Anti-computing punctuates computational advances; the rush ahead, the demand to be wanted, the claims for progress, and for progress as automatically good. Anti-computing is a pause, a stop, a refusal, an objection, a sense, an emotion, a response, a popular campaign, a letter, an essay, a code-work, a theorist, a sensibility, an ambience, an absolute hostility, a reasoned objection, a glitch, a hesitation, an ambient dislike. It may be articulated by a human, a crowd, a network, or by a program that refuses to run. It may also be an element of an assemblage containing many other elements; some of them in conflict.

Anti-computing takes many forms. It is unique to its moments of emergence, but is also characterized by recurrence. Many (though not all) of the components of anti-computing today are familiar even if they arise in response to a specific formation ‘as if new’, even if the claim is that the issues they address are graver than ever before, entailing qualitatively higher – even existential – stakes. Anti-computing may emerge as something felt and believed, something cultivated. It may be presented as a rationally adopted position or critique, a feeling (perhaps a yuck factor), or it may be an operation. It is articulated as an individual view, a collective one, as a one-off, and as a circulating and time-travelling discourse (operating synchronously and diachronically – albeit in the latter case in disrupted ways). As a human response to the global embedding of the computational it materializes in heterogeneous ways, but since humans, after all, are not separate from the machines with which they co-evolve, this materialization entails technologies as well as humans. It is indeed partly technological.

Anti-computing is all of these things at once because it is part of, and responds to, a formation that is itself pervasive, intrusive, ambient, emergent, installing, materially heterogeneous. I am talking, of course, of the computational, and of computational culture; about a history, about informational capitalism now, and about the computationally saturated and only partly foreseeable future. Computers have spread. They have spread further and faster than earlier human generations imagined possible – or that earlier generations of computer appeared to make feasible. There are over two billion computers in the world, and many billions more embedded control circuits. There will be 20 billion sensors in the world by the end of the 2020s, the decade we have already entered as I complete this book. The numbers become meaningless – though their sheer weight remains significant. The extent of the computational has broadened and deepened; networking the world, entering the body, reaching into Space.

In view of this relentless expansion, campaigns against computerization, anti-computing critiques, hostile interventions, writings ‘against’, in the academy and in popular culture, might be judged to have comprehensively failed. Certainly, if the measure of success is less not more computing, then they haven't worked. Anti-computing arises within horizons that have been – and still are – dominated by acceptance of the computational and its expansion, whether this acceptance is constituted as compliance, enthusiasm, indifference, acceptance, disavowal, or fatalism; all these have surfaced as responses over the seven decades or more since early computers first made an appearance in public, during which time computerization has been a live social issue. Anti-computing is, viewed in relation to this massive and apparently unstoppable advent, and in view of mass acceptance of it, an exceptional response to computerization, relatively rare, relatively rarely consistently expressed, even more rarely persisting, or becoming an organized movement. Of course, since the late 2010s and now in the 2020s, there has been an efflorescence of commentary – governmental, cultural, social, academic – expressing or exposing hostility and concern about the digital. Arguably we are in a renewed age of amplified automation anxiety (Bassett and Roberts, 2020). Consider the endlessly multiplying channels across which this anxiety runs the (digital) production processes that make it (increasingly digital), or simply consider the growth in the number of sensors in the world since, say, Jaron Lanier's 2010 You are not a Gadget, which might have been an automation anxiety outlier and the latest internet scare book to be up-marketed on Amazon.

Given the above, it is legitimate to wonder whether critical thinking, writings, campaigns, or interventions of other kinds ‘against computing’ matter. Or how they matter. For instance, do they make a significant intervention, are they significant historically, do they persist; how do earlier forms of anti-computing connect with contemporary anti-computational turns? Or, to introduce another register, do they suggest ways in which a different attitude towards a techno-socially saturated future and present might be cultivated? Could this be one which responds to a situation in which computational capitalism has reached the limits of viable expansion (environmentally) even as its growth continues? This is one of the ways the world is being broken, and one of the ways we are not repairing it. What do different forms of hostility or refusal suggest about the terms of the exchange or interaction between computers and (other) material and immaterial elements: bodies, discourses, texts, and (other) machines, intelligent or not; and what kind of alternative propositions can they open up – even in, or by way of, their intransigent refusal to conform?

The wager of this book is that historically arising and contemporary anti-computing formations matter and have more to do with each other than might at first appear. One example: contemporary surveillance concerns and hostility to comprehensive data capture by states are often explained as arising because of new data-capture techniques (big data), but they are also partly shaped by, and certainly resonate with, at least thirty years of fears of a data society. We have figurations and imaginaries to reach for when we discuss data surveillance, even as we declare it all new when we encounter it in a specific form, or at a specific moment, as a crisis not yet narrativized and thereby contained, let alone simply turned into information (Doane, 1990). You might say we reach for old stories, but we also need to ask how these stories materialize in new ways.

How, then, is anti-computing to be investigated? In particular, how are anti-computing moments of the past to be identified, given the thoroughness with which the relentless march of computerization obliterates traces of earlier resistance, hesitancy, and hostility in its rush to claim the present, and given its designs on, its foreclosure of, the future? Furthermore, once ‘found’, how are such formations to be explored? Addressing this last question demands not only identifying traces of the anti-computational, it also provokes a reconsideration of how the materials and discourses involved in its formation and articulation may be theorized and understood in relation to specific or located instantiations, and more broadly.

A dominant – teleological – understanding of ‘what technology does to society’ doesn't help to address these questions. It buries them. Computerization is often visualized as an inexorable progression. This progression, often taken to be synonymous with human progress, or understood as its leading edge, attracts support; at its most ardent this becomes quasi eschatological – something like the hope and desire for new gods, or a new regime of power to solve the problems of an old order. The latter is a sensibility evident in strong versions of singularity discourse, which see in artificial intelligence (AI) the chance to generate new kinds of post-human super-being. But a sense of the inexorability of computational growth also infuses visions of the coming quotidian, and is widely offered as a matter of fact, the apparently given and unarguable reality of the present and of the future. Computers, it is widely argued, are not only shaping life today but are increasingly also dictating the way the future will go. Granted, it may also be admitted by those who wish to follow it that this path is not entirely foreseeable. The complexity of interwoven processes of innovation, and perhaps also the much-heralded arrival of new forms of generative complexity, notably through advancements in machine learning, mean the logics of computerization are not fully possible to discern (at least by humans – and there is still nobody else). Despite this, however, the trajectory of much commentary on digital futures suggests a belief that – unknowable though they are – what is known is that these imprevisible logics will out. This belief, which is essentially teleological, turns any obstacles to the further advancement of computing – whether these concern narrowly technical issues, bottlenecks, reverse salients, or social tools including legislation – into temporary impediments to what will come, obstructions that will be routed around, sooner or later. Viewed from this perspective anti-computing interventions too, whether they come as intellectual broadsides, or emerge as circulating hostile discourses amongst various publics or as aesthetic or literary interventions, will only ever momentarily stall the disruptive energy of change, the forward march. They are, that is, deemed to be doomed to eventual failure or terminal subsidence. What are raised as problems concerning digital automation/instantiation – digital culture, AI for instance – may be taken seriously whilst maintaining this general perspective, but they are nonetheless also understood as proximate problems, as problems that ‘we’ may expect to overcome in time.

Consider the very different responses to the computational found in Joseph Weizenbaum's critique of the logicality/rationality equation, in Hannah Arendt's attack on the cybercultural idealists of mid-1960s New York, in F.R. Leavis’ attack on technologico-Benthamism, in anti-database elements of the US and UK counter-cultures (all discussed here), or consider the more recent writings of Morozov (2012), Pariser (2012), Lanier (2010), Noble (2018), Hawking (2015), respectively railing against the cloud, the crowd, algorithmically accelerated racial bias, and the rise of computer intelligence. All partake of something I am terming the anti-computational. All are – if the dominant logics of computational progress are accepted – sideshows to the main event. They have been or will be absorbed, set aside, sidelined, routed around, dismissed, put in their proper place. It isn't an accident that futile or mindless resistance is the popular (and distorted) meaning currently widely ascribed to Luddism, ¹ or that the term Luddite is now invoked far more often in relation to computational technologies than to its original contexts^{. 2} Perhaps this is why an avowed technological Luddism, proffered as a form of computational dissent, never had much purchase, though it was demanded in the early years of the public internet (see e.g. Webster and Robins, 1986) and is undergoing something of a recurrence (e.g. Lachney and Dotson, 2018).

A sense of the inexorable expansion of the computational is easily discerned as long standing, part of a structure of feeling emerging alongside the digital as it becomes pervasive in 20th- and 21st-century societies (albeit unevenly distributed). Consider the ‘drift’ (in fact a powerfully flowing current) towards the adoption by nation-states of bulk surveillance techniques no longer based on ‘need to know’ rationales but on exhaustive capture, which may be disliked by many but which are not being widely or systematically contested. Surveillance has already become embedded, an intrinsic part of the wider social architecture (and not only in the West), including the architecture of the built environment and of the self (see e.g. Andrejevic, 2015). Infra-structuralization produces a lockdown, not only cementing technological systems but also constraining the cultural imaginaries that might let us imagine them operating in other ways, let alone not operating at all. The teleological horizon, then, in the end acceding to a form of fatalism, leaves little space for the impertinent intrusion of the anti-computational. Anti-computing, whether understood as worthy or foolish or tragic, hubristic or pathetic, ill conceived, optimistic, or even heroic, is within this horizon ultimately always to be understood as a futile form of resistance, at least if it intends to do more than rectify an immediately correctable, temporally discrete anomaly.

In response, the wager of this book is that teleological readings of the computational, both those found in popular culture and those generated by particular framings of the technological, can be disputed, and that accounts that fail to register the significance of dissent and hostility can also be challenged – and should be. The anti-computational favours a dissonant reading of computational history, demands an alternative understanding of the technological, and can generate an alternative account of the actual and potential impacts of the computational. It does so as an act of writing, but also as an act of excavation.

Acknowledging that anti-computing moments falter, and that anti-computing often fails, and that there are many positive aspects to the computational, I set out to acknowledge and explore dissent, hostility, antagonism, doubt, unease, in many forms. Teleological accounts of computational technologies and the computerization of culture, taking various explicit forms, working in multiple registers, silently informing many accounts, need to be challenged. Both because they operate in powerful ideological ways and because (partly because of this) they tell only half of a larger story; anti-computational formations are an element of, and an element essential to the understanding of, those processes which Bernard Stiegler memorably defined in terms of a co-evolution between humans and technology and explored in relation to medium technological times in terms of grammatization or exteriorization and its potentials for good or ill (Stiegler, 2013).

These are times in which assumptions of unlimited space for growth or progress, or of humans’ ultimate mastery of our world, are revealed – in the flames of Australia, or in the melting of the ice, or in COVID-19 – to be the fictions that they always were. A theory of technical co-evolution, challenging assumptions of mastery and control by ‘the human’ over ‘nature’ by challenging the boundaries between them, specifically challenges technologically enabled visions of automatic or open-ended progress. The engagements between the computational and the environmental are complex and multilayered, but also brutally obvious; land poisoned by tech manufacturing processes, the rise of throwaway culture, labour exploitation around rare metal mining for smartphone components, pollution, the energy consumption of data lakes used in AI. These issues are explored elsewhere (see Solnit, 2007, 2013, Bauman, 2003, for early engagement, and emerging writing in ecological digital humanities for more contemporary treatments, including Yusoff, 2018, and Smith, 2011) and do not directly constitute the proximate subject matter of this book. However, the broad attack on the presumption that unlimited growth represents progress (and perhaps virtue), made for instance by Latour and Haraway in discussions of technology and environmental limits, resonates with many explicit critiques of the computational as the handmaiden of such forms of progress, and is present in a more amorphous form more widely (see Latour, 2011, Haraway, 2016). Work by Timnit Gebru et al. on the high environmental/energy cost of expanding data lakes used in AI sharpens the critique in relation to emerging technologies (Gebru, 2021).

Significance

The anti-computing formations I explore here are significant first of all because their very existence provides evidence that in earlier decades the path ‘forwards’ towards greater computerization has not been as direct as it appears to us in retrospect. Foucault (1972: 14) talked of the ‘tranquilized sleep’ of a conventional history and asked how this might be disturbed. Anti-computing formations punctuate those breathless but still somnolent accounts in which the ‘progress’ of the computerization of culture, viewed as the latest/last work of humans, is assured through the smooth and relentless application of (computer) science. An exploration going by way of the anti-computational exploits this disruption and theorizes its significance. Anti-computing can be understood in this respect as a partly theoretical construction or methodology, one which isn't quite archaeological in Foucault's sense, nor quite a media archaeology, but which shares with both of these some interests in discontinuity, series, scales, irreducibility, and looped connection, seeking out and exploiting the possibilities opened by interruption – and systematizing them where possible.

Anti-computing exposes other ways of seeing the history of computational adoption and expansion, and discerns ratios operating between humans and machines other than those most conventionally remarked upon. Of course, there are multiple ways in which dominant forms of computational culture – computational capitalism and its growth logics – are questioned. The anti-computational moment relates to other forms of resistance, for instance those that can be framed as doing otherwise (hacking ‘fixes’ to reopen apparently finished technologies, the adoption of alternative modes of use, turning data against its owners, all are good examples here and are also much studied; see e.g. Jordan, 2017). Anti-computing makes its point differently. Its modus operandi is dissent. It tends to prioritize refusal and critique over appropriation. This matters partly because, as is increasingly clear in relation to commodity culture and culture jamming, appropriation only too easily produces reappropriation – net autonomy was always going to be temporary, as Hakim Bey (1991) put it long ago. Dissent may (also) jam up the technological imaginary, but this time through forms of distancing, refusal, doubt, rejection. Anti-computing is not about playing with, but rather about not playing at all, or at the very least about questioning the logics of the game.

A certain irritation might be legitimate at this point; what is anti-computing? So far, I have defined it as a series of found and generated responses to the computational that wish to question some of the powerful assumptions about the inevitability of computational technology and computational growth and/as progress. I have suggested that they may be powerful in themselves, and that their existence disrupts a story that is all too easy to tell, that threatens to tell itself. I have also begun to talk about anti-computing as an excavatory methodology. I recognize that these two definitions provide only the beginnings of an answer, and they are intended to be partial. In this book the term is allowed – indeed encouraged – to morph, to expand and contract, and to change case. Indeed it is more than a term, although the matter of terms is where I turn to next.

Computing and anti-computing

To explore anti-computing demands some discussion of terms such as ‘computing’, ‘the computational’, and ‘computational culture’, given the multiple ways in which they are deployed. A computer is a machine that operates on data according to a set of instructions (an algorithm). Digital computing is a particular instantiation of such a machine; this distinction can be clearly made – but it needs to be recognized that the two terms are endlessly switched in popular culture. The computational is not the cybernetic, certainly not as this was defined by Norbert Wiener (1950), but some argue that the latter constitutes something like the basic architecture of the former. Competing definitions are common; computing ‘began’ with Babbage, or with Lovelace and programming, or with Hilbert's work on algorithms, or with Turing, or with the Turing Machine, or with Turing's paper ‘Computable Numbers’ (Fazi, 2018). Or, in deep media accounts, it began centuries sooner (Zielinksi, 2008). These definitions inform more or less contemporary analyses of computerization processes in society and culture; consider Wing on computational thinking (2006), Lanier on the computational (2010), and Berry on the computational turn (2011).

I employ some of these definitions here, chiefly using them in the following ways. First, computing is taken to mean something reasonably close to Turing's original definition of the universal machine, capable of calculating according to coded instructions (Turing, 1950). This draws attention to computing as a process and an operation which may be instantiated in many ways; on this basis the digital (digital computing) is a particular kind of computing, quantum computing another. It also allows that computational operations may be components of assemblages taking many physical forms; this is highly pertinent today as billions of embedded computer controllers animate and endlessly expand the emerging Internet of Things (itself a term now approaching senescence).

Second, computing, the computational, computers: all these are also allowed to be umbrella terms, designating assemblages widely understood to be computers, and/or operations understood to be computational at the time of this writing, or in their time. Today these include that expanded constellation of personal computers (PCs), tablets, cell phones, and cloud storage devices and facilities that constitutes the network central to computer culture: the internet and the platforms. To this are added those things and processes that are computerized (becoming computer controlled, or ‘becoming’ computational operations) and that are, in the process, transformed, often via a process that might be termed ‘sensorship’. The qualifying criterion for what constitutes a computational object or process isn't absolute and is contingent. In 2020 prototype driverless cars are often understood as computerized vehicles; their capacity to automate what has been viewed as a form of human manual and cognitive expertise – driving – is what marks them out. But digital circuitry is of course deeply embedded in all reasonably modern cars (even mine, which still has cassette technology and mechanical windows), even though these are not conventionally recognized as (defined as) computational objects. The sociology of media-technological innovation, with its exploration of the becoming invisible of technology through appropriation, which produces a culturally informed sense of the technical per se, has much to teach here (see e.g. Silverstone, 1994, Hartmann, 2009). Similar ambiguities arise elsewhere. Consider smart homes technologies. Fridges transformed through their computerization (embedded sensors, storage, control circuitry) into informational and media devices continue to perform their traditional chilling and storing functions. But what is milk to me is now information to a supermarket. The computational is a relational category.

Despite the flexibility of these terms computing does mean something, not only as a technical definition, indicating an assemblage with the capacity to undertake particular operations, but also in the shifting vernaculars of the everyday. Allowing for this I bring into my orbit discussions that, despite the vagueness with which they may adumbrate specific objects, or specific computational technologies, nonetheless focus hostility onto computing, or the computational moment, however they define it. This focus is sometimes a part of a larger critique: of technology in general, singularity in general, or modernity in general, for instance, but it is there. This focus shapes my own inquiries. Later in the book, for instance, I explore Arendt's treatment of automation, beginning not with The Human Condition (1998 [1958]), a work of philosophy exploring the viva activa, but by asking how Arendt fared among the cyber-enthusiasts and haters, industrial organizers, civil rights activists, and nascent tech industry professionals who argued with her face to face at a conference exploring the ‘threat’ (or promise) of new waves of specifically computational automation or ‘cybernation’ in the early 1960s.

One more definitional comment. I am wary of using ‘the computational’ as a term to replace ‘the digital’, at least if this switch is taken to mark an exact break point, since I don't believe it does. Terms such as computational, digital, postdigital are invoked as periodizing categories in cultural or critical theory and do have useful heft as aesthetic provocations (on the postdigital see Florian Cramer, 2014). However, their usage isn't precise. In the case of the digital and the postdigital, for instance, there is obvious slippage across registers marking the aesthetic and the sociological – and that produces the peril of an unexamined assumption that we're all in this together, and all in the same place. We should ask of the postdigital, post for whom? post where? post when? – and recognize that the response ‘it's post for me!’ often isn't good enough. Not least because it fails to recognize (post)colonial issues of historical discrimination and normative legibility. In another context the scholar and artist Lewis R. Gordan has discussed these kinds of universalizing labels in terms of the epistemic closures they produce, particularly around race (Gordan, cited in Evans, 2019), and this is apt.

The computational, then, is understood as a formation, one in which computational control logics dominate, but one which invokes and assembles many other actors in operation. The computational cultures that arise as the logics of computational assemblages expand through becoming operational are specific to particular times and places, which is to say they operate in relation to, and as an integral part of, specific political economies. Computational culture today is not necessarily isomorphic with neoliberalism as a political economy and a social order, although the computational is a key mode through which neoliberalism is organized and operates – and is also important in how it sees itself (as an information society, knowledge society, or, in so far as the label ‘computational capitalism’ is accepted, as a computational society).

Anti-computing instances

In this book anti-computing is explored from two directions. First via a series of investigations of anti-computing in specific sites and contexts, scattered across the seventy years or so of post-war computing. In each, the goal is to understand anti-computing in a particular social and material context, to ask what forms of understanding underpinned anti-computational interventions (how the technological was constituted as powerful, how it was understood in relation to political economic conditions pertaining, for instance, or what justifies my labelling of it as anti-computational). These investigations focus on some (of many) instances where aspects or instantiations of the forward march of the computational are attacked.

There are no pretensions to completeness here. They would be frustrated, since anti-computing is found in many places, takes multiple forms, and operates at different scales. Consider these various iterations: campaigns against computer surveillance, for instance early opposition to the Regulation of Investigatory Powers Act 2000 in the UK, were and are anti-computational, as were objections to the potential data capture operations of a single software package, the finance package Lotus 1–2–3 (an early example of a tightly defined anti-computational moment). Then there were the bundles of assembled new technologies that promised to ‘cybernate’ various forms of work in the 1960s, or the electronic page make-up systems central to shifts in the print industry in the UK in the 1980s; both aroused a hostility that was anti-computational. The UK print case is a useful example, since it illustrates how the anti-computational can draw itself up into and articulate antagonisms far exceeding the purely technological; the print industry wars were part of the Thatcherite attack on trade unions and working-class organization in the UK in the 1980s (Cockburn, 1992). The computational informs cybernetic routes to the post-human – and so warnings against these forms of post-human existence may usefully be explored in anti-computational terms. Or – bathetically perhaps – consider those on trains who prefer silence to electronic noise. Or those who prefer thick, rich, ‘real world’ contact to virtual interaction, or those who take detox ‘holidays’ away from the saturated computationality of everyday life (see Harrison, 2020), or who call for organic community over techno-scientific culture, or who critique technocratic rationality as that which will hollow out the political world. All of these positions and arguments are examples of anti-computing. Some of them are explored further in this book.

Hopefully, these examples already suggest that there is something impure about the anti-computational, something incommensurate about the various scales at which it operates, and something partial and rarely complete. This impurity interests me and is partly why I have not set out to research the most obvious cases. I am interested in interventions, often now largely forgotten, aired in very different sites, all of which took forms that, even if they were staged by academics, reached beyond the spheres of the academy as these latter are narrowly conceived of, and certainly beyond critical theory as an entirely contained project sufficient unto itself. Rather, anti-computing constellations, which entail an engagement with theory, are considered as they emerge in real-world contexts. The attempt is to produce critically informed rich descriptions. A series of events, writings, moments, that cross between theoretical and everyday registers, and that are not often explored for what they say about the computational, are reopened and explored as anti-computational, and explored through the optic this taxonomic division provides. This close-up view, somewhat set aside in the rest of this opening chapter, or at least not referred to in specific terms, constitutes the bulk of the work undertaken.

Anti-computing and its travels

The second concern of this book is to explore what connects apparently discrete moments of anti-computing that irrupt in response to the new. It should already be clear that the objects, assemblages, movements, writings, that constitute anti-computing are markedly heterogeneous, a term that, as Fredric Jameson notes, has had spectacular success in late capitalism (Jameson, 2015). Even in this heterogeneity, however, specific iterations of the anti-computational are often regarded as disjunct from one another and their connection to earlier forms often not recognized; it tends to be ‘new’ dangers, ‘new’ forms of sociality, ‘new’ kinds of addiction, ‘new’ kinds of AI that are feared or disliked, after all. This suggests that discontinuity wins out. On the other hand, there are also connections to be identified and explored. Anti-computing is multifaceted and diverse in its materiality, orientation, and scale, unexpected in its forms, unique in its instantiations, but it does exhibit forms of coherence, and exhibits traits held in common. Moreover, although the times of its emergence cannot be predicted, that forms of anti-computing will emerge and re-emerge over time can be anticipated, at the very least.

Thus, whilst discrete forms and moments of anti-computing are explored for their intrinsic interest, the question also tackled here is how to understand anti-computing as a formation that regenerates, takes up, takes on, moves on, and revives in various recurring forms across time, in irregular but nonetheless recurring circuits. How do these forms travel, how do they translate, come to be rematerialized? Does anti-computing itself, as a formation, have a form of coherence over time? Is there a way of doing technological history that can encourage recognition of this coherence? This last begins to ask questions about technology and history, and will return us, in Chapter 2, to an engagement with media archaeology. Briefly prefiguring that here, my argument is that a conditional and partial continuity, complementing the inaugural moments of new forms of computational dissent and confounding what might have appeared to be the terminal disappearance of some others, constitutes one of the defining dynamics of anti-computing. The question, then, is how to identify and account for discontinuity/continuity, to explore what appears to be – and what is – durable in the relationship between computers and the (techno)cultures they intervene into.

This proposition, constituting the second focus of the book and supplementing the concentration on discrete moments, also begins to redefine the anti-computational itself, since the latter becomes recognizable in part as exhibiting a kind of discontinuous continuity. Here work by the historian Valerie Traub, seeking to understand historical recurrence as well as synchronic emergence, has been found insightful. Exploring the periodic rematerialization of various bodily tropes across long stretches of human history, Traub sets out to understand how various perennial logics and definitions remain useful, across time ‘emerging at certain moments, silently disappearing from view, and then re-emerging as particularly relevant (or explosively volatile)’. As Traub sees it, such ‘recurrent explanatory logics … are subject to change’, but also ‘evidence a form of persistence’ (Traub, 2007: 126). Older tropes, apparently consigned to history, may come back to life to ‘trespass’ on the new. Anti-computing may be explored, at least initially, in terms of these dynamics. That is, it can be viewed as a series of perennial logics, operating within the horizon of the computational, or of computationally influenced times, and disrupting them. Within this time frame, which is massively compressed compared to the long historical periods with which Traub is concerned, forms of anti-computing emerge and find force and purchase through various modes of materialization, and then fade in ‘relevance’ before possibly reviving and returning. These oscillating circuits or, borrowing from Traub once again, these ‘cycles of salience’ (Traub, 2007: 126, 133), are irregular and partial, and new irruptions always contain the discontinuous, as it were. However, in the sense that the return is likely to come, the anti-computational moment can be characterized as erratically predictable.

An initial taxonomy of the anti-computational

Anti-computing exists and is found operating in the world; this is the burden of the preceding section. Anti-computing, as I define and use it, is however, not only a real-world formation but also a methodology, a working thesis, and an investigatory heuristic. It is to be developed as a critical methodology or a ‘working concept’ (Foucault, 1972: 9) and the taxonomic operations set out below begin that task. They suggest commonalities, isolate matters of interest, and expose or suggest connections across time, location, and register. The task is to identify a series of initial taxonomies, identifying contingently stable and recurring logics or tropes of anti-computing.

Consistent with this, these categories have been constructed in response to what is found ‘in the wild’ (how popular culture might group various elements), in response to what my investigations of particular anti-computing moments across the decades suggest, and because particular principles of grouping might offer insight into how the formations under investigation can be understood. What results – some possible working taxonomies – are then themselves set to work, the taxonomic being recognized as a mode of categorization that is both descriptive and operational, a suggestive model with some performative force, rather than anything that will be fully realized. It becomes a means through which to consider rising forms of anti-computing, to ask how they are continuous and discontinuous with what has come before (as is further explored in the final chapter).

Polarities/binaries

Taxonomies may be built, or found in operation. Either way, they imply the finding or the identification of a pre-existing order – but are themselves a powerful ordering process, constituting a mode of interested (non-innocent) knowledge production. As Foucault famously taught us, or, rather, Borges did, social and other taxonomies may tame the wild profusion of things by the imposition of categories – which may be apparently incommensurate with each other, or which may overlap in various ways (Foucault, 1972, Borges, 2000). Moreover, the designation of categories may become the subject of contestation, or be part of a political struggle, a demand, as Foucault put in an article exploring critique, not to be governed ‘like that’ (Foucault, 1997).

Building an anti-computing taxonomy, it seems logical to begin by identifying a naturalized or dominant taxonomy of anti-computing, informing much debate around the computational in public culture but also found in theoretical writing, the chief characteristic of which is a blunt binary division. This organizes hostility to the computational into two categories, dividing hostility to computers arising ‘because computers are particular kinds of machines’ from hostility arising ‘because human societies are ordered in particular kinds of ways’. The first category concerns computer power, and the second the amplification of systems of human domination through the use of control technologies.

The first category is essentially ontological. Something intrinsic to computers, the computational in essence, taken variously as instrumental, indifferent, in-organic, in control, or simply as ‘alien’ (see Weizenbaum, 1976), provokes fears, anxieties, hostilities, and dissent. These fears expand as computational cultures themselves expand into new zones. The invocation of a familiar meta-discourse from within this category constructs or produces an equally familiar machine; this is ‘the computer’ that stands against ‘us’ humans because it threatens to control us. It is why, as the technologist Bill Joy put it about AI back in the 1990s, we should be very afraid of the future (Joy, 2000); the same message is now being repeated around AI singularity issues.

The second category, the ‘wrong hands’ category, is centrally concerned with political economy and human determinants. Its central concern is with what those in power can do with computers, and who those in power will be, or are – by virtue of that character of the computational (what Winner called political technology). It fears that those who control computer systems use them to accelerate control over the ‘rest of us’. This category attaches itself firmly to computing, but we also see earlier iterations in other media technological forms. Consider Orwell's 1984, where the evil media was television, or rather the all-seeing telescreen zooming state violence – and, by the way, fitness classes – into the homes of citizens.

Working with this categorical division, two distinctly different ways to account for the recurrence of forms of the anti-computational arise. In the first case a certain ontological consistency is presumed – computers continue to provoke a particular kind of hostile response because of what they are and continue to be. The second category opens the way to account for the recurrence of certain forms of hostility to emerging forms of the computational in social terms – for instance, via the argument that the society that has produced computers is itself, in a fundamental way, static, operating with consistent social logics or taking the same structural forms; that the same threats and problems with the computational, which are really ‘problems with society’, continue to irrupt is, then, not surprising.

This polarized tale, this binary division between computer ontology and social organization, has wide purchase. It emerges in popular discussions of technology and culture and finds echoes surprisingly often in techno-cultural theory and critique. It is powerful, operating as what Erkki Huhtamo (1997) might call a topos, or leading metaphor, in discussions of computer and society. Ultimately, however, these are false dichotomies and explanations that work with this model to account for the anti-computational. To push this somewhat, my claim is that sustaining absolute divisions between the ontological and the politically economic, here articulated as a division between what computers might do to ‘us humans’ by virtue of their ontology, and what (bad) uses humans (or human societal organizations peculiarly stripped of their material supports) might put computers to, is to support the unsustainable – in theory (where it produces a distorted account) and in practice or operation (where it operates ideologically).

To ask the (apparently) social question ‘What it is about the social systems within which computers have arisen that provokes the return of hostile responses to the computational?’ is immediately to provoke questions about the distinction between technology and society, to ask how the constitutive role of the computational within a social system can be understood. Langdon Winner's sense of technology as a political actor (Winner, 1985) would be the critical correlative here. The powerful ideological force of this binary taxonomic division should not be set aside. It is what lets actors such as Mark Zuckerberg claim it's culture, not technology, that produces platform violence. It is also what lets governments that have instantiated anti-Islamic policies and encouraged racial hatred insist that it's the (social media platform) technology, not the governmental policies, the social media gurus, nor the politicians, who are at fault for fundamentalism (which is not to say they are innocent either). However, I also want to insist that even as this division is found in practice, it is also in practice where this taxonomy can be seen to break down. There is rarely, if ever, a purely ontological basis upon which hostility to the computational arises, and no purely social basis either, once it is accepted that the social is techno-social. The interaction between materials and their constitution, and social structures and their operation in specific historical and located contexts, is misrecognized in this binary taxonomy, its purchase notwithstanding. Anti-computing, we might say, always has coeval motivations. Elaborating a useful taxonomy of anti-computing thus means recognizing as extant, but also rejecting as a naturalized division, and therefore rejecting as an appropriate tool for analysis, that binary division between categories of anti-computing that divide political and ontological factors.

Providing a better starting point to think about anti-computing and its characteristic forms demands keeping in play ontological and social factors of the computational, looking at durable and inaugural elements of each, and exploring their tensioned relationship in operation. For instance, to understand the relationship between/difference between hostility to database expansion in the UK in the 1970s, in the late 1990s, and hostility today, two sets of continuities and disjunctions need to be taken into account. Computation took very different forms at each of these periods, but material consistencies, the digital in this case, remain (this might be a matter of computational ontology); this is the first disjunction and continuity set, as it were. Second, what is required to build a useful taxonomy is both recognition of the specificity of the political climate of any relevant period in computing's social history and acknowledgement of underlying consistencies within the social order across this period – capitalism persists, albeit it in remodelled forms, and its logics are dominant. The principles informing this example are now taken up and contribute to the elaboration of a new taxonomy of anti-computing which sets out to crosscut and disrupt the binary logics of the first taxonomy whilst also recognizing its framing power.

A general taxonomy: eight forms of anti-computing

The new ordering now set out is composed of eight categories, constituting a provisional taxonomy of the anti-computational, an anti-computational catalogue. It is generated through consideration of extant instances and/or recurring examples of anti-computing across some decades. New forms of the anti-computational may find their place in this general order, deform or stretch it and thereby remake it, or constitute a global challenge to it.

Anti-computing (i): computer technology as control technology

Computers are control technologies. This category responds to that. It concerns itself both with computer autonomy and threats to human autonomy. Asking ‘who is in control – humans or machines?’, it fears that the answer to this is that computers themselves are ‘out of control’. But it also includes, and on an equal footing, concerns around the social order and questions of social control, power, and domination. Merging ontology and political binaries, it encompasses existential and political concerns. It fears the capacity of computers to deliver totalizing power to the state and to the market, cementing existing authority and power by ceding control of computational resources and foreclosing the open horizons of the future. Left critiques of technocratic rationality come into this category, for instance those reaching back to the later Frankfurt School (Marcuse in particular), but so do conservative variants of anti-computing, those who fear ‘the computational’ as that which will undermine the existing order of things where that order is regarded as just, fair, or simply desirable according to the situated position of those defending it. Hostility to groups who assert their counter-authority and threaten ‘law and order’ through computer use and appropriation often evidence this kind of anti-computing impulse.

Anti-computing (ii): computers becoming more lively

This form of anti-computing responds to fears concerning the displacement, and ultimately the replacement, of ‘the human’ by ‘the machine’. Central concerns include the rise of computer intelligence and the extension of various forms of computer-delivered automation. The existential variant of this category fears that humans have no place in the future, since the advent of advanced forms of AI will mean the replacement of humans, either by their machinic successors or by future humans sufficiently different from us to be unrecognizable as humans. A less elevated set of concerns and anxieties, also coming under this heading, coalesce around issues of replacement as they pertain to labour and work: white-collar unemployment through augmented intelligence of computers, service sector and other labour replaced through the extended reach of computers into areas traditionally regarded as human specialisms because they entail particular forms of emotional intelligence or human imagination are key – home robots taking on care, teaching robots undertaking education, for instance. Many of these concerns are captured as fears around the automation of expertise or the replacing of human expertise with computational expertise in new areas. Anxieties around the displacement of the body, in relation to the bio-digital – for instance, the ever-closer coupling between humans and their devices (partial cyborgization) – can also be fitted in here.

Anti-computing (iii): computerization and the hollowing-out of everyday life and social interaction

This category captures anxieties that arise around the computational transformation of the everyday world, of social interactions, and of everyday life and the ways in which it can be lived. They focus on the surveilled self and life, platform sociality, the hollowing-out of social interactions through their automation, felt perhaps in spatial and temporal terms (slow movements respond to this phenomenological threat), and on the loss of bodily richness and the pleasures of the informing sensibility of a located physical embodiment. They also concern the lack of depth of new forms of everyday life that distributed and virtualized interactions are felt to produce. Concerns about the attenuation of attention – the continuously disrupted life – also fit in here. Latterly, digital detox camps and other refusal programmes respond directly to these concerns.

Anti-computing (iv): computer technologies and the threat to human culture

Issues captured and grouped here include hostility concerning the rise of computational culture and the perceived threat to older cultural forms: narrative versus database structures, the logic of code versus the emotional intelligence of the human, the game over the story, graphic principles over aesthetics, data visualization over human art production, virtualization over materialization, mobile screens versus cinematic projection. One strong version of this says computerization debases culture even as/or even if it expands the latter's sphere of action/operation, or that it enables the completion of the project of technocratic rationality delivered as a whole way of life (the computational culture industries). Adding an ‘s’ here to pluralize human cultures and recognize difference is also important. Doing so identifies a related but distinctly different set of concerns: that the computational reinvigorates a form of binary thinking and calculation that standardizes and universalizes, so that the threats responded to here are of computer-assisted standardization, conformity, and non-situated universalism. Concerns around AI and bias find a place here. Finally, there is a strand of anxiety around computational engagements with the humanities and the arts which fears the replacement of human creativity with computational instrumentality.

Anti-computing (v): the general accident/catastrophe theory

This category groups concerns that neither computers nor humans control the outcomes of increasingly complex processes that computers undertake, particularly in relation to biotech but also in relation to network complexity, neural networks, emergence, AI. The fear is that these developments make the more or less accidental emergence of catastrophic outcomes drastically at odds with the programs’ (or programmers’) original intentions more likely to come about. Relating to cultural theorist Paul Virilio's (1997) theory of the ‘general accident’ and Beck's (1992) discussion of the ‘risk society’, this kind of anti-computing often emerges as a response to genomics and environmental issues, but also in relation to neural networks, AI agency, and algorithmic bias. It is invoked in more partial ways in many other spheres – program trading and market crashes are good examples. Once again, existential fears emerge but here tend to focus on the fate of the environment rather than, or as well as, foregrounding the future of the human.

Anti-computing (vi): horrible humans

This form of anti-computing doesn't ‘blame’ computers for the wrongs it discerns. Rather, it excoriates humans for taking advantage of what computers increasingly enable humans to ‘get away with’. Essentially illiberal, it says that the anti-social grounds of the machine, the distancing from real accountability that pervasive networks and the platforms provide, undermine collective moral responsibility and social norms that other social forms policed. The computer is not only a mirror to ourselves through which we see ourselves clearly and do not like what we see, it also enables or ‘encourages’ forms of action that humans would not previously have undertaken; ‘because they are there’, ‘because I can’, ‘because I can't be seen’. Computers are disliked because they come close to enabling (what is viewed as) a nascent human degeneracy.

Anti-computing (vii): standardization/quantification

This gathers together a cluster of hostilities and anxieties coalescing around epistemology. Computer operations are framed as determinate and reductive, and it is feared their expansion will produce a world organized according to ‘alien’ logics. This includes concerns that computing will render the world into quantifiable chunks so that it may be indifferently operated upon. ‘Solutions’ will triumph over ‘theory’ in the sciences and the arts (Anderson, 2008), quantification over qualitative judgement and hermeneutics, and the imposition of standardization, uniformity, and the ironing out of exception will triumph over variation, difference, and human judgement. This category operates in relation to, for instance, forms of knowledge production, computationally derived managerialism, bureaucratically led thinking. A criterion focused on knowledge issues, it also entails a concern about human ethics and their overruling or bypassing.

Anti-computing (viii): too much information

This category captures concerns over information overload, but also includes anxieties about the fetishization of information capture, the fetishism of ‘facts’, the overproduction of information – whether as data, text, image, and the prioritization of production and circulation over interpretation. This category of anti-computing is one of the most long standing. It finds new articulations around big data. Its obvious connections to other categories make it clear that the divisions between these clusters are not fixed or impervious.

This initial taxonomy is informed by a close engagement with a series of forms of anti-computational thinking over time. However, it does not reflect a robust, data-driven investigation. It is incomplete and can be disputed. Some people I have shown this to have suggested another one or two (or more) categories – although no alternative mapping provided a more total ‘solution’, nor even a more complete map. It is necessarily provisional. It responds to different methodological approaches to material studied, it accepts the hybrid, it assumes crossovers will continue to occur, and that the power and force of vectors linking the various categories will change over time and may rewrite the whole. For these reasons, rather than despite them, it is a substantial improvement over the binary division already discussed, and not only because each category can encompass ontologically directed criticisms and those that relate to instantiated technologies within specific political economies. It enables an exploration of anti-computing, its temporality, character, rarity, and the forms it takes across time, that begins by assuming a complex relationship between the ontological properties of the technologies entailed in computational intervention and the horizons of social power and particular forms of social system (capitalism, late capitalism, neoliberalism) within which these operate, of which they are part.

The emotional register?

There are other registers that could be used to further elaborate a taxonomy of anti-computing, perhaps to produce a taxonomic series. The sort could be by results (effective or ineffective campaigns), genre (theoretical writing, public campaign, popular or public opinion, legal response), material (documents, campaigns, technologies), or along various time lines that might impose forms of periodization (anti-computing before personal computers, without wires, including mobile devices, after 9/11, before mini-computing, after the dot.com crash, since Web 2.0, and so forth. And there is always the emperor's robot dog …). I am not pursuing most of these here, in part because they tend to divide where I wish to join – for instance, I want to trespass between, rather than to demarcate, critical theoretical and other forms of anti-computing, and I am interested in how hostility travels through rather than defines an era, or is defined by it. There are, however, some alternative taxonomies that intersect in productive ways with the one set out above. The first, set out here in Figure 1, albeit in skeletal form, organizes anti-computing by emotional register.

Figure 1 An affective taxonomy of anti-computing

This ordering crosscuts the earlier eight-part taxonomic catalogue, as is evidenced in specific ways in some of the later chapters – for instance, in relation to revelation, revisionism, and expert witnessing in the case of a computer science insider who becomes a hostile witness (Chapter 6); anger in the case of F.R. Leavis, the unlikely subject of Chapter 5; or in relation to joy when the protagonist of a singularity novel celebrates embodiment over AI minds in Chapter 7. This list is suggestive, and again certainly not complete; perhaps no emotional mapping can be otherwise. It is also potentially misleading – at least, if it suggests that anti-computing can be comprehended primarily as an individually felt emotional response or as an individual response per se. What I am trying to capture, even when working with or through individual responses or stories, are forms of collective and ambient unease, the collective sensibilities, perhaps the minor-key structures of feeling, often as they are gathered up and expressed through a single example, or by an individual, that constitute anti-computing as a material social formation, articulated on multiple platforms and through many materials, within a dominant culture that is generally positive or accepting, or fatalistic about computational ‘advance’.

It's clear that an obverse list of categories, one that does not work through emotions but, rather, deals with categories concerning reason, might also be generated; anti-computing may be categorized variously as rational, thought through, analytic, deliberative, irrational, for instance. These cuts, however, rather easily fall into binaries (rational/affective, reason/emotion), again tending to stress division where I wish to notice joins.

A taxonomy of anti-computing computing?

A further alternative or complementary taxonomy could be built specifically around (or perhaps could be restricted to) materials. This would include those forms of computing (following N. Katherine Hayles and others and taking computing as undertaken both by humans and machines) that might be regarded as, or exploited to forward the goals of, anti-computing: glitches for instance, varieties of hacking, simple machine breaking, could be included in a broad spectrum anti-computational mapping. It would also be possible to categorize anti-computing in terms of the forms of technology or material or matter that ‘resist’ the computational (vinyl, shellac, paper, ink, for instance), or as a collection of forms of cultural acts or practices that in their materiality or form frustrate the computational; symbolic language with its ambiguity, the codes of the unspoken or unsaid, that which is only implied, glossolalia, face masks designed to disrupt face recognition, such as those developed by Zac Blas, for instance (Blas, 2012–14, Bassett, 2013). Some of these forms are discussed in later chapters. These groupings, whilst suggestive, are not pursued further here, but they do inform the whole. They might be seen to crosshatch the main grouping.

Categorizing critical approaches?

Anti-computing is heterogeneous in mode of address and form; notably, it is neither solely an intellectual formation nor solely a popular one. Various theoretical orientations, more or less directly engaging with anti-computing, might be permed out and captured under another set of headings (Figure 2).

Figure 2 A taxonomy of theoretical positions

These critical and theoretical orientations inform the eight-part taxonomic mapping and feature in it, but not neatly, often appearing in more than one place. That is, my general taxonomy certainly does not reduce to categories informed by these theoretical orientations, indeed it complicates them. All these other groupings, lists joined by what Bogost (2012) called the ‘gentle knot’ of the comma, can be unjoined and redistributed across all of the eight categories of my overarching taxonomy, where they will not sit still and do not fit neatly. They loosen it up, let it breathe, perhaps tend to let it begin to transform; the taxonomy after all is both a catalogue and a list, and the latter demands endless and ongoing elaboration. As Georges Perec, a theorist of everyday life and a lover of the automated sort, noted:

Critical orientations and situations

I now briefly turn to reflect on my own situated position – and my own take on the computational. My intention has been to develop a cultural study of and with anti-computing that is orientated by broadly historical materialist approaches and that substantially engages with media archaeology and medium theory. This general outlook produces certain starting points. Notably it generates an account framed in material and materialist terms, cognisant of overarching social and techno-social logics, understood or explored in their diachronic as well as synchronic aspects; the complex temporalities and continuities and disjunctures of the computational/anti-computational demand this. Technology, recognized in its materiality, is always also explored in relation to culture, and in relation to social power; that is to say, by a critical theorization of technology and culture that sees the two as co-evolving. This general theoretical/critical approach does not demand or entail either a sustained rejection of the technological, or its vehement adoption. Finding a reconfigured role for technology in the generation of new forms of future possibility – which might stand against the expectations of progress – is certainly not a matter of quantity, more of this stuff or less or that. However, I recognize that where I write from does have consequences. I have more sympathy with some of the anti-computing formations I explore than with others, and perhaps also read their significance differently.

This account turns around shards and fragments of computational culture, involving therefore a series of decisions about what to pick up and what to leave on the ground. For instance, thinking about questions of technologies and bodies, I selected as one object of inquiry the science fiction (SF) writing of Hanu Rajaniemi, read as a defence of embodiment over anti-human uploading fantasies, choosing to explore this rather than – say – concentrating on a discussion of the moral economy of the games-censorship lobby in the UK in the 1990s; Either of these could find a place in the kind of expanded taxonomy of anti-computing under development here. To me, the first seems to provide a more intriguing way of thinking through unease around ‘artificial’ bodies and practices than the second. Similarly, whilst much public attention is being paid to issues falling within the category I designate ‘horrible humans’ – cyber-terror, internet pornography, shaming are often accounted for currently in terms of the unleashing of the inhuman in the human, of letting beasts not gods into the machine – this has not been explicitly addressed through a case study; rather, I return to it in the concluding chapter to ask why and how it has exploded as an operative category and why, as a categorizing operation, it has become so pervasive and so effective in ascribing blame. Other shards were more demanding of attention.

Am I anti-computing? I have elsewhere considered the possibilities of silence and refusal and have certainly critiqued particular computational formations in terms of their neoliberal instantiation, amongst other things (Bassett, 2007, 2013, Bassett and Roberts, 2020). However, against the writing on the computational framed in terms of refusal that I have produced, I would set other work that articulates a conviction that there are other forms of the computational than those that dominate today, and that these are important, significant, creative, affirmative, live giving, and can be used as tools for justice. So, I am not declaring for anti-computing. And this is not a manifesto. Nor is it a call for de-acceleration, for a go-slow, and certainly it is not a call for revived forms of primitivism. On the other hand, an anti-computing impulse of a kind chimes with my desire to upset an established form of thinking about technological innovation, precisely the teleological progress narrative already outlined above, and to protest the multiple ways it is put to ideological use. Anti-computing opens a way to think critically about the claims and instantiations of the computational, and as part of this I seek to think about the discrete formations I explore in critical terms. So, if this is not a manifesto, it is, as Barad would put it, a cut (Barad, 2003). And I am intending to cut in.

To attempt a summation of the issues opened up here, which stand as a rationale for what is to come, and which also set out my own position, I close this chapter by suggesting that anti-computing is useful to develop and explore for six connected reasons.

• Hostility to the computational is a significant strand in the fabric of human–computer engagement as that tissue has developed over seventy years. It is a strand that is largely ignored in histories, and certainly rarely viewed as something systematic. It is worthy of attention because of this. It offers a different perspective on, and suggests other ways to gauge the implications and claims of, computational ‘advancement’. Anti-computing might be, somewhat in the manner of counter-factual history, deployed to raise the possibility of other possibilities.
• Anti-computing moments are part of the history of computing as a material cultural history and are worth excavating, in and of themselves.
• Anti-computing formations are at times strident and gain much popular acceptance. But they are also, measured against the overwhelming tide of acceptance of the computational insertion into everyday life, relatively uncommon …
• … But although there has never been a grand movement, an overarching narrative of computational dissent, nor a single defining moment that has persisted, various arguments recur, repeating and also mutating in each iteration. These circuits, giving insight into processes and forms of intersection between material cultures and their histories and social and political developments over time, demand further investigation.
• Exploring anti-computing can contribute to developing ways of critically understanding media-technological histories – which I read as asking questions about the intersection of questions concerning computational technologies, culture, and power.
• Anti-computing formations do not necessarily set out to directly challenge the power structures within which they are embedded; their orientations are varied. But exploring anti-computing formations can expose the ideological power and force of authorized versions of computational ‘advancement’ that constitute the dominant computational imaginary. The anti-computational turn which I make in this book therefore has critical intent.