The Society of the Like: Hegel, Veblen and the Negative Ontology of Recognition in Cyberspace

Abstract

This article advances a philosophical interpretation of contemporary cyberspace through the categories of recognition, negativity and invidious distinction. Existing accounts attribute digital self-exploitation to surveillance, economic coercion or behavioural design; these readings describe mechanisms without identifying the ontological operation they presuppose. The paper argues that the like is not a reduced form of recognition but its structural contrary, and that the same cancellation operates across the broader architecture of commerce in cyberspace. Drawing on Hegel’s analysis of Anerkennung as a negative operation requiring risk and determinate negation, and on Veblen’s theory of invidious distinction as the preservation of that negativity within commercial society, the article shows that algorithmic cyberspace cancels the three conditions under which recognition is ontologically possible—the risk of non-recognition, determination through scarcity, and the reciprocity of the encounter. What the like offers is pure positivity, and that positivity produces the very exhaustion it claims to soothe. Platform fatigue is therefore structural: the consequence of an institution that monetises recognition demands it cannot satisfy. The paper concludes by situating the diagnosis within the Kantian question of autonomy and the Luddite prehistory of technical resistance to the reorganisation of recognition.

1. Introduction

A peculiar exhaustion pervades contemporary digital life. Users formally free return daily to platforms that produce in them compulsion, comparative anxiety and a seemingly insatiable desire to be seen. The phenomenon is neither confined to a demographic nor limited to a single application. It runs across the entire constellation of likes, followers, subscribers, rankings and engagement metrics that now mediates an increasing portion of social existence. That users are exhausted is no longer controversial; why they return is.

The critique of platform society inherits, whether its contemporary practitioners acknowledge it or not, a genealogy considerably older than the behavioural turn of the 2010s. When the Luddites broke mechanical looms in Nottinghamshire and Yorkshire between 1811 and 1816, they were not—as the received narrative long suggested—defending an idyllic past against the efficiency of the modern factory. They were defending, in the only register available to them, a structure of recognition that the mechanical reorganisation of weaving had already begun to dismantle. The artisan who had recognised himself in the object wrought by his hand—the very structure that Hegel would theorise almost simultaneously in the Phenomenology of Spirit [1], and that Veblen would elaborate on a century later under the name of the instinct of workmanship [2]—found himself confronted with a device that retained the product while eliminating the recognition. The Luddite could not articulate this philosophically. He could only break the machine. Thompson’s historiography has shown how precisely that inarticulate lucidity was subsequently recoded as irrationality by the very economic order that had produced it [3]1.

Neither Hegel nor Veblen is customarily summoned to think cyberspace, and the distance separating their respective vocabularies, centuries and disciplinary frames is real. That distance is also the reason to summon them. Cyberspace does not invent the human compulsion for recognition; it reproduces it, intensifies it, and—as this paper will argue—finally cancels the very ontological architecture through which recognition had been thinkable. Hegel supplies the category in its strict sense: recognition is a structure of negativity, risk and reciprocal mediation, not a transaction of approval. Veblen supplies its commercial afterlife. Once modernity displaces the heroic and clerical-military forms within which recognition had been directly visible, it falls to invidious distinction and conspicuous consumption to carry the negative structure forward, modulated but not abolished [5]. Read together, the two authors do not describe the same phenomenon. They describe two moments of a single dispositive—one ontological, the other economic—that the cyberspace of the like now reorganises from the inside.

The wager of the paper can therefore be stated as a precise claim about a gap. The literatures that currently dominate the philosophical understanding of platform society—surveillance, digital labour, behavioural design, and the phenomenology of positivity—share a common limit: each describes a mechanism that the platform operates or an affect that the user undergoes, and none identifies the ontological operation that the mechanism presupposes and the affect registers. What is missing is not a further mechanism or a more refined description of the malaise, but an account of what recognition is, such that its cancellation could produce the malaise in the first place. Hegel and Veblen are summoned for that determinate reason, and for no other. Hegel furnishes the ontology of recognition as a negative structure; Veblen furnishes the proof that this structure survives, translated, in the commercial medium that platforms inherit. The two together supply the one thing the existing accounts lack: a concept of recognition robust enough to specify what the like removes.

The thesis of this paper can be stated without hedging. The “like” is not a reduced form of recognition; it is its structural contrary. Recognition, in the sense Hegel established and Veblen preserved, is negative through and through. It requires the risk of not being recognised, the determination of the self through what it is not, and the reciprocal cancellation of immediate self-certainty in the encounter with another consciousness. The like cancels the three conditions at once. It cancels the risk, because the absence of a like is not determinate negation but silence without ontological status. It cancels the distinction, because likes aggregate without subtracting and submit the self to no test of scarcity. It cancels the reciprocity, because the other does not distribute the visibility—the algorithm does. What the like offers, therefore, is not weak recognition but pure positivity, and this positivity produces precisely the exhaustion it claims to soothe. The problem that the platform industry treats as a technical matter—attention economy, engagement optimisation, recommender systems—is in fact philosophical of the first order. It concerns the institutional form through which human subjects demand, and are denied, the specific worth that only another can grant.

Two preliminary clarifications are in order. First, the term cyberspace is used here in its broader philosophical sense, as the informational space within which subjectivity is produced and validated, rather than as a synonym for any particular platform2. In this sense it retains its affinity with the tradition that runs from Lévy [6] through Hui [8], even as it designates the concrete devices—Instagram, X, TikTok, LinkedIn—in which the present analysis finds its empirical material. Second, the diagnosis advanced here is structural, not nostalgic. This paper does not lament a pre-digital age of unmediated recognition; no such age existed. It argues, more precisely, that the cyberspace of the like institutes a particular reorganisation of the recognitive dispositive, and that this reorganisation has consequences which neither the technical vocabulary of the industry nor the prevailing critical theory has so far been equipped to register.

The argument proceeds in four steps. Section 2 identifies the limits of the dominant critiques of digital positivity, with particular attention to Byung-Chul Han, whose phenomenology of fatigue describes the symptom but misrecognises its ontological structure. Section 3 reconstructs the Hegelian account of recognition as negativity, drawing on the Phenomenology of Spirit and the Philosophy of Right to show that Anerkennung is irreducibly constituted through risk and determinate negation. Section 4 turns to Veblen, whose theory of invidious distinction and pecuniary emulation preserves the dialectical architecture of recognition in the degraded medium of economic comparison. Section 5, the argumentative core, analyses the like as the operation that cancels negativity altogether and produces a cyberspace structurally incapable of recognising the subjects it absorbs. A concluding section draws the political and anthropological consequences, with particular attention to the Kantian question of whether the algorithmic delegation of social judgement constitutes a new form of that minority of which the Enlightenment took itself to be overcoming [9].

2. Beyond Surveillance and Addiction: The Recognition Deficit

Three families of critique currently dominate the philosophical literature on platform society, and each captures something real. None of the three, however, identifies the ontological operation that makes the phenomenon they describe possible in the first place. A brief reconstruction is necessary, not to dismiss what these accounts have achieved, but to mark precisely the threshold at which they stop and at which the present argument begins.

The first family is the surveillance paradigm consolidated in the work of Zuboff [10]. Its diagnostic strength is considerable. Platforms are described as extractive infrastructures that convert behavioural surplus into predictive products, instituting a new mode of accumulation whose principal raw material is human experience itself. The analysis is precise concerning what platforms do, but silent concerning why subjects voluntarily supply the experience that platforms then extract. Surveillance presupposes a surface of behaviour available for capture. It does not explain the persistent enthusiasm with which that surface is offered. The user who posts, comments, rates and rates again is not, at the moment of posting, a subject undergoing extraction. He is a subject performing a demand. Whatever extraction takes place is parasitic upon a more original gesture, and surveillance theory cannot name that gesture without leaving its own analytical register.

The second family is the digital labour paradigm developed in different versions by Fuchs [11] and Chicchi and Marrone [12]. Here the conceptual move is to recover the classical Marxian vocabulary of exploitation and extend it to the unpaid productivity of platform users. The point is well taken. There is indeed a structural asymmetry between the value generated through user activity and the value retained by users themselves. Yet the labour paradigm faces a difficulty that its proponents tend to acknowledge only obliquely. Users would continue producing the content that platforms monetise even if the monetisation were absent and the surveillance lifted. The teenager who films and posts a dance, the academic who broadcasts a thread on her latest article, the elderly man who shares a photograph of his grandson—none of these subjects is principally motivated by the wage they are not receiving. They are motivated by something prior to the labour relation, something that the category of labour cannot itself articulate.

The third family is the behavioural addiction model derived from neuropsychology and consolidated in the design literature itself [13]. This account locates the explanation in dopaminergic reward circuits, intermittent reinforcement schedules and the affordances of variable rewards. Its descriptive precision is real. What it cannot explain is why the content of the addictive stimulus is, with such structural regularity, social. The pull of the slot machine is the unpredictable money. The pull of the platform is the unpredictable validation. The behavioural model treats the second as a special case of the first. The present paper argues that the relation runs in the opposite direction: validation is not a particular variety of reinforcement, it is the anthropological substrate that makes platform reinforcement legible as such.

The fourth and most philosophically ambitious diagnosis is offered by Han, particularly in The Burnout Society [14], The Transparency Society [15] and Psychopolitics [16]. Han’s analysis comes closer than its competitors to identifying the structure at issue. His diagnosis of the contemporary subject as auto-exploited rather than coerced, as enmeshed in an excess of positivity rather than constrained by negation, as performing visibility rather than enduring discipline, captures with phenomenological accuracy the texture of platform subjectivity. Han is also right to insist that the pathology in question is not external to the subject but constitutive of its very form of self-relation. Where Han does not press the analysis far enough, however, is precisely at the ontological level. He describes positivity as the medium of contemporary suffering, but he does not return to Hegel—from whom the category of negativity descends—to specify what positivity actually cancels and why that cancellation produces fatigue. The result is a phenomenology of the symptom that diagnoses without explaining.

The unease of the present argument with Han’s diagnosis can be stated more precisely. To say that contemporary subjectivity is too positive is to presuppose that there exists a structurally adequate degree of negativity from which the present is a deviation. But what is that adequate negativity? In Han’s pages it remains gestural, evoked through a conceptual repertoire (Heidegger, Buddhism, Aristotle’s vita contemplativa) that does not return to the question of recognition as such. The Hegelian word Anerkennung appears in Han’s work, but it appears as background, not as a structure to be reconstructed. The consequence is that Han’s reader knows what the present lacks affectively (calm, slowness, capacity for negation) but not what it lacks structurally. The present paper takes Han’s diagnosis as a point of departure, accepts its phenomenological correctness, and proposes to supply the ontological argument that Han suspends.

The four critiques sketched above are not wrong. Surveillance does occur. Labour is extracted. Reward circuits are exploited. Positivity is, as Han describes, the dominant atmosphere of platform existence. What none of these analyses provides is an answer to the question that motivates the present paper: what does the like cancel? The remainder of the argument develops the categories required to answer that question. The detour through Hegel and Veblen is not antiquarian. It is the only available route to a diagnosis that goes beyond the description of mechanisms and reaches the ontological operation that those mechanisms presuppose.

It is worth making explicit, before the constructive argument begins, the distinction between the three registers at which the platform can be analysed, since the difference governs everything that follows. An analysis at the level of mechanism describes what the platform does and how: it extracts behavioural data, optimises engagement, and distributes visibility through a recommender system. Zuboff and Srnicek operate at this level, and their findings are correct as far as they reach. An analysis at the level of affect describes what the user feels: exhaustion, compulsion, and the comparative anxiety of the feed. Han operates at this level, and his phenomenology is accurate. An analysis at the level of ontology asks a different question from either: not what the platform does, nor what the user feels, but what kind of thing recognition is, such that a mechanism could cancel it and an affect could register the cancellation. The mechanistic account cannot answer this, because it presupposes the demand for recognition it does not theorise; the affective account cannot answer it either, because it describes the symptom of a cancellation whose structure it leaves unspecified. The ontological deficiency of both is not a defect of rigour but a limit of register. The present paper supplies the missing register, and the categories of Hegel and Veblen are what supplying it requires.

3. Hegel: The Negative Architecture of Recognition

The contribution of Hegel to the present argument can be reconstructed with relative economy, provided that one resists the temptation to treat the master-slave dialectic as a sociological allegory or a pedagogical fable3. The argument of the Phenomenology is more radical than either reading admits, and only the radical version yields the diagnostic instrument that the cyberspace of the like requires. What follows is therefore a strict reading, not an extended commentary.

Self-consciousness, in the Phenomenology, does not arrive at itself through inspection of its own contents. It arrives at itself through desire, and desire—properly understood—is not the appetite for natural objects. Natural appetite is satisfied by the consumption of its object, and in being satisfied it confirms only the dependence of the desiring subject upon the natural world. The desire that constitutes self-consciousness is of a different order. It is desire for another desire, desire to be acknowledged by a being capable of acknowledging, and desire for recognition by another self-consciousness [1] (§175–§176). The structure of this desire is from the outset intersubjective; the subject does not first exist and then encounter others, it comes to exist as itself only through the encounter [18,19].

The encounter, however, does not deliver immediate reciprocity. Each consciousness initially demands recognition without conceding it. The result is conflict, and Hegel describes that conflict in terms whose ontological severity is often softened in commentary. The struggle is one of life and death (§187). What is at stake is not honour, prestige or social standing as later sociology would conceive them. What is at stake is whether the subject can demonstrate, by risking its mere biological existence, that it is not reducible to that existence. Only the consciousness that risks its life can claim to be a consciousness rather than a living thing. The famous asymmetric outcome—one consciousness becoming master, the other servant—is the consequence of this risk being unequally borne. The one who refuses the risk subordinates himself to the one who accepts it. Recognition is thus achieved, but achieved as a structurally one-sided relation that fails to satisfy the very demand that motivated the conflict4.

The reading we owe to Pinkard [19] and to Williams [20] insists, correctly, that the dialectic does not stop at the moment of subordination. The slave, through work, encounters the resistance of the natural world and discovers in that encounter a mediated form of self-recognition that the master, who consumes without producing, cannot attain. Work is the negative held in suspense, the destructive impulse converted into formation. The natural object yields to the worker not as a thing to be consumed but as a thing to be formed, and in that formation the worker recognises himself. The world becomes humanly inhabitable through this productive negativity [1] (§195). Pippin [22] adds the crucial precision that what is at issue here is not a psychological process but the constitution of normative space itself. The categories of free agency emerge through the recognitive struggle, not before it.

For the purposes of the present paper, three structural features of this account must be retained, because they will determine what exactly the like cancels in Section 5.

The first is that recognition is negative. It is not given as a positive transaction in which one party confers approval and another receives it. It is constituted through the determinate negation of the immediate self-certainty of each consciousness in its encounter with the other. The other is not an audience that approves; the other is a counter-subject whose very existence puts the first subject’s self-conception at risk. Recognition that does not pass through this negative moment is not recognition in the Hegelian sense. It is approval, and approval has neither the depth nor the constitutive force of Anerkennung.

The second is that recognition requires risk. The subject who is unwilling to put something at stake cannot be recognised as a subject. What is staked need not be biological life—Hegel’s master-slave scene is a structural illustration, not a literal blueprint—but something whose loss would matter must be exposed. Risk is the index of the seriousness with which the subject claims to be a subject. A recognition demand that costs the demander nothing makes itself unrecognisable, because it has not differentiated itself from any other unstaked claim.

The third is that recognition is reciprocal, or it is nothing. The asymmetric recognition obtained through the master-slave struggle is unsatisfying precisely because it is asymmetric. The master is recognised by a consciousness whose recognition no longer counts, since that consciousness has surrendered its claim to count [1] (§192). Adequate recognition requires that the other be a subject capable of withholding what it grants, and grant only what it could withhold. The freedom of the recogniser is the precondition of the value of the recognition.

These three features—negativity, risk, and reciprocity—are not independent attributes that recognition happens to possess. They are mutually entailing aspects of a single structure. Take any one away and the others collapse with it. A recognition without risk cannot be negative, because nothing is determinately negated. A recognition without reciprocity cannot be risked, because the recogniser bears no symmetric exposure. A recognition without negativity cannot be reciprocal, because what is exchanged is not the seriousness of mutual self-determination but the reassurance of unilateral approval.

The Philosophy of Right extends this structure into the institutional domain. Civil society, in Hegel’s analysis, is the sphere in which the recognitive dynamic articulates itself through the system of needs, the division of labour and the mediated exchanges of property and contract [23] (§189–§208). The point is not that civil society dissolves the negativity of recognition into the harmony of commerce. It is that civil society institutes mediated forms in which negativity persists as the condition of the recognitive relation, even as direct confrontation gives way to the impersonal mediations of the market and the law. The lord and bondsman do not vanish; they are translated into structurally analogous positions within an institutional grammar [24].

This translation is what makes Veblen possible. Where Hegel theorises the ontological architecture of recognition and gestures, in the Philosophy of Right, towards its institutional articulation, it falls to Veblen to specify the particular form that the negative structure assumes once commercial society has displaced the heroic and clerical-military arrangements within which recognition had previously been directly visible.

4. Veblen: Invidious Distinction and the Commercial Preservation of Negativity

Veblen’s The Theory of the Leisure Class [5] is conventionally read as a sociological treatise on the consumption habits of the late nineteenth-century American bourgeoisie. The reading is not wrong, but it is insufficient. What Veblen describes under the categories of conspicuous consumption, conspicuous leisure, pecuniary emulation and invidious distinction is not merely a portrait of class display; it is a structural account of how the recognitive dynamic that Hegel had theorised in ontological terms persists, modulated and translated, within the institutional medium of commercial society.

The category around which the entire analysis turns is invidious distinction [5] (Ch. II). The term is precisely chosen. Veblen does not speak of differentiation, variety or stratification, all of which would describe non-conflictual asymmetries. He speaks of a comparison that injures—invidiosus, in its Latin etymology, denoting that which arouses envy and inflicts damage upon the one compared unfavourably. Distinction in Veblen’s sense is not the descriptive observation that subjects differ. It is the structural insistence that they differ in a way that matters, that one position is gained at the cost of another, that the prestige obtained by some is constitutively unobtainable by all. The leisure class is leisured precisely because the working class works; the wealthy display their wealth precisely because the poor cannot match the display. The distinction is invidious because it is structurally negative: it exists only to the extent that what it confers cannot be conferred upon everyone.

This is Hegel translated into the medium of money. The lord-bondsman dialectic has not disappeared. It has been institutionalised within the system of commercial exchange, where the test of life is replaced by the test of consumption, and where the asymmetry of recognition is achieved not by survival of a duel but by the demonstration of an expenditure that the other cannot match. Pecuniary emulation is the pursuit of recognition by the only route that the commercial order leaves open, since the heroic and clerical-military routes have been institutionally foreclosed [5] (Ch. III). The wealthy man does not need the goods he buys; he needs the visible fact of being able to buy them. What he displays through his purchases is not utility but the very capacity for invidious distinction.

What makes Veblen’s analysis a strict continuation of Hegel rather than a sociological digression is the structural architecture he preserves. Three features in particular must be marked, because they will determine what is at stake when the cyberspace of the like cancels them.

First, invidious distinction requires scarcity. A good that everyone possesses cannot perform pecuniary emulation. The whole point of the conspicuous expenditure is that not everyone can match it. This is, formally, the same condition that Hegel identified at the ontological level: recognition is constituted through differentiation, and differentiation requires that what one has cannot be possessed by all. Hirsch [25] would later formalise this insight in economic terms under the category of positional goods, demonstrating that a class of goods exists whose value resides not in any intrinsic property but in the position they confer relative to others. Positional goods are intrinsically scarce because their value collapses if access is generalised. Veblen had already grasped this; Hirsch supplies the formal vocabulary [26].

Second, invidious distinction requires risk. This is less obvious in Veblen than in Hegel and merits explication. The conspicuous consumer risks two things at once. He risks the resources he expends, which are by definition resources that cannot be expended elsewhere. And he risks failure of the display itself, since conspicuous consumption that is misread as ordinary consumption performs no distinction at all. The signal must be costly to be credible, in the sense later formalised by Spence [27] in the economics of signalling. Cheap signals do not distinguish. A signal whose cost is negligible can be sent by anyone, and what can be sent by anyone confers nothing.

Third, invidious distinction requires the other as judge. The conspicuous expenditure presupposes a witness whose acknowledgement matters. Without that witness the expenditure is not conspicuous at all; it is merely wasteful. The recognitive structure thus retains, in attenuated form, the reciprocity of the Hegelian encounter. The wealthy man depends upon the gaze of those who cannot afford what he displays. He is constituted as wealthy through their acknowledgement of his wealth. The relation is asymmetric, as in the master-slave dialectic, but it remains a relation, and the recognition it confers continues to require an other capable of conferring it.

The Instinct of Workmanship [2] develops the complementary side of the same structure. Where the Theory of the Leisure Class analyses recognition as it operates in consumption, the later book analyses recognition as it operates in production. The instinct of workmanship is, for Veblen, the disposition to find satisfaction in efficient and intelligent activity, in the configuration of materials according to purpose, and in the capacity to produce something that bears the maker’s competence. This is, structurally, the Hegelian moment of work: the negativity of desire converted into formation of the world, the subject finding itself in the product. Veblen’s contribution is to register what happens when the industrial reorganisation of production interferes with this dynamic. The workman who is reduced to repetitive operations on a fragmented task no longer recognises himself in the product, because the product no longer bears his individual competence. The instinct persists as a residual disposition, but it loses the institutional medium through which it had been satisfied.

It is here that the Luddite intuition recovers its full philosophical weight. The Luddite did not lack the theory that would have allowed him to articulate his resistance, because no such theory was yet available. What he sensed practically—and what Hegel was simultaneously formulating ontologically, and what Veblen would formalise economically a century later—was that the mechanical loom was not merely a more efficient device for producing cloth. It was a device that retained the cloth while eliminating the artisan’s recognitive relation to it. The destruction of the loom was not the rejection of efficiency. It was the inarticulate defence of a recognitive structure whose disappearance the artisan correctly perceived even as he lacked the conceptual resources to name it.

Veblen and Hegel together yield the following position. Recognition is a negative structure that requires risk, scarcity and the freedom of the other. Commercial society does not abolish this structure but translates it into the medium of pecuniary emulation, where conspicuous consumption and conspicuous production preserve the architecture of negativity in the degraded but functional form of invidious distinction. What remains to be shown is that the cyberspace of the like operates a further translation, but a translation of a different kind. It does not merely modulate the recognitive structure as commercial society had done. It cancels it.

This further translation does not bear only upon social media in the narrow sense; it bears upon the entire architecture of commerce as it has been reorganised in cyberspace. What Veblen identified in the conspicuous consumption of the late nineteenth century is, with structural fidelity, what the systems of star ratings, verified reviews, top-reviewer badges, seller-reputation scores and recommendation algorithms now generalise across the platforms of Amazon, eBay, Mercado Libre and their competitors. These are not external addenda to the commercial transaction; they are the very medium within which contemporary commercial recognition is performed. Each of them is a translation of the Veblenian dispositive into the algorithmic medium, and each of them is, simultaneously, the locus at which the dispositive itself is cancelled. The like in social media and the rating in electronic commerce are not two distinct phenomena to be theorised separately. They are two manifestations of a single architecture whose ontological structure Section 5 now examines.

5. The Like as Cancellation: The Algorithmic Spectator

The argument now reaches its core. Section 3 established that recognition, in the strict Hegelian sense, is constituted through three mutually entailing features: negativity, risk and reciprocity. Section 4 showed that commercial society preserves this architecture in the translated medium of invidious distinction, with scarcity replacing existential risk, conspicuous expenditure replacing the staking of life, and the witness of one’s pecuniary peers replacing the counter-consciousness of the master-slave encounter. The present section shows that the cyberspace of the like cancels each of these three features and, in doing so, produces the first institutional medium in human history within which recognition is structurally unavailable while its demand is structurally intensified.

A terminological caution is in order before the analysis begins. The cancellation at issue is not a Hegelian Aufhebung. The like does not sublate the recognitive structure, negating it while simultaneously preserving and elevating it; it cancels it privatively, retaining nothing of the negativity, scarcity and reciprocity that it dissolves. Where Aufhebung conserves the negated moment within a higher unity, the like leaves no residue and yields no synthesis, only the pure positivity of the metric.

Before examining each cancellation, it is useful to state the correspondence on which the analysis turns, since the connection between the philosophical categories and the concrete devices of the platform must be made exact rather than suggestive. To each structural condition of recognition there corresponds a determinate platform mechanism that cancels it. The condition of risk, which in Hegel is the staking of life and in Veblen the irreversible expenditure of wealth, corresponds to the like button and the follower count, signals that cost the sender nothing and therefore stake nothing. The condition of scarcity, which in Veblen is the positional structure of the conspicuous good, corresponds to the unbounded supply of likes, a token whose distribution carries no upper limit. The condition of reciprocity, which in Hegel is the free counter-subject and in Veblen the judging peer, corresponds to the recommender system, the mechanism that distributes visibility in the place of any judging other. The like button, the metric counter and the recommender algorithm are not incidental features of the interface; they are the concrete sites at which the three conditions of recognition are, respectively, cancelled. The analysis that follows examines each correspondence in turn.

Consider first the cancellation of risk. In the Hegelian encounter, the subject who would be recognised must put something at stake, and what is staked must be visible to the counter-subject as such. In the medium of invidious distinction, what is staked are the resources expended in conspicuous consumption, the irreversible commitment of wealth to display. In both cases, the absence of recognition has a determinate ontological status: the subject who has staked and lost is recognisably defeated, and has recognisably failed to obtain what was sought. The like cancels this structure entirely. The user who posts and receives no likes has not been recognisably refused recognition; he has been met with silence, which is not negation but the absence of any determinate response. The post that fails has the same surface appearance as the post that succeeds, minus a number. There is no counter-subject who has refused to acknowledge, no recognisable adversary who has prevailed. There is only an algorithm that has not surfaced the post to other users, and the post itself, languishing in the indeterminate space of the under-engaged. The exhaustion that this produces is not the exhaustion of repeated defeat, which would at least confer the dignity of having been determinately denied. It is the more debilitating exhaustion of indeterminacy, of demands that neither succeed nor fail in any ontologically registrable sense. Han’s phenomenology of fatigue [14] describes this affective surface accurately. The present analysis supplies the ontological reason: the user is exhausted because he is engaging an interface that does not even possess the capacity to refuse him.

The cancellation of risk invites an immediate objection, and the objection is instructive. Does the dislike, the downvote, the thumbs-down not restore precisely the determinate negation whose absence has just been described? A dislike, unlike silence, appears to be a registered “no”, an active refusal that the platform makes available. The objection fails, but it fails in a way that confirms the analysis rather than weakening it. The dislike is not determinate negation; it is the simulation of negation, and it cancels the negativity of recognition by the same operation through which the like cancels its positivity. Three features establish the point. The dislike aggregates: it is absorbed into a score, a ratio, a net figure, exactly as the like is, and an aggregated negation is not the determinate negation of a counter-subject but a quantity subtracted from a sum. The dislike is anonymous and reversible: the one who dislikes risks nothing, exposes nothing, and may withdraw the gesture without consequence, whereas the determinate negation of the Hegelian encounter binds the negating subject to its negation. And the dislike is, like the like, distributed within the visibility that the recommender system controls, so that the disliking other is no freer than the liking other. The empirical tendency of the medium corroborates the structural claim. Platforms have progressively withdrawn the public dislike—the removal of the visible dislike count on the principal video platform [28] is the most conspicuous instance—not from squeamishness but because determinate negation is precisely what the engagement architecture cannot metabolise. Where the downvote survives, as in the rating and reputation systems of commerce, it survives only as aggregated quantity, never as the exposure of a subject to a subject. The dislike, in short, does not reintroduce the negativity that the like cancels. It extends the cancellation to the negative pole, completing rather than interrupting the operation.

Consider next the cancellation of distinction. The architecture of invidious distinction, as Veblen and Hirsch established, requires scarcity. A signal whose distribution is unconstrained cannot perform recognition, because what it confers can be conferred on everyone. The like is structurally unconstrained. There is no upper limit to the number of likes that can be given, no zero-sum competition for the supply of recognition tokens, no scarcity that would force the user into the test of pecuniary emulation. Likes aggregate without subtracting; my receipt of a like does not diminish your prospects of receiving one. The economic literature on positional goods [25,26,29] established that the value of certain goods collapses if their distribution is generalised. The like is the asymptotic limit of that collapse: a recognition signal whose generalisation is built into its design from the outset. The platform may, of course, manufacture an artificial scarcity through algorithmic visibility—the post that no one sees cannot be liked—but this artificial scarcity is precisely the inverse of the Veblenian condition. In invidious distinction, the recognition token is scarce because its generalisation would destroy its function. In the platform economy, the recognition token is artificially restricted because its abundance has destroyed its function. The platform produces the scarcity it has itself dissolved, and produces it in a form that is not even visible to the user as such, since what is restricted is not the supply of likes but the distribution of the underlying visibility on which likes depend5.

Consider, finally, the cancellation of reciprocity. Hegelian recognition requires that the other who recognises be a free counter-subject capable of withholding what is granted. Veblenian invidious distinction preserves this requirement in attenuated form: the witness of conspicuous consumption is a peer whose own social position and judgement constitute the medium within which the display registers. The like operates within neither structure. The user who likes my post is not a counter-subject who has freely chosen to confer recognition; he is, more often than not, a user whose own engagement is mediated by the same algorithmic surfacing that produced my post in his feed. He may not have read what he liked. He may have liked reflexively in the expectation of a reciprocal like. He may have been targeted by the algorithm precisely because the system has predicted his likeliness to engage. In none of these cases is the relation Hegelian. The free other has been replaced by the algorithmic other, and the algorithmic other is not free in the sense that the recognitive structure requires. The visibility that determines whose post reaches whom is distributed not by the judgement of subjects but by the recommender system [30,31]. The platform, as Galloway [32] established for the protocol layer of the internet, exercises a form of control that operates through the appearance of decentralisation. The user believes he is engaging with other users; he is in fact engaging with the optimisation of an attention-allocating mechanism. The recognitive other has been substituted with what we may call, with deliberate philosophical weight, the algorithmic spectator6.

5.1. The Algorithmic Spectator as Pseudo-Recognition

The category of the algorithmic spectator deserves explicit elaboration, since it does not merely supplement the analysis of the three cancellations but names what comes to occupy the place vacated by the recognitive other. The deliberate philosophical weight of the term is best appreciated by contrast with the figure it inverts: the impartial spectator developed by Smith in The Theory of Moral Sentiments [33]. Smith’s spectator is internalised through the moral education of the agent and operates as the medium through which the agent comes to judge his own conduct from a perspective that is not strictly his own but is also not foreign to him. The spectator is impartial because the agent has, through sympathetic identification with the judgements of others, achieved the capacity to step outside himself and assess his conduct as another would. The judgement remains the agent’s own, but it has been schooled by an internalised social standpoint that the agent recognises as morally authoritative.

The algorithmic spectator inverts each of these features. It is externalised rather than internalised; it operates not within the agent’s moral self-relation but within proprietary infrastructure whose parameters are determined by the commercial interests of the platform owner. It is partial rather than impartial; its judgements are calibrated to maximise engagement, not moral correctness. And it is not authoritative in the moral sense Smith intended but only in the operational sense that the user, lacking any alternative within the medium, treats the metric distribution of visibility as the index of his significance. Where the Smithian spectator schools the agent in moral autonomy, the algorithmic spectator schools the user in metric heteronomy. The two figures stand in exact philosophical opposition: one is the medium through which the agent recovers his judgement, the other is the medium through which the user surrenders it.

What the algorithmic spectator delivers is therefore not recognition but its institutional simulation—what may be called pseudo-recognition. The simulation reproduces the surface phenomenology of recognitive transactions: visibility, response, gradation of esteem. It cancels, however, the three structural conditions under which such transactions are ontologically possible. The user receives the form of recognition without its substance, the appearance of acknowledgement without the negativity, scarcity and reciprocity that would make acknowledgement what it claims to be. This is why platform fatigue is not a contingent affective state but a predictable structural outcome. The user is exhausted because he is engaging with an apparatus that performs all the gestures of recognition while remaining ontologically incapable of conferring it. The pathology is not that the platform recognises poorly; it is that the platform cannot recognise at all, and yet the demand for recognition is what it most efficiently harvests.

The figure of the algorithmic spectator reaches its limit, and its clearest demonstration, in the automated bot. A substantial and growing share of the likes, follows and engagements that circulate on the major platforms is produced not by users at all but by automated accounts [34], whether deployed to inflate a profile, to manufacture consensus, or to simulate the activity that attracts further activity. The bot is not a marginal aberration of the system; it is the system’s own logic carried to completion. If a like can be produced by a programme that recognises nothing, possesses no self-certainty to stake and no judgement to withhold, then the like never carried recognition in the first place. The bot does not corrupt an otherwise recognitive transaction; it reveals that the transaction was never recognitive. What the analysis of reciprocity established as a tendency—the substitution of the free counter-subject by the algorithmic other—the bot establishes as a possibility internal to the token itself: the recognition signal can be generated in the total absence of any recognising consciousness. The algorithmic spectator, in its purest form, is no longer even the displaced judgement of an engagement-maximising system distributing the responses of real users; it is the manufactured response of no one. That users continue to experience bot-inflated metrics as recognition, and frequently cannot distinguish them from the responses of other users, is the final evidence that what the platform delivers is the form of acknowledgement detached from its substance. Pseudo-recognition does not require a deceived other; it requires no other at all.

The same three cancellations operate, with structural fidelity, in the platforms of electronic commerce. The Amazon star rating cancels the risk by design, since a low rating is not determinate negation but a number absorbed into an aggregate average and rendered ontologically inert. The verified review badge cancels the distinction, since the badge is granted by the platform algorithm rather than earned through a recognisably scarce transaction whose loss would matter to the reviewer. The seller reputation score cancels the reciprocity, since the buyer who rates does not know the seller as a counter-subject and the seller cannot return a meaningful judgement on the buyer beyond a parallel rating. What appears as a sophisticated commercial epistemology is, ontologically, the same dispositive that organises the like in social media, applied to the medium of buying and selling. The cyberspace of commerce and the cyberspace of likes are not two distinct domains but two manifestations of a single architecture that has cancelled, across both, the recognitive structure that earlier commercial society had preserved.

It might be objected that the analysis just sketched proves too much. After all, the absence of likes is experienced by users as a form of negation, and the asymmetric distribution of visibility does produce a structure that resembles invidious distinction. The user who receives few likes feels, phenomenologically, that he has been refused; the user who receives many feels recognised. Why insist that the structure has been cancelled when its phenomenology persists?

The objection is serious and deserves a precise reply. The phenomenology persists because the demand for recognition persists. What has changed is not the demand but the institution that purports to satisfy it. The user feels refused because he carries with him an anthropological structure of recognition that pre-exists the platform and that the platform exploits. But the platform is structurally incapable of satisfying that demand, because it has cancelled the three features without which satisfaction is impossible. The result is the perfect institutional pathology: an apparatus that converts a constitutive human demand into a measurable behavioural surplus while remaining ontologically unequipped to discharge what it elicits. The user is exhausted because he is engaging a device that has all the surface features of a recognitive interface and none of the structural features of one. Han’s positivity is, on this analysis, not a cultural mood. It is the institutional consequence of a recognitive demand routed through a medium incapable of the negativity that recognition requires.

This account also clarifies what is distinctive about the platform economy as a stage of capitalism. Srnicek [30] has shown that platforms are best understood as a particular type of business model whose central activity is the extraction of data through the provision of digital infrastructure. What the present argument adds is that this extraction is parasitic upon a more original anthropological capture. The platform does not invent the demand for recognition; it harvests it. And the harvesting is so productive precisely because the platform converts a structurally inexhaustible demand—recognition is by Hegel’s argument never definitively obtained—into a structurally inexhaustible behavioural surplus. Every refresh is a renewed petition. Every petition unanswered intensifies the next. The platform’s commercial success is the mathematical translation of the ontological pathology it institutes.

A final clarification concerns the relation between the present analysis and the spontaneous order tradition associated with Hayek. There is an obvious sense in which platforms can be described as spontaneous orders that emerged without central design from the interaction of users, developers and infrastructure providers. There is, however, a crucial difference between the spontaneity of the Hayekian cosmos and the spontaneity of the platform. The Hayekian order presupposes subjects capable of self-limitation, capable of submitting to abstract rules, capable of recognising one another within an institutional grammar that does not require central enforcement. The platform produces, on the contrary, subjects whose self-relation is increasingly delegated to the algorithm itself. The spontaneity of the order is real, but the order it produces is one that progressively erodes the subjective capacities on which Hayekian spontaneity had depended. To this extent the platform represents not the triumph of decentralisation but its self-cancellation, since the order that emerges is one within which the recognitive substrate of subject-formation is no longer reliably reproduced.

5.2. The Active User and the Co-Production of the Algorithm

A more serious objection remains, and it must be confronted directly rather than deflected, since it bears on the very structure the analysis has imputed to the platform. The account given so far appears to theorise the algorithm as a force acting upon a passive user, a mechanism that distributes visibility and shapes behaviour while the user merely submits. This picture is incomplete, and its incompleteness is not trivial. The relation between user and recommender system is not unilateral but recursive. Users do not only receive the outputs of the algorithm; they supply its inputs, train it through their choices, and in aggregate determine the very patterns it learns to exploit. They tag, they search, they linger and they scroll past, and each gesture is a signal that reshapes what the system will surface next. The literature on algorithmic culture has insisted, correctly, that the algorithm is a sociotechnical process co-produced by the very subjects it conditions, and that users develop an algorithmic imaginary through which they anticipate, game and contest the system that profiles them [35]. The user nudges the algorithm while the algorithm nudges the user. Any account that omits this reciprocity describes a machine acting on inert material rather than the dynamic it claims to analyse.

The reciprocity is real, and the analysis must accommodate it; but accommodating it does not restore recognition, and the reason discloses the deepest layer of the pathology. The activity of the user is not the exercise of a freedom that escapes the apparatus; it is the precise form in which the apparatus captures him. Every signal through which the user shapes the algorithm is, in the same gesture, a datum through which the algorithm shapes the user, and the loop does not open onto an outside. To train the recommender is to be trained by it. The user who curates his feed, games the metric and cultivates his profile is not thereby a free counter-subject conferring recognition upon others; he is a more thoroughly engaged producer of the behavioural surplus on which the system feeds. Activity, here, is not the opposite of capture but its most productive form. This is the sense in which recommender systems have been described as traps that capture not despite but through the user’s active engagement [36]. The reciprocity between user and algorithm is therefore not the reciprocity that recognition requires, which is the mutual exposure of two free subjects each capable of withholding what it grants. It is the reciprocity of a feedback loop, in which the user’s agency over the inputs leaves wholly untouched the fact that visibility, and with it the distribution of recognition, is allocated by the system and not by any judging other.

The concrete phenomenology of platform use confirms that the user’s activity coexists with, rather than dispels, his subjection, and the structure that reconciles the two is the one that Sloterdijk named enlightened false consciousness [37] and Žižek reformulated as fetishistic disavowal: the subject knows very well what he is doing, and does it nonetheless [38]. The shopper on a commercial platform knows that the recommendations are engineered to maximise his expenditure, and that the scarcity signals are manufactured and the rankings curated, and yet he follows the recommendation and completes the purchase, his lucidity leaving his conduct unaltered. The viewer on a video platform knows that the autoplaying sequence is optimised for engagement rather than for truth or value, and that the next item is selected because it resembles what has held his attention and not because it merits it, and yet he leans into the sequence and follows the descent. In neither case is the user deceived; in neither case does his knowledge emancipate him. The disavowal is structural, not personal. It is the subjective form proper to a medium that has captured the recognitive demand, since the user continues to seek through the apparatus the acknowledgement that the apparatus is constitutively unable to confer, and his knowledge that this is so changes nothing, because the demand is anthropologically prior to the knowledge and the apparatus is the only institution on offer to receive it.

The thesis of an empowered user who masters the algorithm through his activity is, on this analysis, an overoptimistic inversion of the truth. The more actively the user engages, the more finely the system models him, and the more completely his self-relation is delegated to the metric. The active user is not the emancipated exception to the iron cage; he is its most fully integrated inhabitant, and his activity is the mechanism of integration. This does not abolish the possibility of resistance, but it relocates it. Resistance cannot consist in more skilful engagement, and in gaming the algorithm more effectively, since skilful engagement is the form the capture takes. It can consist only in the institutional reconstruction of the conditions of recognition outside the engagement architecture, which is the task to which the concluding section turns.

5.3. Liking as the Production of Identity

The objection from the active user has a counterpart on the side of self-presentation that deserves its own treatment, since it concerns not the recognition the user seeks but the identity he produces. The like, it may be said, is not only a token the user solicits from others; it is an act the user performs, and through the aggregate of his likes he composes and broadcasts a self. To like is to declare an affinity, and the accumulated declarations—the pages followed, the posts endorsed, and the tastes avowed—constitute a public identity that the user actively authors. On this view the like is a genuine mode of social identity formation, expressive rather than merely acquisitive, and the present diagnosis, fixed on the recognition the user fails to receive, has overlooked the identity he succeeds in producing.

The objection identifies a real phenomenon and misdescribes its significance. Identity is indeed produced through the like, but the identity so produced is subject to the same cancellation as the recognition the like withholds. The composition of a self through avowed affinities is, in its classical form, the work of distinction in Bourdieu’s sense: taste is a system of differences through which the subject classifies himself by classifying his objects, and the labour of classification is constitutive of social identity [39]. Distinction, however, requires precisely what the like dissolves. It requires that an affinity cost something, that it commit the subject, that it differentiate him at the risk of misclassification. The like satisfies none of these conditions. It is costless, so it commits the subject to nothing; it is reversible, so it exposes him to no enduring determination; and it is unbounded, so it differentiates him from no one. An identity assembled from costless and reversible endorsements is not the determinate self of distinction but its simulation, a profile of positivities from which every negation—every refusal, every renunciation, every taste defined by what it excludes—has been removed. The user produces an identity in the same sense in which he receives recognition: he obtains the form of the thing and not its substance.

The point can be stated as a strict parallel. Recognition, the analysis has argued, requires negativity, and the like cancels it; identity, this section adds, requires negativity no less, and the like cancels it identically. A self is determinate only through what it negates, only through the affinities it refuses as much as those it avows, and a mode of self-formation that registers only endorsement and never refusal produces a self without contour. Liking is therefore not a suitable mode of social identity formation, however ubiquitous it has become as a paradigm of social engagement, and its unsuitability has the same root as its incapacity to confer recognition. The medium that cannot negate cannot constitute, whether what is to be constituted is the acknowledgement of the subject by another or the determination of the subject by himself. The society of the like is, in this precise sense, a society in which both poles of subject-formation—recognition by the other and identity through negation—are routed through an institution structurally incapable of either.

6. The Algorithmic Minority: Honneth, Kant and the Political Conclusion

The diagnosis advanced in Section 5 has consequences that exceed the strictly philosophical register and reach into the political constitution of the contemporary subject. Two complementary lines of analysis allow us to draw out those consequences with the precision the argument requires. The first is Honneth’s reformulation of Hegelian recognition as a critical-theoretical category capable of registering social pathologies. The second is the Kantian question of autonomy, recovered through the formulation of the Beantwortung der Frage: Was ist Aufklärung? [9].

Honneth’s Struggle for Recognition [40] established that the moral grammar of social conflicts cannot be reduced to the distribution of material resources, and that what subjects demand in their conflicts is, irreducibly, the acknowledgement of their specific worth as the kind of subjects they take themselves to be. His subsequent work [41,42] elaborated a typology of pathologies of recognition—invisibility, disrespect, denial—that have proven analytically powerful in domains as varied as labour, race, gender and nationality. The cyberspace of the like introduces, the present paper argues, a distinct kind of pathology that Honneth’s typology can absorb but that has its own structural specificity. The platform user is not invisible in the sense of going unseen. He is hyper-visible, but visible only as a metric. He is not disrespected, since the medium has eliminated the respect/disrespect distinction along with the negativity that supports it. He is not denied recognition; he is offered a counterfeit of it. The pathology proper to the algorithmic spectator is the substitution of recognition by its formal simulation, and this substitution is institutionally enforced because no alternative is available within the medium itself.

A clarification concerning Honneth is required at exactly this point, since the use made of him here is partial and the partiality is deliberate. Honneth’s mature theory understands recognition as an affirmative relation: love, rights and solidarity are the three spheres in which the subject is granted a positive acknowledgement that enables self-confidence, self-respect and self-esteem. The present argument draws on Honneth’s grammar of recognition and on his typology of its pathologies, both of which are indispensable; it departs from him, however, on the question that the critics of this debate have rightly pressed, which is whether recognition is at bottom positive or negative. The position taken here is that the constitutive core of recognition is negative, in the strict sense established in Section 3: recognition is achieved through the determinate negation of immediate self-certainty, through risk, and through the exposure of each subject to a freedom that may withhold what it grants. Honneth’s affirmative spheres are, on this reading, the institutional results of a recognitive operation whose internal structure is negative, and the affirmative model underestimates precisely the negativity that the analysis of the like requires. The divergence is not a rejection of Honneth but a recovery, against the affirmative emphasis of his reconstruction, of the Hegelian negativity that the present diagnosis cannot do without. It is for this reason that the pathology of the like cannot be captured as one more deficit of positive recognition—as a failure of love, right or esteem—but must be grasped as the cancellation of the negative operation that any such affirmation would presuppose.

The political consequences of this substitution are considerable. Three deserve mention.

The first is the privatisation of the recognitive circuit. Recognition, in any of the forms that political modernity has known, has historically been mediated through institutions whose normative architecture was at least in principle public: the legal recognition of personhood, the political recognition of citizenship, and the social recognition of professional or familial standing. The platform privatises this circuit by routing it through proprietary infrastructure whose design parameters are determined by the commercial interests of the platform owner [30,43]. The user demands recognition through a medium whose recognitive grammar he does not control, cannot inspect, and cannot contest. The political grammar of recognition has been replaced by a commercial grammar of engagement, and this substitution proceeds without democratic deliberation because the very structure that would allow such deliberation has been displaced into a sphere from which it has been institutionally excluded.

The second is the fragility of democratic citizenship. Democracy presupposes subjects capable of recognising one another as fellow citizens, that is, as bearers of equal claim to participate in the determination of common concerns. This recognition is itself a recognitive act in the Hegelian sense; it requires the negativity through which the citizen distinguishes himself from those who have not earned the title and the reciprocity through which he grants to others what he claims for himself. When the recognitive substrate is captured by the algorithmic spectator, the capacity for citizen-to-citizen recognition is correspondingly attenuated. Citizens who relate to one another primarily through metric visibility lose the structural conditions under which mutual political acknowledgement can be performed. The political consequences of this attenuation are now visible in the polarisation of public discourse, in the migration of political identity into performative micro-niches, and in the increasing difficulty of distinguishing political speech from the optimisation of personal engagement [44,45].

The third consequence reaches the deepest level of the diagnosis and connects the present argument to Kant. The user who delegates the judgement of his own social worth to the algorithm has performed an institutional act that Kant identified, two centuries before any of this was possible, as the structural form of unfreedom. Aufklärung, in Kant’s famous definition, is the human being’s exit from the self-incurred minority that consists of the incapacity to make use of one’s own understanding without the guidance of another [9] (8:35). The minor is not the one who lacks the capacity for understanding; it is the one who, possessing the capacity, has institutionally surrendered its exercise to a tutor—prince, priest, administrator. The Kantian critique was directed at the ecclesiastical and political tutelage of the eighteenth century. The present paper proposes that the algorithm is the tutor proper to the twenty-first.

The proposal is not metaphorical. The structure of delegation is identical. The user who consults his metrics to determine the worth of his utterance, who modifies his speech to optimise his engagement, or who treats the visibility distributed by the algorithm as the index of his significance, has done precisely what Kant described: he has substituted an external authority for the exercise of his own judgement on a matter that constitutively requires that judgement to be his own. The difference is that the eighteenth-century tutor at least possessed an explicit authority that the subject could in principle identify and refuse. The algorithmic tutor operates without such identification. Its authority is exercised through the very interface through which the subject performs his most ordinary acts of self-presentation. The minority it institutes is therefore deeper than the one Kant diagnosed, because it has been folded into the medium through which the subject relates to himself.

It is at this point that the double cancellation announced in the introduction acquires its full philosophical weight. The cyberspace of the like cancels both poles of the modern ontology of the subject. From the Hegelian side, it cancels the negative architecture of recognition through which the subject is constituted in the encounter with the other. From the Kantian side, it cancels the autonomous architecture of judgement through which the subject is constituted as the source of his own normative claims. The user who emerges from this double cancellation is neither recognised nor autonomous. He is metrically visible. His social existence has been reduced to the dimension that the platform is structurally equipped to register, and the dimensions that exceed that register—the determinate negation through which recognition is performed, and the autonomous judgement through which freedom is exercised—have been institutionally externalised into a sphere where they are no longer his.

This conclusion is severe, and it should not be received as a counsel of despair or as an argument for technological abandonment. The present paper is not a neo-Luddite tract. It does not propose the destruction of platforms, nor does it imagine that the recognitive demand can be relocated outside the institutions of the contemporary world. What it proposes is more modest and, perhaps for that reason, more demanding. It proposes that the philosophical diagnosis of the platform medium must be conducted at the ontological level—in the sense distinguished in Section 2—not merely at the level of mechanism or affect, and that such diagnosis is the precondition of any political response that would not simply replicate the pathology in inverted form. The Luddite, lacking the diagnosis, could only break the loom. The contemporary subject, equipped with the diagnosis, can begin to ask what institutional forms might restore the recognitive and autonomous architectures that the platform has cancelled—restore them not in a nostalgic gesture, since they were never untouched, but in a deliberate gesture of institutional reconstruction whose historical precedent is the modern constitutional order itself.

The recovery of recognition will require institutions that reintroduce the negativity, the scarcity and the reciprocity that the like has cancelled. The recovery of autonomy will require, as Kant insisted, the courage to use one’s own understanding—sapere aude—under conditions in which the institutional architecture of the contemporary world systematically discourages that use. Neither recovery will be easy, and neither can be accomplished by individual subjects acting in isolation, since the pathology in question is institutional rather than individual. What philosophy can contribute is the diagnosis. The institutional response is the work of the political community that recognises the diagnosis as its own.

The society of the like, in conclusion, is not a degraded form of the recognitive society it has displaced. It is a society in which the recognitive demand persists with undiminished intensity while the recognitive institution has been structurally eliminated. The exhaustion that this produces is not a personal failing of the users who succumb to it; it is the predictable institutional consequence of an arrangement that monetises a demand it cannot satisfy. The philosophical response is not therapy and not regulation. It is the patient labour of reconstructing the categories through which the demand can be acknowledged in its proper depth, and through which the institutional forms capable of acknowledging it can be conceived. Hegel and Veblen, summoned across the distance of two centuries, are not the antiquarian guests their absence from the contemporary debate would suggest. They are, perhaps, its most timely interlocutors.