Trauma as neurological disturbance and experimental obstacle

On Saturday April 17, 1937, an elect group of clinicians, educators, and researchers from numerous disciplines gathered in Washington, DC, to attend a conference on “Experimental Neuroses and Allied Problems” held under the auspices of the National Research Council’s (NRC) Inter-Divisional Committee on Borderland Problems of the Life Sciences (NRC 1937). Among those presenting were the two pioneers of the study of experimental neurosis in the United States: Horsley Gantt of the Johns Hopkins University Hospital and Howard Liddell, chair of the department of physiology at Cornell University’s Medical College. Like the influential experimental psychologist Clark Hull, who also attended the conference, Liddell and Gantt briefly outlined their respective research programs.

The modes of enquiry cultivated in the laboratory, the discussions at the conference show, could not always be squared with the exigencies of the clinic and the investigative strategies favored by clinicians. The medical educator Clifford Farr, for instance, presented the results of an informal survey conducted among psychiatrists about their views of experimental research. On the one hand, most respondents welcomed laboratory studies exploring the etiology of behavior disturbances, especially the neuroses. On the other hand, they also believed that many experimental findings had merely been “occasional byproducts” of other research and that, in any case, neurotic behavior could only be observed in predisposed animals. Therefore, the interviewees suggested, researchers’ efforts should concentrate on the study of genetic strains to elucidate the role of hereditary factors (ibid., 7).

These and similar comments show that clinicians were generally unaware of the specific problems experimentalists struggled with at the time; they also point to a stubbornly persistent misconception about the origins of the study of experimental neurosis (see, e.g., Abramson and Seligman Reference Abramson, Seligman, Maser and Seligman1977). During the discussion that followed, the eminent psychologists and primatologist Robert Yerkes thus felt compelled to point out what he saw as the methodological advantage experimental models provided over clinical case histories: Unlike clinicians, comparative psychologists like himself were not confined to the “predisposed human subject.” “Our attitude in working with the infra-human,” he unequivocally stated, “is just the opposite: We prefer the normal.” In the long run, he explained, the study of “normal animal subjects” would yield bigger dividends than experiments with “specially selected and bred predisposed lines” (NRC 1937, 9).Footnote ⁴ Like other experimental researchers, comparative psychologists hoped to isolate causal mechanisms and wished to control for influences that led to deviations from the norm and limited the comparability of their results. Moreover, to develop methods that convincingly and reliably produced neuroses in animals it was important to demonstrate that the animals that had been rendered neurotic had indeed been normal, initially healthy specimens. To know the abnormal, the argument went, experimentalists had to get to know the normal first.

The practical problem of controlling for a laboratory animal’s idiosyncrasies extended beyond experimental psychology. Turn-of-the century physiologists had long been aware of the individual differences in temperament and affective reactions among their experimental subjects. To arrive at reliable conclusions, laboratory workers knew, an animal’s reactions had to be related to what was known about its typical behaviors as well as its affective state during the experiment (Dror Reference Dror1999). Pavlov, too, had noticed differences in the performance of his animals in the laboratory, particularly, the tendency of otherwise lively dogs to fall asleep during experimentation (Smith Reference Smith1992, 202). He described these differences in anthropomorphic terms and came to understand them as an expression of their differing personalities or nervous types, which he began to study systematically in 1922. In fact, it had been within the context of his research on nervous types that Pavlov and his close collaborator Maria Petrova first decided to deliberately induce neurotic states in two dogs (Todes Reference Todes2014, 497–8). Differences in the dogs’ “personalities,” Pavlov postulated at the time, were caused by the predominance of either one of two opposing neural forces (inhibition, excitation) or their being in balance. As the medical historian Daniel Todes has shown, Pavlov was influenced by Freud’s idea that conflicting unconscious drives lead to neuroses, as he and Petrova subjected dogs, which they had previously diagnosed as two contrasting types, to experimental trials designed to instigate a “clash” between excitatory and inhibitory impulses (ibid., 501).

As is well known, the paradigmatic experiment in the Pavlovian laboratory involved the presentation of a neutral stimulus (commonly the sound of a buzzer or the beating of a metronome) together with a stimulus (usually meat powder or acid) that unconditionally provoked the salivary reflex in a hungry animal. After several trials, the once neutral but now conditional stimulus (CS) would alone elicit the salivating response. In short, a so-called conditional reflex had been established.Footnote ⁵ In the experiments designed to upset the nervous balance of the two dogs, Postrel and Milord, Pavlov and Petrova used electric shocks of varying intensity, which functioned as the CS, while giving them meat powder. As the experimenters raised the current of the shock, the conditional reflex became irregular or disappeared altogether as the dogs, clearly in pain, tried to escape the experimental stand (ibid., 502). In another type of experiment, the so-called difficult differentiation experiment, which would become the leading paradigm for those studying experimental neurosis in the United States, dogs were trained to acquire conditional reflexes to certain prescribed stimuli, but not to other, similar stimuli. During a series of such experiments conducted around the same time by another one of Pavlov’s collaborators, the ophthalmologist N.R. Shenger-Krestovnikova, circles and ellipses were used as CS to study visual discrimination. As she manipulated the shape of the ellipses and circles over several trials, making them more similar and the differentiation increasingly difficult, the behavior of one dog in particular would become erratic to a point where the experiments could no longer be continued. The animal appeared to be, Shenger-Krestovnikova concluded, in a “genuine neurasthenic condition” (Shenger-Krestovnikova 1921, 41; cf. Windholz Reference Windholz1990, 50). In this case also, Pavlov now theorized, a “violent clash” between inhibitory and excitatory impulses had led to a lasting disturbance of the dog’s nervous system and impaired its ability to acquire new conditional reflexes, and reliably reproduce previously established ones (Pavlov 1927, 292–3).

The discovery of experimental neurosis, then, was far from incidental. Pavlov devoted much of his late career to the study of psychopathological phenomena, seemed well-acquainted with the clinical theories of his day, and regularly consulted clinicians of nearby clinics. Moreover, his overarching research aim had always been to explore what he called “higher nervous activity” by using the conditional reflex method, which provided him with a quantitative, albeit indirect measure (drops of saliva) of nervous activity.Footnote ⁶

Still, Pavlov would have agreed, partially at least, with the clinicians interviewed by Farr: Neurotic behaviors seemed to appear only in predisposed animals and practically never in the “well-balanced” type. In the case of the latter, lasting breakdowns solely occurred after exceedingly strong disruptions (Pavlov Reference Pavlov1934).

Trauma as “boundary object”Footnote ⁷

At the time of the Washington conference, experimental neuroses had been studied in animal laboratories in the United States and Canada for roughly a decade and neuroses had reportedly been produced in rats, chicks, goats, dogs, rabbits, monkeys, chimpanzees, doves, cats, and human children and adults (Cook Reference Cook1939a; NRC 1937). Not all of these experiments modeled their interventions on Pavlov’s work. Behavioral disturbances had long been observed in animal experiments – disturbances that could now be classified as experimental neurosis (see ibid.; Pressman Reference Pressman1998).Footnote ⁸ Crucial for the relatively early adoption of Pavlov’s method in America were institutional conditions conducive to explorative experimental research in interdisciplinary settings.

Comparative psychologists and physiologists, for one, saw in the experimental study of animals a promising new method to draw out the similarities between humans and other species. Darwin’s The Descent of Man and, especially, his Expression of the Emotions in Man and Animals, published a year later in 1871, played foundational roles for the discipline; although, comparative psychologists would soon reject the anecdotal approach of pioneers like Darwin and George Romanes in favor of more rigorous, experimental methods (Darwin 1870, [1871] Reference Darwin1998; Greenwood Reference Greenwood2015, 219; see also Dewsbury Reference Dewsbury1984). The aim of comparative psychology, Yerkes stated in 1913, was to arrive at “genetic descriptions of mental life” through the application of experimental methods (Yerkes Reference Yerkes1913, 581). Evolutionary theory in general and, particularly, the notion of the “phylogenetic scale” – believed to stretch from the so-called lower, simpler species to the higher, more complex species – provided comparative physiologists and psychologists with a unifying perspective under which behaviors and their presumed antecedents could be compared, arranged, and analyzed.

The use of experimental methods promised to bring to the study of mental disorders the respectability enjoyed by the natural sciences. Although, it must be noted, in the first two decades of the twentieth century, psychological laboratory research with animals was only undertaken by a handful of investigators in the United States. Wary of the high costs associated with maintaining animal laboratories on university campuses, but also fearing opposition and negative publicity from antivivisectionists, – who had grown in number towards the end of the previous century and came to include physicians and researchers as well as other prominent public figures – university administrators were quick to close many animal laboratories directed by psychologists in the wake of the financial crisis of 1907 (O’Donnell Reference O’Donnell1985). By the 1920s, however, in light of medicine’s recent therapeutic successes, public opposition to animal experimentation had weakened while the experimental sciences had only garnered more prestige (Guerrini Reference Guerrini2003). In 1926, William H. Welch, dean of the Johns Hopkins School of Public Health, declared antivivisectionism a “lost cause” (Lederer Reference Lederer1992, 62). Although this claim was somewhat exaggerated and surely premature, in the United States, at the time, the movement did fail to secure the kind of legislative victory it had carried away in Great Britain, already in 1876, with the Cruelty to Animals Act (cf. Franco Reference Franco2013). During the time period covered here, some British experimental psychologists envied the relative lack of legal oversight enjoyed by their colleagues in the United States (cf. Wilson Reference Wilson2004).

Among those invested in advancing the field of psychiatry, Pavlov’s method also attracted interest because it held the promise of bridging existing institutional and methodological divides that separated the research laboratory from the clinic (Pressman Reference Pressman1998). Under the leadership of Adolf Meyer, since 1913 director of the Henry Phipps Clinic at Johns Hopkins, a new psychiatry was taking shape, one that sought to overcome the restrictions placed on the investigation and the treatment of mental pathologies by a dualist framework that treated body and mind as separate entities both calling for different, at times conflicting, explanatory and therapeutic strategies. By contrast, Meyer’s monistic philosophy and medical doctrine of psychobiology demanded that mental pathologies should be studied from a biological as well as from a psychological standpoint. Meyer also no longer hoped to identify the biological defects that, as many of his colleagues posited, underlie psychopathologies. Instead, therein influenced by philosophical pragmatism and evolutionary thinking, he described the latter mainly in terms of problems of an individual’s adjustment to his or her social environment. Maladaptation or, to use the term preferred by Meyer, maladjustment was thus at the basis of what he called mental disorders, as opposed to “mental diseases” or “insanity” (ibid., 21; see also Lamb Reference Lamb2014).

In Alan Gregg, the director of the Rockefeller Foundation’s medical sciences program, Meyer found an influential devotee to his interdisciplinary vision of psychobiology. Under Gregg’s leadership the Foundation funded interdisciplinary research endeavors – most prominently in the evolving field of psychosomatic medicine – carried out at institutions such as the Chicago Institute of Psychoanalysis and Yale’s Institute of Human Relations, both of which were represented at the NRC conference in Washington. In the 1930s, the study of experimental neuroses directly profited from Rockefeller grants that supported the work of Yerkes and his team at the Laboratory of Comparative Psychobiology, also at Yale, as well as Liddell’s group at Cornell (Pressman Reference Pressman1998, 37). It was also Gregg who had introduced Meyer to Gantt’s work, prompting Meyer to invite him to set up the Pavlovian Laboratory in 1929 (Ruiz and Sánchez Reference Ruiz and Sánchez2016, 3). It was at Cornell, however, where, three years before, a case of experimental neurosis had first been diagnosed in the United States.

To study the effects of thyroxin on organismic functions such as reflexes, muscle activity, and temperature regulation, the physiologist Sutherland Simpson had begun to conduct a series of behavioral tests on thyroidectomized sheep in the early 1920s (see, e.g., Simpson Reference Simpson1924, cf. Kirk and Ramdsen Reference Kirk and Ramsden2018). Simpson and his student Liddell grappled with practical challenges, however, as they attempted to study the influence of hormonal secretions on learning. Their initial method, a maze test, failed to demonstrate any differences. As a promising alternative to this protocol, the researchers turned to Pavlov’s work. Pavlov’s student and translator Gleb Anrep functioned as Liddell’s advisor as he adopted the Pavlovian methods at the Cornell Physiological Field Station in the rural surroundings of Ithaca (Liddell Reference Liddell1957, vii). In the summer of 1926, the same year his mentor Simpson passed away, Liddell also visited Pavlov’s laboratories where he met his compatriot Gantt who had already spent several years abroad studying under Pavlov before accepting Meyer’s invitation.

After applying the conditioning method for four years, behavioral abnormalities roused Liddell’s interest and provoked a profound shift in his research (Liddell Reference Liddell1957, vii). Promisingly, neurotic reactions were first observed in a “normal,” non-thyroidectomized sheep, which he used as an experimental control (Liddell and Bayne Reference Liddell and Bayne1927; cf. Block Reference Block1963).

Subsequently, Liddell and his group studied artificial neuroses in a broad range of species, mainly in goats, pigs, and sheep, but also in dogs, rodents, and, very briefly, rabbits. Whereas Pavlov’s original method focused on the salivary reflex and required that the dogs underwent a surgical procedure to implant fistulas, Liddell analyzed conditioned motor reflexes, which was less invasive and allowed the animals to move more freely in the experimental stand. In their paradigmatic experiment, the animal was given an electric shock to one of its forelegs while a conditioned stimulus – again, the sound of a buzzer or the beating of a metronome – preceded the shock and provoked a defensive reaction: the flexing of the foreleg where the electrode was attached. The researchers then either used difficult differentiation experiments (during which only one of two increasingly similar stimuli preceded a shock) or conditioned delayed responses (during which the shock was administrated after a latency period) to induce neuroses.

Responding to the reservations his students might harbor towards such experiments, in his lectures, Liddell would invite listeners to hold the “battery clips” to experience the electric shock, which provoked, so he insisted, not pain but a “minimal startle stimulus” (Liddell Reference Liddell1957, 5–6). Arguing that the animals did not experience strong physical pain also served to corroborate his claim that neurotic disorders were ultimately caused by emotional distress, which, he further argued, was always linked to an individual’s expectations (ibid.).

Although physiological correlates of neurotic states such as elevated heart and breathing rates were diligently recorded, the indicator Liddell would focus on and explore in the following years was the erratic flexing of the foreleg – and, generally, “hyperirritability” (NRC 1937, 20) – attributed to the overstraining and subsequent breakdown of nervous inhibition that particularly difficult experiments presumably demanded. Inhibition, which already in Pavlov’s thinking had not only referred to a neurological process but also, in a metaphorical sense, to discipline (Todes Reference Todes2001, 352), was conceived of by Liddell and his co-workers as an active process. Early on, researchers concluded that requiring an animal not to respond to a signal “is not mere passivity but a cultivated repression of motor activity” (Liddell, James, and Anderson Reference Liddell, James and Anderson1934, 50; my emphasis). In a certain sense, over the course of the next two decades and by developing a number of experimental protocols, Liddell would explore the emotional and physiological toll this “cultivation” exacted on his laboratory animals. In 1938, with financial support from the Josiah Macy Jr. Foundation, Liddell and his team would move into a 110-acre facility, the Behavior Farm Laboratory. The proximity between experimental and rest spaces on the farm – where the animals could live “the normal life of the flock in barn and pasture” (Liddell 1956, 31) – made the boundaries of the laboratory permeable. The animals’ behavior inside and outside the laboratory could be systematically compared and, to this end, Liddell also began to employ ethological methods (for a detailed account of the evolution of Liddell’s techniques, see Kirk and Ramsden Reference Kirk and Ramsden2018).

The other pioneer of the study of experimental neurosis, Gantt, was more committed to the physiological interpretations and experimental procedures of his Russian teacher than Liddell. Although he included additional physiological parameters as he studied conditioned reflexes, mostly in dogs, his interpretations more clearly derived from, but also expanded Pavlov’s speculative neurological framework for understanding experimental neuroses (see Ruiz, Sánchez, and De la Casa Reference Ruiz, Sánchez and De la Casa2003; cf. Abramson and Seligman Reference Abramson, Seligman, Maser and Seligman1977). Portraying him as a dogmatic Pavlovian, as some of his contemporaries did (Levens Reference Levens2005, 158), would be to overstate his commitment to his teacher, however, and to downplay the influence Meyer and his vision of crossdisciplinary research had on his work.

Encouraged by his superior, Gantt formalized the biographical approach, for instance, by assembling so-called life charts for his dogs, a method for clinical recordkeeping developed by Meyer (Ruiz and Sánchez Reference Ruiz and Sánchez2016). Although, as mentioned, it was by no means uncommon for experimental workers to be concerned about their animals’ lives outside the experimental chamber, they did not publish the kind of extensive case histories that Gantt included in his 1944 monograph Experimental Basis for Neurotic Behavior. The case history of the mongrel Nick, Gantt’s most celebrated account, not only detailed the dog’s travails in the laboratory over more than a decade but also his life outside the experimental camera, including a lengthy stay at Gantt’s farmhouse, which gave him temporary relief from his symptoms. Initially chosen for “laboratory work” for being, “according to general appearances, normal” (Gantt Reference Gantt1944, 77), Nick nevertheless developed what Gantt and his coworkers understood to be neurotic symptoms.

In the foreword of the same monograph, which won him the Lasker Prize in Mental Hygiene, Gantt also expounded on the societal relevance of the study of experimental neurosis. Echoing the aspirations of the Rockefeller Foundation, Gantt hoped that his laboratory studies would contribute to a growing body of research about “emotional imbalances” and the wide-ranging social problems associated with them. Quoting Gregg, Gantt affirmed that Americans’ most pressing problems, after having conquered the “frontier,” concern the social and no longer their natural environment, as the members of the human race must now “’learn to live together.’” In a “periodically recurring chaotic world state,” he argued, it had in fact become imperative to better understand the “destructive tendencies that make adjustments difficult” (ibid., viii). Accordingly, Gantt set out to develop diagnostic tools that could assist physicians in detecting neurotic breakdowns before they were “overtly expressed” (ibid., vii).

Recent experiences of social upheaval and armed conflict did not only provide a rationale for the study of neurosis, underscoring its exigency. The psychiatric casualties of the two World Wars had also amply demonstrated the viability of a potential causal pathway. Among the psychological methods to artificially produce neuroses mentioned in the research appraisal prepared by the Committee on Problems of Neurotic Behavior, “shock through fright or threat” figured at the top of the list (NRC 1938, 8).Footnote ⁹ Also Gantt devoted the first of his empirical chapters of his 1944 monograph to “natural emotional shocks; traumatic and experimental war neuroses” (Gantt Reference Gantt1944, 13). Following the suggestion of one of his colleagues, a psychiatrist who had served as a medical officer during the war, Gantt had begun to study “experimental war neuroses” by exposing three of his dogs to the explosions of fireworks, the sound of actual gun shots, or by placing them in a swinging hammock (ibid., 31). “Elements of war and battle,” he assumed, must have also been experienced by dogs who had escaped their paddocks one Saturday night in 1931: As they roamed the upper floors of the laboratory building, the dogs became involved in altercations and were severely punished by lab personnel the next morning before they were forced back into their paddocks (ibid., 22). The deliberate attempts of re-enacting war experiences in the laboratory as well as these “intercanine fights,” however, did not – to Gantt’s surprise – prove as “potent” as the classical Pavlovian conditioning experiments (ibid., 31). Difficult differentiations held greater promise of exploring the physiological underpinnings of neurotic behavior. Moreover, medical experiments that caused animals pain continued to rouse skepticism among the public and the ire of antivivisectionists, who never ceased to campaign against cruel physiological experiments, especially those involving dogs (cf. Lederer Reference Lederer1992). Investigating difficult differentiations, then, may have also seemed less of a liability.

“Originating situations”: Trauma as pathogenic experience

What constituted neurotic animal behavior besides the breakdown of previously established conditional reflexes? Answering this question was not an easy task, also given the myriad symptoms by now recorded and investigated. The researchers discussed so far, it seems, purposely eschewed it. Gantt, for instance, was warned by Meyer not to apply diagnostic labels to the phenomena studied in his lab since, he argued, the disorders found in humans cannot be equated with those seen in other, less complex animals (Levens Reference Levens2005). The members of the NRC’s Committee on Problems of Neurotic Behavior, which included Liddell, voiced similar concerns. In its research appraisal, the committee considered the term “experimental neurosis” to be unfortunate since it implies “in a limiting sense a specific parallel between an animal’s disturbed behavior and human psychoneuroses” (NRC 1938, 12; my emphasis). Thus, it would be misleading to think of experimental neurosis as an experimental model in a narrow sense, i.e. as an investigative tool to study a specific process hypothesized to be involved in human disease and presumed to operate across different species. Indeed, because of species differences, it was expected that neurotic symptoms would take on different forms in different species. The comparative psychologists first task was to identify the similarities.Footnote ¹⁰

The presumed differences between humans and non-human animals had other methodological implications. Because humans can use language to convey their symptoms and animals cannot, Liddell argued, “originating situations must take precedence over resulting symptoms” when analyzing behavioral abnormalities (Liddell 1956, 59; emphasis in original). In short, to understand neuroses one must try to understand the conditions that caused them. This did not imply, however, that individual influences were not considered or deemed irrelevant. As already mentioned, these laboratory workers did not regard the external conditions as entirely independent of the animal exposed to them.

For one, there was the question of neurotic predisposition, which clinicians continued to raise, and experimenters had to be aware of. Once more following his teacher Pavlov, Gantt considered a dog’s constitutional type, the congenital “stability” of its nervous system, to be a decisive influence on a subject’s “performance.” However, only after consulting both biographical data and physiological measurements did he conclude with sufficient confidence that the dogs with “the stable [nervous systems] are always stable and the labile are always susceptible” to experimental neuroses (Gantt Reference Gantt1944, 31). Liddell, on the other hand, basing his conclusion on animals that had undergone different experimental regimens, explicitly rejected Pavlov’s assumption concerning the role of predisposition in the etiology of neurosis (NRC 1937, 22). More cautiously, the psychologist Norman R. F. Maier after inducing neuroses in rats with an insolvable discrimination experiment in the late 1930s, – more about these experiments shortly – assumed that because the individual animals “differ so greatly in their perceptions of the situation … their stability cannot be compared until we have matched their perceptions;” after all, “a given situation may constitute a conflict for one animal and not for another” (NRC 1939, 6). In a similar vein, Liddell, in one of his later lectures, warned of the “quantitative delusion” of many of his behaviorist colleagues who had become “preoccupied with the measurement of stimuli to the neglect of the meaning of these stimuli to the individual animal” (Liddell 1956, 54; emphasis in original; see also ibid., 34, 55).

Trauma as social conflict

To elucidate the “originating situations,” Liddell and Gantt not only relied on their own observations but also built on the hermeneutic expertise of scholars from bordering disciplines. For his monograph, Gantt elicited – besides a more orthodox reflexological interpretation by a Russian physiologist – psychoanalytical as well as anthropological readings of his cases: The psychoanalyst Leon Saul and the anthropologist and psychiatrist Alexander Leighton both weighed in with comments that highlighted the social aspects of the laboratory situation, including its underlying power dynamics. In his interpretation, Saul described the “threatening and traumatic situation” in the laboratory as opposed to the attachment the animals formed toward their masters (Gantt Reference Gantt1944, 138). Gantt himself would later systematically investigate the soothing effect that the presence of a familiar laboratory worker could have on the animal’s lab performance – the so-called “effect of person,” which had already been noted by Pavlov (Gantt et al. Reference Gantt, Newton, Royer and Stephens1966). Because of the social consequences of neurotic suffering, both Gantt and Liddell believed that the gregarious nature of dogs and sheep, respectively, made them ideal subjects for the study of experimental neuroses (cf. Levens Reference Levens2005, 185). Although, Liddell preferred other domesticated animals over the dog since he believed the latter became too attached to humans and were too eager to please, which could potentially lead to “experimental bias” (Kirk and Ramsden Reference Kirk and Ramsden2018, 24).

In the aforementioned lecture, Liddell explicitly conceived of the experimental situation as a “social process” that imposed conflicting demands on the animals (Liddell 1956). Through their experiments, he and his coworkers “were led to believe,” Liddell had already written in 1944, “that Pavlov’s classical method of the conditioned reflex was not an impersonal observational procedure, but a traumatizing procedure” (Liddell Reference Liddell and Hunt1944, 391; my emphasis). Accordingly, the restraint or inhibition demanded of the animals during experimentation, was not exclusively described in terms of nervous activity, learning, or stress – all idioms Liddell was conversant in as well. While he drew on the nineteenth-century physiologist Henry Head’s notion of vigilance to explain the animals’ mental exertion (Broadhurst Reference Broadhurst and Eysenck1960, 738), he also described his subjects’ labors as “self-imposed restraint” (Liddell 1956, 44). The confusing, monotonous, yet overstimulating experiments, so typical for the Pavlovian method, Liddell inferred, could strain the animals to the point where they would lose “emotional control” (ibid.).

Leighton, the anthropologist conscripted by Gantt, expanded the notion of conflict as a driving force behind the production of neurotic symptoms to the realm of culture: Within complex and rapidly changing societies, he deliberated, ambiguous or contradictory demands and conflicting patterns of behavior are as inevitable as they are potentially hazardous (Gantt Reference Gantt1944, 190).

Trauma as motivational conflict

The Chicago-based psychoanalyst Jules Masserman placed an equally strong emphasis on the role of conflict in his explanation of experimental neuroses. For Masserman, however, the site of the pathogenic conflict was an internal one. Although he had spent his early career working clinically under Meyer at the Phipps Clinic – also conducting neurophysiological experiments in one of its laboratories – Masserman began to embark on the study of experimental neuroses only a decade later, after his arrival in Chicago in the early 1940s. Like Thomas French and Franz Alexander, his colleagues at the Chicago Institute of Psychoanalysis, Masserman wished to expand psychoanalysis’s empirical basis beyond the clinical interview with findings from the laboratory-based sciences. The studies for which he would become best known were performed on cats, although he would later also use primates as well as dogs to confirm some of his earlier findings (see, e.g., Masserman and Pechtel 1953; Reference Masserman and Pechtel1956).

In Masserman’s laboratories at the University of Chicago (and later at Northwestern), cats were placed in a glass cage where the depressing of a lever would trigger a food dispensing mechanism. After they had learned to obtain food in this manner, the experimenters manipulated the situation to precipitate what Masserman believed represented a motivational conflict: Approaching the food tray, the hungry cats received either a moderately painful shock administered through the grid floor of the cage (a “grid-shock”) or an unpleasant air blast. What up to this point had been a positive signal, one that reliably announced the dispensing of food, had now also acquired a negative meaning. Symbolic ambiguity, Masserman posited in his monograph, Behavior and Neurosis. An Experimental Psychoanalytic Approach to Psychobiologic Principles, triggered the breakdown. With this turn to the symbolic dimension of animal life, he believed that he had departed from the Pavlovian vision, which he denounced as overly narrow and “mechanistic.” His own “biodynamic” approach, he asserted, forged the “comparative psychobiological” with the psychodynamic perspective (Masserman Reference Masserman1943, 1–21).

Like his colleagues, Masserman underscored the pathogenic qualities of the “physically harmless but fear engendering ‘emotional’ trauma” of the shock (ibid., 66). He also paid close attention to the social dynamics unfolding between “cage-mates” (e.g., ibid., chap. 9). Unlike the other researchers mentioned thus far, however, he was committed to a particular clinical-theoretical viewpoint. Accordingly, he placed a greater emphasis on relating laboratory observations to clinical techniques. He described, for instance, how some of his cats were cured from their resulting phobias by “transference therapy”; by placing the animal in the experimenter’s arm and petting them gently as they were being fed (ibid., 206).

The trauma of experiment

The laboratory facilities at the Phipps Clinic, the University of Chicago, and the Behavior Farm Laboratory in Ithaca evolved into sites of interdisciplinary research, and because of this were also conceived as social spaces – as sites of interspecies collaboration, confrontation, and conflict. The interdisciplinary approach shaped not just how neurosis was studied but as what it presented to researchers: The wide range of methods and methodologies they employed, their choices of laboratory animals, the collaborations they engaged in, and the consults they sought out both reflected and shaped their respective understandings of traumatizing laboratory conditions. These understandings varied considerably, yet they were all concerned with emotionality and the social conditions that would lead to its disturbance.

Arguing that a clash of conflicting social or motivational forces leads to neuroses, which highlighted that the latter were, indeed, not caused by a particular painful stimulus, carried with it conceptual tensions. Although the laboratory was conceived of as a social space, its hierarchical structure, as well as the traumatizing methods themselves, were never questioned. Indeed, if the laboratory were to provide a fractal image of society and its power asymmetries, the social arrangements that endowed the experimenter with power over his subjects had to remain uncontested. Moreover, the social, motivational, and biographical contexts, at times, threatened to obscure the very causal relationship the researchers hoped to demonstrate. Liddell’s case illustrates this perhaps most dramatically: The ethological perspective he adopted over the years let the identifiable external cause – which the notion of trauma today encapsulates – recede into the background. No longer did he believe that a violent shock or clash, as Pavlov had postulated, caused neurotic behavior: “The responses to … routine conditioned stimuli (food or shock) give us scarcely more insight into the consequences of a long-continued regimen of training in the laboratory,” he argued, “than do punctuation marks in helping us to understand the written page” (Liddell Reference Liddell and Hunt1944, 392). The strenuous conditions of the procedure and the experiment itself were deemed traumatic, not the shock. That Liddell held onto the notion of trauma despite the diminishing importance he attached to a specific traumatic episode only attests to the inevitability of the traumatogenic model of experimental neurosis. Moreover, it demonstrates the importance of observations made about the animal’s “normal” life as a domestic animal, which provided a contrast to the traumatic laboratory conditions. The epistemological status of trauma as an explanatory category began to erode, however, as psychologists came to dominate the experimental study of mental disorders.

Trauma as a theoretical entity

Differences in styles of reasoning, to speak with Ian Hacking (Hacking 1992; Reference Hacking2012), had already surfaced at the conference in Washington, DC. The neobehaviorist Hull, representing psychology’s theoretically most ambitious program, did not engage in “experimental explorations,” the style of reasoning embraced by Liddell, Gantt, and Masserman. Influenced by logical empiricism, Hull espoused the hypothetico-deductive method; his style of reasoning more closely resembled what Hacking termed “hypothetical construction of models by analogy” (ibid.). In his talk, Hull presented recent studies in which he had attempted to experimentally confirm hypotheses distilled from psychoanalytic theory. Expressing the interdisciplinary spirit of his home institution (Yale’s Institute of Human Relations), he urged psychoanalysts and learning theorists to work together – yet according to the rules of science set out by the logical empiricists (NRC 1938, 81; cf. Lemov Reference Lemov2005, chap. 4).

Skepticism of psychoanalysis’s explanatory merits among psychologists, then, was not what impeded the study of experimental neurosis. Initially, the bigger impediment was a recurring practical problem: It proved exceedingly difficult to lastingly induce neurotic behavior in the rat, most psychologists’ laboratory animal of choice. Not that they hadn’t tried, by some accounts successfully (Cook Reference Cook1939b; Peterman Reference Peterman1940). But when, in 1938, the rather spectacular symptoms reported by Maier were unmasked as artefacts, a backlash ensued.

By the 1930s, the albino rat had become a mainstay in psychological laboratories across the United States. Somewhat ironically, it was Adolf Meyer who first started a breeding colony of white rats at Worcester State Hospital in 1897 (Logan Reference Logan2005). Originally bred for the use in neurological experiments, Meyer considered the rat to be an ideal species because of its comparatively slow neural development – which allowed for more nuanced developmental studies – and because their intermediary position on the phylogenetic scale made them useful for comparative analyses. Experimental psychologists, instead, came to value rats because of their short gestation periods – which made the systematic study of potential genetic influences a realizable goal – and, as a handbook author notes in 1933, for the “absence of complicated habits whose history is unknown” (Munn 1933, xii, quoted in Logan 1999, 4; cf. Logan 2002; Reference Logan2005). In short, the mass production of standardized rats promised neurologists and psychologists an extension of experimental control over their model organisms. Standardization implied that laboratory workers would have to be less concerned with individual differences, their animals’ unique learning histories and idiosyncratic reaction patterns (cf. Logan 1999; Reference Logan2001).

In contrast, much of the dogs’ biographies prior to their time in the laboratory often remained unknown. Dogs were obtained from animal shelters or were sold to laboratories from suppliers whose practices, moreover, were often criticized in the popular press. The newspaper magnate William Randolph Hearst, for instance, known for his opposition to animal experimentation, ensured that several reporters covered stories supportive of the antivivisectionist cause. Stories of pets, lost or stolen, turning up in laboratories, frequently circulated in Hearst papers until the early 1950s (Lederer Reference Lederer1992, 63). Also, dogs in particular had become a focal point of the legal advances of the moral critics of animal experimentation. Between 1919 and 1946, numerous bills were introduced in legislatures on both the federal and state levels to abolish experimentation on dogs (ibid.). Although, these legislative efforts, like previous ones, were thwarted by a coordinated response by the medical establishment. Rats, mass-produced for laboratory research, were less likely to rouse similar opposition (cf. Rudacille Reference Rudacille2000).

What made rats an incomparably versatile and cost-efficient species for laboratory research, rendered them at the same time unsuitable for the experimental study of neuroses, however: Their resilience, the rat’s “capacity to recover quickly,” the psychologist Frank Finger wrote in Reference Finger and Hunt1944, could possibly “account for the difficulty of establishing ‘experimental neurosis’ in the white rat with the procedures ordinarily effective in other creatures.” In the years leading up to 1940, only very few of the many attempts to establish “relatively permanent and severe disorders” in the rat had been successful (Finger Reference Finger and Hunt1944, 414; cf. Hunt Reference Hunt1964). Discussing their failed attempts at the conference of the Eastern Psychological Association in 1940, the psychologists John McVicker Hunt and Harold Schlosberg speculated that, since its ability to engage in symbolic behaviors was underdeveloped compared to other species, the rat simply could not “carry its troubles about” (quoted in Finger Reference Finger and Hunt1944, 414; Hunt and Schlosberg Reference Hunt and Schlosberg1950).

In the experiments that roused so much controversy and would soon become a cautionary tale for experimental psychologists, Maier and his colleagues forced several brown, white, and black rats to jump from one platform to another, opposing platform, comprised of two marked windows, one of which opened onto a food reward while the other window was locked. After the rats had learned which marker indicated the reward, the experimenters made the problem unsolvable by assigning the markers randomly or by locking both windows, which, as they were forced to jump, caused the animals to fall from the stand. After several futile attempts, some of the rats became agitated, defensive, and exhibited symptoms of what appeared to be seizures. Then, immediately after having thrown such fits, they appeared unresponsive: Without any sign of struggle, the experimenters could now mold them into seemingly random positions. Maier and his coworkers at the University of Michigan captured these symptoms on film, and a monograph published the following year, in 1939, won him the Newcomb Cleveland Prize by the American Association for the Advancement of Science (Maier and Glaser Reference Maier and Glaser1938; Maier Reference Maier1939; see Dewsbury Reference Dewsbury1993).

Clifford Morgan, a physiological psychologist, led the charge against Maier. The seizures, he argued, had not been the result of a conflict leading to frustration, as Maier had claimed, but an artifact of his experimental design, an “audiogenic seizure” caused by the noise of the compressor used to generate the air blast that forced the animals to jump towards one of the two windows (Morgan and Waldman Reference Morgan and Waldman1941). Although a more detailed historical record points toward the possibility that Morgan’s victory may have been not as unambiguous and clear-cut as it is now remembered, Maier, upset by the hostile response, turned away from the study of experimental neurosis and embarked on what would become a successful career as a work and organizational psychologist (Dewsbury Reference Dewsbury1993). Audiogenic seizures, however, became a research field in its own right (see Finger Reference Finger and Hunt1944).

On the backdrop of the boundary disputes that had erupted in the “literature on rat convulsions,” the psychologist Donald Hebb, in an influential paper from 1947, set out to delimit neurotic from other pathological states (Hebb Reference Hebb1947, 8). Because “neurosis is anthropomorphic,” Hebb began by defining it for humans in a general way that could then be extended to nonhuman animals. Per Hebb, neurotic states are chronic, undesirable, and relatively rare “emotional” mental states that were generalized (i.e. not tied to specific situations) and “not due to a specific, gross neural lesion” (ibid., 9).

Hebb had done what others studying experimental neurosis had been reluctant to do: propose general criteria for neurotic conditions that supposedly transcend species differences. Moreover, he was skeptical of the idea that neurosis always “follows some theoretically ‘traumatic’ experience, such as conflict or frustration” (ibid.). Here, he also drew on his own research at the Yerkes Laboratory of Primate Biology in Florida. In the two cases of “spontaneous neurosis in chimpanzees” presented in the article, he could neither clearly determine when the symptoms started nor identify a specific episode that might have triggered them (ibid., 3). Hebb also outright questioned the possibility of inducing neuroses in rats, since their behavior was “still farther removed from man’s” than that of the “carnivores and ungulates” that had been studied in the laboratories in St. Petersburg, at Hopkins, Cornell and the University of Chicago (ibid., 14).

“Resulting symptoms”: Alternatives to neurosis

Psychologists working with rodents did not concede the field to primatologists, however. In a string of studies, the origins of anxiety were investigated by eliciting what became known as “conditioned emotional responses” (CER) in rats (see, e.g., Hunt and Brady Reference Hunt and Brady1951; Brady Reference Brady1951 add to reference list). This strand of research was shaped by two initially isolated experiments conducted by the radical behaviorists James Watson and B.F. Skinner. Watson famously conditioned “emotional reactions” in a six months-old infant (“little Albert”); and, nearly two decades later, Skinner had observed what he called the phenomenon of “conditioned suppression” in rats receiving electric shocks at a predetermined schedule (Watson and Rayner Reference Watson and Rayner1924; Estes and Skinner Reference Estes and Skinner1941).

The mounting skepticism among psychologists and how Hebb’s criteria were put to use are reflected in the two editions of the Handbook of Abnormal Psychology, published in 1960 and 1973, respectively. In the book’s first edition, Paul Broadhurst, while critically pointing out what he considered insufficient attempts at experimental standardization and a lack of quantitative data in earlier studies, still extensively reviewed the early literature on experimental neurosis. He praised Liddell’s team for their technical ingenuity, but also considered it “not surprising that no new principles emerge from the plethora of their observations” given “that the subjects were of different species, and different breeds within species” (Broadhurst Reference Broadhurst and Eysenck1960, 735). In the first edition of the handbook, Broadhurst came to the conclusion that Hebb’s criteria had not yet been met and that better data was needed. By the second edition, he had given up hope that neurotic behaviors truly deserving the name would ever see the light of day in an animal laboratory (Broadhurst Reference Broadhurst and Eysenck1973). Hoping that the experimental training of psychologists – as opposed to the broader training in medicine – would bring more methodological rigor to the field, he recommended that future researchers should “use more strictly controlled situations” as employed in the study of the CER (Broadhurst Reference Broadhurst and Eysenck1960).

Broadhurst, like other experimental psychologists, had become enthralled by the prospect that a “general learning theory” would explain all experimental observations through the application of a coherent system of behavioral laws (ibid.). The work that epitomized this project was Hull’s Principles of Behavior, published in 1943 (Hull Reference Hull1943). The psychological laws Hull strived to uncover, whenever possible in mathematical form, should be applicable to every form of behavior, “individual and social, moral and immoral, normal and psychopathic,” and as such provide a basis for “all the behavioral (social) sciences” (ibid., v). After his death in 1952, this theoretical edifice was further developed by his students, especially Kenneth Spence. According to Hull and his students, the basic laws of behavior should be extended not only to every form of behavior but to all mammals. “The distance from rat to man is a long one,” the psychologist William Schoenfeld summed up this prevailing attitude, “yet it may turn out to be filled with differences in detail rather than principle” (Schoenfeld Reference Schoenfeld, Hoch and Zubin1950, 95; quoted in Broadhurst Reference Broadhurst and Eysenck1960, 755). To uncover the basic mechanisms of learning, in other words, the exploration of species differences was no longer deemed necessary.

Trauma as experimental technology and theoretical impossibility

The quest for a general learning theory had immediate theoretical implications for the study of experimental neurosis. This is best illustrated by the work of a late admirer of Hull’s, the psychiatrist Joseph Wolpe, who became the fiercest critic of the first generation of US-American Pavlovians and an influential champion of behavior therapy. Not only did Wolpe reject the idea that neurotic behavior could be traced back to the experience of a traumatic conflict, he questioned, more broadly, the disease model that tacitly formed the basis of previous understandings of experimental neurosis (cf. Ruiz and Sanchez Reference Ruiz and Sánchez2012). In his writings he poignantly rejected Pavlov’s assumption that a “break,” a discontinuity of some kind, must cause neurotic behaviors: “The theories” offered by Liddell, Maier, Masserman and others, he wrote in his monograph Psychotherapy by Reciprocal Inhibition from 1958, “differ widely in their details but each one presumes that damage or disruption of some kind underlies the neurosis, and not ‘mere’ learning” (Wolpe Reference Wolpe1958, 67). Still, Wolpe at no point questioned the potential sequelae of traumatic experiences. After all, it had been his treatment of traumatized soldiers during the war that had led also him to the study of experimental neurosis in the first place.

Wolpe was trained as a physician in South-Africa and had completed a doctoral dissertation devoted to experimental neurosis at the University of Witwatersrand in 1948, although he would spend the major part of his career in the United States, accepting a position at Temple University in 1960. By his own account, Wolpe, once devoted to psychoanalysis, had become increasingly doubtful of the central tenets of Freudian theory and the effectiveness of its therapeutic measures during his service in the South African Medical Corps in World War II (Wolpe Reference Wolpe1958). In search of theoretical alternatives to Freudianism, he turned first to the works of Pavlov, only to grow skeptical of his physiological assumptions, eventually rejecting them in favor of Hull’s formalistic theory (Poppen Reference Poppen and O’ Donahue2001, 40).

For his doctoral work in South Africa, Wolpe adopted Masserman’s procedure to neuroticize cats, although he used stronger electrical currents than his colleague in Chicago. The currents were raised deliberately to put Masserman’s as well as Maier’s explanations to the test (Wolpe Reference Wolpe1952; 1953). Wolpe believed to have demonstrated that it was the fear-engendering, painful stimulus alone, the shock administered in the laboratory, not a motivational conflict or frustration that had led to avoidance behavior. The experience of the shock, he further argued, did not entail a potentially damaging event. As Hull had suggested, the same learning principles that account for normal behavior could now also be drawn on to explain abnormal behavior.

The shock, while no longer being thought of as a “break,” nonetheless represented an extreme event: Although Wolpe maintained that, ultimately, all neuroses are caused by fear – essentially equating them with phobias – he considered neurotic reactions of the type produced in his cats to be rare in humans. Only among concentration camp survivors, he surmised, does one witness cases that had been caused by a single traumatic event (Wolpe Reference Wolpe1958, x). He thus harbored similar suspicions as Liddell with respect to the traumatizing nature of his experiments – going further even, by drawing a parallel between his feline subjects and Holocaust survivors. Yet Wolpe no longer deemed the social context of experimentation and the animals’ perceptions epistemically relevant. The painful stimulus became the causal agent that displaced the empirically more ambiguous traumatic laboratory situation, which his colleagues aimed to explore. In Wolpe’s laboratory, instead, the “trauma of experiment” was reduced to the shock that was deployed to debunk their competing psychoanalytically-inspired explanations.

Abnormal behavior under normalized conditions: From disease to dysfunction

Wolpe may have been its most ardent critic, but he was not alone in questioning the disease model of neurosis. In a 1960 Science article titled “Normal sources of pathological behavior,” the psychologist Murray Sidman, detailing the results of experiments modelled on Skinner’s conditioned suppression paradigm, made the case to a broader audience that there was nothing “disorderly” about abnormal behavior. Only recently and gradually have experimental and clinical psychologists begun to realize, he wrote, that pathological behavior can result “from quantitative and qualitative combinations of processes which are themselves intrinsically orderly, strictly determined, and normal in origin” (Sidman Reference Sidman1960, 61); researchers would come to see the “lawfulness” of pathological behavior once they had discerned the crucial variables at play. But if aberrant behavior was indeed predictable, what distinguished it from normal behavior?

In a paper devoted to “problems in the interpretation of ‘experimental neurosis’” from 1964, the psychologist Howard Hunt piercingly questioned conventional definitions of the slippery terms “abnormal” and “maladaptive”: To assess a behavior as maladaptive presupposes a telos, he critically remarked, a predetermined goal an organism must strive toward. What constituted such a goal, however, “may depend almost entirely on the point of view—S[ubject]’s or E[xperimenter]’s” (Hunt Reference Hunt1964, 29).

Previously, the notion of adaptation had been readily applied to the experimental setting itself: An animal’s breakdown under the demands of an experiment signaled pathology. Now, such a presupposition had become suspect. The skepticism of Hunt and others must be understood against the backdrop of a wider reckoning with comparative psychology, gaining momentum in the 1960s and 70s (see also Beach Reference Beach1950).

The discipline’s origins in Darwinism notwithstanding, comparative psychologists had been oddly out of step with more recent developments in biology. The so-called modern synthesis of Mendelian genetics and Darwinian evolutionary theory, which evolved in the 1930s and 40s, had yet to be applied to the evolution of behavioral traits. By the 1960s, psychologists had abandoned the outdated notion of the scala naturae and its static, hierarchical structure, which gave way to the appreciation of the differences between evolutionary lines and the unique selection pressures shaping them (see, e.g., Hodos and Campbell Reference Hodos and Campbell1969). In the following years, laboratory scientists also took note of the work of ethologists who, in the tradition of Nikolaas Tinbergen and Konrad Lorenz, studied instinctive animal behaviors, often in their natural environments. In short, when thinking about adaptation, learning theorists became reluctant to apply it to an individual’s behavior; instead, their attention began to shift towards specific behavioral patterns and their relation to natural environments.

Hunt, who had made significant contributions to the CER literature in the previous decade, now questioned the “episodic model” he had taken for granted. Tellingly, his critique took an ecological turn: Whether an animal showed abnormal behavior in the laboratory or not could only be determined, he argued, if one takes into account what was known about the behavior of the species in its natural environment; he also pondered the question whether the environment had a “cumulative” rather than an “episodic” effect on behavior (Hunt Reference Hunt1964, 33). In this context, he pointed to ethological research like John Calhoun’s much publicized studies of “behavioral sinks,” social pathologies observed among rats and mice in crowded habitats (Calhoun Reference Calhoun1962; see also Ramsden Reference Ramsden2011).

While theoretical uncertainty concerning what should be regarded as abnormal behavior abounded, the question of what constituted a normal specimen or subject had been contained. The standardization of laboratory animals and the use of inferential statistics permitted psychologists to expand their control over laboratory animals and to assess the generalizability of experimental results. Moreover, as Broadhurst could now point out in the second edition of the Handbook of Abnormal Psychology, the combination of factor analytic methods with selective breeding promised the modeling of behavioral syndromes that “mimic those associated with patients” by “literally creat[ing] laboratory populations extreme in this respect” (Broadhurst Reference Broadhurst and Eysenck1973, 728). Not ethology nor the ecological perspective would shape the future of experimental psychopathological research, but the spirit of what the historian of science Philip Pauly has described as the “engineering ideal in biology” (Pauly Reference Pauly1987; cf. Lemov Reference Lemov2005).

Traumatic learning

With the rejection of the disease model of neurosis came the fragmentation of the experimental study of psychological disorders. The exploration of the evolution of neurotic behaviors and their precipitating conflicts was displaced by the study of animal analogues of more delineated syndromes such as behavioral fixations, aggression, addiction, and, most prominently, depression (see Broadhurst Reference Broadhurst and Eysenck1973). Concurrently, as evolutionary theory could no longer provide a unifying perspective for the study of animal behavior and threatened to undermine the hierarchical structure of a system of behavioral laws, the prospect of a general learning theory no longer seemed attainable, or alluring, to most experimental psychologists. Laboratory workers turned to the study of “behavioral phenomena” generated in animal laboratories and the intricate mechanisms they supposedly allowed to investigate, without attempting to integrate their findings into one coherent explanatory system (Amsel and Rashotte Reference Amsel, Rashotte and Zeier1977, 149). One such phenomenon was known as “traumatic avoidance learning” and it was investigated by Richard Solomon and his group of students, also dubbed the “Neo-Pavlovians,” first at Harvard and later at the University of Pennsylvania (ibid., 147).

The persistence with which laboratory animals showed avoidance behaviors in response to a danger signal had long posed a theoretical challenge for experimental psychologists. Trying to explain fear without relying on “mentalist” notions, such as remembering, propelled behaviorists into the theoretical deep end. While Hull, in one of his earlier publications, offered an explanation of how organisms learn to respond to a signal announcing a shock, he was unable to resolve the theoretical “dilemma of the conditioned defense reaction” (Hull Reference Hull1929, 510): Why do fear reactions persist, so the riddle behaviorist struggled with, even if the pain is no longer experienced as the organism learns to avoid its source? In 1939, the psychologist Orval Mowrer had attempted to resolve the issue by reformulating Freud’s second theory of anxiety in behaviorist terms and positing that a state of fear or anxiety – unlike Freud, he used the terms interchangeably – was central to understanding the persistence of avoidance reactions. In a paper presented at the Institute of Human Relations, Mowrer also posited that “infra-human organisms” – he most likely had the rat in mind – will give up the fearful response quickly once the signal is no longer followed by a “so-called ‘traumatic’ (painful) stimulus.” Humans, on the other hand, tend to be “less realistic,” as they are more likely to engage in efforts to reduce “emotional discomfort” by learning to avoid it, which, paradoxically, makes them more susceptible to enduring fears (Mowrer Reference Mowrer1939, 558). As in Wolpe’s, also in the context of this interpretation of anxiety the word “traumatic” denoted the experience of pain.

A persistent empirical challenge to this, as it came to be known, two-factor theory of anxiety (see Mowrer Reference Mowrer1940; Miller Reference Miller1948; Reference Miller and Stevens1951) was the observation that many animals, not just humans, continued to react with fear to danger signals even if they were no longer followed by painful stimuli (see Eysenck Reference Eysenck1979). In a 1954 paper on “traumatic avoidance learning,” Solomon and his co-author, the psychiatrist Lyman Wynne,Footnote ¹¹ pointed to the evidence, including the research undertaken by Liddell, Gantt, and Masserman that suggested “a tremendous qualitative difference and quantitative difference between appetitive and traumatic learning” (Solomon and Wynne Reference Solomon and Wynne1954, 368).

At Harvard’s Laboratory of Social Relations, Solomon began investigating “traumatic learning” by placing his subjects, usually dogs, in a so-called shuttle box. Placed inside this confinement, the dogs could escape a grid-shock by jumping over a barrier or passing through a doorway in the middle of the enclosure to the adjacent compartment where they were safe from harm. Theoretically, Solomon and Wynne explained the persistence of some avoidance behaviors by postulating the existence of a class of powerful stimuli causing “’traumatic’ or very intense ‘pain-fear’ reaction[s]” that the organism would always, also without re-exposure, try to avoid (ibid., 361). “Certainly,” they admitted, “this general notion is not original,” citing also works by Freud, Hull, Robert White, Gordon Allport, as well as Selye’s general adaptation syndrome (ibid., 362). Not surprisingly, Wolpe decried what Solomon and Wynne called the “principle of partial irreversibility” as merely another “variant of lesion theory” (Wolpe 1956, 923).

Solomon’s subsequent work at the University of Pennsylvania, where he was appointed professor in 1960, moved him still further away from an integrated, general theory of learning. In a series of experiments he designed with Lucy Turner, he could demonstrate that instrumental learning and Pavlovian or “classical” conditioning could be distinguished experimentally and that dogs would “transfer” what they learned by association to instrumental learning conditions (Solomon and Turner Reference Solomon and Turner1962). These experiments were to evolve into the study of a behavioral syndrome labeled “learned helplessness.”

Trauma as loss of control

As researchers in Solomon’s lab applied the “transfer paradigm,” they compared the behaviors of immobilized dogs that had to endure shocks with a control group that had been given the possibility to avoid them (cf. Rescorla Reference Rescorla1995, 305). By 1964, Bruce Overmier systematically studied the effect of the “degree of control” subjects were given over the shocks on the onset of behavioral disturbances (Overmier and Seligman Reference Overmier and Seligman1967). Soon upon his arrival at the University of Pennsylvania, Seligman, together with a fellow graduate student, Steven Maier, attended some of the trials. Intrigued and, purportedly, saddened by the dogs’ suffering, Seligman decided to make the passivity he witnessed in the dogs, which he soon likened to depression in humans, the topic of his doctoral research (Seligman Reference Seligman2018).

In their paper “Failure to Escape Traumatic Shock” from 1967, Seligman and Maier described the paradigmatic experiment upon which a series of experimental protocols involving a diverse range of subjects, animal and human, would later be modelled: While being suspended in a Pavlovian harness, one group of dogs was given the possibility to terminate a series of shocks – which, as Seligman would later insist, produced a “sudden startle but no pain” (Seligman Reference Seligman2018, 77–8) – by pressing their heads against a panel, while another group of subjects was immobilized and had to endure shocks of the same duration and frequency. Twenty-four hours later, dogs of both groups were placed in a shuttle box and their ability to learn how to avoid shocks in this new environment was tested. The subjects that had been given the possibility to terminate the shocks by pressing the panel earlier reliably learned to escape, whereas the previously immobilized dogs failed to do so in two thirds of the cases (Seligman and Maier Reference Seligman and Maier1967).

Seligman and Maier’s explanation of this phenomenon both built on and expanded the conventional understanding of associative learning: Besides contiguity, they argued, “independence between events” or a lack of contiguity could lead to learning, namely that instrumental actions were futile (ibid., 8). The expectation of uncontrollability, they further asserted, would transfer onto other conditions and lead to passivity, withdrawal, and diminish the ability to learn avoidance behaviors in novel situations.

Seligman showed a keen interest in the work of the field’s pioneers. During his doctorate he came into contact with Wolpe, attending the monthly salons he organized at his home in a Philadelphia suburb where behavior therapists would gather to discuss recent trends in a colloquial setting. Also, immediately after his doctorate, he visited the University of Michigan to talk to Norman Maier, one of his self-professed “heroes.” Maier discouraged him from accepting a position there; and Seligman went on to Cornell instead (Seligman Reference Seligman2018, 89). Three years later, he returned to the University of Pennsylvania where he completed a two-year clinical training before becoming tenured faculty in the university’s psychology department (ibid.). Influenced by one of his clinical supervisors, Aaron Beck, a former psychoanalyst, often regarded as the father of cognitive behavioral therapy, Seligman began to develop a highly influential theory of depression that associated the experience of uncontrollable events with a loss of the “motivation to act” in the depressive state (Seligman Reference Seligman1975).

In 1975, Seligman published his first trade book, Helplessness: On Depression, Development and Death, which offered a compilation of his by now decade-long research, interwoven with clinical observations and his takes on current social issues. He emphasized that it was not the “shock” or the “trauma in itself” that produces behavioral disturbances but the “inability to control the shock”; “learning that no response at all … can control trauma” (Seligman Reference Seligman1975, 26–7). The idea that a lack of belief in control leads to depression was also compatible with a strand of clinical theorizing that stressed the significance of feelings of helplessness or, as psychotherapists more often called it, hopelessness, in their patients (see, e.g., Miller et al. Reference Miller, Rosellini, Seligman, Masser and Seligman1977, 122).

As Seligman conducted his experimental work, increasingly less beholden to behaviorism and its theoretical conundrums, attitudes towards animal experimentation had clearly begun to shift. In the 1960s and 70s, the public again became more receptive to the arguments of those challenging scientific and moral justifications of experimenting on animals. In February 1966, a photo essay in Life magazine entitled “Concentration Camps for Dogs” portrayed the appalling conditions at a dog dealer’s farm in rural Maryland (Wayman Reference Wayman1966). This widely circulated article together with other media accounts of malicious dognapping, carried out to meet the growing demands of medical researchers, led to the passage of the Laboratory Animal Welfare Act in the same year. Later amendments to the law would more tightly regulate animal experimentation on university campuses (cf. Guerrini Reference Guerrini2003). In the following decade, philosophical critiques of animal experimentation emerged, often incorporating sociobiological principles (Midgley Reference Midgley1973; Clark Reference Clark1977; Regan Reference Regan, Miller and Williams1983). In the most influential of these works, Animal Liberation by the Australian philosopher Peter Singer (Reference Singer1975), learned helplessness experiments were directly attacked as ethically dubious; as lacking any discernable benefit to humans that would justify the animals’ suffering. Inflicting pain on other animals or making them otherwise suffer merely to advance a theoretical program – especially, one might add, after said program had largely been abandoned – was deemed indefensible. In hindsight, Seligman would later be compelled to partially agree with Singer. Much of the basic experimental science practiced by him and his contemporaries, he recently acknowledged, rarely reached “external validity” and often failed to turn up the promised real-world applications (Seligman Reference Seligman2018, 77–8). Also, within a number of disciplines, researchers began to raise the question whether and how findings from animal experiments can effectively be applied and “translated” to humans; animals, it turned out, often are insufficient models for humans (cf. Rose and Abi-Rached Reference Rose and Abi-Rached2013, chap. 3).

As with many other psychological experiments, the epistemic significance of the learned helplessness experiments did not consist in spurring therapeutic innovations. Rather, they placed human experiences under new descriptions, scientifically sanctioning forms of subjectivity. The experimental study of learned helplessness further solidified the idea that traumatic events were uncontrollable events. In 1978, John Kihlstrom and Susan Mineka, a former student of Seligman’s, undertook a re-evaluation of the literature on experimental neurosis in light of the more recent experimental work. What had “sometimes been labeled experimental neuroses,” they assured, can readily be explained in terms of the effects of uncontrollable or – as earlier experiments by Robert Lockard and others had shown – unpredictable “life events” (Mineka and Kihlstrom Reference Mineka and Kihlstrom1978, 266). One year later, the clinical psychologist Leonore Walker drew on Seligman’s model in The Battered Woman, a widely read book detailing the emotional toll of spousal abuse (Walker Reference Walker1979). Finally, a decade later, Edna Foa, in developing her influential cognitive-behavioral model of PTSD, would appropriate the experimental neurosis literature as it had been passed down to her by Kihlstrom and Mineka (see, e.g., Foa, Steketee, and Olasov Rothbaum Reference Foa, Steketee and Rothbaum1989).

Seligman himself loosely applied his insights to a number of social ills that came to the attention of the general public in the 1960s and 70s. In his 1975 monograph, he discussed the negative motivational effects of “maternal deprivation,” “helplessness in the classroom,” and poverty (Seligman Reference Seligman1975, chap. 7). While the experience of poverty itself may lead to a belief in a lack of control, the “welfare system, however well intentioned” along with the persistent problem of “crowding” in urban centers may only exacerbate the issue (ibid., 161), making the inoculated belief in the lack of control a “central problem of race and poverty today” (ibid., 164). By pointing to the presumably hazardous motivational effects of governmental assistance, Seligman echoed the neoconservative criticisms gaining currency at the time. Clearly, the syndrome he had identified hinted at a different kind of solution to the social problems of his day than the remedies offered by American postwar liberalism. In the introduction to a later book from 1993, Learned Helplessness: A Theory for the Age of Personal Control, Seligman and his co-authors noted that, in the previous two decades, a fundamental shift had taken place in American society: After the assassinations of political leaders in the 1960s and a growing distrust in the effectiveness of governmental institutions – that supposedly undermined the belief in control through collective action – the authors observed a resurgence of the value of controlling one’s destiny (Peterson, Maier, and Seligman Reference Peterson, Maier and Seligman1993). The underlying tenet of this and many of Seligman’s later best-selling books about how to expand personal control was that, because the negative effects of uncontrollable events were due not to the experience itself but to generalized assumptions about the futility of one’s actions, a sense of loss of control could be counteracted through the stringent application of psychological techniques.Footnote ¹² Learned Optimism was the title of another one of Seligman’s books that skillfully combined popular science with self-help advice (Seligman Reference Seligman2006).

In a later application of the learned helplessness model to posttraumatic stress, the clinical psychologist Raymond Flannery developed a “stress management approach” designed to help victims of trauma to become “survivors” (Flannery Reference Flannery and van der Kolk1987). The popular trauma narrative that was to evolve over the years and which came to include both victimization and its opposite, the overcoming of obstacles and the regaining of control, appears as the elaboration of a theme that had already been articulated through the study of learned helplessness: Helplessness could be unlearned, and control over one’s environment, by implication, the ulterior goal of behavior, regained.