The Diagnostic and Statistical Manual: Mental Disorders (1952), now generally known as DSM-I, attracts little attention in the current literature. The received view is that this system was a relatively unimportant, psychoanalyticly oriented classification. For example, Allan Horwitz (Reference Horwitz, Cautin and Lilienfeld2015) claimed that
All of the approximately 100 diagnostic definitions in DSM-I were short, cursory, and infused with psychodynamic assumptions. (Horwitz, Reference Horwitz, Cautin and Lilienfeld2015, p. 2)
In her history of DSM-III, Hannah Decker stated
Little attention was paid to the first two DSMs, which were published mainly for psychiatrists in state mental hospitals who were interested in compiling a variety of statistical information on their patients' lives and deaths. (Decker, Reference Decker2013, p. xvii)
Gerald Grob, whose (1991) ‘Origins of DSM-I: a study in appearance and reality’ remains the seminal discussion of this classification, noted
The publication of DSM-I marked an internal transformation that mirrored the growing dominance of psychodynamic and psychoanalytic psychiatry and the relative weakness of the biological tradition. (Grob, Reference Grob1991, p. 421)
The goal of this paper is to challenge this received history. We will argue that DSM-I was not psychoanalytic; rather it was eclectic, owing as much to Kraepelin as Freud. In addition, DSM-I was used frequently in its time. Copies circulated widely, and the manual went through 20 printings (APA, 1968, p. ix). In short, DSM-I was much more like recent editions of the DSM than generally thought. This matters because modern writers have often severed later editions of the DSM from their historical context. Conceiving of DSM-I as psychoanalytic and professionally unimportant has led to it being dismissed as a mere artifact of a long-gone era. DSM-I and DSM-II are presented as having only their name in common with later editions, and the current epoch in psychiatry seems to start in 1980. For some this has resulted in an assumption that the classificatory strategies of the more recent DSMs delineate the only ways in which ‘right thinking’ psychiatrists have ever divided up the domain of psychopathology. For others (critics of the DSM) the idea that the current DSM categories appeared from nowhere around a committee table in the late 1970s has facilitated their too-easy dismissal (compare, for example, Maxmen, Reference Maxmen1985, p. 35; Whitaker, Reference Whitaker2010, pp. 269–271; Shorter, Reference Shorter2015, p. 2). [see the Supplementary material (Historiographical Note) for further discussion.] In our history, the continuities between the first DSMs (and earlier classifications) and the more recent DSM editions become much more visible.
DSM-I was not a psychoanalytic classification
We start by raising four initial points that should make readers doubt that DSM-I can be characterized as psychoanalytic. With a prima facie case established, we then turn to consider the way in which DSM-I was developed. Demonstrating the influence that various prior classifications had on DSM-I will further support our argument.
DSM-I was not a psychoanalytic classification – the prima facie case
Psychoanalysts did not write DSM-I
DSM-I included a list of 28 individuals who served for some portion of time on the APA Committee of Nomenclature and Statistics between 1946 and 1951 (APA, 1952a , p. xii). These are the individuals who were credited with producing DSM-I. The Biographical Directory of the American Psychiatric Association for 1950 contained details regarding 25 of these individuals (the other three had died prior to 1950). Of these 25, only seven (28%) had either trained as psychoanalysts, or were members of psychoanalytic societies (specifically, George N. Raines, Moses M. Frohlich, Franz Alexander, Norman Q. Brill, Jacob H. Friedman, Nolan D. C. Lewis and George S. Sprague). Of the three committee members whose details cannot be found in the 1950 Biographical Directory, the obituary of Jacob Kasanin listed him as a member of the American Psychoanalytic Association (Anon, 1946). Clarence O. Cheney and James V. May were one-time presidents of the APA and the American Journal of Psychiatry contains extensive obituaries for them (APA, 1949a ). These obituaries make no mention of any psychoanalytic training, or membership of psychoanalytic societies, and these two were thus almost certainly not psychoanalysts. This means that only 29% of the 28 individuals responsible for DSM-I can be characterized as psychoanalysts.
The 1950 Biographical Directory also provides details of place of work and publications. The 28 committee members had diverse interests (as catalogueed in Anon, 1946; APA, 1949a , 1950). Many had military experience (notably Norman Brill, John Caldwell, J. P. S. Cathcart, Jacob Friedman, Moses Frohlich, Baldwin Keyes, George Raines). Others had worked for the Veterans' Administration (notably John Baird, George Brewster, Moses Frohlich, Harvey Tompkins). At least three had worked in state mental hospitals (Clarence O'Cheney, Jacob Kasanin, James May). There were notable neurologists (John Baird, Walter Breutsch, Norman Brill, Houston Merritt, Robert Schwab). Neil Dayton and James May had long-standing interests in classification and epidemiology. Baldwin Keyes and Mabel Ross were child psychiatrists; Lawrence Kolb was a specialist on addictions; Abram Bennett and Walter Breutsch worked on somatic therapies.
Members of the committee were assigned to small groups, each of which was responsible for some section of the manual, for example, one group was responsible for the nomenclature of psychoneuroses, another for the psychoses of unknown origin, and so on (Menninger, Reference Menninger1963, p. 475). Committee members would most likely have been allocated to groups where their expertise was greatest, making it probable that psychoanalysts would have greater input into certain sections of the manual. Still, given that overall less than one third of the committee members were psychoanalysts, and that the committee members had a wide range of interests, DSM-I cannot be considered a psychoanalytic classification in the sense of having been written by analysts.
Psychoanalysts did not use DSM-I
DSM-I cannot be considered a psychoanalytic classification in the sense of being a classification used by psychoanalysts, as psychoanalysts in the 1950s tended not to employ DSM-I. Between 1954 and 1958 a committee of the American Psychoanalytic Association tried to collect statistical data with the aim of measuring the effectiveness of psychoanalysis in treating particular conditions (Weinstock, Reference Weinstock, Katz, Cole and Barton1965). They decided to use DSM-I to code the diagnosis of the patients. Psychoanalysts participating in the study were each sent a free copy, suggesting that psychoanalysts were unaccustomed to using DSM diagnoses. Early in the course of the study it became apparent that the psychoanalysts were unwilling or unable to use the DSM diagnoses. Fewer and fewer psychoanalysts returned their forms. Many of those who did continue to participate either left the diagnosis unspecified or used the same diagnosis, most often passive-aggressive personality, compulsive personality or pseudoneurotic schizophrenia (which was not a DSM-I diagnosis) for all their patients. Others abandoned the diagnostic section of the questionnaires altogether. In addition to finding the DSM unusable, the psychoanalysts were uninterested in developing their own diagnostic system. The committee invited the membership to form groups to study the problem of diagnosis, ‘but as far as could be determined, only one group met and that for only one discussion.’ (Weinstock, Reference Weinstock, Katz, Cole and Barton1965, p. 68).
DSM-I was not ‘infused’ with psychoanalytic theory
Horwitz claimed that DSM-I was ‘infused’ with psychoanalytic theory, and provided examples of the influence of psychoanalytic thinking in DSM-I to prove his point (2015). One of the major families of disorders in DSM-I were the ‘Psychoneurotic Disorders’. The first sentences from the DSM-I description for these disorders stated:
The chief characteristic of these disorders is ‘anxiety’ which may be directly felt and expressed or which may be unconsciously and automatically controlled by the utilization of various psychological defense mechanisms (depression, conversion, displacement, etc.). …
‘Anxiety’ in psychoneurotic disorders is a danger signal felt and perceived by the conscious portion of the personality. It is produced by a threat from within the personality (e.g. by supercharged repressed emotions, including such aggressive impulses as hostility and resentment), with or without stimulation from such external situations as loss of love, loss of prestige, or threat of injury. (APA, 1952a , pp. 31–32)
Here the psychoanalytic influence is plain to see. If the definitions of all or even most of the mental disorders read this way, Horwitz's claim would be correct.
The hierarchical organization of DSM-I was a top-down system that was designed to reflect clinical thinking about psychopathology at the time. DSM-I subdivided mental disorders into (I) those with an organic basis (‘Disorders caused by or associated with impairment of brain tissue function’), (II) those for which an organic basis was unknown (‘Disorders of psychogenic origin or without clearly defined physical cause or structural change in the brain’) and (III) the ‘mental deficiency’ disorders. The organic disorders were split into (I.A) acute (13 diagnoses) v. (I.B) chronic forms (26 diagnoses). The (II) non-organic disorders were split into (II.A) psychotic disorders (20 diagnoses), (II.B) psychophysiological disorders (10 diagnoses), (II.C) psychoneurotic disorders (eight diagnoses), (II.D) personality disorders (22 diagnoses) and (II.E) transient situational personality disorders (10 diagnoses). Notably, the DSM-I class of ‘personality disorders’ was wide and heterogeneous, including not only what would now be considered personality disorders, but also sexual deviations, addictions, and ‘special symptom reactions’, such as speech disturbances and enuresis.
Of the 127 diagnoses in DSM-I (including non-diagnostic terms), there were 44 categories that could be said to have partly psychoanalytic definitions: the psychophysiological disorders, the psychoneurotic disorders and the personality disorders (minus six diagnoses under sociopathic personality disturbance). Thus, 35% of the diagnoses in DSM-I could be said to be psychoanalytic. In contrast, 46 (36%) of DSM-I diagnoses descended directly from Kraepelin's classification (e.g. general paralysis, schizophrenic reaction (dementia praecox), manic-depressive psychosis, etc., see Aragona, Reference Aragona, Zachar, St. Stoyanov, Aragona and Jablensky2015, for a useful chart showing similarities between DSM-I and Kraepelin's classification). As a result, categories with psychoanalytic terminology in their definitions did not represent the majority of diagnoses. DSM-I drew as much on Kraepelinian concepts as it did psychoanalytic concepts.
Of course, not all DSM-I diagnoses were of equal importance; some diagnoses were used far more frequently than others. If the psychoanalyticly-influenced DSM-I diagnoses were those most used in practice then perhaps the DSM-I-as-used might be considered psychoanalytic. In 1955, a year when DSM-I had been widely adopted and statistics are available, many psychiatric patients in the United States continued to be treated as inpatients. That year, public mental hospitals housed 559 000 patients (excluding those in VA facilities), while an estimated 379 000 were seen as outpatients (Kramer & Pollack, Reference Kramer and Pollack1958; Bahn & Norman, Reference Bahn and Norman1959). Among inpatients, the most common diagnoses were the schizophrenic reactions (51%), cerebral arteriosclerosis (9%), and mental deficiency (8%) (National Institute of Mental Health, 1964). Outpatient clinics at the time saw more children than adults (Bahn & Norman, Reference Bahn and Norman1959), and the most common diagnoses given to children were transient situational personality disorders (36%), personality disorders (which in DSM-I included diagnoses for learning disturbance, speech disturbance and enuresis) (21%), and mental deficiency (18%). The most common adult diagnoses in outpatient clinics were personality disorders (which included sexual deviations and addictions) (32%), psychoneurotic disorders (31%), and psychotic disorders (20%) (Bahn & Norman, Reference Bahn and Norman1959). Thus, both in terms of the numbers of categories included in DSM-I, and in terms of the numbers of patients diagnosed, the psychoanalyticly influenced diagnoses, although significant, were not dominant.
Contemporaries of DSM-I did not consider it psychoanalytic
Contemporary characterizations of DSM-I support our claim. George N. Raines was chairman of the committee that produced DSM-I. In his view,
The present new nomenclature of The American Psychiatric Association follows … the general nomenclatural scheme of Adolf Meyer, utilizes the names originated by Kraepelin and Bleuler, and incorporates the dynamics developed by Freud and later analysts, wherever these are applicable. (emphasis in original, Raines, Reference Raines1953a , p. 425)
Commentators from the time of DSM-I saw both a Kraepelinian and psychoanalytic influence,
It seems to be inconceivable that the Diagnostic and Statistical Manual: Mental Diseases could have been written without the preceding work of Kraepelin. (Kahn, Reference Kahn1955, p. 395)
The modern taxonomy adopted by the American Psychiatric Association is a direct descendant of the old Kraepelinian classification of mental disorders as definite disease entities. (Blinder, Reference Blinder1966, p. 259)
The definition of terms in the new nomenclature are largely descriptive. … In addition to the classical descriptive material, however, the revised nomenclature includes in its definitions a certain amount of psychodynamic theory as to the nature and origin of the conditions described. (Bowman & Rose, Reference Bowman and Rose1951, p. 165)
With a prima facie case established that DSM-I was not a purely psychoanalytic classification, we turn now to consider the way in which DSM-I was developed. Demonstrating the influence that various prior classifications had on DSM-I will further support our argument.
Development of DSM-I
In producing the new manual, the American Psychiatric Association's Committee on Nomenclature and Statistics, drew on two series of earlier classifications. The first of these systems was the Statistical Manual for the Use of Institutions for the Insane (1918) and its successor volumes (called Statistical Manual for the Use of Hospitals for Mental Diseases), which was used to collect statistics on mental hospital populations (American Medico-Psychological Association, 1918; APA, 1942). The second influence was classifications designed for use with military personnel during World War II (Army, Navy, and VA).
DSM-I as descended from the Statistical Manual for the Use of Institutions for the Insane (1918)
Gerald Grob (Reference Grob1991) reported that in 1908 the Bureau of the Census asked the American Medico-Psychological Association (later the APA) to appoint a Committee on Nomenclature of Diseases to assist with the collection of data. The Federal Census took a particular interest in those citizens who were dependent upon governmental care for their well-being, as it was hoped that statistical knowledge could inform the development of appropriate public policy. The American Medico-Psychological Association created the requested committee in 1913 and the first classification system, produced in collaboration with the National Committee for Mental Hygiene, was finally published in 1918, as Statistical Manual for the Use of Institutions for the Insane (American Medico-Psychological Association, 1918).
The 1918 Statistical Manual organized mental diseases into 21 groups (families of disorders). Included in these 21 were groups such as traumatic psychoses (with three subdivisions), general paralysis (no subdivisions), psychoses due to drugs and other exogenous toxins (four subdivisions), and dementia praecox (four subdivisions). Thirteen of these 21 groups were ‘associated with organic brain disturbance’ (in DSM-I terms). Of the remaining eight groups, four were non-organic psychotic disorders.
James V. May sat on the committee that produced the 1918 classification. His textbook Mental Diseases (1922) discussed the origins of the classification. May noted that Kraepelin's classification had been introduced to the United States by Adolph Meyer and August Hoch and became widely accepted (May, Reference May1922, p. 244). The 1918 classification sought to reflect the contemporary consensus within US systems (May, Reference May1922, p. 247) and was thus a modified Kraepelinian classification.
The 1918 manual went through a number of different editions. By 1929 the APA Committee of Statistics celebrated the success of the manual (now renamed Statistical Manual for the Use of Institutions for Mental Diseases).
The classification has been officially adopted by the Federal Census Bureau, The United States Public Health Service, the Surgeon-General of the Army, the United States Veterans Bureau, and by practically all the state hospitals for mental diseases in the United States. It is being taught in courses in psychiatry in many medical schools and is given with approval in the newer American text-books in psychiatry. (1929 report of APA Committee on Statistics, cited in Pollock, Reference Pollock1945, p. 10)
The Statistical Manual sold in fair quantities. The eighth edition (1934), for instance, sold 2700 copies in the first 6 months of availability (APA, 1935, p. 472), at a time when there were only 1510 APA members (APA, 1935, p. 461). Compared to the 1918 edition of the manual, in the eighth edition the section for psychoneuroses was greatly expanded (going from four to 17 sub-divisions) (Dayton, Reference Dayton1935). The tenth edition, published in 1942, remained very similar to the eighth, although the ordering of some classes was revised (APA, 1942).
Published in 1952, DSM-I was a continuation of this series of manuals for compiling hospital statistics. In comparing itself to the 10th edition of The Statistical Manual, the DSM-I described itself as a continuation of an existing series rather than as a paradigm-breaking document. The introduction to DSM-I suggests that ‘Perhaps the greatest change in this revision from previous listings lies in the handling of the disorders with known organic etiologic factors’ (APA, 1952a , p. 9); DSM-I introduced a distinction between acute and chronic brain disorders. Other key changes were said to be as follows: the schizophrenic reactions increased in numbers and type (including a schizoaffective type, and childhood type, not included in earlier classifications; p. 27); the manic depressive reactions were reduced in number, but a new code ‘000-x14 Psychotic depressive reaction’ was included (p. 25); and new, or greatly expanded, sections of ‘psychophysiologic autonomic and visceral disorders’ (psychosomatic disorders), personality disorders and transient situational personality disorders (reactions to stress) were added. Morton Kramer, Chief of the Biometrics Branch of the National Institute of Mental Health (NIMH), prepared comparison sheets for contemporaries charting the main changes between the 10th edition of the Statistical Manual for the Use of Institutions for Mental Diseases and DSM-I. In Kramer's view ‘there were not … many basic differences’ (Gottlieb, Reference Gottlieb1952, p. 7).
DSM-I as descended from Armed Forces classifications
During World War II each of the US Armed Forces, and the Veterans' Administration, developed its own classification for mental disorders. William C. Menninger was chiefly responsible for the Army classification, published as Technical Bulletin Medical 203 in 1943 (Houts, Reference Houts2000). The classification caused some controversy. Menninger claims that he was initially discouraged from producing Medical 203 by the then chairman of the APA Committee for Nomenclature and Statistics (who would have been either Neil Dayton, chair 1942–1946, or Nolan Lewis, chair 1946–1948), and that the American Journal of Psychiatry refused to publish the classification (Menninger, Reference Menninger1947, p. 582 fn. 5).
Medical 203 was heavily influenced by psychoanalytic theory. The introduction explained:
The term ‘disorder’ is used for the designation of the generic group of the specific reactions, while the specific reaction types have been termed ‘reactions’. The classifications of the psychoneuroses are based on the dynamics of the psychopathology. Of necessity, a few terms remained descriptive (symptomatic). [Office of the Surgeon General, 2000 (orig. 1946), p. 925]
Compared to the classifications intended for hospital use, the emphasis of Medical 203 was reversed. The classification started with ‘Simple personality reactions’ (which encompassed transient personality reactions to acute or special stress, combat exhaustion, and acute situational maladjustment). Neurotic reactions, ‘Character and behavior reactions’ and ‘Immaturity reactions’ also received prominent billing. By contrast, the sections on psychotic disorders, especially affective disorders, were fairly brief. Very little was said about organic conditions. While organic disorders made up at least half of the Statistical Manual for the Use of Hospitals for Mental Diseases, the organic disorders in Medical 203 were dealt with in a single paragraph, relegated to the end of the classification.
In 1947 the Veteran's Administration adopted a classification closely following Medical 203 (Veterans Administration, 1947). In turn, a pamphlet titled ‘Joint Armed Forces Nomenclature and Method of Recording Psychiatric Conditions’ was released in 1949 (US Army/Navy/Air Force, 1949). This sought to provide a common system for use in the Army, Navy and Air Force. Much of the wording remained the same as in Medical 203 and the VA classification – although the ordering of disorders differed. Given the similarities between Medical 203, the VA classification, and the Joint Armed Forces classification in what follows we will talk of ‘Medical 203/VA/Joint Forces’ when all three classifications are being discussed.
In places, especially in the schizophrenic reactions, psychophysiological autonomic and visceral disorders, psychoneurotic disorders, and personality disorders, the text of DSM-I borrowed heavily from Medical 203/VA/Joint Forces. However, the influence of Medical 203/VA/Joint Forces was limited. In 1949, the APA Committee on Nomenclature and Statistics reviewed the VA classification but was not overly impressed: it worried that the VA classification reflected the requirements of ‘expediency’ and ‘enthusiasms of a temporary nature’ (APA, 1949b , p. 930). Where DSM-I borrowed phrasing from Medical 203/VA/Joint Forces, often the most ‘psychoanalytic’ wording was deleted. In practice, it looks as if the DSM-I committee wrote sections of DSM-I by taking Medical 203/VA/Joint Forces text and crossing out what they considered to be controversial psychoanalytic claims. Compare, for example, Medical 203/VA/Joint Forces and the DSM-I on obsessive compulsive reaction. The former described this reaction as follows:
In this reaction the anxiety may be observable in connection with obsessional fear of uncontrollable impulses. On the other hand, the anxiety may be under apparent control, through a mental mechanism (isolation), by which the emotional charge becomes automatically separated from the main stream of consciousness and manifests itself in a displaced form through useless or excessive, and often repetitive activity. In the latter instance, the patient is utilizing the mental mechanisms of ‘undoing’ – a symbolic act which temporarily protects the patient against a threat – and ‘displacement’. The patient himself may regard his ideas and behavior as unreasonable and even silly, but nevertheless is compelled to carry out his rituals. The diagnosis should specify the symptomatic expressions of such reactions, including touching, counting, ceremonials, handwashing, recurring thoughts, accompanied often by compulsion to repetitive action. This category includes many cases formerly classified as ‘psychasthenia’. [Office of the Surgeon General, 2000 (1946), p. 928; Veterans Administration, 1947, p. 5; US Army/Navy/Air Force, 1949, p. 7; italics and underlining added]
In DSM-I, the last three sentences were almost identical (shown by underlining). However, the initial part of the definition (shown in italics) was distinctly altered. The DSM-I version was more descriptive of the signs/symptoms of the disorder and was less about the theorized mental mechanisms that allegedly explained those signs/symptoms.
In this reaction the anxiety is associated with the persistence of unwanted ideas and of repetitive impulses to perform acts which may be considered morbid by the patient. The patient himself may regard his ideas and behavior as unreasonable, but nevertheless is compelled to carry out his rituals. The diagnosis will specify the symptomatic expression of such reactions, as touching, counting, ceremonials, hand-washing, or recurring thoughts (accompanied often by a compulsion to repetitive action). This category includes many cases formerly classified as ‘psychasthenia’. (APA, 1952a , p. 33)
In contrast to the avowedly psychoanalytic orientation of Medical 203/VA/Joint Forces, the DSM-I presented itself a classification that ‘recognizes the present day descriptive nature of all psychiatric diagnoses’ (APA, 1952a , p. 9). Still, not all psychoanalytic phrasing was deleted before text was inserted into DSM-I. For example, both Medical 203/VA/Joint Forces and the DSM-I described the paranoid personality as being characterized by a ‘tendency to utilize a projection mechanism’ [APA, 1952a , p. 36; Office of the Surgeon General, 2000 (1946), p. 929]. In addition to borrowing wording from Medical 203/VA/Joint Forces in describing disorders, the DSM-I also adhered to the terminology of the earlier classifications by often describing disorders as ‘reactions’ (APA, 1952a , p. 9).
In summary, the influence of Medical 203/VA/Joint Forces on DSM-I was limited. Most obviously, about half of the DSM-I concerned organic disorders; Medical 203/VA/Joint Forces had no influence on these disorders. Among the non-organic conditions there were also a few notable differences. Hypochondrical reaction was included in Medical 203/VA/Joint Forces but excluded from DSM-I. Schizophrenic reaction, schizoaffective type and schizophrenic reaction, childhood type were in DSM-I but did not appear in Medical 203/VA/Joint Forces.
A brief note should also be made of ICD-6, published in 1948 (WHO, 1948). US psychiatrists played a key role in shaping proposals for ICD-6, thus accounting for the similarities between ICD-6 and DSM-I (Frohlich, Reference Frohlich and Zubin1961, p. 87; Kendell, Reference Kendell1975, p. 92). Although there are some differences between DSM-I and ICD-6, attempts were made to ensure that DSM-I codes could be converted into ICD equivalents (as outlined in Supplementary Appendix A of DSM-I). The direct influence of ICD-6 on DSM-I appears to have been minimal. Schizophrenic reaction, schizoaffective type is the only ICD-6 category included in DSM-I that cannot be found in earlier US classifications (either the military classifications or the Statistical Manual for the use of Hospitals for Mental Diseases).
DSM-I was influential
As noted in the foreword of DSM-I, one of the innovative steps in creating this classification was that the DSM-I committee distributed advance copies to 10% of the members of the APA and solicited feedback in the form of questionnaires. Of the returns, 93% expressed general approval of the new classification. Raines noted that a number of the state hospitals and mental health clinics held staff meetings to discuss the new system and its changes. The efforts by Raines and his committee to pre-test DSM-I probably helped this classification in achieving its goal of becoming accepted by the clinicians who were its audience.
During the 1950s, the American Journal of Psychiatry printed an annual note on the accounts of the APA. These recorded sales of DSM-I as being $1465 to March 1952, $10 305 from April 1952 to March 1953, and $9825 from April 1953–March 1954. From 1955 and 1965 no sales figures were recorded (APA, 1952b , 1953, 1954). DSM-I thus made at least $21 595. The DSM-I sold for $1.50 (Felix, Reference Felix1956, p. 405). This implies that the DSM-I sold over 14 000 copies in its first 3 years of publication. This equate to roughly two copies for each member of the APA [in 1952 the APA had 7105 members (APA, 1952a , p. 210)]. The DSM-I classification was also reprinted in a number of textbooks (Menninger, Reference Menninger1952; Masserman, Reference Masserman1955; McCartney, Reference McCartney1956) and was used as the organizational basis for popular undergraduate textbooks, on abnormal psychology (Coleman, Reference Coleman1956), and psychiatry (Freedman & Kaplan, Reference Freedman and Kaplan1967).
DSM-I was used for a variety of purposes. A year after publication, one half of public mental hospitals, one third of private mental hospitals, and three-quarters of general hospitals were using the new nomenclature for patient records (Raines, Reference Raines1953b ). The annual statistical surveys, ‘Patients in mental institutions’, employed DSM-I categories (Public Health Service, 1958). Outpatient clinics likewise came to use the DSM-I; in 1956, 39% of State and VA outpatient clinics were reporting statistical data that included a DSM-I diagnosis (Bahn & Norman, Reference Bahn and Norman1959). Researchers also used the nomenclature of DSM-I categories (notable examples include Pasamanick et al. Reference Pasamanick, Roberts, Lemkau and Krueger1957; O'Neal & Robins, Reference O'Neal and Robins1958; Greenblatt et al. Reference Greenblatt, Grosser and Wechsler1964). Stengel wrote an influential review of international psychiatric classifications for the WHO, who were concerned by the relative lack of impact of ICD-6 on the international psychiatric community (Stengel, Reference Stengel1959, p. 605). In this review Stengel notes that ‘some other countries of the Western Hemisphere’ had been considering adopting DSM-I, and dedicates a special chapter to the classification ‘in view of its special importance’ (Stengel, Reference Stengel1959, p. 605). Gruenberg's summary comment in his ‘Foreword’ to DSM-II appears to have been an accurate assessment, ‘The first edition of this Manual (1952) made an important contribution to U.S. and, indeed, world psychiatry. It was reprinted twenty times through 1967 and distributed widely in the U.S. and other countries’. (APA, 1968, p. ix).
In summary, DSM-I sold in large numbers and circulated widely. This being said, its use was by no means universal. When researchers found the DSM-I unsuitable for their purposes, they could ignore it and define the categories they studied descriptively in their papers. Some textbooks employed non-DSM diagnoses (for example, Arieti, Reference Arieti1959). State hospitals in New York State never moved to the DSM-I system, but moved straight from using the 1942 edition of the Statistical Manual to DSM-II (Spitzer & Wilson, Reference Spitzer and Wilson1968). Although DSM-I was available to psychiatrists to use as a common reference point, it could be ignored whenever a different approach seemed preferable.
Conclusion
The generally accepted view of DSM-I is that this system was psychoanalytic. This view is incorrect. The roots of DSM-I came from (1) World War II military classifications that did contain psychoanalytic concepts, and from (2) earlier classifications by the APA that were Kraepelinian in focus and intended primarily for state hospital inpatient settings. Although US psychiatry in the 1950s is often characterized as psychoanalytic in orientation it is better described as eclectic. Articles in the American Journal of Psychiatry during this period were as likely to discuss somatic therapies, such as psychosurgery and ECT, as they were to discuss psychotherapy. Individual psychiatrists often combined ideas from a variety of schools in their own thinking. DSM-I reflected this eclecticism and was a common-sense compromise among the schools of thought in American psychiatry at the end of World War II. DSM-I sold well and circulated widely. The current tendency to dismiss DSM-I as unimportant is an error.
Supplementary material
For supplementary material accompanying this paper visit http://dx.doi.org/10.1017/S0033291715002093.
Acknowledgements
We are grateful to the following individuals, who read drafts and made many helpful comments: Sam Fellowes, A. Kenneth Fuller, John Gunderson, Allan Horwitz, Arthur Houts, Jared Keeley, Martyn Pickersgill and Geoffrey Reed.
