‘Ghastly marionettes’ and the political metaphysics of cognitive liberalism: Anti-behaviourism,language,and the origins of totalitarianism

Arendt and the psychology of totalitarianism

War is often boon to the sciences of the psyche. As historians of science have demonstrated, psychiatry was epistemologically and institutionally cemented at the turn of the century, owing much to its use in the First World War (Grob, 1994; O’Donnell, 1985; Pressman, 2002). Psychology then proved itself in the US as a nationally useful discipline during the Second World War, offering the military a means of evaluating the intelligence of conscripts, measuring psychological constructs such as ‘morale’, and understanding the interiority of the enemy (Herman, 1995; Morawski, 1988). The American postwar infatuation with the sciences extended to psychology, where behaviourism reigned as the supreme (if contested) hegemon whose aura of scientific bona fides motivated the rapid incorporation of its approach in the social sciences (Lemov, 2005; Samelson, 1981). Although the full extent of the Nazi death camps was as yet unrevealed, both US and immigrated European psychologists and social scientists inaugurated a scholarly cottage industry, using their new-found status as experts to offer explanations of the psychology of Nazism, the death camps, and the newly coined concept of totalitarianism (Arendt, 1950; Baehr, 2010; Miller, 1986).

Arendt intervened in this discourse by resisting the very premise of the question: For her, there could be no ‘psychology of totalitarianism’, precisely because the operation of totalitarianism was to destroy the psyche. The inverse of the psychology of the free individual was not, for Arendt, the psychology of totalitarianism. Instead, Arendt saw totalitarianism as the eradication of the psyche itself, mechanizing the human and stripping it of spontaneity:

We attempt to understand the behavior of concentration-camp inmates and SS-men psychologically, when the very thing that must be realized is that the psyche can be destroyed even without the destruction of the physical man; that, indeed, psyche, character, and individuality seem under certain circumstances to express themselves only through the rapidity or slowness with which they disintegrate. (Arendt, 1973: 441)

The role of language in cognitive liberalism

Arendt was not a cognitivist, nor could she have been one. Yet, I suggest, the contours of her thought demonstrate the conditions of possibility of both the cybernetic turn and the ultimate ascendance of cognitivism over behaviourism. A close reading of Arendt’s prioritization of language in the final section of The Origins of Totalitarianism yields the shape of the contestation between mechanistic and transcendental conceptions of the human in the postwar period, a struggle that was parlayed into an epistemic-political proxy war against behaviourism. The contours of the spontaneity/mechanism binary in which Arendt’s account of language took shape illuminate why Cold War liberalism found the cybernetic escape from this impasse so appealing. Yet they also demonstrate why the cybernetic manoeuvre ultimately foundered on the rocks of an internal incoherency around language, its failure to unmoor itself from behaviourism’s metaphysics spelling its ultimate subsumption by cognitivism. In short, it is no accident that the defeat of behaviourism by cognitivism – a cognitivism that emerged dialectically through the cybernetic mediation of the mechanism/spontaneity impasse – came precisely through the aperture on which Arendt both hinged her critique of totalitarianism and reintroduced a transcendental metaphysics of spontaneity: language.

For Arendt (1973: 458), language was directly implicated in the ability to think, by which she meant specifically the capacity to think creatively. In order to be able to truly think, the individual had to be present to itself in an inner dialogue mediated by language: ‘All thinking, strictly speaking, is done in solitude and is a dialogue between me and myself’ (ibid.: 476). It was only on the condition that this inner dialogue was present that the individual might escape from loneliness. A technical term, loneliness described the condition of being unable to dialogue with others as a result of an inability to dialogue with oneself. Anything that could not be articulated explicitly – anything like sensation, Umwelt, or affect – played no part in psychic life proper. Dialogue with others and oneself refused the identity that made them speak with ‘the single voice of one unexchangeable person’ (ibid.).

Despite her critique of neo-Kantianism (Keedus, 2012), Arendt’s identification of spontaneity as the core attribute of human freedom retained a fixation on the possibility of undetermined ethico-political action derived from Kant. ² This philosophical manoeuvre prioritized subjective human operations such as cognition, feeling, and language over physical operations; the human began precisely where mechanized nature ended. This emphasis on language as coterminous with thought, subjectivity, and political action inherited problematics that Karl Jaspers had used to reckon with the legacy of both Kant and phenomenology. As Arendt recounted in ‘What is Existential Philosophy?’ (originally published in 1948), Jaspers had been trying to bridge two seemingly conflicting commitments. He had declined to fall in line with what she diagnosed as the characteristic tendency of modern philosophy (whether pragmatism or phenomenology), namely ‘giving up Kant’s basic concept of human freedom and dignity’. Simultaneously, he had resisted Kant’s violent extraction of man from the world, by which ‘at the same time that Kant made man the master and the measure of man, he also made him the slave of Being’ (Arendt, 2005: 170). The self-evident path out of this Kantian antinomy between human and world had been phenomenology, which had swept the world into the capsule of human apprehension, accentuating the ‘phenomenally verifiable content of an event’ rather than its genesis in a chain of natural causality (ibid.). But for Jaspers, the phenomenologists’ approach had overemphasized feeling at the expense of thinking. Jaspers’ attempt to split the difference between reuniting man and world, on the one hand, and preserving a sphere of spontaneous action, on the other, had produced what we might call a cognitive phenomenology, in which the human could enact its free spontaneity only through the freedom of thought, always mediated through communication. In Arendt’s gloss, ‘in the concept of communication lies a concept of humanity new in its approach though not yet fully developed that postulates communication as the premise for the existence of man’ (ibid.: 186). Arendt inherited Jaspers’ emphasis on communication and freedom into a ‘new approach’ that would become a cognitive liberalism rooted in language, designed to restitch the subject into the world while maintaining a space of spontaneous action.

The reticulated quality of language and thought for Arendt was exemplified in her later depiction of Eichmann, whom she figured as an exemplar of the individual who had been rendered unable to think by virtue of a compromised quality of language. Arendt noted Eichmann’s inability to respond to the prosecutor’s questions in anything but rote, evacuated words: He repeated ‘word for word the same stock phrases and self-invented clichés’ (Arendt, 1964: 49). Eichmann’s other defining characteristic – his inability to ‘look at anything from another’s point of view’ (ibid.: 48) – was not simply an additional failing, but was integrally connected to the vapid, robotic quality of his speech. ³ If for Arendt all thought occurred in language, then the potential for creative thought had to reside in the possibility to re-signify language, to use it inventively, to make it say something new. Eichmann’s desiccated language indicated that he was able to think only within the cramped terms of a signification that had been dictated to him. The eradication of his internal duality was guaranteed by his dead language.

The political enemy of totalitarian government, then, was not an ‘external enemy’ in the sense of individuals within the population who were actively fomenting dissent. Rather, its true foe was the ‘internal enemy’ found in this internal duality and unpredictability of the authentically thinking person. In this sense, Arendt’s totalitarian government’s relation to the political subject mirrors what Peter Galison has called cybernetics’ ‘ontology of the enemy’: What constitutes the threat is the opacity of the enemy’s mind and the difficulty in predicting his next move (Galison, 1994). Arendt (1973: 421) famously drew the distinction between dictatorship and totalitarianism by specifying that while dictatorship located its enemy in the acts of subversives through political domination, the enemy of totalitarianism was interiority itself. Thus, the political mode of totalitarianism was terror, the sort of political repression that began where any real external political threat ended. The object of terror was the elimination of the internal enemy, the human that conversed with itself and consequently posed the threat of introducing something new into the world: ‘Simply because of their capacity to think, human beings are suspects by definition, and this suspicion cannot be diverted by exemplary behavior, for the human capacity to think is also a capacity to change one’s mind’ (ibid.: 430). Thus, the ultimate political threat to totalitarianism was not any particular thought, but thought in its very form. Because the ability to think meant the power to creatively introduce something unpredictable and spontaneous into the world, the ‘internal’ enemy of the auto-dialogic individual needed to be eradicated, and with it, the essential capacity of language to mediate thought that could not be predicted determinatively (ibid.: 341, 464). This is not to say that for Arendt, language in the totalitarian universe did not play a central role. It remained a crucial mode of creating internality and exercising power, but its mode and effectivity changed entirely: Its operations were transformed from potentiating to eradicating politics (or, for her, thought specifically as the capacity for undetermined action). ⁴ If in a free society language created subjects, in totalitarianism language created a reality of which subjects were merely the effects.

Arendt was clear on the fundamental quality of the reality totalitarianism sought to create: consistency. Before a totalitarian movement seized power, its communication to the masses took the form of propaganda that exhibited ‘an extreme contempt for the facts as such, or even for invented facts’ (Arendt, 1973: 351). After gaining power, the totalitarian leader began the task of remaking the world in the image of the fictional world they had posited (ibid.: 349). Totalitarian language prophesied an entirely consistent reality into existence, because ‘what convinces the masses are not facts…but only the consistency of the system’ (ibid.: 351). The balm totalitarianism offered was that of a consistent fictionalized world in which the possibility of a disruptive event was eternally foreclosed: ‘Totalitarian movements conjure up a lying world of consistence which is more adequate to the needs of the human mind than reality itself’ (ibid.). In order to achieve this, totalitarianism needed to castrate the power of a word to participate in the spontaneous occurrence of a new thought, and, by implication, a spontaneous political action (ibid.: 458). All language had to bend to the task of flattening reality into such homogeneity; the possibility of any dialectical synthesis – any emergence of an unpredictable difference out of contradiction (ibid.: 470) – was utterly foreclosed. In place of dialectics, totalitarianism offered deduction.

As Arendt (1973: 462) repeatedly underlined, the creation of an utterly consistent world sealed from the possibility of dialectical emergence was the actualization of a fiction. The world produced by totalitarianism’s instrumentalization of language was marked by an absence of the kind of contradiction inherent to a heterogeneous world in which spontaneous events could occur. The impossible could become possible, but not spontaneously; the impossible was achieved not through human striving, but because the fiction of totalitarianism posited the impossible as the necessary outcome of deterministic Natural Laws (either the laws of Nature for Nazism, or of History for Stalinism). Here, we can observe the contours of a curious paradox at the heart of Arendt’s account. If, following the binary of determination or freedom erected by Kant, postwar political theory maintained that language/cognition/subjectivity had to be undetermined in order to register in the ethico-political sphere of ‘the human’, in Arendt’s totalitarianism we see the opposite: the world (rather than the sign) became arbitrary to the extent that language itself became determined and mechanical. ⁵

In this subtle chiasmatic reversal, the stakes of the cybernetic synthesis of mechanism and spontaneity for Cold War liberals come into relief. In the wake of the revival of metaphysics that marked the critique of early 20th-century politics, intellectuals like Arendt relied on a binary opposition between the ‘ethical’ realm of human subjectivity (including language, the psyche, and cognition) framed as uncaptured by mechanistic determination; and the realm of the natural universe (including physical causality and deterministic causation) comprising the zone of the animal, the natural, and the machine. ^6,7 In a free society (Arendt and Kohn, 2006: 169), the human realm was non-utilitarian (engaging in ‘work’ in addition to ‘labor’, to cite only one instance of this contradistinction in Arendt’s thought), while the realm of physical reality remained bound by the laws of utilitarian, deterministic causation. Yet totalitarianism inverted this relationship: the human became incapable of acting out of non-utilitarian concerns, while reality became marked by non-utilitarian concerns. In an inversion of Kantian ethics, the human became unable to behave ethically (that is, for the sake of an ‘unmotivated’ end), and the ethical function was taken up by totalitarian reality itself. Arendt (1973: 322) adduced numerous examples of this thing-ification of the human: Himmler’s men were created so as to ‘never do a thing for its own sake’; Stalin recognized that even a seemingly unthreatening activity like enjoying chess as a hobby posed a threat because it was done for its own sake (ibid.); political work was all flattened out into careerist labour as the Nazi work of killing became bureaucratized. As the ultimate embodiment of this process, the death camps created ‘ghastly marionettes with human faces, which all behave like the dog of Pavlov’s experiments’ (ibid.: 455), as humans were reduced to a ‘bundle of reactions to the environment’ (ibid.: 441).

Yet in Arendtian totalitarianism, even as the realm of the subject (that is, the political) was encased within deterministic causality, reality itself became liberated from utilitarian concerns to the benefit of the imposed fiction. Arendt (1973: 444) noted the grotesquely non-utilitarian nature of the death camps, in which labour was extracted from inmates for no reason other than to increase their misery: ‘The concentration camps existed for their own sake’, and ‘as an institution was not established for the sake of any possible labor yield’. ⁸ Similarly, the periodic purges of the population served no economic or political utility other than the fictionalization of reality into a zone in which ‘anything is possible’ (ibid.: 440). Reality itself took on an ‘ethical’ quality as the realization of suprahuman laws operated beyond the checks of self-interest or utilitarian motives. In its full realization, the totalitarian government no longer operated on the principle that ‘the ends justify the means’, because such logic limited the potentiality of this ‘reality’. While in the zone of ‘the political’ proper it was human subjectivity that was arbitrary (in the sense that it was capable of generating an undetermined event), in totalitarianism it was reality itself that became arbitrary, capable of being remade instantaneously (ibid.: 425). ⁹ Here, then, is the point that cannot be missed: reality could take on this liberated quality in totalitarianism only insofar as language took on a deterministic function, thereby collapsing the individual into mechanistic bodily behaviour. In sum, for Arendt, a totalitarian project’s success hinged on the mechanization of language as a way of effecting the mechanization of the human. Arendt’s account of totalitarianism as a behaviourist laboratory was not a metaphor; her claim was that totalitarianism and behaviourism were both engaged in a functionally contiguous undertaking by which human beings were reduced to the status of roboticized animals.

‘Machine is a great and good god; praise be to Machine!’

With this understanding of the position of language in Arendt’s figuration of the political in place, we can turn to contextualizing Arendt’s account within the problem space of post-Weimar political theory as it attempted to salvage the possibility of politics from the wreckage of early 20th-century liberalism. By the end of the mid-century war, it was clear that liberalism would have to perform an internal pivot in order to purge itself of the deficiencies that had resulted in its apocalyptic failure earlier in the century (Luebbert, 1991; Mazower, 2009: 30–40). These frailties were diagnosed by two intimately related lines of critique. The first, mobilized between the turn of the century and the mid-1930s by the labour movement, pinned the mechanization of the human to industrialized modernity in general, and the role of behaviourism in Taylorist production specifically. The second was the post-Weimar critique of early 20th-century liberalism that diagnosed its failures as stemming from having ceded metaphysical territory to scientific positivism and pragmatism. According to this line of thought, the failure of liberalism to grapple with metaphysics had produced geriatric, ineffective technocracies incapable of functioning as zones of politics, much less of withstanding the vulnerability of liberalism to collapse under the weight of its internal contradictions into totalitarianism. Together, these critiques of early 20th-century liberalism called for the reintroduction of a transcendental metaphysics to anchor the politics of freedom. In staging this reintroduction of metaphysics into the postwar reconfiguration of liberalism, Arendt’s 1951 writing augured the nature of the political synthesis cybernetics had to attempt – and ultimately fail to achieve – in order to remain a plausible intellectual project in Cold War America.

The role of behaviourism in modernity, in industrialized production, and in liberal technocracy was embedded within the rise of psychiatry and psychological expertise as legitimate fields of scientific endeavour. Their enshrinement hinged on their ability to leverage their forms of expertise for the diverse fields of turn-of-the-century governance. The successful institutionalization and diffusion of psychiatric and psychological knowledge, in turn, depended on the legitimate scientific status conferred by behaviourism’s methodological-epistemological stance.

At the beginning of the 20th century, psychiatry as a unified profession did not exist. The care of asylum-bound patients was the purview of physicians called ‘alienists’, whose training and praxis shared little in common with either private physicians or neurologists. All of this began to change under the leadership of Adolf Meyer (1866–1950), an American neuropathologist who rose to prominence following the Great War by championing a new vision for what he called ‘psychiatry’. Meyer’s vision was to transform a variety of approaches into a rigorous science of mind, to be forged in the nation’s fledgling university system and turned toward solving a vast array of social ills ranging far beyond insanity within the asylum (Braslow, 1997; Geroulanos, 2014; Lamb, 2014). In 1921, the American Psychiatric Association was created out of the American Medico-Psychological Association, and quickly established itself as a leader in the push to frame psychiatry as a field of expertise that was uniquely capable of solving a diverse array of social problems (Fernberger, 1932). In his 1922 presidential address to the newly formed body, psychiatrist Albert Barrett hailed the new purview of psychiatry: No longer confined to the asylum, psychiatry should extend its interest to whatever ‘disturbed the smooth course of social progress’ (Barrett, 1922: 12). The psychiatrist of the late 1920s saw himself as a technician of adjustment in a universal sense: All mental problems were caused by poor adaptation of an individual to his or her circumstances. The task of the psychiatrist, then, was not limited to treating individuals with severe psychiatric disorders; rather, it was to engineer social progress by facilitating adjustment between the individual and the environment in every sense (Canguilhem, 2008[1952]: 116; Pressman, 2002; Ross, 1972). Psychiatry had billed itself as able to address a wide range of social problems not through political processes, but by viewing the citizen as first of all a somatic entity whose lack of adjustment to a political milieu could be framed as medical pathology. The role of behaviourism within this gambit for expanded authority was to guarantee scientific psychiatry’s wider purview as the technical means of managing the plasticity of human behaviour. In place of political contestation over the question of the good, then, psychiatry offered itself as a technical means of managing behaviour and effecting a medicalized good (Napoli, 1981).

Behaviourism emerged within this new scene of psychiatric claims to expertise with a barrage of sweeping claims both about how to make psychological research methodology properly scientific, and about the purview of psychological knowledge itself. While the psychology of the late 19th century had been marked by a focus on introspection and the examination of mental life (Danziger, 1979, 1980; Leahey, 1981), the strictures of establishing the sciences of mind and emotion within the realm of scientific expertise in the age of industrial production – and the consequent visions for population-wide ‘mental hygiene’ – required new methods that were amenable to quantification, experimentation, and statistical rigour (Burnham, 1988; Camfield, 1973; Coon, 1993). William Wundt’s Leipzig laboratory, in which a generation of US psychiatrists was trained, had introduced experimental methods into psychology (Harper, 1950; Krauss, 2017; O’Donnell, 1985). But although Wundt’s experimentalism had cemented the laboratory as an epistemic site, by the turn of the century US psychiatrists were sceptical of the German introspectionist focus on mentation, dismissing it as overly metaphysically dualist and insufficiently rigorous (Ash, 1998; Kroker, 2003). As a German colleague remarked to William James in 1865 about Wundt’s laboratory-introspectionism, ‘How often already has psychology been made “exact” in this way, only to be led back into the path of psychological psychology!’ (quoted in O’Donnell, 1985: 44). It was not just the laboratory, but the behaviourist laboratory that would remake psychiatry and psychology as respectable scientific endeavours (Buckley, 1989; Coon, 1993). In return for scientific bona fides, behaviourism demanded that the scientist of the mind renounce any interest in psychological internality, reorienting their focus only to observable behaviour. By implication, and lending benediction to the introduction of the rodent as an experimental proxy for the human (Boakes, 1984; Ramsden and Adams, 2009), behaviourism stipulated that any idea of qualitative difference between man and animal be relinquished (Greenspan and Baars, 2005; Samelson, 1981). Drawing on the aspirations to exhaustively technologically harness nature that had animated his mentor Jacques Loeb (Pauly, 1987), and inspired by Franco-Russian psychology’s ‘objectivist’ focus on the reflex, the most radical young evangelical of behaviourism would be John B. Watson. Announcing the new tenets of scientific faith in a widely circulated 1913 speech titled ‘Psychology as the Behaviorist Views It’, Watson claimed,

I believe we can write a psychology, and…never use the term consciousness, mental states, mind, content, introspectively verifiable, imagery, and the like.…It can be done in terms of stimulus and response, in terms of habit formation, habit integration and the like.…My final reason for this is to learn general and particular methods by which I may control behavior. (Watson, 1913: 166)

While some historians have ascribed the formation of cybernetics’ ‘ontology of the Enemy’ to the specificities of the war laboratory (Galison, 1994) or treated the cybernetic ‘ontology of the black box’ as an invention of the cyberneticists (Pickering, 2010), it is clear that the conditions of this metaphysics were assembled by the behaviourist disavowal of internal states. ¹⁰ For behaviourists, the ‘prediction and control of behavior’, rooted entirely in physiological stimuli and response, was a methodological and epistemological triumph over an ontology of the human characterized by free will, language, and a qualitative separation from animals. ¹¹ Cybernetics grafted onto an ontology of the mechanized human that emerged out of behaviourism, but was by no means its originator. ¹² The cybernetic innovation was to keep this ontology in place while brokering a synthesis of the binary opposition between machinic and spontaneous behaviour. The novelty of cybernetics was not that it mechanized the human or organism, but that it offered concepts like ‘feedback’ that created a conceptual third space between deterministic mechanism and organic spontaneity (Kline, 2015: 37–67; Mindell, 2002: 307–22).

The rapid ascent of behaviourism in the US was spurred not only by its promise to extend scientific legitimacy to the fledgling disciplines of psychiatry and psychology, but more fundamentally by the signal role that psychiatry and psychology would play – precisely on the basis of their newfound scientific status – in early 20th-century Progressive Liberal visions of social engineering (O’Donnell, 1985; Ross, 1972; Scull, 2011). Sequela to the re-entrenchment of capital and monopoly power in the Gilded Age, early 20th-century liberalism reckoned with the industrialized and rapidly urbanizing social landscape by unhitching itself from both a Jeffersonian suspicion of capital and Lockean individualism (Livingston, 1997; Lustig, 1984; Sklansky, 2002). No longer committed to the primacy of the individual or to laissez-faire sensibilities suited to small enterprise, liberalism became enamoured of collectivity and planning (Furner, 1993; Repp, 2000; Rodgers, 1998). Derived from the pragmatism concentrated at the University of Chicago, the mode of governance espoused by so-called ‘New Liberals’ in the era of industrial capital would be mediated by the technocratic rule of scientific experts in the role of priest-cum-engineer. Their management – exercised over spheres ranging from economic planning, to industrial relations, to eugenics, to social dysfunction – would ensure that society functioned as a harmonious whole (Kloppenberg, 1988; Sklansky, 2002). The scientific technocrat would operate as a hybrid priest-cum-engineer. Albert Wiggam’s widely read 1923 New Decalogue voiced the prevailing sentiment of the new liberal orientation away from individual freedom and toward positivistic science: ‘Government and social control are in the hands of expert politicians who have power, instead of expert technologists who have wisdom. There should be technologists in control of every field of human need and desire’ (Wiggam, 1923: 16–17). Within this context, behaviourism’s appeal lay in its utility for rationalization and control to a regime of technocratic governance that was increasingly annexing the political as its rightful purview. Given the fervour of early 20th-century liberalism to leverage the plasticity of the organism for social ‘progress’ (Leonard, 2009) – visible in the imbricated projects of eugenics, education, and hygiene – behaviourism was quickly adopted by the newly minted suite of social and psychological sciences for the purpose of depoliticizing industrial modernity’s political contestations.

Behaviourism was not simply a movement in psychology, but a manifestation of a general reconfiguration of liberal politics around scientific positivism and technocracy; in industry, its analogously rigid form was Taylorism. Frederick Winslow Taylor’s The Principles of Scientific Management (1911) was published within two years of Watson’s 1913 behaviourist manifesto; ‘scientific management’ and behaviourism are better understood as twinned manifestations of a shared technocratic ethos rather than mere contemporaries (Nelson, 1980). Both Taylorism and behaviourism were predicated on ‘black-boxing’ the question of human agency, apprehending the worker or subject as a purely physiological organism. For Taylorism, this resulted in the prioritization of control of the worker’s bodily movements while the mental or imaginative aspects of labour were relocated to the emerging class of technocratic management, which in turn used the behaviourist methods of the laboratory (or in the factory as laboratory) to maximize efficiency (Akin, 1977; Braverman, 1974; Gillespie, 1993). While scientific management was meant to obviate tensions between labour and capital, in practice, it often elicited robust resistance from the shop floor; workers objected to their loss of autonomy over the labour process and the imposition of the artificial idiocy of repetitive tasks (Nelson, 1980).

In the resistance to the onslaught of industrial modernity – a front comprising industrial capital, scientific positivism, behaviourism/scientific management, and new liberalism – one objection was predominant: that the very spontaneous core of the human was being mechanized as the worker was subordinated to the machine and the utilitarian logics of technocratic control. While historians have attended to the connections between resistance to behaviourism and the valorization of liberal ideals like ‘open-mindedness’ (Cohen-Cole, 2014), it has gone relatively unremarked that the early 20th-century critique of behaviourism was a critique of industrial capitalist production. ¹³ Arendt’s anti-mechanistic vision of the human ramified into both her labour/work distinction and her anti-behaviourism, both of which drew on discourses that were highly developed well before 1951. Indeed, the basic joints of this problem space were already elaborated in the 1930s, as discourses about labour’s Janus-faced capacities (as mechanistically dictated by the needs of biological bodies, or as creative and spontaneous generation) were immanently seen as emerging from the twinned liberal perils of mechanization and behaviourism.

The critique of mechanization and behaviourism that was established by the mid-1920s was equally a critique of the dehumanizing depredations of industrial capital. The notion that industrialization mechanized the human was not new (Rabinbach, 1990), but by the 1920s, many critiques of capital had come to focus specifically on the role of behaviourism in calibrating the worker to the machine. Philosopher Josiah Morse began his introduction to a 1930 popular volume of anti-behaviourist essays titled Behaviorism: A Battle Line! as follows:

Sciences are the products of human minds. Scientists are human beings, subject to the same limitations and responsive to the same psycho-social forces as other human beings. It is necessary to emphasize these platitudes because in the minds of many, science…has been elevated to the level of a cult and its votaries are regarded as belonging to a species of supermen who speak with almost supernatural authority. (King, 1930: 11)

This vision of behaviourism as the product of industrial capital, urbanization, and an atheistic materialism that evaporated the moral soul out of the human was a widely held popular reaction to the encroaches of behaviourism. In the same year that Behaviorism: A Battle Line! was published, the labour rights advocate Abba Hillel Silver published an essay explicitly attacking mechanization (as both technology and ideology) as incompatible with liberal freedom: ‘The third foe of modern liberalism is the Usurping Machine. The rapid industrialization of society has not only surrounded man with machines but has given the machine terrible power over man’s physical, intellectual, and social life’ (Silver, 1930: 141). Silver’s anxieties about mechanization were legible in the context of an American anxiety about the destruction of individuality by mass society, which in the prewar period took the form of hostility to the Taylorist fitting of man to machine:

The machine which standardizes production also comes to standardize the producer…to mechanize man, to absorb him into the process.…As a result, the thoughts of men who serve the machine are frequently driven into grooves. They move with the alignment of pistons. Men carry over into their social and cultural life the machine discipline—the passion for uniformity and organization. They grow to dislike and to distrust any manifestation of individualism. Thus our human values are becoming machine values, our judgments—mass judgments, and our reactions—mechanical. (ibid., empahsis added)

The American discourse of creativity and anti-mechanism in the problem space of behaviourism, industrialization, and determinism was also tied to shared European concerns regarding the failure of liberal politics in modernity. Between the crises of the late 1920s and the early 1940s, political theorists attacked the regnant consensus around technocracy and pragmatism. The charge they levelled against technocratic liberalism was that it had abandoned metaphysics, uncritically accepting scientific positivism’s creep (Purcell, 1973). Newly arrived at the temple of liberal positivist social science at the University of Chicago in 1930, Robert Maynard Hutchins argued that metaphysics must act within the modern university as theology had operated for the university of the Middle Ages, providing a transcendental foundation for democracy (Hutchins, 1936: 99). Political theorist John Hallowell opined in 1947 that liberalism had ‘committed suicide’ in Germany, its collapse caused by its ‘inability to defend transcendentally its own substantive values and convictions’. For Hallowell, liberalism could reconstitute itself only via a philosophical and religious regrounding in modern theology. These concerns resonated with a broader disagreement, unfolding in the late 1930s and early 1940s, between pragmatist positivists and political theorists about the foundations and character of liberal democracy, and its relation to metaphysics and theology (Guilhot, 2017: 69–115; Gunnell, 1993: 126–45).

The turn toward metaphysics via ‘political theology’ resulted from the structuring influence post-Weimar intellectuals exercised on political theory in the US academy. Notwithstanding significant differences in analysis, post-Weimar philosophers – from Marcuse’s Reason and Revolution (1941), to Horkheimer’s Eclipse of Reason (1947), to Morgenthau’s Scientific Man and Power Politics (1946) – were united in narrating the collapse of liberalism in terms of its failure to ground itself in a metaphysics of transcendence. Unsurprisingly, given liberalism’s endorsement of scientific positivism, Leo Strauss’ influential 1936 book The Political Philosophy of Hobbes traced the genesis of liberalism not to Locke but to Hobbes, whose hard-line materialist metaphysics, Strauss argued, formed the deep philosophical armature of modernist liberalism (Keedus, 2012). Against this reduction of all politics to utilitarian calculations, and countering Karl Mannheim’s embrace of centralized scientific planning as the cure for liberalism’s failure, many post-Weimar intellectuals were convinced by Carl Schmitt’s insistence on the absolute autonomy of politics from science. Schmitt rejected the technocratic claims of neutral management and the encroachment of materialistic liberalism, insisting on the existence of ‘the political’ as a space of absolute autonomous decision: ‘The kind of economic-technological thinking that prevails today is no longer capable of perceiving a political idea.…The political vanishes into the economic or technical-organizational’ (Schmitt, 1985: 65; see also the articles by Bates and Guilhot in this volume). The insistence that the metaphysics of ‘the political’ could not be derived merely from reductionist scientism was a mainstay of a shared post-Weimar critique of modernity on both the left and the right. The turn to metaphysics generally motivated the backlash against neo-Kantian ‘mechanism’, notably in the postwar work of Heidegger and Jaspers, but also for critiques of psychology such as those mounted by R. G. Collingwood’s Essay on Metaphysics (1940), and Wilfred Sellars’ ‘Philosophy and the Scientific Image of Man’ (1963).

Yet it would be a mistake to understand the mid-century problem space of the political metaphysics of freedom as purely one of a mechanistic ‘science’ versus the free space of ‘the political’. Rather, it was a question of what sort of science and what sort of vision of nature would prevail. While Arendt and her post-Weimar contemporaries were motivated to defend ‘the political’ as a space discrete from mechanistic nature, within mid-century philosophies of science a resurgent vitalism provided a natural ally for the political fixation on spontaneity, as both shared a deep suspicion about the mechanization of the capacities of life itself (Swift, 2013). To the same extent that political theory was defined by the struggle to demarcate itself from the technocratic social sciences, 20th-century vitalist philosophies of life resisted the epistemic arrogation through which physics and chemistry annexed life into their purview. Writing in 1952, Georges Canguilhem identified the battle over behaviourism as a manifestation of the struggle between vitalist and reductionist theories of life. At the centre of this contest was the nature of the relation between the organism and its milieu. Linking behaviourism to Taylorism (Canguilhem, 2008[1952]: 96), Canguilhem called behaviourism a form of ‘exorbitant Cartesianism’ in which every operation of the organism was reduced to a reflex, erasing the interior boundaries of the organism and rendering the biological functionally indistinguishable from the geographical. ¹⁴ For behaviourism, ‘since the milieu is given, the organism gives itself nothing it doesn’t in reality already receive. The situation of the living, its being in the world, is a condition or more exactly a conditioning’ (ibid.: 116).

What connected these three intellectual movements – the vitalists, the post-Weimar political theorists, and the critics of Taylorism/behaviourism – was a shared sense that only a metaphysics of exception – ‘the political’, life (as crisis or negotiation), and (for Arendt) language – could defend the human against the encroachment of behaviourism and its related menaces of modernity. As we have seen, Arendt’s elevation of language can be read as a gambit to salvage liberalism from utilitarianism by reconfiguring it around a transcendental metaphysics capable of opening a space of exception – that of spontaneity – within mechanized nature. In the interest of remedying earlier failures, her ‘cognitive liberalism’ would eschew technocratic scientism and ground the political metaphysics of spontaneity in language itself. There was a marked isomorphism of this Arendtian logic around language with vitalist notions of the organism as means of resisting determination by external forces, one further underlined by the shared framing of the behaviourist laboratory as the site where this interiority was destroyed. Canguilhem invoked neurophysiologist Kurt Goldstein, for whom life that was determined entirely by external relations of force was ‘pathological’; hence the situation of the laboratory, in which life was ‘totally commanded from the outside’, was the ‘archetype of a catastrophic situation’. This laboratory situation was pathological in a Kantian sense, in that the operations of life had lost their spontaneous capacity, thereby rendered pathological to the extent that Kant’s pathological motivation for the ethical act was overdetermined by exterior causality. For postwar vitalists, an inner logic organized the interiority and vitally creative capacities of the organism: ‘To live is to radiate; it is to organize milieu from around a center of reference, which cannot itself be referred to without losing its original meaning’ (Canguilhem, 2008[1952]: 114).

As the cognitivist turn would demonstrate, the incorporation of a metaphysics of spontaneity within the very nature of language was not a triumph of ‘politics’ over nature as much as a politicization of nature itself. Arendt maintained this insistence on a logic within the human that resists determination from the outside, but located it not in the somatic or the vital capacities of the organism, but in the spontaneous capacities of language itself. This move, in which language became analogic to the vitalist logic of interiorization in the organism, would be crucial for Chomsky’s defeat of behaviourism in the question of language. Given the internal pivot to a metaphysics of transcendence required to resuscitate postwar liberalism, the cognitive turn would leverage the ‘political theology’ of the spontaneous subject into a science whose anti-mechanistic vision of nature – one visible in the generative capacity of language – comported more easily with the politics of freedom. With this understanding of the complex conditions of Arendt’s critique of behaviourism in place, we can turn to Arendt’s account of the catastrophic situation itself: the totalitarian laboratory in which a new, mechanical man was brought forth.

Arendt and the laboratory of behaviourism

While scholars have remarked on Arendt’s suspicion of ‘automatization’ in the 1950s and 1960s (Simbirski, 2016), a careful reading of the final section of The Origins of Totalitarianism demonstrates that the basis of Arendt’s thinking on mechanization and the reduction of the political subject to the biological was already in place in 1951. At the centre of Arendt’s ontology of the human in totalitarianism was a laboratory in which a new type of man was invented. This laboratory was the death camp, which Arendt (1973: 441) described as ‘the most consequential institution of totalitarian rule’. Within the camp, a horrifying experiment was conducted: the ‘destruction of the juridical person’ (ibid.: 449) accomplished by wresting the human out of language and reducing it to ‘a bundle of reactions’ (ibid.: 441). The laboratory of the camps was an experiment in making spontaneous language impossible. The camps mass-produced corpses, in the form of both bodies without life, and bodies without language; man in the camps ‘can no longer be psychologically understood’, because he had been ‘reduced to a bundle of reactions [that] separate him as radically as mental disease from everything within him that is personality or character’ (ibid.). Crucially, Arendt was describing not only the inmates of the camp, but every human who was encompassed by this laboratory (see ibid.: 442, footnote 132).

Arendt was clear that the camp was literally an ongoing experiment: While the earlier categories of experimental subjects had been ‘criminals, political opponents, Jews, and homosexuals on whom the early experiments were made’ (Arendt, 1973: 451), the ultimate goal was the creation of a totalitarian psyche (or non-psyche) in a general, populational sense. To be cast out of language was of a piece with a simultaneous move to render the human nothing more than a body (ibid.: 453). But in each case – those of the camp prisoner, Himmler’s guard, and Eichmann – the goal was to eradicate the linguistically mediated, psychologically internal capacity for creative thought from which genuine political action might arise. ‘Human nature’ was, for Arendt, the opposite of a vision of ‘nature’ in which cause and effect unrolled a deterministic sequence of causation. Rather, the conditions in which human nature was free were artificial, an exception to nature, and had to be carefully guarded from the totalitarian-behaviourist drive to render man an ‘animal-species man’ (ibid.: 457).

Arendt was unequivocal that the camp was a behaviourist laboratory: ‘The concentration camps are the laboratories where changes in human nature are tested, and their shamefulness therefore is not “just the business of the inmates and those who run them according to strictly ‘scientific’ standards; it is the concern of all men”’ (Arendt, 1973: 458). In the camps, human nature’s fundamental characteristic of spontaneity was destroyed:

For to destroy individuality is to destroy spontaneity, man’s power to begin something new out of his own resources, something that cannot be explained on the basis of reactions to environment and events. Nothing then remains but ghastly marionettes with human faces, which all behave like the dog in Pavlov’s experiments, which all which all react with perfect reliability…which do nothing but react. (ibid.: 454; emphasis added).

Within the problem space of postwar American anxieties around behaviourism and totalitarianism, the terms of the ‘questions that seemed worth asking’ were not in contest: Even the behaviourists accepted that they had to argue that behaviourism was not equivalent to totalitarianism. B. F. Skinner’s 1948 novel Walden Two described the establishment of a utopian community built on behaviourist principles, and the latter half of the book was spent with one character after another ventriloquizing Skinner’s apologia that happiness needed to be engineered and could not be left to the illusion of human freedom. By the mid-1950s, behaviourism’s star was on the wane as criticisms mounted that behaviourist ideology was too inflexible and dogmatic (Cohen-Cole, 2014: 145–51). The revelation of population-wide schemes for behavioural control in US prisons and schools in the 1960s spurred a national backlash against the assault behaviourism seemed to represent on the postwar liberal citizen (Rutherford, 2006). Behaviourists began to be accused by their colleagues of authoritarian methods; Milgram was lambasted by his contemporaries for having created a totalitarian situation in addition to shedding light on totalitarian phenomena (Nicholson, 2011). In the Cold War panic over brainwashing (Carruthers, 1998; Selisker, 2016), behaviourism seemed paradigmatically a technique of the enemy, antithetical to American values of liberty and self-determination. In 1954, the New York Times reported, in an article typical of the behaviourist panic,

The totalitarians have misused the knowledge of how the mind works for their own purposes. They have applied the Pavlovian technique to produce the reflex of mental and political submission of the humans in their power. The Nazis employed these methods. The Russians have made these techniques even more thorough and ruthless. (Meerloo, 1954)

The metaphysics of exception and the turn to language

The cybernetic moment took up the mechanistic metaphysics of behaviourism while obviating the binary between determinism and spontaneity. Yet it was precisely insofar as cybernetics maintained its metaphysical commitment to behaviourism that it was overtaken and absorbed in the turn to cognitivism, which would in turn become the hegemonic new paradigm for the experimental sciences of mind. In this final section, we will read Arendt’s linguistic exceptionalism against Chomsky’s, and the suspicion both maintained toward the behaviourist metaphysics cybernetics had inherited. Cybernetics could not have emerged without the metaphysical and technical groundwork laid by behaviourism, but the cybernetic move was inherited by a cognitive liberalism that shed the cybernetic black-boxing of mind that was behaviourism’s remainder. In staging this observation, I call for more nuance in the way cybernetics is invoked in the history of sciences: The ascent of the computational metaphor for mind was facilitated by the rise of cybernetics, but the cognitivist turn indicates the dialectical subsumption of cybernetic metaphysics rather than its continuation. Last, I urge attention to the significance of a common element of miraculous exception to mechanism at the heart of vitalist, cognitivist, and postwar visions of the human/organism. In the context of the metaphysical pivot required for liberalism’s reinvention by post-Weimar thinkers like Arendt, it was cognitivism, not cybernetics, that could survive the turn toward the political metaphysics of spontaneity and language.

Although it may have appeared to be a radical break with the hegemonic mode of thought in linguistics, Chomsky’s 1959 review of Skinner’s Verbal Behavior rose out of a muted but persevering faction in linguistics that was increasingly unable to reconcile their results with the regnant behaviourist theories of language. Watsonian behaviourism held that language was nothing more than ‘laryngeal habits’ in which speech was merely the body’s determined habit of physically producing sounds, and held that thought itself was simply ‘silent speech’ (see Baars, 1986: 477). In the wake of Karl Lashley’s decimation of the behaviourist claim of a neurological basis for all stimulus–response pairings as reflexive, Skinnerian behaviourism expanded its definition of conditioning to include any form of stimulus–response reaction that reinforced particular behaviours. This concept of reinforcement predated the cybernetic notion of ‘feedback’ while being suggestive of its future conceptual utility, allowing behaviourists to expand the field of stimulus and response to the world at large (Schneider and Morris, 1987). At the heart of all behaviourist accounts of language was the absolute resistance to abstraction. Following the strictures of logical positivism and operationalism, behaviourism held that a concept was always and only the instance of the concept occurring with specificity in time and space; there could be no abstract understanding of a word or notion, but only the operational effect of a word in its embodied, specific instance (Petrie, 1971). Throughout the late 1940s, behaviourists performed studies measuring the meaning of words, or engineering the use of language through reinforcement conditioning (Baars, 1986: 70–6, 360; Cohen-Cole, 2015).

The inability of behaviourists to account for novelty was already noted in the early 1950s, when it was clear that the behaviourist model of information processing resembled a Markov chain more than a Turing machine (see Baars, 1986: 360). But even within behaviourist linguistics itself, animal communication ethologists were beginning to want to think about the internal structure of the organism as a means of explaining behaviour (Radick, 2016), and the Group for Verbal Behavior met at the fringes of linguistics conferences to discuss phenomena like free recall and categorical clustering that behaviourism could not explain (Mandler, 2002: 347). As Gregory Radick has shown, Chomsky’s review of Verbal Behavior was a salvo aimed not only at Skinner, but at the entire field of behaviourist linguistics, particularly at Hockett’s attempt to merge information theory with behaviourism. Hockett held that communication was the instance of the ‘acts by which one organism triggers another’ (quoted in Radick, 2016: 61), and that each feature of human language was shared by at least one animal communication system: He argued that ‘to the extent that various phases of human behavior…can be “explained” in terms of language, which is in turn “explained” in terms of conditioned response and the specific structure of the human organism, a part of the problem of sociobiology can be regarded as solved’ (Hockett, 1977: 15).

But Chomsky had abandoned his previous behaviourism as he became convinced that language did not emerge from evolutionary contiguity with animal communication, but must have appeared all at once in a kind of impossible miracle (Chomsky, 1968: 70). The transformational grammar Chomsky posited was an innate structure of potentiation encoded into the human that comprised an exception to anything else seen in nature; the mental structure in humans (which, for Chomsky, was equivalent to language) both operated as a Turing machine and had the capacity to form abstractions from specific instances to transform instances of grammatical structure into meaningful rules. With this, Chomsky demonstrated that Skinnerian behaviourism had no plausible way to account for the human capacity to generate and assimilate novelty in meaning. Chomsky’s attack was equally an attack against what cybernetics had inherited from behaviourism: an operationalist metaphysics that did not allow for abstraction, insisting that the rule was always in the world and incapable of being abstracted (Moore, 1985). Chomsky’s reaction against behaviourism’s and cybernetics’ operationalist rejection of abstraction was political as well. As he later recounted,

I think you could say that there’s a certain feeling about human nature, about human capacity, that lies at the core of my political ideas and my linguistic ideas. Something having to do with innate structure, with creativity, with the fundamental need for creative work and activity. (Baars, 1986: 350)

While cybernetics had offered a way out of the behaviourist inability to account for teleology in the organism, it was ultimately precisely the inability of cybernetics to escape from the metaphysical strictures of behaviourism that contributed to its overtaking by cognitivism. While cognitivism could import from computers a rich vocabulary of mental concepts dependent on abstraction, the cybernetic commitment to the metaphysics of behaviourism resulted in cognitivism taking up the mantle of information theory. While cognitivists were able to harness the computer as both an experimental and a metaphoric resource, the cybernetic moment sputtered as it failed to translate into a proper experimental science. As Robert Morrison, a natural sciences programme officer at the Rockefeller Foundation, noted in 1953 in an internal memo after several times declining the cybernetic group’s requests for conference funding, ‘After ten years of observing the development of cybernetics one cannot help beginning to wonder when the boys are going to get around to reducing their theories to the form of questions that can be answered by experiment’. ¹⁶ While the rise of cognitivism was hitched to the empirical demonstration of physical computational machines to exhibit abstraction, and the transformational ability of language to mediate novelty, cybernetics remained too much in the thrall of behaviourism’s failed metaphysics. Arendt remained suspicious of cybernetics well into the 1960s, warning that the insistence that there was nothing exceptional about the human capacity for thought meant that machines could replace man. Arendt insisted that there had to be something exceptional about the capacity of human thought to produce something new if freedom was to remain a valid political horizon. ¹⁷

In this article, I have tried to illuminate what I have called Arendt’s cognitive liberalism, and her exceptional prioritization of language as a means of staving off biological reduction and mechanism, within the intellectual environs of the fraught terrain of behaviourism in the 1950s. Read in this way, Arendt’s fears over what she would later call ‘the tradition of organic thought in political matters’ (Arendt, 1970: 116) become clear responses to a discursive milieu in which human nature itself seemed to be reinvented through behaviourist techniques of conditioning and control. Her excoriation of behaviourism and insistence on the necessity of human spontaneity illuminate both the binary between spontaneity and mechanism that cybernetics attempted to address, and the metaphysical deadlock of liberalism that cybernetics failed to solve. My goal in this exercise has not been merely to mount an exegesis of Arendt, but also to demonstrate the co-implicated metaphysics of behaviourism and cybernetics, and to excavate the role of language in resisting the mechanization of life, politics, and the human in the latter half of the 20th century. By examining Arendt’s prioritization of language and insistence on exception against its mirrored form in both early 20th-century discourses against mechanization and the Chomskyan turn, I suggest that there is some internal structure to the metaphysics that cognitive liberalism requires: that of a transcendent exception within nature.