Reverse-Engineering Touch: Sense-Making and Making Sense with Prosthetic Neurostimulation

Abstract

At the frontier of research in neuroprosthetic limb technology, experimenters are developing systems for sensory feedback (prosthetic touch). Drawing upon two years of ethnographic fieldwork chronicling neuroprosthetic clinical trials, I interpret neurostimulation experiments as a reverse-engineering: in which efforts to engineer sensory feedback recursively inform basic scientific understanding about touch itself. In this article, I analyse reverse-engineering as technoscientific practice, phenomenological experience, and mode of knowledge-making, in which gaps between natural and artificial (or ‘electric’) touch get sustained and undone. In tracing the ways touch becomes constructed, abstracted, and experienced – including through phantom sensations and syn-aesthetic description – I examine how multiple coinciding versions of touch get produced at the level of the nervous system. I analyse the consequences of this multiplicity on theorizations of human and nonhuman touch, haptic experience, and touching subjects, sustaining epistemological and ontological openness amid efforts to pinpoint touch as a site of knowledge-making.

Céline, a neuroscientist researching human touch, and Anders, an electrician who lost his left arm in a high-voltage accident in 2011, perched side-by-side at a long desk in the laboratory, experimenter and experimentee, both oriented towards a bevy of computer screens. On the desk, amplifiers, cords, and other miscellaneous engineering equipment were strewn haphazardly, a maze of electrical channels. Anders removed his prosthesis, placing it on the tabletop next to him. Directly into the titanium abutment protruding from Anders’ residual arm, where it ended just above the elbow, Céline plugged a single cord, connecting his internal wiring to the external amplifier. Anders’ body was, in that moment, literally ‘hooked up’ to the computer.

Absent of their no-longer hand, yet still retaining remnants of the nerves that used to innervate it, the anatomy of patients’ postamputation bodies offered a canvas for sensory experimentation. Their ‘wires’, as I heard one experimenter call the nerves, were cut in a way that could couple with another electrical system. Céline and Anders were engaging in experiments aiming to understand the neural mechanisms and felt sensory experiences of peripheral neurostimulation, ¹ delivering different kinds of electrical currents (waves) directly to Anders’ nerves. These neurostimulation experiments were both informing basic neurophysiological research on touch and aimed towards engineering prosthetized sensory feedback for a new generation of neuroprosthetic limbs.

Anders was one of six patient-subjects in a highly experimental clinical trial run by a biomedical engineering laboratory in Sweden developing neuromusculoskeletal prostheses – surgically integrated with nerve, muscle, and bone, steered intuitively through a person’s neuromuscular contractions, with the capacity for sensory feedback (touch) (Ortiz-Catalan et al., 2020). For 24 months, I conducted ethnographic fieldwork chronicling these trials, following these patient-subjects and their devices across laboratory, clinical, and home sites as they lived with their intimately integrated prostheses – both subjects of experimentation and subjects of their own daily lives. The larger project from which this article stems examines the ‘everydayness’ of frontier science by tracing how the making of neuroprosthetic science depends upon the mundane intelligence of the domestic lives and embodied sensory knowledge of its first patient-subjects. In this article, I focus on how embodied sensory knowledge – touch specifically – was constructed and abstracted in the laboratory through a series of neurostimulation experiments. Observing these experiments throughout my fieldwork, I witnessed experimenters’ expectations intersect with patients’ articulations of their sensory experiences, a process of constant revision and titrating across the inchoate terrain of human touch.

Each of the patient-subjects in the trial had an osseointegrated implant (titanium rod) protruding from their residual limb, transecting the skin, connecting their skeleton to the device. ² In addition to offering mechanical attachment for a prosthesis, the rod was also a conduit, a ‘plug’ of sorts, making the body’s interior electronics accessible to the exterior. This plug connected to the wiring of surgically implanted electrodes sutured onto patients’ muscles and ‘cuff electrodes’ wrapped around their ulnar and median nerves (Ortiz-Catalanet et al., 2020). While several teams around the world had experimented with tactile feedback by placing electrodes on the surface of the skin, just two – this biomedical engineering laboratory in Sweden and another in the United States – implanted electrodes beneath patients’ skin, communicating directly with the body’s nervous system. Sensory feedback could be delivered from an external stimulus (i.e. a sensor in a prosthetic thumb, or a signal amplifier in neurostimulation experiments) by way of an electrical current. These electrical currents produce different sensory qualia – subjective sensory experiences – in amputees’ bodies, often perceived as occurring in their phantom limb. Neurostimulation, in this way, engaged with the already-electric currents conducted by the body’s nervous system to produce a felt sensation. While relatively new to the field of prosthetic science (Sakas et al., 2007), such experiments exist within a long, somewhat fraught, history of experimenting with, studying, and communicating touch by way of electrical stimulation (Parisi, 2013, 2018; Paterson, 2021). ³

‘Ultimately, we’re looking to compare natural and artificial touch’, Céline had explained to me, emphasizing the words ‘natural’ and ‘artificial’ ⁴ as if to mark them apart.

By natural touch, I mean directly stimulating the mechanoreceptors of the skin. Versus artificial, electrical stimulation which bypasses these mechanoreceptors. There’s a huge gap in awareness and understanding between these two. Reducing that gap? Who knows how long it could take!

Céline’s characterization of the gap between natural and artificial touch – a distinction long inherent to experiments in touch research (Katz, 1925) ⁵ – points to ‘awareness’ and ‘understanding’ as capable of closing this gap, as opposed to widening it. Yet the touch experiments, together with what patients articulated as their felt phenomenological experiences with neurostimulation, called into question both the undergirding assumptions of that gap and, therefore, the work of closing it. While Céline and the experimenters mark this gap as something to be closed, I take the presence of the gap itself as an analytical and theoretical opening.

In this article, I use the gap as a framework to extend upon a body of critical work in sensory studies and touch research (Ladewig and Schmidgen, 2022; Parisi, 2018; Salter, 2022; Schmidgen, 2022) by further unsettling binaries such as natural/artificial, passive/active, human touch/prosthetic touch, surface/subdermal, tactile/haptic, and inside/outside, lingering productively with the in-between. It is precisely this undoing of binaries that sustains the complexity of sensory experience and human touch, underscoring the still-unfolding co-constitution of knowledge about touch with experimental attempts to render, model, and engineer it. In what follows, I ethnographically track the ways touch becomes constructed and abstracted through experiments in prosthetic neurostimulation. I do so to trace how ‘sense-making’ – and I mean this literally: making, or engineering, artificial sensation via prosthetic sensory feedback – is inextricably entwined with ‘making sense’: modelling and communicating touch in a way that is intelligible, justifiable, and practicable, not only to the patient but also to the experimenters. I am interested in the discursive nature by which patients, as both subjects and informants in the touch experiments, articulate not only their experiences with prosthetic sensory feedback, but also recursively inform – and destabilize – scientific, social, and phenomenological understandings about touch at the level of the nervous system.

As phenomenologist Maurice Merleau-Ponty observed, the dehabituation of senses (i.e. sensory input loss or restoration, as in amputation and prosthetization) offers the chance to ‘rediscover phenomena’ (Merleau-Ponty, 1962: 57), peeling back taken-for-granted layers in the realm of perception. Rediscovery, according to Merleau-Ponty, allows us to ‘reawaken perception and foil its trick of allowing us to forget it as a fact’ (Merleau-Ponty, 1962; Paterson et al., 2007: 21). The sense-making experiments Céline and Anders were conducting can be interpreted as attempts at rediscovery. The phenomenon they were rediscovering, which I follow ethnographically in this article, is touch on the level of the nervous system, a mode of haptic perception. This rediscovery sets in motion a process I analyse as a reverse-engineering: in which efforts to engineer a sensory feedback system recursively inform basic scientific understanding about touch itself. Reverse-engineering as a practice happens in and through what I describe as syn-aesthetic description – the recruitment of other senses beyond touch to describe, model, and create a language for the tactile.

These two analytics – reverse-engineering and syn-aesthetic description – are primarily empirical, emerging ethnographically. In conversation, they extend upon and permutate theoretical understandings of touch as ‘symmetrical’ (Ladewig and Schmidgen, 2022), ‘bipolar’ (Katz, 1925), and even ‘reversible’ (Merleau-Ponty, 1962), which underscore how a subject simultaneously touches and is touched by the world. Increasingly, as Ladewig and Schmidgen (2022) theorize, tactile agency extends from a property exclusively constrained to the human, to incorporating nonhuman actors, technical objects, and media (Schmidgen, 2022). In the ethnographic context I trace, basic research on the neurophysiology of touch meets engineering aspirations to facilitate prosthetic sensory feedback, producing a technology that both restores and creates anew haptic experience. In this sense, human and nonhuman touch are brought into close contact by way of the experiment as ways of knowing and learning about one another, revealing their similitudes and co-constitution. Concomitantly, this illuminates the multiplicity of what gets called ‘human touch’. In considering how neurostimulation contributes to theorizations of touch, I stay close to the articulations of the people, bodies, and subjectivities enmeshed in this experiment, sustaining epistemological and ontological openness amid efforts to pinpoint touch as a site of knowledge-making.

In taking seriously the sites of not-knowing, experimenting, and learning re-opened by reverse-engineering touch, I address an additional gap: between scientific and sensory studies of haptic touch (Jones, 2018; Parisi, 2018; Park, 2022), and the more limited tendency within cultural studies (Lafrance, 2018), psychoanalysis (Anzieu, 1989), and cultural anthropology (Howes, 2018) to primarily attend to touch as a skin-bound phenomenon, part of what David Howes has called a ‘dermalogical turn’. A conflation of touch with skin carries consequences for how touch is spoken about and understood as lived experience. These consequences gain significance in contexts of limb loss and traumatic injury, where anatomical rearrangements unsettle biologically normative encounters between bodily surfaces and objects. Neurostimulation, in which touch is engaged as an electrochemical nervous system phenomenon, highlights the limitations of this conflation, resonating with a long history of research in haptics dating back to the late 1800s asserting how tactility exceeds the surface of the skin (Parisi, 2018; Paterson, 2021). This requires theorizations of touch that sustain this complexity, critically incorporating the subdermal, electric nature of haptic sensing into understandings of the touching subject (Parisi, 2018). While sense-making entails a mode of world-making in and through the senses, knowledge-making about the senses can also have a world-unmaking effect: reformulating the way we understand the sensing subject and its exchanges with the world. In what follows, I trace how multiple versions of experiencing and knowing touch intra-act (Barad, 2003) through the practice of reverse-engineering touch, lingering with their multiplicities.

Traversing the Gap: Reverse-Engineering Touch

‘Surprisingly, we really don’t know that much about human touch’. Céline looked at me imploringly, as if making a confession. We were sitting at the lunch table in between neurostimulation experiments with Anders. Céline took the occasion to explain the background of the experiments to me, which spiralled into a larger conversation about a gap between ‘what we know and don’t know about touch’.

‘The first project I was ever involved with asked “how do we humans encode information about texture?” You would expect we already understood that, wouldn’t you? Sounds pretty basic, right?’ Céline asked me, searching my face expectantly. I wasn’t sure that kind of knowledge fit into my conception of what Céline called ‘basic’, but admittedly, I had never explicitly pondered it before. Her question appealed to the taken-for-grantedness of medical knowledge about the most seemingly basic of human sensory functions, despite persisting explanatory lacunae. My internal musings manifested in a disappointing silence. Céline delivered the anticipated punchline anyway: ‘But . . . we really don’t know!’ The gap, it seemed Céline was telling me, was not only due to the nascent emergence of engineered touch. It was also, if not equally, conditioned by the pervasive mysteries of biological touch itself: this, too, a moving target.

Céline enumerated for me ‘a bit of what we do know about touch’. That our bodies are attuned to perceiving different types of touch. That the skin of the human hand has four different types of mechanoreceptors that respond to changes in texture, pressure, vibration, and any stimuli of physical (inter/intra)action. That the hairless skin of the human palm contains the body’s highest concentration of mechanoreceptors. Tellingly, she used her own body to demonstrate. In ‘active touch’, the subject moves their skin freely along a surface, perceiving its nuances and textures (Céline drew her finger along the wood table). In ‘passive touch’, the skin is fixed and an object moves autonomously beneath it (Céline placed a sheet of paper below her finger and pulled it away with the other hand). In ‘intra-active touch’, a subject touches themselves (Céline stroked the surface of her forearm with her index finger). In the ‘loading phase’, the skin stretches upon contact with an object, then slides, creating a lateral force. A particular fibre encodes the skin’s stretching, which researchers believe correlates to the material’s friction.

Céline’s distinction between passive and active touch mirrors a discernment in touch research between the tactile (passive) and the haptic (active), where the former is often confined to the skin’s surface, while the latter incorporates the kinaesthetic perception of muscles, tendons, and joints, capable of sensing position and force (Jones, 2018; Katz, 1925). While Céline’s enumerations of ‘what we do know about touch’ began with the surface of the skin, they quickly moved subdermal, particularly when she recounted recent neuroscientific breakthroughs surrounding the mechanisms of social touch. ‘Now, I’m gonna get nerdy about it’, she said, leaning forward in her chair excitedly,

but there are actually nerve fibres just for social touch. There are C-tactile (CT) afferents that only activate when you are touching someone, at a certain force and speed – slow, like a caress – and they feel rewarding. They don’t activate for force, or vibration. So they are in our bodies for social touch.

Unlike the fast-responding nerve fibres, the CT afferents were found to only be activated by very light touch, the seeming opposite of pain fibres, their difference in speed likely an evolutionary adaptation to respond to threat (Olausson et al., 2010).

But even with all that. We still have quite a basic understanding of human touch. They’re still very much building the sensory cortical map . . . Every time the technology improves, we have to repeat the basics. And even still. There is so much we don’t yet know.

Such caveats and qualifiers, reigning-in the spectacle of the nervous system and its phenomena, were characteristic of the conversations I had with scientists and engineers during my fieldwork, evidencing the contingency of experimental knowledge production. Every result came with a disclaimer: ‘But we don’t really know X yet’. Updates and revisions were inherent to their epistemological process. While Céline’s conception of not-knowing about touch differed greatly from my own, my conversations with the neurophysiologists underscored for me the inherent multiplicity of inputs – down to the level of individual nerve fibres, and their (intra)action with sensory inputs, infinitesimal in their combinations – that comprised the sense of human touch. Recreating, or mimicking, such complexity seemed daunting.

Touch has been referred to as both the ‘forgotten’ (Maurette, 2018) and the ‘least understood’ (Price et al., 2000) sense, notoriously deprioritized in an Aristotelian sensory hierarchy governed by the visual, often undertheorized and taken-for-granted in relation to other senses (Classen, 2005: 1). Yet unlike other senses, touch ‘is not physically or anatomically assigned to a distinct sensory organ’ (Ladewig, 2022: 91). Instead, a complex choreography of skin, muscle, nervous system, and joints produce a range of experiences – from temperature to pressure to proprioception – composing what we call ‘touch’. Within this lacuna – this mystery of human touch – exists the somewhat-paradoxical pursuit of engineering prosthetic touch, even amid the enigma of its biological counterpart: the phenomenon it intends to mimic.

For my neuroscientist and engineering interlocutors in the laboratory, engineering and experimenting with sensory feedback, then, was performing a double work: (1) pursuing an engineered mimesis of biological touch, while also (2) generating information about underlying mechanisms of biological touch itself. In this way, the sense experimenters were engaged in a parallel sense-making, one in which the pursuit of creating what they called ‘artificial’ touch at least partially informed the understanding of what they called ‘natural’ touch. It was both an engineering and a reverse-engineering. By reverse-engineering, I mean that the endeavour to engineer an ill-understood ‘natural’ touch actively crafts a model for understanding a natural that is invariably produced by the technical. Here, the tired natural-technical (or natural/artificial, or natural/electric) dichotomy is turned on its head (Haraway, 2003). This effort of closing the gap between the natural and artificial had just as much to do with the pervasive mysteries of the natural as it did with the nascence of the artificial.

(Not-) Knowing Touch

Earlier that morning, Céline had explained to me a crucial shift over the past decade of touch research: from monkeys to humans. Monkeys, she told me, had their limitations.

Monkeys don’t have the same skin as humans. They have different receptors and fibre types. They can’t independently articulate their fingers. We can’t direct them to move objects. For monkeys, it’s all about the bananas, the reward. It takes years to teach them,

she explained, her tone of exhaustion suggesting she knew from experience. ‘And, most importantly’, Céline emphasised, ‘monkeys cannot tell you if it feels good or not, rough, soft . . .’

Her statement referred to the prevailing index for understanding sensory experience: language, something which monkeys lacked (at least in a way intelligible to humans), and humans possessed. Indeed, language is intimately entwined not only with felt experience (Das, 1996) and representational practice (Wittgenstein, 1922) but also with ontology (Moder, 2019). This turn from sensation towards language was familiar to me ethnographically (Middleton, 2020), but in a way that stirred unrest. Many of the experiments I observed emerged from the field of psychophysics, established by Gustav Fechner in 1860 to codify sensations (like touch and pain) using instruments of measure (Paterson, 2021: 1). In more recent years, psychophysics researchers have grappled to establish a lexicon for touch, an alternative to peoples’ free associations that generated ‘poor data’ (Guest and Essick, 2016). They reasoned that without being provided a language for touch, the individual may ‘neglect to report a sensation or emotional reaction’ or ‘the degree’ of the word (‘how soft did it in fact feel?’), or feel that the ‘nature of each word is ambiguous’ (Guest and Essick, 2016). Their ‘new lexicon for touch’ yielded a taxonomy of 262 terms pared down to 33 ‘sensory’ and 16 ‘emotional’ words (Guest and Essick, 2016: 142), on the premise that the two could be split. This split strikingly contrasts the outpouring of psychological and physiological research indicating that emotions often trigger bodily sensations (Nummenmaa et al., 2014) and vice versa. Language as a metric for sensory experience is often fickle and distorting, sometimes intractably entangled with somatic experience itself (Middleton, 2020). Still, it remained the primary tool at my interlocutors’ disposal to gauge the shifts in sensory experience taking place during neurostimulation.

Céline seemed to read my internal musings. ‘It’s hard, sensation. It all comes back to language . . . the challenge is to express sensation with language. We are not taking measurements in these experiments’, Céline acknowledged. ‘[Asking patients is] only about perception. The problem is everything is super soft’. I couldn’t help but notice the irony that Céline used a tactile metaphor (‘soft’) to describe the refusal of tactile experience to cohere to something measurable (which was, assumedly, ‘hard’). Sensory perception was soft, not hard: malleable, fleshy, and subjective. As another neurophysiologist in the laboratory explained to me, ‘It’s very hard to understand affective sensations. There’s nothing for us experimentalists to hold on to’. I imagined his words, ‘nothing to hold on to’, like hands grasping sand. Upon first grasp there is something, but its substance is ephemeral, fleeting; over time, it slips through the cracks between fingers until only granules are left.

In The Cultural Politics of Emotion, Sara Ahmed (2004) writes about the metaphors of ‘softness’ and ‘hardness’ as entanglements between language and affect, where softness becomes equated with weakness. Ahmed (2004) tends to these metaphors in the context of ‘how emotions work to shape the surfaces of individual and collective bodies’ (p. 1), how a nation’s political touch can be construed as ‘hard’ or ‘soft’ and how these characterizations are bound up in a social hierarchy of emotions. Here, Céline’s invocations of ‘softness’ and ‘hardness’ speak to the tactility of knowledge production – soft perception; hard measurement. While the context of Ahmed and Céline’s invocations differs – Ahmed was speaking of social collectives and Céline of individual bodies – language and emotion operate here in conversant ways. When Céline says ‘It’s hard, sensation’, she is speaking in sensory metaphor, as in: difficult, elusive, escaping the grasp of knowledge. Not only a social hierarchy of emotions, but a scientific hierarchy of measurement and observation. Knowledge-making about touch required scaling the jagged borders, brushing up against rough surfaces, and sometimes (often) failing to grasp the slippery substance of touch itself. ‘I’m always like, “Are you sure what you feel?” to patients’, said Céline, looking aghast. ‘How can we even ask them that?’ she asked, reflexive about the limitations of her own tools. Indeed, moving from monkeys to humans yielded its own challenges. While the presence of language afforded a tool for knowing touch, it introduced its own ambiguities, calling into question whether language could bridge the terrain between touch-as-perceived and touch-as-measured.

The Packet

The neurostimulation tests followed a predictable pattern: Céline selected the parameters of the wave, clicked a mouse to deliver the stimulus (electrical signal) to Anders’ nerves, and asked him to explain what he felt. With each delivery, Céline independently modulated the amplitude, pulse width, and frequency of the wave to understand how these shifts changed sensory experience. ‘Here is a questionnaire with a series of descriptors based on what patients have said it feels like’. Céline presented a paper packet, sliding it across the table to Anders. By ‘it’, I took Céline to mean Anders’ sensory experience of the signals delivered to his nerves, which he experienced as occurring somewhere on his phantom hand. Anders opened the packet with his biological hand, creasing the stapled corner against the laboratory table. I peered over his shoulder.

Page 1 featured the outlined sketch of a hand (dorsal and frontal), asking the patient to ‘locate the sensation’. Page 2, titled ‘Size’, depicted variously sized circles, labelled from 1 to 10 corresponding to their diameter. Page 3, headed ‘Form/Shape’, offered five options to identify the shape of the sensation’s area: circle, oval, line, ‘undefined’, and multiple small dots interspersed in space. Page 4, labelled ‘Borders’, enumerated four visualizations of how clear or undefined the borders of the sensation were as follows: (1) a clearly bordered, self-contained circle, (2) the same circle, with slightly fuzzy pixelated edges, (3) like 2, but fuzzier still, larger pixels creating a rougher edge, and (4) multiple discrete spots on a diffuse background. Finally, page 5, titled ‘Movement’, asked if the sensation moved across the skin, grew, shrank, or ‘telescoped’, connoting a shift from distal (near the hand) to proximal (near the body’s core).

The packet seemed to intend to equip Anders with tools of translation. The spatial dimensions indicated – sensations as shapes and symbols, description as drawing – offered a visual language beyond words, beyond common indexical representations of the body. A multisensory transformation occurs from the register of touch to the visual. Any act of translation requires a degree of abstraction and cannot be taken as a 1:1 representation; experience must inherently be distorted to be rendered inhabitable by another. The provision of choices allotted by the packet – what size? what shape? – also drew its own boundaries. In enumerating five qualia of sensation – location, size, shape, borders, and movement – it established and organized experiences into categories.

The packet also served as a performative (Barad, 2003) primer. Before the first stimulus was delivered, Anders was already anticipating his sensations in terms of location, size, shape, borders, and movement. ‘Sometimes it [the feeling] can move. And it can be more or less accurate’, Céline added. She seemed to be preparing Anders for the vicissitudes of what he might feel. I wondered if this might actually precondition Anders’ own anticipations, or even interpretations, of a not-yet sensory experience. I also wondered what Anders might have generated absent of the packet’s pre-established touch lexicon.

And yet, the packet had its own edges and limits, which the experiment in-practice repeatedly came up against. This friction surfaced new forms of language not already enumerated. When Céline delivered a low-frequency signal to Anders’ nerves and asked him to describe the feeling, Anders responded: ‘It’s like a bubble . . . in the blood’. His words – bubble and blood – were defiant; they did not exist in the packet. They turned to page 4 (‘Borders’), where Anders selected a level ‘3’ of diffusion (‘pretty diffused’), a fuzzy border around a circle.

‘Pressure? Vibration?’ Céline prodded.

‘It’s like . . . electric’, said Anders the electrician, who lost his arm due to electric shock.

I always wondered if electrical sensations were, for this reason, traumatic or especially painful for Anders. When I asked him, he always insisted they were not – he even once reported preferring an electric sensation over a ‘grabbing’ one. The notion that a person might prefer an electric sensation to a so-called more ‘natural’ one defied experimenters’ expectations of the desires surrounding artificial touch, which were always assumed to be as ‘natural’ as possible.

‘But no pressure or vibration?’ Céline seemed intent on these two words – pressure and vibration – spelled out in the packet before them.

‘If I had to choose’, Anders hesitated, ‘. . . vibration’.

Céline changed the wave frequency, holding its amplitude steady. ‘Now tell us if something changes’, she directed. At the new frequency, Anders reported, ‘Growing in area, more like a line/oval in shape now’.

From this point forward, Céline began to refer to this amplitude as ‘The Oval’. Shape here took on an additional meaning, becoming a name for certain wave properties.

Céline increased the frequency of The Oval again. ‘Same size?’ she prodded.

‘A little bigger, it stretches out’, Anders responded.

‘Still the same quality? Or more natural?’ Here, a built-in anticipation of the goal: re-creating a natural sensation.

‘Nah, the same . . . electric’, Anders replied.

‘Electric . . . uncomfortable?’

‘Nja…’ Again, the electrician shrugged, using a Swedish term of ambivalence incorporating the words no – nej and yes – ja.

Together, Céline and Anders discovered that changes in pulse width created a ‘jumping’ (Anders’ word) sensation between the lower and upper part of Anders’ phantom thumb. ‘Jumping’ was another sensation Céline hadn’t heard of until Anders articulated it; the packet provided no provisions for recording or expressing this, so Céline made a note. At a certain point, Céline strayed from the packet, asking Anders to point to his other, anatomically intact hand to identify a sensation’s corresponding location. She directed him to freely draw the sensation on an outline of a hand on paper. Céline even prompted Anders to direct her to draw on his anatomically intact hand with a marker, visualizing the sensation’s contours as if tattooing the sensation onto his skin. Céline and Anders continued this process, titrating the electrical current’s properties and titrating, too, sensation itself, represented by and beyond ovals, circles, squiggly borders.

Syn-aesthetics: Touching with the Senses

In practice, the touch experiments generated creative modes of description beyond language, straying from the representative tools initially available to them. When Anders traced his felt sensation on a paper palm, drawing its shape, borders, and movement, he was producing a guided visualization of touch distinct from verbal self-report. Drawing the shape of the felt sensation was just one point of departure. In addition, visualization extended beyond the packet. Anders invoked the visualization of touch differently during a break between experiments, musing, ‘If the brain were to see what I was touching . . . maybe it would feel a little less artificial or electric’. I asked him if the reverse could also be true; might it feel electric because he was literally hooked up to electrical wires, with an altimeter in view, his brain (to use his own words) seeing all this electrical mediation? ‘That could very well be’, Anders agreed. Anders’ invocation of what counts as visual evidence includes environment, surroundings, with the phenomenon of study. In drawing these contexts in, Anders raises the possibility that seeing the hyperelectrical set-up of the experiments could also contribute to a feeling of electricity; that visual feedback could entangle with felt sense of touch.

Sound, too, was recruited to apperceive and describe the quality of touch. Céline described to me the technique of microneurography, where thin needles probe single nerve fibres. ‘Touch is amplified. We can literally hear the sound of the nerve activity, the action potentials, as we probe for the nerve’, she explained. ‘What does an action potential sound like?’ I asked. ‘Schooooooo . . . oooo.’ Céline made a swooshing noise, her lips pursed. ‘And then you get a little closer, and you start to hear the sound of the nerve fiber . . . Bliiiiiippp.’ Onomatopoeia also works as translation, becoming nonverbal by way of sound (made first by the probe, and again by Céline’s vocal cords). Céline’s onomatopoeia reminded me of the ‘jooooo’ another patient made when describing to me the inaudible, invisible symphony of felt sensations in her phantom limb, which intersected with the sensations delivered to her nerves, sometimes making it difficult to distinguish between them.

Outside the touch experiments, in everyday life, patients were already relying on sound and sight as proprioceptive tools to perceive their prosthesis’ movement in space. Olof described how the hum of the prosthesis’ motor, as he extended and contracted the elbow, sounded distinct from the hand opening and closing. Seeing the proximity of the target object to the prosthetic hand grasping it was another ‘useful’ form of sensory feedback, as Anders put it. It is as if an inchoate sense of engineered touch must transcend to other sensory modalities to be apprehended. The packet, the onomatopoeia, the drawing, the recruitment of sight and sound, altogether point to a syn-aesthetics of touch underway. The experiments’ process of approximation and representation was syn-aesthetic not only in its recruitment of alternate sensoria to render touch legible, but in its attention to the aesthetics of touch itself.

If phenomenology is the philosophy of perception and embodied sensory experience, then aesthetics can be described as a philosophy of affect felt in the body. In Greek, aesthema means ‘feeling’ and synaesthema describes elements of emotion that can be sensed in the body. Before ‘aesthetics’ was abstracted into a philosophy of beauty, form, and taste, aesthetics was, in the words of Terry Eagleton, ‘born as a discourse of the body’ (Eagleton, 1990). Mark Paterson, in The Senses of Touch, describes aesthesis as ‘an important first step of conceptualizing sensation, movement and affect . . . of thinking of the ambiguity of touching and feeling, and of expanding the notion of the “aesthetic”’ (Paterson, 2007: 18). In her meditation on ‘aesthetics and anaesthetics’ through an analysis of Walter Benjamin’s work on art in the age of mechanical reproduction, philosopher Susan Buck-Morss (1992) defines aesthetics as ‘a form of cognition, achieved through taste, touch, hearing, seeing, smell – the whole corporeal sensorium’ (p. 6). In all these configurations, aesthetics is deeply enmeshed with somatic experiences of being in a body-in-the-world, as opposed to a disembodied gaze which takes bodies as decontextualized subjects. In relation to knowledge production about the senses, one mode of sensory experience invariably involved synthesizing the experience of the others, a corporeal sensorium.

Aesthetics in the context of haptics offers yet another alternative to linguistic representation, what I construe as syn-aesthetic description. Syn-aesthetic description emerges when a sense beyond the one in question is recruited to describe the fullness of a body’s sensory experience. It occurs when Céline makes the sound of the nerve’s action potential, or Anders tattoos the parameters of a sensation on the skin of his other hand. ⁶ Syn-aesthetic description reflects the ways sensory integration and convergence are entangled in the corporeal sensorium and the plastic nervous system, undoing the representative tendency to silo the senses into their respective modalities. Reverse-engineering touch – not as packet but as practice – operationalizes syn-aesthetic description to reveal the coexistence of multiple types of touch, that touch is felt in a myriad of sensory ways. It also serves as a reminder of lacunae in our explanatory and measuring models to sustain this complexity.

Phantom Sensing

An additional lacuna emerges between the physically ‘there’ and ‘not there’. In addition to these surgical-engineering interventions, and unlike anatomically normative subjects, amputees have a phenomenological counterpart with which to sense: the phantom limb. Amid one sensory deprivation and re-constitution, another sensory capacity surfaces, albeit one often marked by pain (Flor et al., 2006). Signals continue to fire from the resected nerves, perceived in a location not materially there, but still experienced by most amputees as phenomenologically real. The phantom limb as it is lived subverts a dichotomous division of ‘phantom’ and ‘real’, an implicit hierarchy which, as amputee and media theorist Vivian Sobchack writes, ‘privileges the latter over the former on the basis of solely objective criteria’ (Sobchack, 2010: 53). For some people who have lost a limb, particularly those who have lost their limb in a traumatic accident, their phantom limbs are often ‘frozen’ or ‘stuck’ in a certain position (Collins et al., 2018). This is how Anders reported experiencing his phantom when he wasn’t wearing his prosthesis: his phantom bent at the elbow, the fist clenched tightly.

‘The phantom hand is there the whole time, and it sends signals too. It vibrates and pulsates’, Anders explained. ‘When I do the neurostimulation tests in the lab, sometimes it’s like, “okay, do I feel the sensory feedback, or did my phantom hand just do something?” Anders’ sensory experience with his phantom limb is one example of other sources of sensory feedback that may coexist with, if not contradict, the sensory feedback delivered by neurostimulation.

In a cunning twist of language, Anders used the word ‘noise’ (in English) – often used to describe electrical static from the ambient environment – to characterize the sensations produced by his phantom hand. Phantom noise distracts from the electrical signals just as electrical noise may intercept with the properties of a wave, changing it, along with its intended effect. This created additional challenges of discernment for patients. Another clinical trial patient-subject, Malin, described to me a ‘threshold level’ at which the delivered sensory stimulation must overcome her ongoing phantom sensations to be perceptible. In this sense, the phantom works both with and against neurostimulation: at times a conduit for the experience of electrically stimulated sensation, but also a potential interrupter of the signal itself. Distinguishing between and among these functions and distractions is, in the words of patients, ‘confusing’.

Olof, another trial subject, framed the ambiguity of phantom sensation and neurostimulation with a question: ‘Is the sensation made by the machine, or is it my brain?’ This cunning question slices to the heart of the issue of engineering touch, blurring boundaries of whether a sensation originated from within (nervous system), from outside (machine), or their interplay. It also highlights the experimental imperative facing both experimenters and patients to differentiate between the two, despite their blurring. Yet the two groups reacted differently to this indiscernibility. The inability to tell – machine or brain? – could be confusing and disorienting for patient-subjects. The experimenters interpreted this confabulation between phantom and neurostimulation sensations as a promising sign of their integration: the brain adapting to the sensory feedback from neurostimulation, interpreting it as its own. For them, it wasn’t just random noise. Here, the phantom troubles the bifurcation of natural and artificial touch. Noise, in all its static, seems to actualize the fuzziness that is a property of both the body’s and the machine’s electrical systems.

Olof’s question remains provocatively unanswered. It brings us face-to-face with the instability of these two categories (machine and nervous system) in practice, called into question by the phantom limb. It implores us to consider whether a phenomenon like touch could occur in the space between – a property not entirely of machine nor of nervous system, but rather of their interplay.

Nature in-the-Making: The Body Electric

Having gone through each page, Céline looked up from the packet at Anders. ‘There’s one final question’, she said. ‘Does it feel natural?’ The packet’s questions had been pointing here all along: towards the ultimate goal of differentiating between natural and artificial. This goal was embedded in what touch experimenters in the broader neuroprosthetics field aimed to elucidate: can a natural-feeling sensation be produced by artificial means? In other words, what were the limits (and possibilities) of engineering human touch?

The central question – about the feasibility of one day engineering ‘the natural’ and the relative importance (or not) of striving to do so – was one I discussed countless times with experimenters throughout my fieldwork. They openly debated the utility of extrapolating information gathered from amputees to ‘healthy’ (anatomically normative) subjects. As Céline explained,

The big difference between the control [non-amputees] and amputee patients is that in amputees, a nerve has been cut. And they don’t have a hand. So they have different properties, and are not entirely reproducible. So then does it even make sense to do it?

In calling into doubt whether such extrapolations could be made, Céline’s question seemed to insinuate doubt as to whether the engineered sensation could ever be considered a stand-in for organic sensation.

The principle investigator of the trials – an engineer who, in contrast to Céline’s research efforts to elucidate the basic neuroscientific principles undergirding human touch, was more concerned with building a functional sensory feedback system to improve prosthetic control – often voiced hesitancy, even scepticism, when he spoke about the pursuit. He regularly distinguished touch from sensory feedback, a nod to the limitations of the technology in which cruder dimensions of touch (pressure, vibration) were perceptible while its subtler vicissitudes (temperature, pain) were not. ‘To create a natural sensation’, he explained to me, ‘you need a convergence of multiple factors. To be completely biomimetic would require converging a wide array of different nerve fibres, rather than stimulating just one. Only then, after some time . . . will it become natural . . . or as some say’. His rendition of how the body produced natural sensation sounded to me like an impeccable engineering feat in itself.

Rarely are dichotomies as neat as they appear; the natural/artificial is no different. The dichotomy of natural and artificial (exchanged with ‘electrical’) is troubled by a well-established fundament of basic neurophysiology that emerged around 1850: ⁷ the human body is itself electric (de la Peña, 2003; Finkelstein, 2013; Helmreich, 2013). The body’s internal wiring is meticulously orchestrated, and its electricity is not only passive but generative. Cells conduct electricity, pumping differentially charged elements (sodium, calcium, magnesium, potassium) across their membranes. Muscle fibres, when contracted, produce action potentials. The neuron is often described in electrical parlance: axon, dendrites, and synapses channel electrical impulses, charged by exchanging electrons.

Therefore, collapsing the artificial with the electric in the context of prosthetic neurostimulation is misleading. Reverse-engineering human touch by way of neurostimulation operates upon the very electrical wiring of the body’s central and autonomic nervous systems. The natural is defined, if not produced, by its electricity. Acknowledging this turns the dichotomy – natural and electric – on its head, revealing striking similitudes. This deconstruction creates the sliver of an opening: one touch experimenters could step into, exploiting the already-existing electrical channels of the human nervous system.

Thinking with Reverse-Engineering Touch

Reverse-engineering touch acts upon what was already a vitally electrochemical way of being-in-the-world (Barad, 2012; de la Peña, 2003; Jones, 2018). In the absence of fleshy fingertips or nerve endings, human touch became approximated by engineering interventions with the nerves (sensory feedback). This ethnographic context of reverse-engineering reinforces just how much goes into producing touch: the corporeal sensorium (with syn-aesthetic modes of sensing), phenomenological memory (the phantom), and electricity (the nervous system). Still, this electrochemical, haptic, and ontologically multiple way of understanding touch (Jones, 2018; Ladewig and Schmidgen, 2022; Parisi, 2018) remains somewhat at odds with discursive tendencies in cultural studies (Lafrance, 2018), psychoanalysis (Anzieu, 1989), and cultural anthropology (Howes, 2018) to speak of touch primarily in relation to the skin. ⁸ The skin (Anzieu, 1989; Montagu, 1971) has been referred to as the ‘organ of touch’, much like the eyes are the organ of seeing, the ears the organ of hearing, the nose the organ of smelling. Touch has been described as ‘incarnated in its atypical, all-encompassing organ, the skin’ (de la Bellacasa, 2017: 99). Conversely, neurophysiological and engineering (Park, 2022) as well as historical (Classen, 2012) accounts of kinaesthesia, proprioception, and haptic sensing remind us the skin was never an all-encompassing organ-mediator of touch (Ladewig, 2022; Parisi, 2018: 17). Touch at the level of the nervous system opens-up ways of conceiving how and where we touch beneath the skin: a subdermal touch.

These discursive differences, I argue, carry more than semantic consequence. Insistence on the skin as the organ of touch performs similar foreclosing conceptual work as insistence on natural touch as assumedly opposite to the electric. It also forecloses other ways of being a touching body always-already underway, regardless of the presence (or absence) of a surface. My intervention builds upon a growing body of critical work in sensory studies of touch (Ladewig and Schmidgen, 2022; Paterson, 2022; Schmidgen, 2022) by asking what gets reified – about touch, but also about the sensing subject – when touch discourse glosses over the subdermal, electrochemical nature of touch. In this glossing, certain dichotomies (natural/electric in particular) and reifications (about the touching subject as bounded) are encased.

‘Reclaiming touch’, as Puig de la Bellacasa (2017) calls the epistemological project, demands a critical reorientation to the skin as boundary object. In A Cyborg Manifesto, Donna Haraway (1991) posed an evocative question that would inspire generations of future feminist theorists of embodiment: ‘Why should our bodies end at the skin, or include at best other beings encapsulated by skin?’ Haraway’s question signals to how framing the skin as a boundary object marks interiority and exteriority – separating not only the body, but also the subject, from the outer world. The question I want to follow, made ever more pertinent by the contexts of amputation and neuroprosthetization, fomented by over a century of research in haptics, is a permutation on Haraway’s question. Why should touch end (or begin) at the skin? This resonates with the ways in which human touch has, as Ladewig and Schmidgen (2022) have argued, ‘remain[ed] bound to the common “Cartesian” notion of the material and/or corporeal’ (p. 7). A Cartesian conception of touch – and its inherent distance built-in between the object and subject of touch – collapses as its electrochemical nature is foregrounded.

Indeed, touch itself constitutes a gap, a distance (Schmidgen, 2021); ‘not connection but interval’, as Marshall McLuhan writes, producing ‘a space which resonates’ (McLuhan, 2003: 280). This gap complicates a Cartesian understanding of touch as a collision between surfaces, skin and object, human body with outside world, given further resonance in ever-multiplying ‘contact zones between biology and technology’ (Ladewig and Schmidgen, 2022) permeating the digital everyday. In this ethnographic case of reverse-engineering, human and nonhuman touch are brought into intimate contact, a way of knowing and learning more about their symmetries (Ladewig and Schmidgen, 2022), but also of elucidating the multiplicities and complexities undergirding what gets called ‘human touch’. In the process, their entanglement is revealed.

‘Theorizing, a form of experimenting’ Karen Barad (2012) reminds us, ‘is about being in touch. What keeps theories alive and lively is being responsible and responsive’ (p. 207). To be more responsible and responsive to touch, cultural studies, psychoanalysis, and cultural anthropology can think more synergistically with its haptic and plastic qualities (Jones, 2018; Parisi, 2018; Paterson, 2022), that complicate the ‘dermalogical turn’ (Howes, 2018). This requires not only an undoing (or at least un-privileging) of the skin as the thing that does the touching, but also as the container of the subject. A lens on reverse-engineering helps us see that this flattening does not only exclude other ways of touching and exchanging touch (as in the case of amputation and prosthetization). It also confines theorization of all touching bodies and subjects, in vibrant electrical entanglement with their worlds.

What inroads, then, does reverse-engineering touch offer to knowledge about human touch – the elusive ‘natural touch’ to which experimenters compared their results, phenomenological experiences of touching, and the messy natural-electrical terrains in between? And how do these impact the ways we might theorize an electrotactile subjectivity, what Parisi (2018) has called a ‘haptic subject’ (p. 18)? The import of reverse-engineering extends far beyond neuroprosthetics and engineering fields, contributing to a growing body of work in critical theory, sensory studies, and touch studies on the wide range of touch encounters suffusing the digital everyday (Ladewig and Schmidgen, 2022; Salter, 2022), where bodies and haptic technologies co-inhabit and co-create touch encounters. Reverse-engineering touch becomes not only a way of knowing touch epistemologically, but a site for critiquing normative and ingrained categories of self and other, natural and electric, skin and subdermal.

Engaging touch at the level of the nervous system, reverse-engineering in practice reveals many coinciding versions of touch underway, experientially and epistemologically, sustained by a plethora of intra-actions. With language, through its affordances and limits. Beyond language, through syn-aesthetic description. Through visualization, in drawing touch as a shape on paper or tattooing it onto residual skin. In multisensory convergence, by recruiting sound and sight to confirm its presence. By way of the phantom, its neurological memory made alive through its pulsation and pain in the present. Within and among people, created in the social, relational, intersubjective space between touching bodies and personhoods. Recognizing these different strata and sensory vicissitudes of touch is a way of undoing the tendency to flatten touch: onto the surface of skin, into a particular organ, or inside the boundaries of the normative contained subject as somehow separate from the electrochemical entanglements of being-in-the-world. Neurostimulation, phantom sensations, reverse-engineering, syn-aesthetic description, all serve to remind us of the many ways touch can open up to us as ways of being and knowing, even if these ways are yet-to-be-known or resist explication. Even as they pulse in each of us.