Abstract
What if cognitive neuroscience had more to do with psychoanalysis than we think? In the wake of the 13 November 2015 Paris attacks, neuroscientists have conducted a biomedical study of traumatic memory in survivors. The REMEMBER project, standing for ‘
Keywords
Introduction
Together with Marx and Nietzsche, Paul Ricœur once called Sigmund Freud one of the three ‘masters of suspicion’. In discovering the unconscious, the mechanism of censorship and the existence of ‘screen memories’, Freud forever cast doubt on the transparency of consciousness. As Ricœur writes: The philosopher trained in the school of Descartes knows that things are doubtful, that they are not such as they appear; but he does not doubt that consciousness is such as it appears to itself. . . . Since Marx, Nietzsche, and Freud, this too has become doubtful. After the doubt about things, we have started to doubt consciousness. (1970: 33)
Or so we thought. For once brain scans made the inner workings of the brain newly available for visual grasp, the dynamics of trauma and memory shifted grounds again. It became possible to frame traumatic memories as authentic flashbacks of the original trauma stored in specific locations in the brain (Van der Hart and Van der Kolk, 1995; Young, 1995). In doing so, however, neuroscientific representations of trauma favored the passive, defenseless victim deprived of agency over their own memories as the ideal-type of survivor (Hacking, 1995; Herzog, 2017; Leys, 2000; see El-Haj, 2022). It is the dominant view in contemporary trauma studies anyway that neuroscientific accounts of the way that traumatic memory works played a crucial part in ushering in the ‘condition of victimhood’ as the moral economy in which we still continue to live (Fassin and Rechtman, 2009; Leys, 2007). On this account, the neuropsychology of trauma was born out of a clean break from Freud’s model of repression and censorship, where traumatic memories are notoriously revisable. While such a treatment encourages certain lines of novel inquiry, it closes off others. Crucially, it downplays the forms of life made possible by the affordances of neuroscientific imagery and their role in shaping the interpretation of trauma.
Over the past decades, the analysis of scientific imagery has found new resources in the adjacent field of art history. In a seminal book, Picturing Science, Producing Art (Galison and Jones, 1998), Caroline Jones and Peter Galison devoted a special section to the issue of style. As this innovative approach to scientific pictures shows, forms of imagery are of no less import than the content and objects they show. In my view, this does not simply mean that the transformation of observations, findings, and insights into images by technical means, apparatuses, instruments, or by hand partakes actively in the constitution of knowledge, as other scholars have already argued (Daston and Galison, 2007; Latour, 1987; Lynch and Woolgar, 1990). What is at stake is a new way of apprehending the visual agency of scientific representations. In the wake of David Freedberg’s (1989) pioneering book on ‘the power of images’ in the US, and parallel with Hans Belting (2011 [2001]) in Germany, art historian Horst Bredekamp (2021 [2017]) has recently taken the debate on scientific imaging yet another step further. Building on Aby Warburg’s psychology of art, his theory of ‘image acts’ explores the capacity of images to produce effects of their own. Such an approach to visual agency also applies, beyond the confines of traditional art historical imagery, to science and its technologies for producing images: X-rays, fMRIs, PET-scans, and graphs (Bredekamp et al., 2015 [2008]).
In this article, I build on the new resources offered by this scholarship at the intersection of science and the arts to analyze the impact of imaging technologies’ affordances on the ways the ideal-type of survivor is interpreted. Rather than locate agency in images alone and isolated from their captions and textual context, however, I focus a decidedly intermedial approach on the analysis of brain scans, computer-generated diagrams, and drawings of the mind from 19th-century psychopathology. I draw attention to the interplay of text and image as a motor of aesthetic and scientific innovation. As I argue, reading the scientific literature on trauma for pictures conceived as complex interplays between visuality and discourse runs counter to the established scholarly relation between the rise of neuroscience and brain scans, on the one hand, and the advent of the condition of victimhood starting in the 1980s, on the other.
I will begin by analyzing how the visualization of psychological trauma has evolved from the early 1990s, when PET-scans of the brain first became available, to the latest innovations brought about in the field of neuroscience by the computerization of biomedical imaging. Here I engage in a comparative analysis of the use of scientific images in two notable scientific studies. First, I look back on Bessel van der Kolk’s pioneering study of Post-traumatic Stress Disorder (PTSD) at the Harvard Medical School in 1994, which is often seen as the epitome of the victimhood paradigm. In a second section, I turn to ‘REMEMBER’, a large-scale study of PTSD in survivors of the 2015 Paris attacks. Coupling state-of-the-art detection techniques such as fMRI with algorithms and complex computational systems, the study showcases colorful brain pictures of ‘resilience’ that stand in stark contrast with van der Kolk’s black-and-white scans from the 1990s. A survey of the REMEMBER study reveals an unexpected conceptual genealogy linking current experimental protocols of memory suppression to a dynamic model of intrusion and repression first elaborated by Freud: a mechanism known as Bahnung, or ‘grooving’. Against the grain of the classic split between psychoanalysis and cognitive neuroscience, I rehabilitate Freud’s concept of Bahnung as a precursor to contemporary connectionist models. I do so by looking at the series of drawings that Freud relied on for his signature innovation, from the late 1890s to his last ‘Ego-Id-Superego’ diagram from 1933. At the nerve center of this article, therefore, lies an impulse to bring visual agency to bear on the stories that we tell about trauma, psychoanalysis, and the neurosciences.
The Victimhood Paradigm
In the mid-1990s, Bessel van der Kolk studied PTSD at the Harvard Medical School by having patients listen to a ‘trauma script’ while lying in a Positron Emission Tomography (PET) scanner. The experience is reported on in detail in a 1996 article which van der Kolk co-authored with psychiatrist Scott Rauch, the first Director of the Massachusetts General Hospital Neuroimaging Laboratory (Rauch et al., 1996). As the methods section reads, all eight participants in the study (six women and two men) were selected based on a structured clinical interview testing them for the entry criteria for the nosographic category of PTSD as described in the third revised version of The Diagnostic and Statistical Manual of Mental Disorders, or DSM-III-R (1987). In DSM-III-R, PTSD is defined by two main criteria: 1) the individual has experienced a ‘traumatic event’ (exposure criterion or ‘trigger’); and 2) he or she is besieged by intrusive memories or ‘flashbacks’ of the original traumatic event (see American Psychiatric Association, 1987, quoted in Young, 1995: 117). These intrusive memories, and what they might explain about brain function more generally, are what van der Kolk and Rauch’s brain imagery is meant to elucidate.
Once in the scanner, participants were played a voice recording re-creating a fragment of their traumatic experience. Van der Kolk’s assistant, Rita Fisler, prepared the scripts in collaboration with each patient. She then audiotaped each script ‘in a neutral voice’ for playback in the laboratory settings. Each patient was instructed to lie still and allow herself or himself to focus on the contents of the scripts until after the PET scan was completed. All the while, van der Kolk’s team monitored their blood pressure and heart rate so that these physiological signs could be compared with brain activity. As any part of the brain became metabolically active, its rate of oxygen consumption would immediately change, which would be picked up by the scanner. After the procedure was complete and the scans were made available for interpretation, van der Kolk found that Broca’s area, which neurology assumes to be one of the speech centers of the brain, went offline whenever a flashback was triggered. 1 The scans offered visual proof, he believed, that ‘the effects of trauma are not necessarily different from – and can overlap with – the effects of physical lesions like strokes’. Moreover, he noted that as they emerged from the scanner, participants looked defeated, drawn out, and frozen. His patient Marsha’s ‘breathing was shallow, her eyes were opened wide, and her shoulders were hunched – the very image of vulnerability and defenselessness’ (Van der Kolk, 2014: 42–3).
The main breakthrough of the experiment, according to Bessel van der Kolk’s own analysis, was to offer visual evidence of a phenomenon that had long remained unavailable to the scientific eye for lack of instruments capable of penetrating the inside of the living brain: traumatic experiences are sometimes so overwhelming that they cannot be integrated into existing mental frameworks. They cannot be integrated into current recollection schemes through Broca’s area, where memories are usually processed through verbal description. Instead, these ‘unassimilated scraps of overwhelming experiences’ are stored, or ‘engraved’, as literal imprints in the hippocampus or limbic area (Van der Hart and Van der Kolk, 1995: 176). Hence it creates a signature pattern in which traumatic memories return intrusively and in an untimely manner as sensory and distressing fragments of the event as if it was experienced anew in the present. 2
Van der KoIk’s theory of engraving is a textbook case of neuroanatomy. Not only are emotions locatable in specific areas of the human brain, as exemplified by the PET scans, but traumatic images themselves are theorized as ‘flashbacks’ that retain the imprint of the original traumatic experience. As Ruth Leys comments in her influential Trauma: A Genealogy (Leys, 2000): ‘According to van der Kolk’s view, trauma creates a structural deficit, wound, or “hole” in the mind where representation ought to be, but where literal memories of the trauma are inserted instead and permanently “etched” or “engraved” in the form of an iconic “imprint”’ (Leys, 2000: 249–50). The result is that traumatic memory, in its literality, is not integrated into ordinary awareness but is cut off or ‘dissociated’ from consciousness and hence is made unavailable for normal recollection. The literal nature of the flashback means that the memories can now be viewed as pristine rather than affected by psycho-social factors – rather, in other words, than notoriously revisable, as in psychoanalytic notions of repression. PTSD was thus born when, beginning in the 1970s in the US, the neuroscience of memory displaced psychoanalysis in the form of a clash between cognitively oriented scientists, such as Van der Kolk, who conceptualize the mind in information-processing, neurophysiological terms and reject the Freudian concept of repression or unconsciously motivated forgetting, and more psychoanalytically oriented clinicians, such as Lansky, who retain a Freudian commitment to notions of psychic conflict and repression and who tend to oppose a literal interpretation of the traumatic dream. (Leys, 2000: 244)
Beyond Leys, the scholarly literature on PTSD as a diagnostic category is unanimous on this point: the abandonment of psychoanalytical conceptions in favor of neurological retheorizations of traumatic memories as literal replays of the past was crucial in grounding victimization in science; and in doing so, it paved the way for financial compensation (see especially Fassin and Rechtman, 2009; Herzog, 2017; Young, 1995). The implication is twofold. For one, the neuroscience of memory was key to the emergence of the helpless victim as the ideal-type of the survivor. Second, the engraving theory constituted a sharp rupture with the late Freud’s psychodynamic model of the intrusion and repression of memories, otherwise known as ‘censorship’. The REMEMBER study, I argue in what follows, offers a more contrasting view of the genealogy of trauma on both counts.
Toward Resilience
In the summer of 2019, when I visited the November-13 Program team of neuropsychologists and computer scientists at the Cyceron Centre for biomedical imaging in Caen, France, the atmosphere was much less fatalistic than at Harvard in the 1990s. In the wake of the 2015 terrorist attacks on the Bataclan concert hall, café terraces in Paris, and the Stade de France in Saint-Denis, neuropsychologists at the French National Institute of Health and Medical Research (Inserm) have been conducting a large-scale study of PTSD in survivors. As I have already noted, the biomedical research program is entitled ‘REMEMBER’ (REsilience and Modification of brain control network following novEMBER 13). In scientific articles as well as in newsletters addressed to the broader public, REMEMBER proudly showcases composite brain pictures of ‘resilience’ that stand poles apart from van der Kolk’s PET scans of neural vulnerability. 3 These type specimens emphasize brain connectivity over localization. They do not foreground victimization. Rather, they portray a different type of survivor: the ‘resilient’ ones, that is, those who have recovered from PTSD and developed increased brain plasticity compared with the ‘control group’. ‘We were amazed’, the lead scientist on the project, Pierre Gagnepain, said to me, ‘to find that the dorsolateral prefrontal cortex (DLPFC), which is a core hub of the brain control system, showed increased activity in trauma-exposed individuals without PTSD compared not only with participants with PTSD, but also with healthy individuals who were never exposed to the 2015 attacks’. 4 Between van der Kolk, whose PET scans promoted victimhood, and Gagnepain, whose composite technical drawings move beyond any simple identity between location and function to emphasize brain connectivity and resilience, the difference is both tiny and total.
REMEMBER squarely locates itself within the ‘connectomic era’ of neuroscience, a broader trend in the field that has ceased to nurture unambiguous localizing tendencies (D’Angelo, 2012: 77). The key word for this venture is not ‘center’ but ‘circuit’ (Marshall and Fink, 2003). In the era of computational systems of neuroscience and digital imaging, the attempt to localize higher functions of the mind, such as speech, has given way to a distinct endeavor: an effort to map out with ever greater complexity the connectional architecture of the brain (Beaulieu, 2000; Dumit, 2004; Joyce, 2008; Saunders, 2008; see Guenther, 2015). The REMEMBER protocol, too, is strikingly at odds with that of the pioneering biomedical study of PTSD conducted at Harvard in 1994. Participants were recruited among voluntary survivors of the 13 November 2015 attacks, based on structured clinical interviews. The questionnaire was established by neuroscience-trained psychoanalyst Jacques Dayan, using the diagnostic criteria for PTSD included in the French translation of The Diagnostic and Statistical Manual of Mental Disorders in its revised version from 2013 (DSM-5). Updated standards of ethical clinical research, however, prohibited participants’ exposure to sensory fragments of their trauma. Instead, the REMEMBER protocol involved a highly complex procedure in which subjects were tested for their ability, or inability, to suppress so-called ‘neutral and inoffensive intrusive memories’ (Mary et al., 2020: 1). Unexpectedly, it borrowed from a previous paradigm meant to ‘provide a mechanism for the voluntary form of repression proposed by Freud’ (Anderson and Green, 2001: 366).
Before joining the November-13 Program, Pierre Gagnepain had studied in the department of neuroscience under Michael Anderson at the University of Cambridge. Shortly before then, in 2001, Anderson had co-authored with cognitive psychologist Collin Green an influential paper in the field of cognitive neuroscience, laying out the think/no-think paradigm (TNT), an experimental protocol for making the mechanism of repression available for the study of broader mechanisms of memory, attention, and cognitive control. With the REMEMBER study, Gagnepain brought the think/no-think procedure to bear on the study of trauma. The protocol, he explained to me, aimed at creating ‘an experimental form of intrusion’. It did so by using neutral material entirely disconnected from participants’ traumatic experience. In practice, participants were overtrained to associate neutral (non-traumatic) words with particular (also non-traumatic) images during a three-hour practice session prior to entering the MRI scanner. This overtraining procedure was intended to ensure that the word cue would automatically bring back the artificially created intrusive image. Once in the scanner, they would be instructed to suppress intrusive images when presented with the associated word cues. As they did, the fMRI would measure the dorsolateral prefrontal cortex’s regulation and suppression of memory activity.
Based on medical imaging, the connectivity between the brain’s prefrontal areas (involved in control) and the memory sites (located at the back of the brain) appeared to decrease during successful inhibition of unwanted images. Moreover, additional analyses focusing on the directionality of neural flow communications revealed that this decrease in connectivity arose from an active inhibitory control mechanism, which was preserved in healthy individuals but disrupted in people who developed PTSD. In the process, the central feature of PTSD was thus thoroughly reframed: PTSD, it turned out, did not lie in the persistence of vivid and distressing intrusive memories per se, but in the inability to inhibit associations between memories. What is more, this capacity was found to be more developed in ‘PTSD-’ (i.e. those who had recovered from PTSD) than in the control group – hence the colorful brain drawings, showing that ‘resilient’ subjects developed increased brain plasticity compared with the ‘control group’.
The computer-generated diagrams such as those included in a Science article from 2020, ‘Resilience after Trauma: The Role of Memory Suppression’, are key to the argument. 5 In Picturing Personhood (2004), Joseph Dumit has analyzed the practice of actively constructing images for publication as key to making particular points. Scientific images, he writes, are ‘explicitly rhetorical’ (p. 16). This is emphatically true of REMEMBER’s technical drawings. The arrows connecting the ‘control system’ and the ‘memory sites’ in diagram B are particularly intriguing in this regard. One slightly looping line in light gray terminated by arrows on each end indicates a relation of ‘functionally [inter-]dependency’ between the hippocampus and the prefrontal cortex. The other rounded line, with only one arrow to the right end, in turn is captioned to depict a one-way ‘causal influence’ of the prefrontal cortex over the hippocampus. Connected to this imagery, the Science article has a better explanation for post-traumatic stress impairment than the simple expression and persistence of vivid and distressing intrusive memories, considered to be the central feature of PTSD. Not surprisingly, it is grounded in connectional anatomy. The untimely revival of intrusive images in patients with PTSD, it turns out, is caused not by an illness of memory but rather by a dysfunction of the neural networks underlying the control and suppression of memory retrieval. The ideal survivor is no longer the passive victim. On the contrary, it is the person who has overcome his or her trauma by developing greater cerebral plasticity than ordinary people.
Freud’s Invention of Intrusion, or Bahnung
As we have just seen, nothing predestines neuroscience a priori to favor victimization – least of all the mechanism of imagistic intrusion. The resilience paradigm, I now wish to argue, runs counter to a reading of neuroscience that would pit it squarely against Freudian psychoanalysis. On the contrary, the REMEMBER study needs psychoanalysis and its crucial framing of the mechanism of intrusion in order to thrive. More generally, I suggest, the pervasive contemporary use of the vocabulary of ‘intrusion’ and ‘intrusive memories’ may be more faithful to Freud’s concept of ‘grooving’ or ‘fraying’ (Bahnung) than we have been accustomed to think. The part played by graphs and technical images in scientific innovation, too, calls for more attention than has so far been devoted to it. This includes Freud’s own diagrams of psychic topography, to which I come now.
In ‘Blacked-Out Spaces: Freud, Censorship, and the Re-Territorialization of Mind’ (2012), a captivating account of Freud’s path to the ‘discovery’ of the unconscious, Peter Galison emphasized just how instrumental Freud’s drawings were to what is arguably the most crucial invention in the history of modern psychiatry. Galison’s core argument is that Freud’s experience of print censorship and caviar ‘blacked-out spaces’ during the First World War, when his correspondence with Karl Abraham and others was on occasion inspected twice (first in Austria, then in Germany or Britain, and vice-versa), eventually paving the way for his framing of the mechanism of psychic censorship. As he writes: ‘Seeing the raw, unhidden deletion of German text gave Freud a language in which to work out his original formulation of psychodynamic censorship in the years from 1897 to 1900’ (2012: 242). Freud’s use of diagrams is key to the argument. Shortly following his own private experience of literal censorship during wartime, Galison shows, Freud began to develop a new series of ‘topographical’ pictures of the psyche. He did so as part of his ‘Unconscious’ paper, which he drafted starting in early April 1915. In lieu of a literal topography of brain functions, where psychic elements are mapped onto physically located topoi of the brain, Freud came progressively to develop a new topographic study of relations between psychic components. Galison (2012: 256–7) contends that [Freud’s] intense reasoning about borders, secret packets turned back, softenings, codes, pre-censorship and censors helped constitute the re-territorialization of the mind toward a new, less literal psychic topography. This reasoning set the ground for the diagram he began sketching three years after the June 1919 Treaty of Versailles. For it was in July 1922 that he began framing ‘The ego and the id,’ his theoretical treatise on the structure of mind, that appeared in April 1923. In that work, Freud recoded anatomy into psychiatry.
While the notion of place (topos) is maintained in the drawing from 1923 (see Figure 1), these places are no longer cortical sites. Instead, they are depictions of brain functions. In breaking with Freud’s earlier neurological drawings, this diagram marked a first step in psychoanalysis’s ‘re-territorialization of the mind’ away from anatomy and toward non-literal, psychical maps of relations, not proportions.
‘The Ego and the Id’. Sigmund Freud, Das Ich und das Es (Leipzig: Internationaler Psychoanalytischer Verlag G.M.B.H., 1923).
The full pictorial representation of censorship being twice applied at frontiers (first to delete offending content, and then again to hide the omissions) was achieved in 1932, when Freud drafted a first sketch of ‘The Psychical Apparatus’ diagram, adding the ‘Superego’ to the ‘Ego’ and the ‘Id’. His ‘Anatomy of the Mental Personality’, which was published in 1933, included an ‘unassuming sketch’, in Freud’s own words, meant ‘to portray the structural relations of the mental personality’ (Freud 1933: 78). And indeed, the double arrow accompanying the Superego and its structural model of the psyche, now depicting a system of relations, evokes a drawing style akin to Ferdinand de Saussure’s Course in General Linguistics and its use of the double arrows to depict language as a system of differences (see Figure 2). 6
‘The Psychical Apparatus’. Sigmund Freud, Neue Folge der Vorlesungen zur Einführung in die Psychoanalyse (Leipzig, Wien, Zürich, Internationaler Psychoanalytischer Verlag G.M.B.H, 1933).
All the while shedding original light on Freud’s diagrams, Galison tends nonetheless to frame them as secondary, ancillary, and ultimately inessential to Freud’s scientific breakthroughs and discoveries. Somewhat inexplicably given the new attention paid to the role of visualization in scientific research, visual representation always comes after verbal depiction. In 1915, Galison notes, ‘the topographic did not yet have its diagram’ (Galison, 2012: 246, emphasis added). Thus, the discovery process goes as follows: at first, Freud merely developed verbal descriptions of the topographical ‘picture’ of mind – in letters to Fliess and in the ‘Unconscious’ paper from 1915. Only in 1923 did he begin drawing graphs on paper (Figure 1). The double-ended arrow itself did not emerge until 1932, only after the fact, to go along with the newly found Überich, or ‘Superego’ (Figure 2).
On this view, Freud’s psychic diagrams figure as mere ‘carry-overs’ from the material topography of the brain, infused with gross anatomic meaning. I suggest an alternative reading of this captivating sequence. I want to propose that the activity of drawing might have played an even greater role in Freud’s discovery of the unconscious—and perhaps to slightly alter, in the process, our understanding of his signature breakthrough.
In between his first sketch from 1922–3 and the full-fledge 1933 diagram with the ‘Überich’ and the double arrows in it – in 1925, that is – Freud wrote `A Note upon the “Mystic Writing Pad”’. The ‘Note’ is the text in which Freud first came up with the model of the unconscious as a ‘writing machine’, as Derrida (1972 [1966]) pointed out in his own extended commentary on Freud’s discovery, entitled ‘Freud and the Scene of Writing’. Another possibility, therefore, may be more likely: Freud could just as plausibly have been drawn to what came to be known as the ‘psychodynamic’ model of the intrusion and repression of memories in the very process of picturing the mind, rather than the other way around.
Let me dwell for a moment on Derrida’s close reading of `A Note upon the “Mystic Writing Pad”’ so as to clearly delineate what the implications are of the above suggestion for how we think about Freud’s watershed breakthrough. I will put the question as follows: What sort of ‘writing machine’ is the unconscious? In ‘Freud and the Scene of Writing’, Derrida displaces the focus from censorship as such to the dynamics of Bahnung, or ‘frayage’ in his original translation of Freud’s terminology into French. Derrida’s contention is that Freud started to move away from the idea of a strictly neurological lesion at an earlier date than Galison assumes, namely when he first developed the notion of Bahnung in the Project for a Scientific Psychology (1895). As I mentioned, Derrida’s premise brings him to identify a different text for Freud’s main invention: not the ‘Anatomy of the Mental Personality’ from 1933, but instead ‘A Note upon the “Mystic Writing Pad”’ from 1925. For he draws a direct line between the mechanism of Bahnung as first developed in the Project and its further exploration in the ‘Note’ which, per Derrida’s reading, contains Freud’s truly original conceptualization of the unconscious as a ‘writing machine’. This would situate the diagram from 1922–3 in an intermediate position in the discovery process rather than as a mere afterthought (or a carry-over from the 1915 ‘Unconscious’ paper).
Now, what did Freud exactly discover with the Bahnung mechanism? Something akin, as I will show, to contemporary connectionist models of intrusion. 7 In The Standard Edition of the Complete Psychological Works of Sigmund Freud, Bahnung is translated as ‘facilitation’. However, as Alan Bass has noted in his translation of Derrida’s Margins of Philosophy, ‘facilitation’ erases the sense of breaking open, its distinctive original meaning. Instead of ‘facilitation’, Alan Bass therefore coined the word ‘breaching’ to translate both Bahnung and frayage, the French translation he found in Derrida’s (1982 [1972]: 18n. 21) text. The difference between facilitation and breaching is what I explore in what follows. For only if we conceive of Bahnung as breaching, and not as mere facilitation, as I contend, can we begin to identify its trace in the pervasive contemporary use of the vocabulary of ‘intrusion’, either in noun or in adjective form, as in ‘intrusive memories’.
Freud did not himself coin the word Bahnung. In a recent book, Localization and Its Discontents (2015), Katja Guenther has traced the early days of the connective model of brain function to Viennese neuropsychiatrist Theodor Meynert’s failed attempt to adapt the lesion model to higher functions of the mind. As she demonstrates, the mechanism of Bahnung was first developed by Viennese physiologist Sigmund Exner (1846–1926) and psychiatrist Theodor Meynert (1833–92) in order to describe a physiological process explaining the basic law of association by simultaneity. In contrast to other medical specialties, psychiatry lacked a clear route to somatization and a widely accepted somatic model for brain functions. Contrary to most organic paralyses, hysterical symptoms, in particular, did not correspond to observable lesions in autopsy. What was to be ‘facilitated’, then, were associations between two or more sensory impressions (typically, one visual, the other acoustic) so as to create an audiovisual memory of a particular experience. Once this process could be explained in physiological terms, then mental diseases such as aphasia or hysteria, which were particularly ill-fit for the lesion model, could be explained by assuming that associations were simply lacking – that connections were ‘cut off’, so to speak, rather than centers damaged, as in the hypothesis of an organic lesion. In developing such an associationist model through an early concept of Bahnung as ‘facilitation’, Guenther’s renewed exploration of the concept reveals, Meynert was able to explain aphasia, a language disorder previously thought to be caused by damage in a specific area of the brain that controls speech, without recourse to a neurological lesion. Similarly to Meynert, Freud worked from the 1890s on to come up with a physiological mechanism to explain hysteria and other nervous diseases without recourse to the idea of the lesion. Unlike him, however, he reframed neurological newness as a ‘breach’ of consciousness in lieu of the model of association as facilitation he found in Meynert.
This is the point at which Derrida’s reading of this sequence is most enlightening, for the model of Bahnung as neuronal facilitation did not come without problems. Freud, Derrida observes, was dissatisfied with Meynert’s mechanism because it did not explain how, once a previous association had been ‘facilitated’, what Freud called the ‘writing pad’ of the psyche could be erased and made available for new memories. It did not explain, that is, how the writing pad could be made magic or ‘mystic’. Here is Derrida’s laying out of Freud’s problem already in the Project for a Scientific Psychology from 1895: The crux of such an explanation, what makes such an apparatus unimaginable, is the necessity of accounting simultaneously, as the ‘Note’ will do thirty years later, for the permanence of the trace and the virginity of the receiving substance, for the engraving of the tracks and the perennially intact bareness of the perceptive surface: in this case, of the neurons. (1972: 78)
Freud later solved this problem in ‘A Note upon the “Mystic Writing Pad”’ (1925) by forging the hypothesis of ‘fraying’, now synonymous with the breaking of a path (Bahn). Contra Meynert, he developed a new model of Bahnung as ‘breaching’ or intrusion, not ‘facilitation’. In Derrida’s description, Freud’s invention of the ‘fraying’ mechanism goes as follows: The path is broken, cracked, fracta, frayed. Now there would be two kinds of neurons: the permeable neurons (φ), offering no resistance and thus retaining no trace of impression, would be perceptual neurons; other neurons (ψ) would oppose contact-barriers to the quantity of excitation and would thus retain the printed trace: they ‘thus offer a possibility of representing memory.’ First representation, first staging of memory. (1972: 78)
Freud’s ‘concept of fraying’, as Derrida calls it, thus reframed the tracing of a trail – the opening up of a conducting path, which Freud’s forerunners in neurology such as Exner and Meynert used to conceive as an innocuous and uncomplicated mechanism of neural facilitation, or habituation – as involving a certain violence and a certain resistance to the effraction. Crucially, he did so in a way that made the model of a breach in the psychic economy integral to the way in which the psyche functions. The same economy of fraying now defined all memories, be they ‘traumatic’ or simply perceptual. It defined memory in its regular mode, with no essential difference made between the normal and the pathological. What started as a heuristic model to explain mental disease mutated into a description of the complex workings of memory tout court.
In Freud’s model of intrusion, therefore, memories are not ‘intrusive’ because they are pathological. They are intrusive because the origin of the psyche as (conscious or unconscious) memory in general can be described only by taking into account the economy of differences between breaches. Which is to say, the term ‘traumatic’ has no other sense than an economic one. The double-ended arrow in Saussure’s style in the 1933 diagram is there to describe precisely this: Thoughts and psychical structures in general must never be regarded as localized in organic elements of the nervous system but rather, as one might say, between them, where resistances and frayings provide the corresponding correlates. (Freud, 1925: 615–16 as quoted and translated in Derrida, 1972: 97)
Freud’s invention – the great discovery that cemented his status as a ‘founder of discursivity’ in Michel Foucault’s phrase – thus calls for a work of redescription that may allow us to better capture its unexpected revival in contemporary connectionist models of brain functions (1998 [1969]: 218). It might be best described not as ‘the unconscious’ conceived as a psychic topos (a topographic metaphor), but rather as a fully physiologized, dynamic model of (traumatic) memory born out of a revision of Meynert’s notion of neural association (known as Bahnung). In ‘On the Couch’, her chapter on Freud, Katja Guenther puts it this way: Though the practice, and then the institutions that were built upon it, seem to locate psychoanalysis on the psychological side of the soma-psyche divide, the development that produced the practice, the rejection of the lesion model, was motivated by an attempt to rid neuropsychiatry of its psychological elements. (2015: 94)
In view of this evidence, another reading of Freud’s series of drawings and writings is not only possible but becomes more probable: Only once Freud could come up with the picture (the image-text) of the ‘unconscious’ in 1922–3 was he able to fully develop the psychographic metaphor in ‘A Note upon the “Mystic Writing Pad”’ (1925), and subsequently to tweak and refine the topographic diagram by including the Superego in 1933.
This argument is supported by a noticeable pictural fact: the persistence, in the representation of the ‘Ego-Id-Superego’ from 1933, of the same double-lined link on the right middle side of the picture which was already visible in the diagram from 1923 under the caption: ‘verdrängt (repressed)’, corresponding precisely to the association mechanism. Art curator Lynn Gamwell and psychoanalyst and neuropsychologist Mark Solms have assembled the most exhaustive collection to date of Freud’s image production in From Neurology to Psychoanalysis (Gamwell and Solms, 2006). The exhibition catalogue compiles Freud’s drawings of nerve cells and diagrams of the mind, beginning in 1876, when he was a 20-year-old student, until his last image, the ‘Ego-Id-Superego’ diagram he drew in 1933, four years before his death. In particular, it contains a diagram from 1891 showing an early version of the Bahnung mechanism (2006: 94). In this diagram, the connection between the word sound-image and visual object associations is represented through a double-lined link, the same two parallel lines which continue to appear in Figures 1 and 2. Had Freud wanted to abandon physiological associations altogether, he could easily have done away with this atavistic feature (this ‘carry-over’ from his earlier neuropsychological work). Instead, it still stands as the defining feature of the repression mechanism, sitting as it does at the border between the preconscious and the unconscious. Remarkably, it does so even after the location of the acoustic system, which mimicked the speech center of the brain on the left side of the 1923 diagram (Figure 1), was removed from his final ‘Ego-Id-Superego’ drawing (Figure 2). In other words, anatomical detail is no longer necessary for Freud to portray the functional mechanisms that underpin such complex processes as those involved in the psychodynamic model of the unconscious. But physiological association – now reframed as fraying or ‘intrusion’ – remains (or rather, it becomes) the defining feature of the dynamics of repression.
In the reading of Freud’s path to discovery that I propose, drawings and pictures do not come at the end of the process of scientific innovation to merely illustrate its results or ‘findings’. They are core to the creative process. What the diagram from 1923 affords, for instance, is the possibility of visualizing the association mechanism (Bahnung) as located at the interface between the Ego and the Id (see the two parallel lines with ‘verdrängt’ (repressed), the German legend discussed above).Once made available through visual representation, this new opportunity makes it possible a couple of years later for Freud to come up with the concept of Bahnung-as-fraying in ‘A Note upon the “Mystic Writing Pad”’ (1925). It was this move and the dynamics of memory-as-intrusion which ultimately gave way to Freud’s original conceptualization of ‘the unconscious’, which in turn materializes in the 1933 diagram in the double-ended arrow and its associated caption: Überich, or ‘Superego’, linking together the ‘Preconscious system’ (Vorbewusst) and the newly introduced ‘Unconscious’ (Unbewusst).
Such an intermedial approach to the productive relationship between verbal and imagistic description compels us to rethink our current understanding of the relationship between Freud’s invention of psychoanalysis and contemporary neuroscience. ‘[Freud’s] formulation of psychoanalysis’, as Guenther writes, ‘was not so much a break from earlier brain science, but rather can be more productively understood as a radicalization of its principles’ (2015: 72). Nor, I suggest, is contemporary neuroscience most productively understood as a neat break from Freud’s psychodynamic conceptualization of traumatic memory. At any rate, reading Freud for continuities with earlier brain science illuminates entire swathes of contemporary neuroscientific literature that otherwise remain unfathomable – at best peculiar and sibylline, at worst esoteric and mystifying.
In the particular case that concerns us here, Freud’s mnesic mechanism of ‘pathological’ association made integral to the structural economy of the psyche – the psychodynamics of ‘trauma-as-breaching’ in lieu of the lesion paradigm – provides us with a clear path from which to understand the REMEMBER study’s use of ‘neutral intrusive memories’ (a contradiction in terms from the vantage point of contemporary trauma studies, as we have seen). If the same mnesic economy of fraying defines traumatic and perceptual memories alike, then it makes sense to go ahead and ‘implement neutral and inoffensive intrusive memories’ in order to study broader memory processes (Mary et al., 2020: summary). And while the REMEMBER study’s overtraining procedure rests on a model of physiological ‘facilitation’, its overall argument about brain connectivity instead borrows from a Freudian model of intrusion as ‘breaching’. Wanting it both ways, REMEMBER emphasizes the paradoxes of contemporary neuroscience. 8
Once more, this is perhaps best seen in its diagrams. At first glance, the computer-generated diagram B from the Science article, with its lateral views of the human brain, might seem to back up the claim that the successes of neuroimaging have unambiguously nurtured localizationist tendencies. Areas in color indicate clearly delineated locations in the brain. Orange-colored parts at the bottom of the brain, in particular, are linked by 2D plain lines, together with a text that indicates: ‘memory sites’ (where traumatic engrams are stored). The areas in red at the top are likewise connected to another text: ‘control system’. Here the color chart (red over orange) further denotes the higher status of the dorsolateral prefrontal cortex over the limbic system and the hippocampus. And yet at the same time the overall caption for diagram B reads: ‘Brain connectivity during memory suppression’, unambiguously stressing the ‘connectome’ (Seung, 2012; Sporns, 2012). The interplay of text and image thus also displays a more dynamic view of brain function, a testimony to the revival of Freud’s psychophysiological maps and description of functional relations between the systems and agencies of the mind in the contemporary connective model of brain function.
How Newness Comes into the World
In this article, I have argued that the resilience paradigm runs counter to the classic split between Freud and cognitive neuroscience. To do so, I have relied on the intermediality of text and image as a source of scientific innovation. In so arguing, let me make clear, however, I do not aim to suggest a one-way relationship between diagrams and verbal description, only in reverse. It is the productive interface between drawings and captions, I maintain, that should be the focus of our attention. Peter Galison already mused about how Freud availed himself of the ‘plastic’ nature of the topographic in order to carry out his discoveries – a word Freud used to justify his use of psychical topography, however inaccurate or even arbitrary this form of visualization may be from a strictly referential standpoint: What does [Freud] mean here by the adjective ‘plastic’? Derived from the Greek Plastikos, the term indicates being capable of shaping or moulding. The psychical topographic offers the possibility of shaping and altering the concept as if it were material, distinguishing and altering (non-brain) locations, allowing motion, connection, interruption or distortion between them. (Galison, 2012: 248)
By bringing the rich literature on intermediality to bear on an analysis of scientific discovery, I have aimed to take this argument a step further. The intermedial approach emphasizes the complex interplay between two or more media, such as text and image, as a motor of aesthetic innovation (Edwards, 2017; see also Mitchell, 1994, 2005; Müller, 2017; Wolf, 1999, 2017). Freud’s invention of the psychodynamics of repression, I suggested, is best read against this grid.
Further, the revamped form of genetic reading that I proposed has major consequences for what we understand Freud’s main innovation to be. In the reading that I offered, Freud did not so much discover the unconscious as a new metaphorical topos as he invented the mechanism of fraying (Bahnung), or of ‘pathological’ association, as the ordinary process by which the psyche works out memories. This, in turn, compels us to revise our assessment of his legacy. True, the REMEMBER study’s use of Freud is largely heretic, even unaware of itself for the most part. But is this not the signature mark of a discursive formation, or épistémè? In Michel Foucault’s ‘What Is an Author?’ (1998 [1969]), Sigmund Freud figures prominently as an initiator of discursive practice. Foucault writes: When I speak of Marx and Freud as founders of discursivity, I mean that they made possible not only a certain number of analogies but also (and equally important) a certain number of differences. They have created a possibility for something other than their discourse, yet something belonging to what they founded. To say that Freud founded psychoanalysis does not (simply) mean that we find the concept of the libido or the technique of dream analysis in the works of Karl Abraham or Melanie Klein; it means that Freud made possible a certain number of divergences – with respect to his own texts, concepts, and hypotheses – that arise from the psychoanalytic discourse itself. (1998 [1969]: 218)
The same is true of the ‘think/no-think’ protocol and its affiliated diagrams: Try to read them outside of Freud’s invention of the economy of fraying, and they become downright unfathomable. Reading Freud for diagrams, with a view to how innovation occurs at the fringe of contact between media, tells us a less progressive story about the birth of psychoanalysis on the ashes of 19th-century neuropsychology. It enables a more dynamic approach to the history of science – and to the way that newness comes into the world – than neat antagonism between psychoanalysis and the neurosciences has traditionally allowed for.
Footnotes
Acknowledgements
I thank all members of the November-13 Program for generously sharing their time, knowledge, and insights. In addition to audiences at Columbia University in New York, at the Max Planck Institute in Halle, and at the School for Advanced Studies in the Social Sciences in Paris, I am indebted to many insightful suggestions by Nadia Abu El-Haj, David Scott, Brink Messick, Jérôme Sackur, and the reviewers for TCS.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research for this article was supported by a Mellon International Dissertation Research Fellowship by the Social Science Research Council, a Doctoral Dissertation Research Improvement Grant by the National Science Foundation, and the Camargo Foundation in Cassis, France.
