What Judgment Forgets: Affective Heteronomy and the Asymmetry of Answerability in Digital Capitalism

Abstract

Contemporary critiques of digital capitalism have examined behavioural data extraction, algorithmic governance, attention capture, and platform regulation, while debates about responsibility have generally focused further downstream on consent, moderation, and individual conduct. The relation between these two moments nevertheless remains under-conceptualised. Digital infrastructures increasingly help organise the pre-reflective conditions under which significance, urgency, and practical orientation emerge; yet conduct formed within these environments is subsequently attributed to users as though those formative conditions were neutral background. To conceptualise this relation, the article introduces affective heteronomy: the systematic organisation of affectively charged orientations from which conduct arises, whereby solicited responses are experienced, and later attributed, as the subject’s own. Anchored in Adorno’s account of freedom, unfreedom, and the bodily addendum, the argument proceeds through conceptual reconstruction and immanent critique, bringing that account into conversation with surveillance and data colonialism, societies of control, algorithmic governmentality, affect theory, platform governance, and neoliberal responsibilisation. The central problem is neither influence nor responsibility considered alone, but an asymmetry between formation and judgment: the same order that helps organise the conditions of conduct later disregards those conditions when demanding an account. The article concludes that critique cannot abolish the entanglement of freedom and unfreedom, but it can require practices of answerability to account for the infrastructures that help form the conduct they attribute.

1. Introduction

Within critical analyses of digital capitalism, questions of subject formation and responsibility have often been pursued along separate lines. One body of work examines surveillance, behavioural prediction, extraction, attention, and platform power; another tends to encounter responsibility further downstream, through consent, moderation, sanction, and individual conduct. Contemporary digital life makes this division increasingly difficult to sustain. The architectures that organise how individuals encounter the world also multiply the scenes in which they are required to answer for what they do. The relation between these moments is the subject of this article.

Two movements must therefore be considered together. The first is the increasingly systematic organisation, under digital capitalism, of the conditions through which significance, urgency, preference, and practical orientation are experienced. The second is the proliferation of procedures through which subjects are called upon to account for conduct: clicking, agreeing, posting, purchasing, refusing, violating, or otherwise acting within platform environments. Neither movement is unprecedented when taken alone. The distinctive problem lies in their convergence. An order that participates in forming the conditions of conduct subsequently treats those conditions as neutral background when demanding an account.

Influence, of course, is not new. Human beings have always acted under conditions they did not choose and have always been shaped by institutions, histories, and relations that precede them. The argument does not rely on a fantasy of an unconditioned subject. Its concern is rather with the historically specific intensification and industrialisation of formative power. Recommendation systems, personalised environments, predictive analytics, and adaptive interfaces do not simply present objects to already settled agents; they help arrange salience, anticipation, preference, and impulse before deliberation fully takes hold [1,2]. Yet the subject formed within such environments reappears downstream as an answerable party, addressed through consent mechanisms, moderation practices, reputational metrics, and systems of sanction. The formative role of the order itself recedes from view.

To conceptualise this relation, the article develops the idea of affective heteronomy. The term names the systematic shaping, under digital capitalism, of affectively charged orientations through which experience becomes significant, urgent, desirable, threatening, or negligible, such that responses elicited within digital environments are experienced, and later attributed, as the subject’s own. Affect is not a synonym for private emotion. It names embodied and socially organised ways in which things come to matter and solicit practical response [3]. Heteronomy does not imply that agency disappears. It identifies the organisation of the conditions from which agency proceeds, while subsequent practices of judgment continue to treat those conditions as neutral background.

The article’s contribution is relational rather than merely additive. Accounts of surveillance and data colonialism explain why intervention in lived experience becomes economically valuable; theories of control and algorithmic governmentality explain how environments can be organised in advance of explicit decision; and platform-governance scholarship shows how the same infrastructures increasingly adjudicate conduct [1,4,5,6,7,8]. Critical-legal work on autonomy and heteronomy, meanwhile, exposes the fragility of responsibility once the social conditions of agency are taken seriously [9,10]. Affective heteronomy names the asymmetry disclosed when these moments are read together. Its novelty lies neither in claiming that digital systems influence conduct nor in denying that subjects remain agents, but in showing how infrastructures that help weight the orientational field also preside over scenes in which the resulting conduct is attributed to the individual as though it originated wholly in them.

The argument is deliberately limited. It offers neither a comprehensive theory of digital capitalism nor a doctrine of excuse, and it does not claim that every platform-mediated action is determined by architecture. It asks instead whether practices of responsibility can remain indifferent to conditions the same order has helped to organise. The subject formed within digital environments does not remain merely an object of power. It returns as a subject required to answer. To understand that return, the article first clarifies the relation between freedom, unfreedom, and the conditions of agency.

Framework and Course of the Argument

The principal theoretical anchor is Adorno’s treatment of freedom and unfreedom in Negative Dialectics, read through its critical-legal reconstruction in the work of Reeves. Adorno matters here not because he supplies a ready-made theory of digital technology, but because he refuses to purify agency conceptually of the social and bodily conditions through which willing becomes possible. His account of the addendum further identifies an impulsive, somatic moment within willing that cannot be reduced to formal rational self-legislation. This opens the conceptual space for asking what happens when technical and economic architectures increasingly solicit and channel the orientations through which willing passes into conduct.

The analysis proceeds through conceptual reconstruction and immanent critique. It neither tests a causal hypothesis nor claims empirical comprehensiveness. Instead, it reconstructs concepts already presupposed by digital practices—choice, preference, authorship, consent, responsibility, and violation—and sets them against the formative operations of the infrastructures in which they are invoked. The critique is immanent because it arises from a contradiction within the order’s own normative vocabulary: platforms rely upon the image of an answerable user while organising conditions that complicate the very self-authorship their judgments presuppose. Its negative-dialectical commitment is not to import a method from outside, but to resist resolving freedom and unfreedom into a harmonised formula in which either formation or agency disappears.

The article’s account of affect follows Wetherell’s practice-oriented approach rather than treating the autonomy thesis associated with Massumi as the final word; the distinction is developed in Section 4.

The argument proceeds in five stages. Section 2 reconstructs the social and embodied conditions of agency through Adorno, critical-legal theory, and a limited comparison with critical realism. Section 3 examines the political economy and mechanisms through which digital capitalism reaches those conditions, moving from surveillance and data colonialism to algorithmic governmentality, societies of control, attention, and unequal forms of datafication. Section 4 defines affective heteronomy and distinguishes it from manipulation, coercion, discipline, control, and related accounts of digital influence. Section 5 follows the movement from formation to consent, moderation, platform judgment, and neoliberal responsibilisation. Section 6 develops the asymmetry of answerability and clarifies the practical significance of critique without pretending that the contradiction between freedom and unfreedom can be conceptually abolished.

2. Freedom, Unfreedom, and the Conditions of Agency

Before asking how digital capitalism shapes and judges its subjects, a more basic question must be confronted: under what conditions does agency become possible? Kant’s account of autonomy binds freedom to rational self-legislation and thereby makes responsibility intelligible as answerability to a law the rational subject gives to itself. Adorno does not simply reject that aspiration. He asks what becomes of autonomy when rational self-legislation is abstracted from the bodily, social, and historical conditions through which a subject becomes capable of willing. Critical theory thus shifts attention from freedom as a formal possession to the forms of life in which freedom is enabled, damaged, or rendered increasingly abstract.

Adorno’s treatment of freedom in Negative Dialectics matters precisely because it refuses to separate freedom and unfreedom into two simple alternatives. In the model on freedom, he writes that the two are “not a simple antithesis; they are interwoven” [11] (p. 220). The contradiction at issue is therefore social and historical rather than primarily metaphysical. Rather than asking only whether causation abolishes freedom, the analysis turns to the conditions through which freedom becomes possible or is damaged. On the reading pursued here, reason becomes ideological when it presents socially and bodily mediated agency as though it were self-originating. Horkheimer and Adorno’s Dialectic of Enlightenment had already traced the ambivalent social history of enlightened reason, including the ways formal rationality can pass into domination rather than emancipation [12]. Negative Dialectics returns to freedom at the level of the subject and refuses to allow rational self-legislation to erase the unfree conditions within which willing occurs.

Adorno’s metacritique of Kant matters also because it refuses to reduce action to pure practical reason. The will cannot be understood as rational self-legislation standing above natural and social mediation. Adorno’s discussion of the addendum names the non-identical moment without which willing would be emptied into abstraction: a bodily, impulsive moment, “an impulse” without which “there would be no real will at all” [11] (pp. 228–229). The significance of this bodily moment for the present argument will become clear below. Freedom, on this account, depends upon the relation between reflection, impulse, embodiment, and social form. It is not a property that can be cleanly detached from the conditions under which subjects are formed. Adorno’s account of freedom and its damaged conditions has been developed in different directions in recent scholarship [13,14]; the reading pursued here follows its critical-legal reconstruction.

Reeves argues that this Adornian shift from metaphysical dispute to social-historical critique exposes the limits of understanding the determinist challenge exclusively in metaphysical terms. He therefore presents Adorno’s intervention as a metacritical reformulation of that challenge: a social-historical, rather than metaphysical, thesis concerning forms of life that undermine the conditions of self-determination [10] (p. 323). What matters, then, is less whether conduct has causes than whether the capacities through which freedom becomes possible are themselves historically damaged.

A neighbouring critical-realist line of thought reinforces this refusal to choose between sovereign agency and structural dissolution. Reeves’s critical-legal reconstruction is itself developed in conversation with Bhaskar. Bhaskar’s transformational model of social activity likewise treats agents as acting within pre-existing social structures that condition, but do not mechanically determine, the practices through which those structures are reproduced or transformed [9,15]. Archer’s morphogenetic account further distinguishes conditioning structures from agents who reproduce or transform them over time [16]. Her later account of human agency likewise emphasises embodiment, practical engagement, and emergent personal powers [17]. The article does not adopt critical realism as a second framework. It draws only on the shared insistence that conditioned agency remains agency, and that recognising its conditions is not equivalent to denying its normative significance.

These implications become especially visible within the law. Modern practices of responsibility depend upon what Reeves describes as the autonomy assumption: an image of the person as a self-governing and answerable agent [9]. Such an image is indispensable; without it, responsibility would lose its footing. Yet it also carries a danger. In the criminal-law context, Reeves argues that retributivist logic abstracts actions from the “context of their genesis” and treats them as “freely willed actions” [10] (p. 324). Responsibility itself is not the difficulty. The difficulty lies in the ease with which its own conditions disappear from view.

Recent work by Reeves and Sinnicks carries these concerns into the context of technological capitalism. Their account of administered heteronomy draws attention to the increasing penetration of administration into work, leisure, and political life, while insisting that autonomy and heteronomy cannot simply be divided along the line between individual and society [18] (pp. 289–290). Existing heteronomy must be investigated as a function of “concrete social arrangements,” not as the mere fact of social relatedness.

The present argument proceeds in a similar spirit, but at a different level. Administered heteronomy concerns the organisation of spheres of life—work, leisure, politics—and the institutional forms through which conduct is channelled. Affective heteronomy concerns the orientations through which those spheres are encountered as attractive, urgent, normal, threatening, or self-chosen. The former names the structured terrain of social life; the latter, a mode of being oriented within it. The distinction is not between external structure and inner psychology, but between two levels of mediation. If forms of unfreedom are historically specific, the mechanisms through which they operate must be so too. How, then, does digital capitalism reach the conditions of agency themselves?

3. How Digital Capitalism Reaches the Conditions of Agency

Historically specific forms of unfreedom require historically specific mechanisms. The question, then, is not simply whether digital capitalism influences conduct, but how that influence reaches the conditions from which conduct proceeds. Explanations centred on persuasion, coercion, or isolated acts of manipulation capture only part of the picture. What matters here are the ways contemporary systems increasingly organise the environments within which significance, preference, and practical orientation emerge.

Shoshana Zuboff’s account of surveillance capitalism provides the broader political-economic setting for this development. Ordinary digital activity generates what she terms behavioural surplus: forms of data extracted from experience and transformed into predictive products aimed at markets in future behaviour [4]. Yet information is not the end of the process. As Zuboff argues, surveillance capitalism increasingly seeks not merely economies of scale but economies of action [4]. Behaviour is not only observed; it becomes something to be anticipated, optimised, and ultimately shaped.

Couldry and Mejias extend this political-economic diagnosis by describing contemporary data relations as a form of data colonialism: an appropriation of human life through infrastructures that render ordinary activity continuously available for extraction and capitalisation [5]. Their account prevents behavioural data from appearing as a neutral informational residue. What is appropriated is an expanding range of lived relations, habits, movements, and interactions. Data extraction then supplies the material through which environments can be recalibrated around the subject, while its normalisation makes formative intervention appear as an ordinary condition of connection.

Where Zuboff explains why such intervention becomes economically desirable, Antoinette Rouvroy and Thomas Berns offer a more precise account of how it operates. Their conception of algorithmic governmentality describes forms of power that act less through commands than through the organisation of environments and probabilities [1]. Rather than addressing subjects as reflexive agents, algorithmic systems seek to anticipate and modulate conduct before it becomes the object of explicit deliberation. As Rouvroy and Berns put it, algorithmic governmentality “produces no subjectification” and “circumvents and avoids reflexive human subjects” [1] (para. 10).

Rouvroy and Berns further suggest that algorithmic governmentality complicates or postpones the moment of reflexivity, critique, and recalcitrance necessary for subjectification [1] (para. 10). Reflection increasingly takes place within environments that have already been selectively arranged. Power enters earlier, participating in the configuration of the field within which some possibilities become more salient and others less so.

This mode of operation can be situated within the transition Deleuze describes from disciplinary societies to societies of control. Discipline organises conduct through relatively bounded sites of enclosure—the school, factory, hospital, or prison—where subjects are trained through fixed sequences and norms. Control, by contrast, operates through continuous modulation across more open environments: access, opportunity, visibility, and risk are recalculated as the subject moves [6]. These are not cleanly successive epochs. Disciplinary institutions persist, but increasingly coexist with forms of modulation that extend across and between them. The individual does not disappear; it is decomposed into data points, profiles, probabilities, and adjustable permissions.

This genealogy clarifies algorithmic governmentality without reducing it to a technologically intensified discipline. Its characteristic operation is not simply to impose a stable norm after conduct occurs, but to rank, recommend, filter, personalise, accelerate, delay, and differentially expose in advance. Rouvroy and Berns specify how control can be automated through correlations that seek to anticipate conduct without first addressing a reflective subject. Affective heteronomy is not a fourth historical stage after discipline, control, and algorithmic governmentality. It names the lived and orientational dimension of the present configuration, especially where anticipatory modulation later feeds into practices of attribution and judgment.

Bernard Stiegler directs attention to one of the capacities through which such organisation becomes effective. Attention, for Stiegler, is itself socially and technically mediated. Contemporary systems increasingly invest in its capture and management, threatening the conditions necessary for sustained and critical forms of engagement [2]. More than distraction is at stake: what is being organised are the conditions under which significance appears. Deliberation presupposes attention. If attention itself becomes vulnerable to industrial organisation, power reaches still closer to the conditions of agency.

Taken together, these accounts point towards a transformation that cannot readily be reduced to external regulation. Digital capitalism increasingly acts upon the conditions through which conduct emerges: attention, salience, anticipation, and orientation. These mechanisms are affective in the social sense because they help organise the field in which some objects, risks, images, invitations, and futures appear urgent, desirable, irritating, threatening, or irrelevant. Following Wetherell, affect concerns embodied and socially organised practices through which the world becomes practically meaningful [3].

The conversion of lived qualities into data does not affect all subjects in the same way. Beller’s account of datafication and racialised and colonial-gendered abstraction emphasises that computational systems do more than assign numbers to pre-existing qualities; they participate in the commodification and differential valuation of social being [19]. A full account of racial capitalism and colonial gender formation lies beyond the scope of this article. Nevertheless, the qualification matters for any universal description of affective formation: the fields of salience, risk, visibility, and answerability organised by digital systems are historically stratified, and the burdens of being rendered legible, governable, or disposable are unevenly distributed.

In an Adornian vocabulary, the apparent naturalness of these arrangements can be understood as second nature: historically produced conditions acquiring the immediacy of a background that no longer appears made [11] (pp. 357–358). Surveillance and data colonialism clarify the political economy of appropriation and intervention; algorithmic governmentality and societies of control describe anticipatory modulation; Stiegler identifies attention as one of its vulnerable conditions. Still insufficiently conceptualised is the form of unfreedom that arises when these mechanisms weight the pre-reflective and affectively charged conditions of agency, while the conduct that follows remains attributable to individuals as though those conditions were neutral. A more precise concept is therefore needed.

4. Affective Heteronomy

The preceding analysis indicates why a further concept is needed. The relevant unfreedom is not exhausted by being watched, predicted, nudged, or disciplined. It concerns the organisation of the orientational field in which subjects come to care, fear, desire, attend, and respond. The concept proposed here is intended less as a comprehensive account than as a way of bringing a particular tendency within contemporary forms of unfreedom into view. I refer to this tendency as affective heteronomy.

By affective heteronomy I mean the systematic shaping, under digital capitalism, of affectively charged orientations through which subjects encounter experience as significant, urgent, desirable, threatening, or negligible, whereby solicited responses are experienced, and later attributed, as the subject’s own. The concept concerns the point at which possibilities acquire urgency and come to matter. What becomes susceptible to organisation are not simply actions, but the orientations from which actions emerge.

Because affect is used here in a social rather than narrowly psychological sense, affective heteronomy does not require the location of a private emotional state behind conduct. It concerns embodied orientations through which the world becomes practically salient: the ways possibilities acquire intensity, attraction, aversion, or felt relevance before they are fully thematised as objects of deliberation. This is why the term affective is necessary rather than ornamental. The argument concerns not only which options become visible, but how possibilities press upon the subject as worth pursuing, fearing, resisting, or ignoring.

Massumi’s account of the autonomy of affect is indispensable to this problem because it distinguishes affective intensity from its conscious qualification as a named emotion or narratively intelligible meaning [20]. His emphasis on the temporal and bodily priority of intensity converges with the present argument’s concern for the pre-reflective moment. Affective heteronomy begins from the same proposition: practical orientation is not exhausted by explicit reasons, beliefs, or decisions. Power may become effective before a subject can render its own orientation fully articulate.

The divergence concerns what follows from affect’s irreducibility to conscious qualification. Massumi’s autonomy thesis names the excess of intensity over the meanings through which it is subsequently recognised. The present argument does not deny that excess. It asks, rather, whether what cannot be fully captured in meaning is therefore beyond social patterning or technical intervention. Following Wetherell’s practice-oriented account, the interval can be understood as both emergent and organised: embodied in its occurrence, yet sedimented through practice, repetition, and relation [3]. Digital architectures need neither determine affect nor secure a predictable outcome in order to work upon this interval. They can arrange the timings, sequences, and objects through which tendencies are solicited and intensified, while leaving open the possibility of refusal, surprise, and resistance. Affective heteronomy therefore names neither the elimination of spontaneity nor its conversion into programmable causation. It names the unequal organisation of the field in which spontaneity must find its form.

The issue, then, is not whether affect can be fully captured—it cannot—but what political significance attaches to the attempt to operate upon what exceeds conscious narration. Digital capitalism increasingly treats this interval as a practical site of prediction, experimentation, interface design, and commercial modulation. Affective heteronomy names that project without mistaking its ambition for accomplishment: infrastructures need not master affect for its pre-reflective movement to become consequential for accumulation and governance.

The relation between affect and ideology also becomes clearer at this point. Ideology need not operate only by supplying false propositions that subjects consciously believe. It can work through the naturalisation of organised orientations: a historically produced urgency appears as personal impatience; architectural amplification, as spontaneous interest; repeated calibration, as authentic preference; and designed friction, as an individual failure of resolve. Not every unnoticed influence is ideological. The ideological moment arises when the organised character of an orientation systematically disappears and presents itself as natural or self-originating in ways that benefit—and simultaneously exempt—the order that arranged it. In this form, ideology concerns the misrecognition of genesis. Adorno’s category of second nature therefore names the achievement through which mediated dispositions acquire the appearance of immediate inclination [11] (pp. 357–358).

The term “affective” must also be connected to the impulsive moment within willing. Here Adorno’s account of the addendum is more than illustrative. The will reduced to pure practical reason is, for him, an abstraction; the addendum names the bodily, impulsive moment that this reduction eliminates—an “impulse” that denies “the Cartesian dualism of res extensa and res cogitans,” and without which “there would be no real will at all” [11] (pp. 228–229). A power that organises impulse and felt inclination therefore works not around the will, but upon the moment in which willing passes into act. Schematically, attention determines what comes into view; affect determines what moves the subject. It is this movement that downstream practices of answerability will treat as the agent’s own.

This does not mean that the impulsive moment is simply available for administration. For Adorno, the addendum also marks the non-identical residue within willing: the somatic and spontaneous element that cannot be fully absorbed into rational self-legislation. The point, therefore, is not that digital capitalism abolishes this moment, but that contemporary architectures increasingly solicit, anticipate, and channel it, thereby weakening the spontaneity through which resistance might otherwise appear.

Several neighbouring concepts illuminate aspects of the phenomenon without exhausting it. Manipulation undoubtedly captures important dimensions of influence, including within digital environments [21,22]. Yet affective heteronomy is not simply another name for manipulation. Manipulation ordinarily identifies a problematic influence upon judgment, belief, desire, or choice. Affective heteronomy concerns the practical field within which those distinctions themselves become available: the quieter accumulation of patterned orientations across repeated encounters, through which some possibilities acquire force before any discrete manipulative act can be isolated.

Nor is affective heteronomy the same as coercion. Coercion ordinarily appears as pressure against the will, or at least as a constraint experienced as external obstruction. Affective heteronomy is subtler. Its point is not that subjects are forced to do what they do not want, but that the field in which wanting itself takes shape has been organised. It may include coercive pressures, but its distinctive concern is not external compulsion. It is the patterned formation of the orientations through which compulsion, inclination, and choice are experienced.

For this reason, affective heteronomy is not exhausted by surveillance. Observation, extraction, and prediction matter because they feed back into the shaping of salience, urgency, and orientation. Surveillance is one condition of the process, not the process as a whole.

Nor is the phenomenon adequately captured by disciplinary power alone. Foucault’s analysis remains indispensable for understanding how institutions form subjects through surveillance, examination, normalisation, and embodied routines [23]. Deleuze’s society of control marks a shift in emphasis from bounded sites of discipline to continuous modulation across environments, without implying that discipline disappears [6]. Algorithmic governmentality specifies one contemporary form of that modulation: correlations are used to anticipate and arrange possibilities without first producing a normatively self-reflective subject [1]. Affective heteronomy names the lived and orientational dimension of this configuration, not a fourth historical stage. Discipline trains conduct; control continuously recalibrates access and possibility; algorithmic governmentality automates anticipatory modulation; affective heteronomy describes how the modulated field is encountered as felt urgency, attraction, aversion, or apparent self-direction—and how those orientations later enter scenes of attribution.

Nor can the phenomenon be assimilated to ordinary social influence. Human beings have always been shaped by institutions and relationships. What distinguishes contemporary digital capitalism is not the existence of influence, but its industrialisation. The shaping of orientation becomes continuous, personalised, predictive, and economically interested.

Nor is affective heteronomy simply Stiegler’s account of attention under another name. Stiegler names the industrial capture of attention and desire; affective heteronomy names that capture in its relation to answerability, where organised orientations later reappear as the individual’s own responsible conduct.

Affective heteronomy must also be distinguished from the account of administered heteronomy developed by Reeves and Sinnicks [18]. Their concern is with the organisation of spheres of life under technological capitalism. The present concern is with the orientations through which those spheres are encountered. Administered heteronomy asks how life domains are structured; affective heteronomy asks how subjects are disposed within those domains so that some paths feel obvious, urgent, pleasurable, risky, or unavoidable. The two concepts are complementary rather than competing.

Affective heteronomy is best understood relationally. It does not compete with the accounts examined above; it identifies what becomes visible when upstream formation is brought into relation with downstream attribution. What is distinctive is not influence as such, but the organisation of the impulsive and orientational conditions from which conduct emerges, followed by the purification of that conduct as the individual’s own at the moment of judgment. The next section turns to this second movement.

5. From Formation to Answerability

The conceptual point is clearest when the analysis moves from formation to answerability. Digital infrastructures do not merely help shape the conditions under which users act; they also establish the conditions under which users become accountable. The movement from affective formation to answerability is therefore central to the form of digital power at issue here.

This is visible first in the environments within which consent is solicited. Karen Yeung’s account of hypernudge governance illuminates their significance [24]. Hypernudges are adaptive forms of choice architecture capable of modifying the presentation and salience of options in response to data [24]. The same data-driven environments that shape which options appear salient, urgent, or default also generate the record against which the resulting choice is later read as the user’s settled preference. Formal choice remains, but the field within which it is exercised has been configured by the same platform order that will subsequently treat the choice as binding. Consent does not stand outside the process of formation. It is produced within it and then redescribed as its origin. This is the first form taken by the asymmetry: the moment of affective formation and the moment of attributed agreement coincide, yet only the second is allowed to count.

The same structure governs answerability beyond consent. As Kate Klonick has shown, major digital platforms increasingly exercise governance functions over speech and participation through moderation procedures and community standards [7]. Tarleton Gillespie’s work further demonstrates that moderation is not a peripheral feature of platform life, but one of its ordinary conditions [8]. What matters here is that the architecture which adjudicates conduct is continuous with the architecture that elicited it. The feed that organises salience and the ranking system that rewards certain responses cannot be separated from the standards against which those responses are later judged; they are administered by the same order, often through the same infrastructure. The user thus encounters as neutral evaluation an environment that has already inclined the very conduct being judged.

The circuit at issue can be described without treating any single platform as paradigmatic. A user enters an environment in which the order of appearance, rhythm of notification, prominence of images, friction of exit, and rewards of visibility have already been arranged. Some posts appear first; some controversies gather speed; some invitations become easier to answer than to ignore; and some affects are intensified because the architecture has learned that they hold attention. The user’s response may remain recognisably their own: a click, a share, an insult, an impulsive purchase, a consent decision, a refusal, or a moment of self-display. Yet it emerges from an environment that has made certain objects urgent and others negligible. When the same response is later treated as evidence of preference, agreement, violation, or risk, the formative circuit disappears. Affective solicitation appears as personal inclination; architectural amplification, as spontaneous interest; designed friction, as ordinary choice. The point is not that the user has been reduced to an automaton. It is that the authorship of conduct is purified at the moment of judgment. The platform order can present itself as merely recording, moderating, or enforcing, even though it has participated in arranging the field within which the conduct became likely, attractive, or difficult to resist.

This illustration also clarifies why the asymmetry is not adequately captured by familiar doctrines of notice, consent, or rule violation. Such doctrines ordinarily stabilise responsibility by fixing a decisive moment: the user clicked, agreed, posted, watched, bought, or breached. These moments remain important, but they come too late to describe the whole process. By the time conduct becomes visible as a discrete act, the field of visibility, urgency, and reward has already been structured. The problem is temporal as well as normative. Formation occurs upstream, through continuous calibration; answerability occurs downstream, through the isolation of a particular act. The interval between the two is precisely where forgetfulness enters. A settled preference may have been cultivated through repeated exposure. Incivility may have been intensified by architectures that reward provocation, while consent may have been solicited through fatigue, default, or designed inconvenience. None of this removes agency. It changes what must be included in any adequate account of agency.

This is not to collapse platform governance into criminal law. Platform sanctions are usually private, contractual, infrastructural, and reputational rather than penal in the strict sense. The point is more limited, but still significant: platforms organise quasi-juridical scenes of answerability in which conduct is attributed to a user, measured against rules, adjudicated by human or automated procedures, and followed by consequences such as deletion, demotion, suspension, exclusion, or reputational loss. What makes these scenes quasi-juridical is not their resemblance to a courtroom, but their structure of attribution: a response is isolated from the conditions of its emergence, assigned to the user as its author, and made the basis of a consequence. This is sufficient for the asymmetry to matter. The conditions that helped form the response are precisely those the scene of judgment sets aside, and they are set aside by the same platform order that helped produce them.

These developments form part of what Jack Balkin describes as an algorithmic society, in which governance becomes increasingly distributed across public and private actors [25]. Julie Cohen’s account of informational capitalism carries the concern further. The significance of digital infrastructures lies not simply in the behaviours they regulate, but in the conditions under which selves emerge and develop [26]. Platform governance therefore reaches beyond conduct narrowly understood. It concerns the forms of life through which subjects become capable of conduct in the first place.

At the same time, contemporary forms of governance increasingly address individuals as responsible managers of themselves. Wendy Brown’s analysis of neoliberal reason captures this tendency with particular force. Subjects are called upon to optimise their choices and assume responsibility for outcomes even as the broader conditions under which those outcomes emerge become increasingly opaque [27]. The demand for answerability intensifies precisely where the conditions of self-determination become more difficult to discern.

Chun’s account of programmability helps explain how this demand for self-management is sustained by a conjunction of visible agency and infrastructural opacity. Software appears at the interface as an instrument through which users act, while the algorithms, hardware, institutional relations, and historical assumptions that generate visible effects remain largely inaccessible. Programmed visions seek to orient futures through past data, yet the environment presents itself as immediately responsive to individual command [28]. This is a form especially suited to neoliberal reason: the subject is invited to experience itself as the sovereign operator of choices whose conditions are programmed elsewhere. Causality is personalised at the interface even where power is distributed through relations the user cannot inspect.

A peculiar reversal follows. The subject formed within digitally organised environments does not remain merely an object of shaping. It returns as a responsible user: the apparent operator of interfaces whose causal architecture remains opaque, the manager of preferences that have been continuously inferred and fed back, and the author of conduct elicited within environments calibrated to intensify response. Conduct formed under conditions of affective heteronomy becomes the basis upon which judgments are issued, sanctions imposed, agreements enforced, and obligations assigned. What matters is neither formative influence nor answerability considered alone, but the asymmetry through which the first disappears when the second begins.

6. The Asymmetry of Answerability

Neither influence nor responsibility is historically novel. Human beings are socially formed, and collective life depends upon practices through which persons answer to one another. Were conditioning alone sufficient to cancel responsibility, responsibility would become unintelligible. Equally, if every demand for an account were treated as domination, the normative distinction between justification and indifference would collapse. The difficulty lies in neither term considered alone.

The problem lies in the convergence of two organised movements. Digital capitalism increasingly arranges the conditions through which practical orientation is experienced, while digital governance proliferates procedures through which discrete acts are isolated and attributed to users. Formation is continuous, distributed, and infrastructural; judgment is episodic, individualised, and declarative. The same order that acts upon the field from which conduct emerges can later encounter a click, post, purchase, consent, refusal, or violation as though its conditions were simply given.

This is the asymmetry of answerability. What the order helps produce at the level of formation is disregarded at the level of attribution. Conditions that are historically and technically organised reappear as personal properties: an amplified reaction as temperament, a cultivated preference as authentic desire, a friction-laden consent as settled agreement, or an architecturally rewarded incivility as isolated misconduct. The injustice lies neither in acknowledging agency nor in imposing consequences. It lies in purifying agency of its genesis while exempting the formative order from answering in turn. Judgment proceeds as though the conditions upon which it depends had no history.

Here Adorno’s primary text strengthens the point. In Negative Dialectics, freedom is not treated as an abstract problem of free will, but as a problem of socially mediated subjectivity. Adorno rejects the isolation of action presupposed by abstract moral judgment: “There is no way to weigh an isolated action” [11] (p. 296). He writes that freedom is “really delimited by society, not only from outside but in itself” [11] (p. 297). A practice of judgment that treats historically shaped conditions as neutral background therefore repeats the abstraction that a critical theory of responsibility ought to resist.

Nor does the asymmetry described here presuppose conspiracy. One of the distinctive features of digital capitalism is the impersonal character of its forms of domination. Competitive pressures, optimisation imperatives, and architectures of engagement generate patterns that no individual actor fully controls. The asymmetry is therefore structural rather than intentional.

The result is a curious one-sidedness. Subjects are required to answer, while the conditions that shape their answerability remain largely exempt from answering in turn. Users are expected to justify conduct, while the architectures that influence it remain opaque. Individuals are called upon to exercise self-management within environments whose organisation they neither determine nor meaningfully contest. Responsibility flows downward; upward answerability remains attenuated.

Adorno’s negative dialectics is helpful here because it refuses reconciliation by conceptual fiat. The contradiction between freedom and unfreedom cannot be resolved by declaring the subject either wholly autonomous or wholly determined. A critical concept must remain faithful to the non-identity between these poles. It asks how subjects can remain agents within conditions that are not of their own making, and how judgment can acknowledge agency without converting those conditions into a natural background.

This refusal of conceptual reconciliation should not be confused with political resignation. To say that critique cannot abolish the contradiction is not to say that institutions can do nothing. It means that no doctrinal formula can finally detach responsibility from social formation. The practical demand follows from that limit: because attributed conduct cannot be separated without remainder from the environments in which it emerges, institutions that organise those environments must bear justificatory burdens of their own.

Upward answerability would require more than transparency in the narrow sense of disclosing that an algorithm exists. It would require platforms to account for how ranking, recommendation, notification, interface friction, monetisation, profiling, and engagement incentives participate in eliciting the conduct they later interpret or sanction. Moderation and appeal procedures would ask not only whether a rule was breached, but whether the architecture systematically amplified the form of conduct under review. Consent practices would be assessed not only by the presence of a click, but by the conditions of solicitation, repetition, fatigue, default, and asymmetrical knowledge. Preferences would no longer be treated as self-originating facts where platforms have continuously inferred, cultivated, and recirculated them. These are philosophical markers of what a less forgetful practice of responsibility would have to notice, not legislative proposals in any complete sense.

Such demands do not excuse every wrong action, assign platforms sole authorship of conduct, or replace moral judgment with causal explanation. Their purpose is to prevent responsibility from becoming a technology of abstraction. A less forgetful practice of answerability would preserve the distinction between explanation and excuse while refusing the fiction that only individuals possess histories of formation. The infrastructures that organise practical fields would themselves become objects of attribution, justification, and contestation.

The concept of affective heteronomy clarifies why this institutional demand is necessary. Surveillance, modulation, and attention capture become normatively acute not simply because they influence behaviour, but because the orientations they organise later return in the purified form of the subject’s own responsible conduct. Massumi’s insistence on the non-identity of affect and conscious meaning is crucial here [20], but the autonomy of affect leaves open the question of how what exceeds conscious qualification can nevertheless be patterned. Deleuze’s account of control [6] and Rouvroy and Berns’s analysis of algorithmic governmentality [1] clarify the anticipatory modulation of possibilities; affective heteronomy identifies how that modulation is lived as urgency, attraction, aversion, and apparent self-direction. In this way, the concept joins the extractive political economy analysed by Zuboff and Couldry and Mejias [4,5] to downstream practices of attribution and judgment, while giving an affective specification to the administered heteronomy described by Reeves and Sinnicks [18]. Its contribution is not to replace these accounts, but to disclose a relation that each, taken alone, leaves incomplete: the formation of orientations and the judgment of the conduct to which they give rise. The injustice lies in severing that relation at the moment responsibility is assigned.

No pristine subject waits to be liberated from every condition of influence, and no technical design can finally separate freedom from unfreedom. The contradiction cannot be abolished or dissolved into a harmonious concept. But critique can alter what institutions are permitted to forget. It can refuse the movement by which historically produced orientations appear as self-originating and by which the order that helped arrange them presents itself as a neutral recorder, moderator, or judge. To insist that an order participating in the formation of conduct must answer for those conditions when it demands an account is not to eliminate responsibility. It is to make responsibility less one-sided. For a critical theory of digital capitalism, that is the practical significance of keeping the contradiction in view.