith in his ideas went undisturbed even when cures failed to materialize.

Perhaps even more important, the therapeutic argument has played only a minor role in the broad appeal of psychoanalysis. Freud did not become one of the most influential thinkers of the twentieth century because he convinced the world that he had found a cure for mental illness. As Cioffi suggests, Freud’s ability to persuade can be attributed largely to the distinctive character of his interpretations. Although the issue is more complex than Cioffi recognizes, the notion of narrative coherence rightly affirms that the secret of Freud’s success is an intellectual rather than a therapeutic matter. Grünbaum’s massive philosophical attack is thus curiously irrelevant. Whether one feels (as I am inclined to) that Grünbaum wins the therapeutic argument or (with Marshall Edelson) that he loses it, the heart of Freud’s appeal, as both Frank Cioffi and David Sachs show, lies elsewhere. He seduces us as a thinker, not as a doctor.

Freud and the Empiricist Tradition

In the end, what are we to make of Adolf Grünbaum’s critique of Freud? I’ve tried to suggest something of the difficulty of evaluating it. Perhaps the central problem—beyond its sheer density of expression and general disorganization—is Grünbaum’s curiously inconsistent attitude toward Freud. On the one hand, there is the impassioned defense of Freud’s methodological acuity and the scathing dismissal of his ill-informed and philosophically inept critics. On the other hand, there is the repeated complaint that Freud’s creation is “fundamentally flawed” because it fails to measure up to the inductivist standards of modern science—standards whose legitimacy, Grünbaum argues, Freud himself not only fully understood but embraced.

As we have seen, Grünbaum’s critique hinges on his construction of the Tally Argument and its supposed demise. But, despite extraordinary ingenuity and special pleading, Grünbaum is unable to show that the Argument figured centrally in Freud’s thinking. Without question, this failure is the most disabling fault in Grünbaum’s entire treatment of Freud. I am inclined to agree with David Sachs that the Tally Argument as such never existed in Freud’s own mind—certainly not in the elaborate, self-conscious form that Grünbaum gives it. Hence the question of the Argument’s validity is essentially moot, as is the question of whether its demise is attributable to Freud’s abandonment of it or to disconfirming evidence produced by recent experimental studies or (as I believe, and as Grünbaum himself argued in 1978) to a simple error of logic in the Argument’s formulation.

The evidence that the Tally Argument provided the sustaining intellectual justification for Freud’s ideas from 1896 to 1926, the central decades of his career, is even less persuasive. Rather, the Argument seems to have been more of a passing thought, tossed out almost nonchalantly in the “Analytic Therapy” lecture and perhaps hinted at elsewhere (in the passage Grünbaum cites from the Little Hans case, for example), but never thought through systematically or relied upon consistently. No amount of textual bullying can force Freud’s random and contradictory statements about the relation between his ideas and their therapeutic effect into conformity with the methodological principles supposedly enunciated in the Tally Argument. Likewise, both David Sachs and Frank Cioffi are certainly right when they maintain that Freud’s confidence in his ideas derived in large part not from cures—not, in fact, from clinical evidence at all—but from his interpretation of such disparate nonclinical phenomena as dreams, slips, jokes, fairy tales, and works of art. Neither Freud himself nor, just as important, those who have been persuaded by his ideas rest their conviction on the ability of psychoanalysis to cure the mentally ill. Rather, that conviction has much more to do with the peculiar charm of the ideas themselves and with their power to illuminate a wide array of psychological and cultural behavior.

In short, the central thesis of Grünbaum’s critique is poorly sustained. His arduously constructed (and deconstructed) image of Freud as a methodological sophisticate gone wrong ultimately collapses under its own weight and sheer improbability. In this respect, it is very much like Frank Sulloway’s image of Freud as a crypto-biologist—again a product of great textual ingenuity but no less mythical than the Freud of Grünbaum’s Tally Argument. Just as Sulloway tries to persuade us that, beneath his psychological cloak, Freud was a clandestine Darwinian, a precursor of the sociobiologist Edward O. Wilson, so Grünbaum insists that Freud, in his heart of hearts, was as self-conscious a philosopher of science as Francis Bacon or John Stuart Mill and a far better scientific methodologist than anti-Freudians like Karl Popper or Hans Eysenck have cared to recognize. But the evidence of Freud’s writings provides no better support for Grünbaum than it does for Sulloway. Rather, Grünbaum has created a Freud after his own philosophical fancy. For all the industry and intelligence he has invested in his critique, its actual payoff is remarkably meager.

Grünbaum’s judgment of Freud as a scientist contrasts interestingly with Sulloway’s. On first blush, one might expect Grünbaum and Sulloway to adopt similar positions. After all, both approach Freud from a scientific perspective, Grünbaum as a philosopher of science, Sulloway as a historian of science. But instead they offer radically opposite views of Freud’s scientific credentials. Sulloway’s long book contains not so much as a word to suggest that Freud falls short of the empirical standards of modern science. On the contrary, Sulloway speaks of Freud as a scientific genius of the first order. Why, one wonders, is Sulloway unconcerned with the methodological shortcomings that inspire Grünbaum’s obsessive elaboration of the Tally Argument and its unhappy fate?

The answer lies in the disparate conceptions of science that Grünbaum and Sulloway entertain. When Grünbaum speaks of science, he has in mind classical physics, the subject of his earlier work in the philosophy of science. Thus, when Grünbaum says that psychoanalysis is “fundamentally flawed,” he means that it does not meet the evidential standards routinely expected in physics. By way of contrast, the science against which Sulloway measures Freud is historical biology, which necessarily has a more indulgent conception of proof—a more latitudinarian notion of what is authentically scientific—than is generally tolerated in physics. Because Darwin is Sulloway’s model, Sulloway finds Freud’s intellectual procedures far less heterodox than does Grünbaum. In historical biology the demand for laboratory comparisons and group control studies to justify causal assertions is, if anything, even more a counsel of perfection than it is in psychoanalysis. Darwin’s science is at once empirical and hermeneutic: it works by deciphering clues—such as the fossil remains of extinct species—in order to reveal a hidden reality, namely, the workings of natural selection. Freud proceeds in exactly the same fashion when he deciphers the evidence of dreams, slips, and neurotic symptoms to reveal the hidden reality of the unconscious.

The sharp contrast between Grünbaum’s and Sulloway’s judgments of Freud as scientist is a potent reminder that “science” encompasses a wide range of intellectual practices. Science is in fact a continuum, with psychoanalysis occupying an honored place toward the Darwinian end. Such is the view taken by I. Bernard Cohen in his magisterial and authoritative study Revolution in Science, which treats Freud, alongside Darwin, as a major scientific innovator.

Grünbaum is best understood, I believe, as the latest and most sophisticated spokesman for the long-standing empiricist hostility to psychoanalysis. More generally, he is a representative of the well-established tradition of analytical philosophy—especially popular in Britain and North America but also boasting important adepts in Freud’s own Vienna—that takes a dim view of the speculative and metaphysical habits of Continental thinkers. This affinity is obscured by Grünbaum’s squabble with Karl Popper, who has been one of the leading voices of analytical philosophy in the twentieth century. But despite their disagreement about whether Freud comes to grief for inductivist or falsificationist reasons, Grünbaum and Popper share the same basic philosophical prejudices. For both of them, psychoanalysis fails because, unlike science, it is not genuinely empirical. Grünbaum differs from Popper mainly in being much cleverer and much better informed about the enemy.

Seen in this perspective, Grünbaum’s critique of Freud has a historical flavor quite distinct from that of the nearly simultaneous critiques of Frank Sulloway and Jeffrey Masson. What is most striking about Sulloway and Masson is their clear dependence on intellectual and, in the case of Masson, social developments of the past two decades. Sulloway’s Freud obviously presupposes the rise of sociobiology, or, more generally, what Carl Degler has called the “return of biology” in the 1970s.^[72] Masson’s Freud is a product of our newfound consciousness about child abuse in the 1980s. Grünbaum, by contrast, can be linked to no significant intellectual movement of the immediate past. If anything, he is at odds with the prevailing philosophical temper of recent times, when the linguistic turn and the enthusiasm for French imports like deconstruction have made Grünbaum’s hard-nosed empiricism seem decidedly old-fashioned. Hence my claim that his critique of Freud should be viewed as the end product of the astringent philosophical tradition that has dominated British and American philosophy for most of the century and whose roots in English intellectual history reach back into the classical and medieval eras. This accounts for Grünbaum’s profound hostility to Jürgen Habermas and Paul Ricoeur, both of whom speak for the Continental philosophical tradition so opposed to the radical empiricism (Habermas and Ricoeur would say positivism) espoused by Grünbaum, and both of whom, of course, argue that Freud belongs firmly in the more capacious and intellectually supple school of interpretation that has flourished on the Continent.

Above all else, Grünbaum’s critique serves to heighten our awareness of Freud’s tense and richly dialectical relation to the ideals of modern science. By placing the Tally Argument at the center of Freud’s thinking, Grünbaum points up Freud’s commitment to those ideals: the belief in observation, the demand that generalizations be supported by a large number of individual instances, the recognition that a single contrary instance can discredit such generalizations, and the need to eliminate alternative explanations for one’s findings. In this respect, Grünbaum, like Sulloway, offers a valuable reminder of Freud’s profound identification with the scientific tradition, as well as a useful corrective to the misguided efforts of hermeneutic interpreters like Habermas and Ricoeur to obscure that identification. Freud, Grünbaum shows us, puts up enormous resistance to being treated as an artist rather than a scientist.

At the same time, Grünbaum’s equal insistence on the failure of the Tally Argument draws our attention to the opposite but no less important truth that Freud also stands at odds with the scientific tradition. Freud refuses to trim his imagination to suit the strict empirical and skeptical canons of modern science. Like the great artists, he knows that the deep and important truths about human experience are complex, ambiguous, and, alas, often obscure. Perhaps unwittingly, Grünbaum thus enriches our sense of the paradox of Freud’s thought—its aspiration to be a science like any other and its simultaneous refusal to settle for the kinds of quantifiable insights that can be readily verified through laboratory comparisons and group control studies (such as have been pursued by American academic psychology, thereby guaranteeing its aridity and its irrelevance to modern intellectual life). The Freud who emerges from Grünbaum’s critique is decidedly more exigent and empirical than his hermeneutic students like to allow, but he still insists that knowledge of human beings will always remain interpretive: it cannot be made to resemble the kind of knowledge we have about the physical universe. Of course, Grünbaum wants to push Freud further in the scientific direction than Freud can comfortably tolerate without sacrificing his distinctive intellectual achievement. But precisely Grünbaum’s failure to prove his case testifies eloquently to Freud’s refusal to budge from the rich interstices of art and science. Freud remains a border thinker, neither fish nor fowl, always at risk of seeming caught in a hopeless contradiction and destined to be fought over perpetually by the representatives of the two great intellectual traditions that have dominated modern culture.

Notes