This is the pre-print of an article forthcoming 2024 at Philosophy, Psychology and Psychiatry. Please do not distribute without permission. In describing a long career, the psychiatrist Joel Paris recalls that (2005, p.123):It was not unusual in the 1960s for academic psychiatrists’ offices to be routinely fitted with an [psycho]analyst’s couch. In my own hospital, couches were regarded with awe as the tool of a powerful and arcane method of treatment. Over time however, couches began to be used only as a convenient place for a quick nap.Inevitably, when I tell people I research contemporary therapy, they assume I am referring to psychoanalysis. The quintessential image of the analyst’s couch is still so tied up with our cultural ideas of therapy, we seem not to have noticed that the landscape of therapy is now very different. Whilst psychoanalysis continues to be popular in certain parts of the world, such as continental Europe — largely in its contemporary iteration, psychodynamic psychotherapy — during the latter half of the 20th century, psychoanalysis declined in popularity and was supplanted, particularly in the anglophone West, by Cognitive Behavioural Therapy (CBT). As a Washington Post article stated (Simon, 2002), ”For better or worse, cognitive therapy is fast becoming what people mean when they say they are ’getting therapy’.”If CBT had an origin story, the way Spiderman has his radioactive spider and the theory of evolution its finches, it would be the moment that Aaron Beck, often described as the ’grandfather’ of CBT, began to question the psychoanalytic account of depression. As the story is described in CBT texts, in the late 1950s, Beck, who was at the time a psychoanalyst — as were most psychiatrists at the time — at the University of Pennsylvania came to question the psychoanalytic understanding of depression in terms of inverted hostility, as anger did not seem to be a prominent emotion expressed by his depressed clients. Although Beck’s first manuscript published in 1967 foregoes commentary on this moment in favour of describing the details of the various studies he conducted as result of his scepticism, by the time we reach the preface of his third book published in 1979, Beck is more forthcoming about the significance of this moment:Although the initial findings of my empirical studies seemed to support my belief in the specific psychodynamic factors in depression, namely, retroflected hostility, expressed as a ‘need to suffer’, later experiments presented a number of unexpected findings that appeared to contradict this hypothesis. These anomalies lead me to a critical evaluation of the psychoanalytic theory of depression and ultimately the entire structure of psychoanalysis. The anomalous research findings ultimately led to the conclusion that depressed patients do nothave a need to suffer […] This marked discrepancy between laboratory findings and clinical theory led to an ‘agonising reappraisal’ of my own belief system.Beck’s description of CBT’s inception embodies a narrative familiar to scholars of science concerned with theory-change as well as those interested in the pedagogy of science — a pivotal experiment proves its predecessor wrong, setting the scene for the theory that follows. These are the sorts of stories that introduce neophytes into a field and explain the provenance of that field, the means by which the current theory supplanted its predecessor.In later interviews, Beck continues to distance CBT from psychoanalysis, describing his early experiments in terms that suggest they resulted in a clean break or departure from psychoanalysis. What drives this break is a certain kind of inquiry, one that can be described as ’scientific’, involving empirical studies and the development and revision of theories in light of the data collected from these studies (as cited in Weishaar, 1993, p.18):When the other studies did not fit in, I went back to my dream studies and I thought, ‘maybe there is a simpler explanation and that is that the person sees himself as a loser in the dreams because he ordinarily sees himself as a loser’ […] Dreams could simply be a reflection of the person’s thinking. If you take motivation and wish fulfilment out of the dream, this undermines the whole motivational model of psychoanalysis. I started looking at the motivational model all the way through, and its manifestations in behaviours, everyday slips, and so on. It seemed to me that the motivational model did not hold. Once that collapsed, and I inserted the cognitive model, I saw no need for the rest of the superstructure of psychoanalytic thinking.The characteristics that define CBT’s origin story — an emphasis on scientificity and empirical grounding — permeate throughout CBT. As Rachael Rosner (2012, p.8), one of the few academic historians writing on CBT, puts it:This [scientific] faith gave shape to the entire opus of his new cognitive model of mental illness, from how [Beck] did his science to how he taught his patients to help themselves (that is, to become scientists in the laboratory of their own lives), to his idea of what good living should look like (rational, pragmatic, scientific).Given the centrality of this origin story not only to CBT’s account of its early history but its therapeutic approach, it should be unsurprising that it is a staple of contemporary CBT texts, at times taking on the air of hagiography when extolling the virtues of Beck (Dryden, 2012, p. 25):Beck was a somewhat sceptical recruit into both psychiatry and psychoanalysis. His open and enquiring mind has served the generation of Cognitive Behaviour therapists who followed his work well. He was particularly interested in trying to establish the validity of psychoanalytic concepts by empirical research and set out to test the hypothesis that depression was caused by anger turned inward on the self. His research found that this theme was not reflected in the content of dreams reported by depressed patients. At the same time he discovered the existence of a constant stream of nonconscious negative thoughts in depressed clients. Eventually he concluded that the content of dreams of depressed patients was better explained by the nature of this negative thought stream than by any unconscious ‘need to suffer’ — a completely extraneous concept, he suggested.In their introduction to the Freud Files , Borch-Jacobsen and Shamdasani (2011, p.7) describe the ’legend’ that Freud and proponents of psychoanalysis presented of themselves when offering their own origin story and offer the following caution, “Legenda is a story meant to be repeated mechanically, almost unknowingly, like the lives of the saints that were daily recited at matins in the convents of the Middle Ages.” Similarly, the clear-cut, straightforward story CBT proponents offer of their origins imbues CBT with a certain legend, one that supposedly sets it apart from psychoanalysis and additionally explains its successes: we are told that Beck came to question the accuracy of the psychoanalytic account of depression and eventually prove it false via scientific experiment. Other popular histories of the transition describe psychoanalysis as buckling under the weight of demands for evidence of its efficacy from a number of parties—disillusioned analysts, disgruntled patients, critiques from behaviour therapists and challenges posed by psychopharmaceuticals and biological psychiatry. In contrast, ”evidence-based” is now almost invariably mentioned in the same breath as CBT and touted as one of the reasons, alongside its scientificity, for CBT’s runaway success. The legend projected in this origin story is one in which CBT is a field that takes empirical evidence and scientific investigation to be of central importance, unlike its predecessor.So sparse has the attention been to contemporary psychotherapy, and so great the preoccupation with psychoanalysis, that the history of CBT is only now starting to be written by professional historians rather than CBT practitioners themselves. The vast majority of work covering the period of transition is found in: introductory sections of CBT texts such as training manuals (e.g. J.Beck, 2011); memoirs of clinicians (e.g. Paris, 2005); biographies written by colleagues of key figures such as Beck (e.g. Weishaar, 1993), and amateur historical writing by clinicians for a general audience (e.g. Tallis, 1998). The danger of simply repeating the legend of CBT in recounting its early history is thus very great — philosophers, such as myself, are particularly susceptible as it is now near impossible to think of psychoanalysis without Popper’s (1963) declamations of psychoanalysis as unscientific ringing in our ears. Indeed, CBT practitioners and other clinicians critical of psychoanalysis themselves make reference to Popper when distancing themselves from psychoanalysis and justifying their own scientificity. For example, Isaac Marks, a central figure in the development of behaviour therapy, from which the ’B’ in CBT is derived, quips in reference to early attempts to test psychoanalysis, ‘we were Popperians before we even knew of Popper’ (as cited in S.Marks, 2012, pp. 3–4).In this series of two papers, I will chart the ascendance of CBT. I make no pretensions towards thoroughgoing novelty nor claim that the following is a complete account of this period. Some of the way points may already be familiar to readers; my goals are to bring the history of contemporary psychotherapy into contact with more well-known histories of medicine and psychiatry and to highlight a central theme regarding evidence that has been hitherto neglected – a theme that continues to shape the landscape of contemporary talking therapy. The two papers are an exercise in integrated history and philosophy of science. Such an interdisciplinary approach is vital as the period spanning CBT’s inception and its rise to prominence coincides with a period where the notion of evidence became contested then settled in a manner which disadvantages psychoanalysis. Though historians of medicine (see H. M. Marks, 2000) have noted this shift, I believe it additionally requires a philosophical eye to unpick the conceptual implications that accompanied it.Part I complicates the origin story of CBT. It is not that psychoanalysis failed to provide evidence for its theories and its efficacy as a treatment, but rather that what constitutes evidence differed between these two schools of psychotherapy. Nor is it obvious that psychoanalysis failed to be scientific. Rather, this shift in the notion of evidence rippled through a cluster of conceptual issues and questions that bear on the transition from psychoanalysis to CBT: what is the relationship between evidence and scientificity? What is mental illness and its aetiology? What does it mean to treat mental illness? What does therapeutic success, or efficacy consist in?In Part II, I will offer a retelling of the transition away from psychoanalysis by charting a common theme in both the decline of psychoanalysis and the rise of CBT — a shift in what was considered evidence — a shift that CBT navigated successfully while psychoanalysis floundered. I will demonstrate that CBT gained traction as a result of a notion of evidence introduced and popularised, not by CBT itself, but largely by psychopharmaceuticals. It is only once this alternate account of evidence and its accompanying understandings of treatment and therapeutic success became entrenched that CBT truly gained a foothold in the landscape of contemporary psychotherapy.Below, I draw on a number of sources, including but not limited to: histories of CBT written by clinicians of the sort described above; histories of CBT by academic historians (sparse as they are); psychology, psychiatry and clinical research papers on the efficacy and accuracy of these therapies; and philosophical debates on psychoanalysis, namely by Karl Popper and Adolf Grünbaum. I believe the range of sources demonstrates both the necessity of an interdisciplinary approach and the need to examine broader or background histories when scrutinising a field such as psychotherapy which sits at the cross-roads of multiple fields (see Rosner, 2018a for discussion of this point). In concluding the two papers, I will also offer some reflections on the methodology of integrated history and philosophy of science adopted herein.A Note On TerminologyThe main charge against psychoanalysis expressed in CBT’s origin story is that it presents an inaccurate picture of mental illness, namely the psychoanalytic understanding of depression in terms of inverted hostility. Whilst there were some early support for the psychoanalytic model in his research, Beck concluded based on further studies that the psychoanalytic account was not borne out, necessitating the need for a new understanding of depression (a detailed account of this research can be found in A. Beck, 1967, pp. 174–96). Beck spent the next couple of years further observing his depressed clients, culminating in the development of Cognitive Therapy (CT), which provided an alternate understanding of depression as a thinking disorder, referred to as the ‘cognitive model of depression’, and a novel therapeutic approach which aimed to address these issues in cognition. CT later merged with behavioural therapy to become CBT.11In considering the early history of CBT, this article primarily concerns the development of CT. As CBT and CT share an origin story, we will not delve into the theoretical differences between the two and I will speak of them interchangeably, unless clarification is necessary.I will take critics such as Beck to be claiming that psychoanalysis lacked evidence for its theoretical adequacy 22”Adequate” is intended to signal that ”accuracy” in the definiens is to be understood in the context of Van Fraassen’s (1980) notion of empirical adequacy, where the aim is not to secure truth per se, but rather to ’save the phenomena’ or to ‘fit’ the available empirical evidence. Accuracy is often understood in a straightforwardly realist sense and the reader may choose to interpret it as such. However, beyond alignment with my philosophical inclinations, adequacy is, I believe, a better epistemic standard to use, given the emphasis on empiricism we see stressed in CBT’s origin story. For further clarification of the two notions see XX.— that is, whether it presented an accurate theoretical account describing mental illness, its aetiology and its maintaining factors. Theoretical adequacy is to be distinguished from what I will refer to as therapeutic efficacy – whether it was able to successfully treat mental illness. Naturally, the two are related. Also, we shall see that the two terms revolve around other core concepts such as ”evidence”. The need for the distinction will become clearer by and by.We shall consider the charges against the theoretical adequacy and the therapeutic efficacy of psychoanalysis in turn.The Theoretical Adequacy of Psychoanalysis33In what follows I will offer a broad characterization of psychoanalysis rather than referring to any particular school or interpretation, as that is not vital for the discussion. Insofar as any particular interpretation is relevant, Freudian psychoanalysis is the most pertinent given the historical period. It is also the main school of psychoanalysis discussed by Adolf Grünbaum, whose work we will examine in more detail shortly. Grünbaum believes most of the central theoretical claims were simply inherited by later theorists and further states that it is the Freudian account rather than the later hermeneutical accounts that present themselves as scientific. Additionally, note that I will resort to inelegant turns of phrase such as ‘Psychoanalysis states…’ or ‘CBT can respond by…’ for reasons of brevity. Such phrases are to be understood as shorthand for ‘The theoretical account of psychoanalysis states that…’ or ‘CBT proponents can respond by…’Let us turn to the origin story as it appears in the work of Judith Beck, the daughter of Aaron Beck and a prominent CBT clinician in her own right (J. Beck, 2011: 5):In the late 1950s and early 1960s, Dr. Beck decided to test the psychoanalytic concept that depression is the result of hostility turned inward toward the self. He investigated the dreams of depressed patients, which, he predicted, would manifest greater themes of hostility than the dreams of normal controls. To his surprise, he ultimately found that the dreams of depressed patients contained fewer themes of hostility and far greater themes of defectiveness, deprivation, and loss. He recognized that these themes paralleled his patients’ thinking when they were awake. The results of other studies Beck conducted led him to believe that a related psychoanalytic idea—that depressed patients have a need to suffer—might be inaccurate … At that point, it was almost as if a stacked row of dominoes began to fall. If these psychoanalytic concepts were not valid, how else could depression be understood?The described methodology resembles a Popperian empiricism, where theories are actively tested and falsified in light of empirical data. A prediction was made based on the psychoanalytic theory of depression in terms of inverted hostility – namely that the dreams of depressed patients will be characterised by hostility – and was not observed, bringing the theory into question.Based on Beck’s correspondence at the time, it is clear that he was not only aware of Popper’s work but drew upon it in trying to develop and test his new understanding of depression (Rosner, 2018b, p.42). Unfortunately for Beck, the immediate fruits of his labour during this period — the results of these various tests and the tool he developed to test the psychoanalytic understanding of depression, the Beck Depression Inventory — did not make much of an impact (A. Becket al ., 1961; A. Beck, 1963; 1964).Rachael Rosner, currently working on the first biography of Beck by a historian, claims that the failure of Beck’s research to gain a foothold during this period was due to the fact that psychoanalysis tended to downplay empirical research and instead favoured theorising. Rosner states (2014, p.739) that ‘theory trumped data’ during this time, describing analytic research as focusing on making theoretical adjustments to, and reinterpretations of, psychoanalysis rather than subjecting it to empirical testing. Seemingly contradictory data was accommodated into the pre-existing psychoanalytic theoretical account, or ’interpreted away’, rather than precipitating a questioning of the theory.This seeming resistance to revision and the desire to account for all cases and observations within the existing theory, including ones that seem contradictory, was central to Popper’s infamous claim that psychoanalysis is unfalsifiable and thus unscientific (1962, p.34):It was precisely this fact—that they always fitted, that they were always confirmed—which in the eyes of their admirers constituted the strongest argument in favour of these theories. It began to dawn on me that this apparent strength was in fact their weakness.For Popper, rather than accommodating all observations, it is the ability of a theory to make predictions that could prove it false if they are not observed thatmakes the theory scientific. Thus, the apparent reluctance or inability to revise its theoretical account in response to evidence cemented Popper’s view of psychoanalysis as unscientific, a sentiment echoed in the origin story of CBT.This supposed lack of interest in evidence is often reported with reference to a particular exchange between Freud and the psychologist Saul Rosenzweig, who wrote to Freud with news that he had experimental evidence supporting the theoretical account of psychoanalysis. Later authors describe Freud’s response to the letter as disinterested at best and hostile at worst (Fancher, 1995). The highly publicised ’breaks’ between Freud and Adler, as well as Freud and Jung, also accords with a picture of psychoanalysis as uninterested in providing evidence for its theoretical account and quelling dissent and criticism (see Rogow, 1971; Grosskurth, 1991 for exposés of psychoanalytic culture).Beck’s own experience echoes aspects of this not entirely undeserved image of psychoanalysis akin to a cult of personality, resistant to revision and uninterested in securing evidence for its theoretical account. Beck appears to have been actively discouraged from publishing his depression inventory and the results of his work on depression. Rosner recounts how upon presenting his findings to select colleagues, he was told that it was too ‘heavily empirical’ and ‘not sufficiently psychiatric’ (Rosner, 2014, p.752), synonymous with the criticism that it was ‘insufficiently analytic’. Following his disappointment with the early reception of CT, Beck took a sabbatical from 1962–67.We will pick up the thread of Beck’s career again in Part II. For now let us dwell on the lacklustre opening act and evaluate how it is mythologized in CBT’s origin story. Certainly, Rosner’s descriptions of this period, exemplified in the early responses to Beck’s work as well as the oft cited Rosenzweig incident, lend support to the image of psychoanalysis projected in CBT’s origin story — a field uninterested in testing or evidence, thus unscientific according to the likes of Popper.However, in highlighting psychoanalysis’ seeming lack of interest in evidence and the way in which this affected the early development of CBT, one must be wary of unthinkingly promulgating CBT’s legend-making of itself as uniquely scientific. As Borch-Jacobsen and Shamdasani caution, philosophers in particular are susceptible to awarding scientificity to the ‘victors’ in historical transitions. Though Borch-Jacobsen and Shamdasani (2011: Chapter 3) themselves challenge the adequacy of Freud’s theories — not least by drawing attention to the various omissions and outright fraud found in analytic case studies — they remind us that Freud launched the psychoanalytic project with the belief that he was bringing psychology into the realm of science (as cited in Borch-Jacobsen and Shamdasani, 2011, p.4; see also Grünbaum, 1984: 2–3):The intellect and mind are objects for scientific research in exactly the same way as any non-human things. Psycho-analysis has a special right to speak for the scientific Weltanschauung at this point . . . Its special contribution to science lies precisely in having extended research to the mental field. And, incidentally, without such a psychology, science would be very incomplete.Against this background, the confidence with which psychoanalysis is declared unscientific by the likes of Popper and in CBT’s origin story, should be viewed with caution.Simply corroborating the victor’s view of events is also a peril of biography or oral history, the worst examples of which are historical canons written by colleagues and proponents (e.g. Dryden, 2012, Weisnaar, 1993). Though professional historians such as Rosner offer a more complex view, for instance, by drawing attention to the continuities between Beck’s theoretical account and psychoanalytic concepts which challenges the legend of CBT as making a ’clean break’ from psychoanalysis (2012), when describing Beck’s struggles to get his work accepted, there remains the danger of presenting them as straightforward confirmations of the CBT legend. In her defence, Rosner has also challenged (2018c) the myth of CBT as uniquely scientific. However, she does so by drawing attention to the theoretical orientations of researchers trying to test psychoanalytic theory rather than an appraisal of the theory itself, to which we shall now turn.The most thorough challenge to Popper’s critique of psychoanalysis is to be found in the work of philosopher Adolf Grünbaum, a contemporary of Popper’s. Though once a well-known figure, Grünbaum has to some extent faded from the philosophical canon; my own introduction to Popper and psychoanalysis as an undergraduate did not include Grünbaum. Bringing Grünbaum back into dialogue with Popper will reintroduce nuance to our understanding of psychoanalysis and leave us better equipped to appraise CBT’s origin story.It is difficult to do justice to the painstaking detail with which Grünbaum identifies and dismantles Popper’s various claims about psychoanalysis throughout his oeuvre, the most notable being, The Foundations of Psychoanalysis (1985) and Validation in the Clinical Theory of Psychoanalysis (1993). In a particularly scathing criticism of Popper’s claim that psychoanalysts were unwilling or unable to revise their theoretical account, Grünbaum cites (1985, p.39) the title of one of Freud’s papers, ‘A case of paranoia running counter to the psychoanalytic theory of disease’, demonstrating Freud’s obvious interest in revising his theory in light of empirical evidence. Only a minor example, but instructive in terms of how readily we have come to see psychoanalysis through Popper’s eyes. Elsewhere, Grünbaum documents a much larger repertoire of instances where the psychoanalytic theory was revised in light of recalcitrant evidence, including revisions to the core analytic accounts of dreams and anxiety (1985, pp.108–112). He also brings to light detailed discussions by Freud and other analysts concerning the question of what would constitute falsifying instances of psychoanalysis (1985, pp.120-125).Grünbaum’s most enlightening contribution is the clarification that psychoanalysis was interested in evidence but that this evidence was of a particular kind. Both Freudians and psychoanalysts more broadly took evidence and empirical data relevant for the evaluation of their theories to be based on the therapeutic encounter — that is data and observations from therapy sessions themselves. The primary method of establishing theoretical adequacy was based on generalising from therapeutic evidence and ongoing ’fit’ with what was observed during sessions. In short, the method of case studies, which is all but synonymous with psychoanalysis. As a result, Grünbaum (1985, pp: 93-95; 279-287) takes psychoanalysis to favour induction rather than falsification, which he claims is a stricter standard than falsification when followed rigorously. Furthermore, Grünbaum (1985, p.97) states that though psychoanalysis was not uninterested in falsification, this evidence too was also to be gathered from therapy sessions.I do not wish to adjudicate between induction and falsification here, though note that many philosophers of science, particularly Wesley Salmon (1968) believe that induction is indispensable to science; Popper’s insistence that falsifiability solely demarcates science from non-science has not weathered the test of time. Furthermore, it is unclear whether the method of case studies is to be understood inductively (the seminal work on the intellectual history of the case study is Forrester, 2016). For instance, contemporary psychoanalytic psychotherapist and philosopher Richard Gipps (2019) understands case studies in illustrative terms and offers a different interpretation of psychoanalysis’ claim to scientificity. Regardless of whether we demarcate science in inductive or falsificationist terms, or how exactly the method of cases is to be understood, it is clear that insofar as psychoanalysts would accept falsifying data that would necessitate a revision to the account, or evidence in support of its theoretical account in an inductive (or some other) sense, this would be based on evidence from therapy sessions.Grünbaum’s explanation of how psychoanalysis secured theoretical adequacy and what this signifies about its scientific credentials shows that the historical story promulgated by CBT proponents and those viewing psychoanalysis through Popper’s eyes is overly simplistic. Psychoanalysis was not uninterested in securing theoretical adequacy and was open to revision and falsification. However, what was considered evidence for these purposes was to be drawn solely from therapy sessions. In fact, the unenthusiastic response to Beck’s early work can be explained in part due to the fact that it was not all ’evidentially kosher’ by psychoanalytic lights. In his first manuscript, Beck states (1967, pp.174–96) that his research into the psychoanalytic account of depression drew on diverse evidential sources:1) The thematic content of dreams (‘dream studies’) in depressed individuals. 2) The thematic content of early memories in depressed individuals. 3) An experimental ‘fantasy test’ scrutinising the sorts of narratives favoured by depressed individuals when presented with prompts. 4) A self-report inventory interrogating how depressed individuals conceived of themselves and their attributes. 5) Experimental studies examining the relationship between mood and success on tasks, such as card games, where the outcome was manipulated. 6) Experimental studies concerning the relationship between probability of success and success on tasks, where the outcome was manipulated. 7) Thematic patterns in the verbal material of depressed individuals, including free associations. 8) Longitudinal studies examining the presence of early negative experiences in the lives of depressed individuals, i.e. the loss of a caregiver.It is difficult to glean how much of this evidence originated from therapy sessions. For instance, dream studies are discussed in various places, sometimes as simple observations during therapy sessions (1967) and elsewhere (1963) as interviews with patients who were at the time undergoing analysis. It is unclear whether the latter would count as appropriate evidence by psychoanalysts as they were not gathered directly from therapy sessions. Being generous, of the above, 1), 2), 7) and 8) would count as evidence because they were either gathered directly from therapy sessions or could feasibly have been. Whereas the others, such as psychological experiments, are somewhat further from an analytic understanding of evidence. As Grunbaum claims, Freud’s lack of interest in Rosenzweig’s experimental data can also be diagnosed in light of this disagreement about what constitutes evidence (see Grünbaum, 1985, p.101).Now for the customary caveats and clarifications:I am not suggesting that the evidence and methods used by psychoanalysis to secure theoretical adequacy were superior to those that eventually supplanted them. Even Grünbaum himself, having carefully defended psychoanalysis from Popper, falls upon it with the same fervour — a review of Grünbaum’s book has the telling title, ‘With a friend like Professor Grünbaum, does psychoanalysis need any enemies?’ (Caplan, 1986). Grünbaum describes a multitude of problems with using evidence from therapy sessions as support for theoretical adequacy. The main worry is neatly summarized by Robert Fancher (1995, pp.110-111), another philosopher of science: the analyst is not making neutral observations during a therapy session. When a client presents, say, a string of free associations which might be considered part of the data gathered from the therapeutic encounter, it is not recorded or evaluated in a neutral manner, but rather considered through the theory of psychoanalysis. Indeed, even the client may understand their verbal expressions through a psychoanalytic lens and describe their thoughts in these terms. Given that the analyst and plausibly the client are both working within the theoretical account of psychoanalysis, the evidence is already laden with the theory it aims to support or falsify. A point interestingly conceded by Freud, though never quite solved (as discussed in Grünbaum, 1985, pp.246-50; 1993, pp.182-3; see also Glymour, 1980).It should be kept in mind that Grünbaum is not the last word on psychoanalysis. For instance, Michael Lacewing (2012; 2018) takes issue with much of his characterization of psychoanalysis, leaving open the possibility that a more charitable understanding will survive the objections levelled by Grünbaum. Alternate understandings of psychoanalysis may also present a different perspective on CBT’s origin story. Gipps (2017) suggests that Beck misunderstood the psychoanalytic account of depression, such that the origin story does not describe a falsifying instance of psychoanalysis at all. Here, I have deliberately eschewed the exegesis of psychoanalysis in favour of illuminating a different thread regarding evidence and its relevance to the early history of CBT.However, I am not claiming that the poor reception of Beck’s work can be put down solely to a disagreement about what constituted evidence. In addition to the mixture of evidential sources there were the cultural tendencies documented by Rosner (2014, pp.738-9) such as revering key figures and prioritising theoretical adjustments over empiricism. More importantly, I suspect the difficulties are to be put down to the more commonplace problem of a novel theory coming up against an older, more established one. This is particularly so when the novel theory has only a small amount of recalcitrant evidence under its belt in comparison to the older theory, which appears to have a great deal of confirming evidence and institutional entrenchment.Regardless, we have found strong reasons to believe that the presentation of CBT as uniquely interested in scientific inquiry and empirical evidence is, at the least, overstated. In Part II I will show that a better explanation of the transition from psychoanalysis to CBT is that what constitutes evidence and thus scientificity shifts during this period. Before that, let us consider criticisms of psychoanalysis in terms of its therapeutic efficacy.The Therapeutic Efficacy of PsychoanalysisConcerns about the efficacy of psychoanalysis are often cited as a reason for the waning of analytic influence. For instance, Joel Paris, a former psychoanalyst, describes (2005, p.12) a collective feeling of disillusionment about psychoanalysis grounded in its ‘failure to keep its promises’. Paris recalls (2005, pp.145–49) questioning whether psychoanalytic treatment was in fact helping his clients and his growing concern that there was no evidence for the efficacy of psychoanalysis beyond the heavily theorised publications of individual case studies.Much like the theoretical inadequacies of psychoanalysis, worries about efficacy also seemed to be theorised away, with failures being attributed to lack of skill in the analyst, errors in case formulation and the belief that more analysis was required to resolve the underlying issue. The increasing length of a course of psychoanalysis — growing from months in the time of Freud to years in American settings, where psychoanalysis became deeply entrenched (Burnham, 2012) — together with the fact that analysts largely treated private clients from the general population who tended to be moderately distressed rather than seriously mentally ill, further contributed to the worry that psychoanalysis simply did not work as a treatment for mental illness. These concerns about the efficacy of psychoanalysis are echoed elsewhere (for example Tallis, 1998, pp.134–35) and are often repeated in contemporary explanations for the decline of psychoanalysis. Therapeutic efficacy is one of the main ’selling points’ of contemporary CBT, mentioned alongside the term “evidence-based”. Though it is not usually included in the origin story, there are occasional concerns about the efficacy of psychoanalysis to be found in Beck’s early work (1967, preface):Concurrently, I became somewhat painfully aware that the early promise of psychoanalysis observed in the early 1950’s was not borne out by the middle and late fifties — as my fellow psychoanalytic students and other colleagues entered their sixth and seventh years of psychoanalysis with no striking improvement in their behaviour or feelings!If Popper is the poster boy for challenges to psychoanalysis’ theoretical account, Hans Eysenck was the primary antagonist concerning the efficacy of psychoanalysis. A prominent psychologist at Maudsley Hospital in London, Eysenck was not a disinterested figure in these debates as he himself founded a school of psychotherapy — behaviour therapy. Furthermore, Eysenck’s critiques were not directed solely at psychoanalysis. Though psychoanalysis did draw much of his ire (Eysenck, 1985), the target of his critiques was the efficacy of psychotherapy as a whole. Eysenck’s critique was based on the idea of ‘spontaneous remission’ or ‘self termination’ (Eysenck, 1990, 145–46):I looked at the published evidence concerning the efficacy of psychotherapy and psychoanalysis as methods of treatment and came to the conclusion that there was nothing to suggest that it did any better than Placebo treatment or no treatment at all. Comparing all the studies that have been done I found that neurosis was a self-terminating kind of disorder rather like the common cold; even without any psychiatric treatment, most sufferers from neurotic disorder would get better over time.The belief that mental illnesses would eventually disappear without the need for therapeutic intervention was first and most forcefully expressed in a 1952 paper, written in response to an initiative by the American Psychological Association aiming to increase the number of psychologists trained in psychotherapy—at the time largely the province of psychoanalytic psychiatry. Eysenck used pre-existing published results regarding discharge rates from mental hospitals (where patients received little to no therapy) and disability insurance claims as a baseline or control group to estimate the rate of spontaneous remission. He then compared this rate to results in studies of psychoanalysis, other schools of eclectic therapy and interactions with general practitioners (n=17, 7000 case studies). The results (Eysenck, 1952: p.661) suggested that generic interactions with a general practitioner had the best effect but were on par with estimated spontaneous remission rates. Psychoanalysis performed on par with eclectic psychotherapy, but both had lower rates of efficacy than the baseline remission rate. Beyond lending support to the spontaneous remission hypothesis, the fact that the particular schools of therapy were not uniquely efficacious and performed comparably to generic care from general practitioners suggested that insofar as there were efficacious elements to analysis, they were factors common to all schools of therapy and clinical interactions more generally.In light of the results, Eysenck suggested (1952, p.662) there was no point training psychologists in psychoanalysis and raised the need for ‘carefully planned and methodologically more adequate trials’ of psychotherapy. As Eysenck’s autobiography (1990) and the recollections of ex-analysts such as Paris indicate, the dominant culture of psychoanalysis seemed resistant to challenges regarding its efficacy. However, as with theoretical adequacy, it is unclear that from this we should conclude that the theory was lacking in evidence for efficacy. Rather, the disinterest in the sort of trials Eysenck proposed raises a more complex question regarding what constitutes evidence for efficacy in psychoanalysis. In examining this question, here again, we will find reason to doubt the easy story of the transition from psychoanalysis to CBT.Though Grünbaum focuses largely on theoretical adequacy rather than efficacy, his point that evidence in psychoanalysis comes solely from therapy sessions is of relevance here in light of the notion of mental illness and therapeutic success presupposed by psychoanalysis. Mental illness in psychoanalysis is understood to stem from various defence mechanisms levelled against thoughts or feelings that arise from the unconscious. Defence mechanisms may have once been helpful in some way, for instance, by protecting the individual from overwhelming emotions, or thoughts they would not typically lay claim to, but have since become detrimental resulting in mental illness. Consequently, therapeutic success44It is unclear that psychoanalysis takes therapeutic success to consist merely in ’treating’ mental illness. Rather the aim could be described in terms of cultivating a number of characteristics — for example, better self-regulation strategies and changing internal working models of others— that foster something we might refer to as ’mental health’ (to use a currently overused and ambiguous term). To put this another way, we should be wary of assuming that the absence of mental illness is synonymous with mental health. These issues, interesting as they are, are not directly relevant to our discussion of CBT’s early history so I will lay them aside. consists of acknowledging unconscious thoughts and feelings and addressing the problematic defence mechanisms (for example, by adopting different self-regulation strategies) rather than simply the alleviation of symptoms. As such a psychoanalyst may claim that it is not enough for an individual to say, simply stop being anxious of certain seemingly innocuous situations. Additional conditions of success can include the client understanding what initially precipitated the anxiety, how their defence mechanisms regulated these negative feelings in the past and how this now results in anxiety in these situations (see Grünbaum, 1984, p.10-11).The upshot of the psychoanalytic understanding of mental illness and therapeutic success for evidence of efficacy is that simple observation of whether the symptoms of the mental illness had abated cannot constitute success, or therapeutic efficacy, for psychoanalysis. It is additionally necessary to probe whether the client has understood and resolved their previously unacknowledged unconscious phenomena and defence mechanisms. Such evidence can only come from the likes of discussions in therapy sessions and case studies, rather than the sort of comparative trials favoured by Eysenck and the behaviour therapists, as the latter rely on narrow measures of symptom change.Furthermore, the spontaneous remission hypothesis advanced by Eysenck would not be readily accepted by psychoanalysts. Even if the symptoms of neurosis were alleviated, psychoanalysts could claim that the underlying issues which constituted the neurosis may remain with the possibility of recurrence. For instance, an analyst might point out that a client who represses thoughts that they find uncomfortable will continue to feel negative feelings each time these thoughts arise, until they are suitably acknowledged. Interestingly, CBT makes a similar claim in asserting that achieving long-term change to the behaviour and emotional symptoms of mental illness involves altering the patterns of cognition that underlie them (Beck et al ., 1979, p.168). Behaviour therapy in contrast saw mental illness to be nothing more than a cluster of symptoms, and therapeutic success to involve only the alleviation of these symptoms. Eysenck himself recognised the disagreement concerning what constituted mental illness or the target of the therapeutic interventions in discussing the difficulty of identifying a common measuring stick across various therapies which differed in their methodology and understanding of mental illness (Eysenck, 1952; 1990, p.130).Again, the claim is not that the type of evidence or the standard of success favoured by psychoanalysis was suitable or superior to alternatives. It is not important for my purposes to spell out the many problems with the sort of evidence for efficacy and standard of therapeutic success that psychoanalysis favoured, but here is one of them: by distancing therapeutic success from readily observable symptoms, it becomes difficult to identify when the mental illness has been addressed. That is, at what point do we consider unconscious thoughts and feelings to have been truly acknowledged and unhelpful defence mechanisms to be fully resolved?However, it is a mistake to think that the therapies which supplanted psychoanalysis such as CBT or behaviour therapy were unique in that they were interested in evidence for efficacy. Here again we find reasons to doubt the narrative of straightforward transition from psychoanalysis to CBT. Rather, there is a far more complex disagreement regarding what constitutes therapeutic success or efficacy, and consequently, divergence about what is considered appropriate evidence.SummaryChallenges to psychoanalysis in terms of its theoretical adequacy and its efficacy were part of the zeitgeist by the time Beck’s concerns about the psychoanalytic account of depression arose. However, the opening act of Beck’s early work closed with an unenthusiastic response to the alternate understanding of mental illness and therapy he was developing. This alone suggests that the straightforward account of theory change expressed in the legend of CBT, whereby its predecessor psychoanalysis was decisively replaced as a result of Beck showing the analytic account of depression to be theoretically inadequate, is overly simplistic.Instead, key questions around what constitutes mental illness and therapeutic success, revolved around this contested notion of evidence, as did the question of scientificity. In Part II we turn to the influential notion of evidence introduced by psychopharmaceuticals, before returning to Beck’s narrative.WORKS CITEDBeck, A. T. (1963). Thinking and Depression: I. Idiosyncratic Content and Cognitive Distortions. Archives of General Psychiatry ,9 (4), 324–333.Beck, A. T. (1964). Thinking and Depression: II. Theory and Therapy.Archives of General Psychiatry , 10 (6), 561–571.Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects . Hoeber Medical Division, Harper & Row.Beck, A. T., Rush, A. J., Shaw, B., & Emery, G. H. (1979).Cognitive Therapy of Depression . Wiley.Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An Inventory for Measuring Depression. Archives of General Psychiatry , 4 (6), 561–571.Beck, J. (2011). Cognitive Therapy: Basics and Beyond (2nd ed.). Guilford Press.Borch-Jacobsen, M., & Shamdasani, S. (2011). The Freud Files: An Inquiry into the History of Psychoanalysis . Cambridge University Press.Burnham, J. C. (Ed.). (2012). After Freud left a century of psychoanalysis in America . University of Chicago Press.Caplan, A. (1986). With a friend like Professor Grünbaum does psychoanalysis need any enemies? Behavioral and Brain Sciences ,9 (2), 228–229.Dryden, W. (2012). Cognitive Behaviour Therapies . SAGE.Eysenck, H. (1952). The effects of psychotherapy: An evaluation.Journal of Consulting and Clinical Psychology , 60 (5), 659–663. https://doi.org/10.1037/0022-006X.60.5.659Eysenck, H. (1985). Decline and fall of the Freudian empire . Penguin, Viking.Eysenck, H. (1990). Rebel with a cause . WHAllen.Fancher, R. T. (1995). Cultures of Healing: Correcting the Image of American Mental Health Care . W. H. Freeman.Gipps, R. (2017), Does the cognitive therapy of depression rest on a mistake? British Journal of Psychology Bulletin , 41(5):267-271.Gipps, R. (2019) ’A New Kind of Song: The Art of Psychoanalysis’ inOxford Handbook of Philosophy and Psychoanalysis (eds. Gipps R. and Lacewing M.), Oxford University Press.Glymour, C. N. (1980). Theory and Evidence . Princeton University Press.Grosskurth, Phyllis. (1991). The secret ring: Freud’s inner circle and the politics of psychoanalysis . Cape.Grünbaum, A. (1984). The foundations of psychoanalysis: A philosophical critique . University of California Press.Grünbaum, A. (1993). Validation in the clinical theory of psychoanalysis: A study in the philosophy of psychoanalysis . International Universities Press, Inc.Kuhn, T. S. (1962). The Structure of Scientific Revolutions . University of Chicago Press.Lacewing, M. (2012). Inferring Motives in Psychology and Psychoanalysis.Philosophy, Psychiatry, and Psychology 19 (3):197-212.Lacewing, M. (2018). The Science of Psychoanalysis. Philosophy, Psychiatry, and Psychology 25 (2):95-111.Marks, S. (2012). Cognitive Behaviour Therapies in Britain: The Historical Context and Present Situation. In W. Dryden (Ed.),Cognitive Behaviour Therapies . SAGE.Paris, J. (2005). The fall of an icon: Psychoanalysis and academic psychiatry . University of Toronto Press.Popper, K. (1962). Conjectures and refutations: The growth of scientific knowledge . Basic Books.Rogow, A. A. (1971). The psychiatrists . Allen and Unwin.Rosner, R. (2012). Aaron T. Beck’s Drawings and the Psychoanalytic Origin Story of Cognitive Therapy. History of Psychology ,15 (1), 1–18.Rosner, R. (2014). The “Splendid Isolation” of Aaron T. Beck.Isis , 105 (4), 734–758.Rosner, R. (2018a). History and the Topsy-Turvy World of Psychotherapy.History of Psychology , 21 (3), 177–186. https://doi.org/10.1037/hop0000102Rosner, R. (2018b). Manualizing psychotherapy: Aaron T. Beck and the origins of Cognitive Therapy of Depression. European Journal of Psychotherapy, Counselling, and Health , 25–47.Rosner, R. (2018c). Three myths and truths about Beck’s early years. InScience and practice in cognitive therapy: Foundations, mechanisms, and applications (pp. 13–28). Guilford Press.Simon, C. C. (2002, September 3). A Change of Mind. Washington Post . https://www.washingtonpost.com/archive/lifestyle/wellness/2002/09/03/a-change-of-mind/19573ec9-9b36-4d11-874f-eb6dbc5d9164/Tallis, F. (1998). Changing minds: The history of psychotherapy as an answer to human suffering . Cassell.Van Fraassen, B. (1980). The scientific image . Clarendon Press.Weishaar, M. E. (1993). Aaron T. Beck . Sage.

This is the pre-print of an article forthcoming 2024 at Philosophy, Psychology and Psychiatry. Please do not distribute without permission.