Reaction value: affective reflex in the digital public sphere

ABSTRACT

The concept of ‘reaction’ has been frequently viewed pejoratively in the history of social and moral theory, as unthinking and often resentful. But ‘reactions’ of various kinds play a central role in the contemporary digital public sphere, of a sort that deserve attention from sociologists of value and valuation. The article identifies three forms of reaction: as ‘feedback’, as ‘content’ and as ‘chains’. It then argues that reactions exert their force within the contemporary public sphere, because they encompass some central ambiguities of the present. Firstly, as reactive beings, people hover in a space between the ‘human’ and the ‘non-human’ (or ‘post-human’), responding to stimuli but in a culturally illuminating fashion. Secondly, reactions sit in a space between judgement and artefact, combining elements of both, and thereby revealing a key ambiguity of platform infrastructure, in which audiences, critics and performers are all constantly morphing into one another. The paper concludes by reflecting on what it might mean to take reaction seriously as an object of sociology.

KEYWORDS:

• Reaction
• feedback
• platforms
• social media
• value
• critique
• public sphere

Sociologists and cultural economists have paid ample attention to practices of valuation, both within and without spheres of market exchange (e.g. Beckert and Aspers 2011; Karpik 2010; Lamont 2012; Stark 2009; Stark 2020). For those working in the tradition of ‘pragmatist sociology’, these studies have often highlighted the importance of moral frameworks and metaphysics to how quantitative measures are designed and implemented (Boltanski and Thévenot 2006). One way of conceiving this empirical orientation towards judgement, measure and valuation is as an inversion of ‘critical sociology’ to produce a ‘sociology of critique’, whereby sociology abandons the aim of establishing the normative grounds of critique, and instead aims to map and interpret the plurality of moral spheres in which everyday action is conducted (Boltanski and Thévenot 1999; Boltanski 2011; Blokker 2011). This builds directly on philosophical perspectives that see demands for justification and recognition as inherent features of everyday ethical life (Forst 2012; Honneth 2018).

Implicit in such philosophies and methodologies is that situations and institutions are periodically punctured by moments of normative and evaluative uncertainty (crises), which open up opportunities for deliberation and moral reasoning, through which the question of value can be debated and some fresh consensus arrived at. This optimistic view is at odds with critical perspectives, that assume that capitalist societies have in-built power differentials (principally those of wage relations), which are not anchored in normative consensus, but rest on domination (see Fraser and Honneth 2003).

It also crucially underestimates the power of contemporary digital technologies to coordinate social and economic activity, undisturbed by human deliberation, critique or consensus-formation. A distinguishing (and frightening) feature of these technologies is that they collect, process and circulate data amongst themselves, with the capacity to bypass the sphere of human perception and evaluation altogether (Hansen 2015). Decision-making can then shift into post-human spaces of exception (Amoore 2013), while humans become governed merely as additional machines, as witnessed, for example, in the algorithmically-governed ‘gig economy’ (Lazzarato 2014). Deleuze’s diagnosis of ‘societies of control’, in which behaviour is tracked constantly across once-separate domains of life, uninterrupted by disciplinary routines of inspection and evaluation, has been borne out by the rise of the giant platforms of the twenty-first century (Deleuze 1992). But even prior to this, the digitization and globalization of financial markets from the 1980s onwards had meant that certain questions of valuation (in particular, of pricing) had already been lifted from spheres and rhythms of deliberative judgement, and automated and accelerated to the point of being practically impervious to normative or political intervention. The presuppositions of a sociology of critique, namely that critical deliberation and ‘tests’ have the time and authority to determine outcomes, and that quantification is downstream of moral principles of value, no longer hold (Davies 2021a).

What becomes of valuation and evaluation practices once the punctuation of critical judgement is overwhelmed or bypassed? Two avenues are immediately apparent. Firstly, these practices may become constant features of everyday life, mutating from judgement (a critical capacity) to feedback (a cybernetic capacity). The constant demand for rating and affirmation of ‘experiences’, even while they are underway, is a dominant feature of consumer and user life today. ‘Likes’, ‘shares’ and star review systems are prominent features of algorithmic governmentality in the platform economy, helping to distinguish gradations of value – a brief moment of human intervention in non-human systems of quantification and differentiation. Feedback of this sort differs from a ‘judgement’ in being potentially impulsive and immersed in a flow of experience, rather than based (as in the liberal ideal of critique) on some aspiration to critical distance and measurement. It’s because real-time feedback does not require the user to achieve any evaluative distance that feedback mechanisms are often as simple as binary ‘like’ or ‘dislike’, ‘happy face’ or ‘sad face’ (Davies 2017). Moral ‘orders of worth’, governing what should be done, mutate into aesthetic questions of what is satisfying or pleasurable in the moment. Or in Habermas’s terms, the validity of action is no longer assessed in terms of compliance to a norm, but ‘dramaturgically’, that is, in how feelings and desires are impressed upon observers (Habermas 2015). Touch-screen interfaces have allowed for even more immersive, instantaneous collection of feedback, such as ‘swiping’, that require less conscious reflection than scoring systems (Kaerlein 2012).

Secondly, once power no longer seems to be constrained by norms of judgement and critique, valuation and evaluation may take on a kind of spectacular and purely affective quality, such that they become public performances. Critique becomes ‘cacophonous’ because it is no longer decisive (Boland 2018). Sloterdijk observes how, when the quest for justice appears fruitless, individuals may instead turn towards ‘rage’, appointing themselves as ‘wild judges’ whose exteriorization of negative affect serves as a wholesale denunciation of established institutions (Sloterdijk 2012, 65). Populist political leaders may be one manifestation of this (Cossarini and Vallespín 2019) but one might also think of celebrity judges on reality TV shows, whose job it is to entertain with the passion and clarity of their valuations (Muniesa and Helgesson 2013). The passing of judgement becomes a performance, convertible into engaging ‘content’, to be valued in its own right.

This article offers a contribution to the sociology of value and valuation, by exploring a category that has become ubiquitous in the digital public sphere: reaction. The idea of ‘reactions’ is associated with both of the post-critical forms of valuation just outlined. On the one hand, we can think of reactions as forms of real-time feedback, as in the case of Facebook ‘reactions’, the seven emojis with which a user can acknowledge another’s post. But on the other, reactions have become the focus of an entire genre of online content, in the form of ‘reaction videos’ (in which a user films themselves or someone else reacting to a film, game, song, news item or unexpected experience) and various split-screen formats, which allow for gamers to film themselves playing a game, or for a sports coach to be observed reacting to events on the field. Such videos turn affective expression into a spectacle and a public performance. These two different ideal types of reaction frequently collide, where (for example) YouTube users might ‘like’ a reaction video. Successful media commentators and populist leaders understand that it is through the spectacle of their reactions (especially rage) that they will achieve the most feedback (or ‘engagement’), be it positive or negative. Such cycles of reaction to reaction are a prominent feature of how value is sought, distinguished and celebrated in the digital public sphere.

These technologies and rituals of reaction are of interest to sociologists of value and valuation, because of the peculiar looping effects that they generate, whereby reactions are both means of valuation (that is as feedback) and forms of content to be valued (that is as performances), very often at the same time. On a structural level, this is due to the ‘synoptical’ arrangements facilitated by social media platforms, whereby everyone can be watching everyone else at the same time, and dramaturgical distinctions between performer, audience and critic all dissolve. In that respect, the economy of reaction resembles the financial economy, in which actors are all watching and valuing the valuations of everyone else (Vogl 2022). The public sphere takes on real-time interactive qualities that have long been associated with markets, and which were previously obstructed by the technologies of (analogue) ‘mass media’ of the twentieth century (Luhmann 2000). A central reason why sociologists ought now to attend to reactions is that the usurping of those ‘mass media’, with the rise of giant platforms, creates a new overlap between the logic of finance and that of civil society, from which it is difficult to escape.

The rest of this article is organized as follows. In the next section, I review and taxonomize different varieties of reaction in the digital public sphere. As already introduced, these include reactions as ‘feedback’, reactions as ‘content’, and reaction ‘chains’. I then turn to two key ambivalences in the culture of reaction. Firstly, reactions hover in a space between impulsive behaviour and judgement, posing a question as to where agency really lies in the reactive body, and whether valuation is still a ‘human’ practice at all. Secondly, reactions hover in a space between critique and cultural artefact, in which we find ourselves gazing at the affective behaviours of others as an alternative (or supplement) to apprehending artefacts ourselves, signalling uncertainty regarding where value lies. I conclude by reflecting on what this rise of reactions means for sociology of valuation, and sociology as a discipline that has sought to attend to action as opposed to reaction. ‘Reaction’ retains a lingering trace of ‘action’, but absent the kinds of validity claims that Habermas saw as definitive of the latter.

The forms of reaction

Where the ideals of reason and rationality are associated with notions of deliberation, measurement and objectivity, the idea of ‘reaction’ has long been associated with affect, impulse and emotion, sometimes in a pejorative sense. Reactions have long been treated as the empirical object of behaviourism, and therefore outside of the scope of social theory or sociology (Brighenti and Sabetta 2024). A capacity to ‘react’ is one that human beings share with animals, raising the concern that questions of moral responsibility and freedom are thereby suspended (Fromm 1995). Reason (if not rationality) implies some moment of reflection, possibly of dialogue, prior to a judgement, which is missing in reaction. As ideals of valuation, ‘reason’ is what is deployed by a restaurant critic when judging food, whereas ‘reaction’ is an instantaneous expression of pleasure or disgust when tasting it. Nietzsche’s critique of ressentiment treats it as a reactive emotion, which revels in feelings of powerlessness (Nietzsche 2013), and feelings are ‘involuntary’ (Nietzsche 1994, 54). The idea of political ‘reaction’, as a breed of conservatism, stems from a feeling that the forces of modernity and progress are active, and that they need resisting via a reactive opposition, that is often associated with a resentful affective disposition (Capelos et al. 2021; Robin 2018). Anger in particular has often been interpreted as a distinctively reactive emotion, that is prompted by an injustice or perceived sleight, but which can set off chains of anger and retribution, that only justice or forgiveness can bring to a close (Arendt 1958; Nussbaum 2016). Affect theorists and psychologists of emotion have explored a wider range of reactions that can be discerned in the human face and behaviour, raising questions as to how cross-cultural these signifiers might be (Leys 2017).

These moral and political questions will be relevant to the discussion that follows. But before we attend to them, I want to first distinguish some different ways in which the category of ‘reaction’ has become prevalent in the digital public sphere. The rise of social media platforms and subsequently smartphones in the first decade of the twenty-first century has produced distinctive conventions and rituals, in which the category of ‘reaction’ is integral. Pejorative connotations or unflattering contrasts with ‘reason’ are largely absent in the discussion that follows, in which three ideal types of reaction are distinguished.

Reaction as feedback

Reaction ‘buttons’ of one kind or another are a ubiquitous feature of the kinds of ‘control’ societies that Deleuze diagnosed in his 1992 essay, though efforts to establish forms of audience or consumer ‘feedback’ pre-date digital computers (Beniger 1986). Use of ‘like’ and ‘dislike’ buttons to glean audience opinions was first pioneered in the late 1930s (Schwarzkopf 2010). Such interfaces offering a choice between a smiley face and a frowning face now appear in various settings, from airport check-in desks, to libraries or public restrooms, offering a way of capturing some kind of feedback regarding a user’s experience. In 2009, Facebook (which had originally offered a ‘share’ button for links that users valued) introduced the famous ‘like’ button, offering a way to offer a simple positive reaction to a piece of content (Gerlitz and Helmond 2013). Then in 2016, it rolled out the explicitly-named ‘Reactions’, six different emojis colloquially known as ‘like’, ‘ha ha’, ‘love’, ‘wow’, ‘sad’ and ‘angry’. In the context of Covid-19, a seventh ‘care’ Reaction was added in April 2020.

Even if they’re not formally named ‘reactions’, similar facilities are available on other platforms. Reddit allows users to ‘up-vote’ and ‘down-vote’ others’ posts; YouTube allows for videos to be ‘liked’ or ‘disliked’; Twitter offers the option to ‘like’. WhatsApp (owned by Meta which also owns Facebook) includes a function, colloquially know as ‘reaction’, in which a message can be acknowledged with a small emoji appearing alongside the message being responded to, rather than as a whole new one. Email platforms, such as Microsoft Outlook, have introduced similar functions, that allow messages to be acknowledged (with an emoji or ‘like’) without involving the ‘reply’ function. Across these various platforms, the brevity and ambivalence of emojis provides an efficient means of signalling receipt of a message, and some brief recognition of (possibly sympathy with) its contents, which doesn’t require the sender to exercise much conscious thought in the crafting of a textual reply. Greater emotional nuance, specificity and effort is involved where gifs and memes (very short clips or stills from famous films or TV comedies, for example) are used for purposes of a reaction. ‘Anime reactions’ are clipped from Japanese anime cartoons, and shared online (often as memes) to provide a gallery of subtly differing emotions.

The idea of ‘feedback’ originates in cybernetics, where it was viewed as the information that allows organisms and machines to be brought under some kind of ‘control’, in the sense of purposive influence towards some predetermined goal (Beniger 1986; Halpern 2014; Hayles 2008; Pickering 2010). It is thanks specifically to negative feedback that behaviour can be adjusted as necessary. Hence, the driver of a car constantly adjusts the position of the accelerator in response to feedback regarding speed, as conveyed in the dashboard; the animal pursues food in response to feedback emanating from their nervous system indicating hunger. In order for ‘feedback’ to be possible, there must be some kind of communication ‘return path’, conveying to the controller the success or otherwise in the pursuit of its goal. A cybernetic system requires some kind of ‘closure’, such that its ‘output’ is also its ‘input’, that is, it is constantly learning from itself. Dating back to the late nineteenth century, techniques of market research can be understood as examples of control technologies, which elicit feedback on the wants and satisfaction of a customer segment (Beniger 1986). Analogue broadcasting or publishing involves no inbuilt feedback mechanism, though early efforts at audience research in the 1930s saw ‘audimeters’ added to a random selection of radios, to track what people were listening to. But where markets or other social relations become embedded in digital platforms, feedback paths can be established by default. Thus, the optimism that widespread computation might provide the means of efficient centralized economic planning (Medina 2011). Hence also the constant invitations today to rate and ‘like’ services and products, and to ‘register’ consumer goods with the manufacturer, so as to establish a feedback loop with the owner.

From the cybernetic perspective, feedback is a binary form of communication: it can be positive (indicating no behavioural change required) or negative (indicating behavioural change required). It is from the latter that organisms and machines learn valuable information, that will help them pursue their goal. Hence, the organism learns that the negative feedback of pain is a lesson to change course in some way. Reactions frequently take the form of positive feedback, inasmuch as they typically convey no new information at all, other than that a message has been received; a simple ‘thumbs up’ (that is, a ‘like’) is often a way of acknowledging receipt, but declining to say anything more. For this reason, the absence of any reaction (say in a WhatsApp group) may be a more concerning piece of ‘information’, where it becomes experienced as negative feedback. To many of their enthusiasts, markets are an exemplar of a real-time feedback system at work, seeing as prices offer their own binary reaction to events, either rising or falling (Mirowski 2002). Financial markets in particular serve the function of reacting to, amongst other things, political events either positively (as when bond yields fall) or negatively (as when they rise).

What is distinctive about digital reactions of the sort we are focusing on here is their entanglement with forms of affective expression largely derived from the human face. The face, and the cartoon derivations of it that produce emojis, provides an entire linguistic system via which reaction feedback is provided. What the face represents, in cybernetic feedback loops, is the interface of the human and the machine, where brain interfaces with machine (Hansen 2003). In contrast to other forms of feedback (such as a rating system, or written comments) the information conveyed by an emoji reaction is that communication has been successful, that content has successfully been conveyed from one user to another, via the interface of the screens. The choice of reaction (where there is a range of options) is thus less representative of how the user is actually feeling, than confirmation that communication is successful (hence, for example, ‘angry’ or ‘sad’ face signals that bad news has been successfully received), and the feedback loop is intact.

These practices, and the use of these facial semiotic systems and interfaces, fall within the broad terrain of what McStay terms ‘empathic media’, which aim to produce cybernetic feedback loops that integrate human affects of one kind or another (McStay 2018). For the most part, designating feedback as ‘reaction’ is to classify it as affective (frequently as facial), as opposed to textual or numerical in nature. Affective expression (and symbolic derivations of it, such as emojis) is part of a drift towards non-representational media, in which narrative and reports are replaced with a state of being-in-touch, sometimes in the literal sense of haptics and smart technologies that take feedback directly from the human body or brain. The turn towards facial expression and ‘body language’ as the conveyor of meaning has been interpreted as one more symptom of a ‘decline of symbolic efficiency’ (Andrejevic 2013; Dean 2009), and while reaction feedback clearly does not manage to bypass the realm of the symbolic or the representative altogether, it endorses a cybernetic ideal of communication as instantaneous, pre-conscious and physiological.

Reaction as content

Watching the unstaged impulsive and affective reactions of others has provided entertainment long before the launch of YouTube in 2005 or TikTok in 2016. The hit US TV show Candid Camera (first aired in 1948) was famous for surreptitiously filming people being subjected to pranks, for the comedy value of their reactions, with the traditional climactic moment being when they were informed ‘you’re on candid camera’, to their great shock (Clissold 2004). The boom in reality television in the early twenty-first century, which in cases such as Big Brother (launched in 2000) also exploited the surveillance potential of cameras, allowed for forensic scrutiny of the unrehearsed, impulsive behaviour of contestants. Talent shows, meanwhile, have been built around the capacity of celebrity judges to emote in ostentatious ways, and the affective impact of judgements on contestants.

The mass amateurization of film and broadcast technology in the twenty-first century has led to a great expansion in these genres of content, in the form of what are known as ‘reaction videos’ of one kind or another. These have a noted propensity to virality, suggesting a particular alluring quality of watching the facial, vocal and bodily responses of others. ‘Reaction video’ is a loose term that encompasses various forms of content, but two particular set-ups are worthy of note, and can be distinguished from each other in terms of how the camera intervenes in everyday settings. Firstly, there are those users who film themselves reacting to films, songs, video games, sports contests or news, allegedly for the first time, with a view to generating content with viral potential. Such videos are most successful when they have an air of ‘authenticity’, that is, the reactor appears to be genuinely moved or excited by what they’re experiencing, and not simply acting. Nevertheless, an element of exaggeration and comedy in reactions can sustain audience attention. Certain YouTubers become expert in producing engaging reaction-based content, achieving vast followings and earning large incomes as a result (McDaniel 2021). In this category, we could add high-profile gamers who command loyal audiences on Twitch, watching both the gamer’s screen and their face simultaneously, and sending tips in exchange for shout-outs (Taylor 2018; Wulf, Schneider, and Beckert 2020). The most successful Twitch gamers have developed a kind of hyperactive, highly expressive online persona, in which their reactions to the game are as alluring as possible (Woodcock and Johnson 2019).

Secondly, there are those videos which (like Candid Camera) exploit a power imbalance between the person with the camera and the person being filmed, or what might otherwise be called ‘pranks’, and which are set up to capture real-time reactions. Many of these may be personal and comparatively private events, such as someone filming their partner’s reaction to receiving some piece of happy news, or filming a family member as they arrive home for a surprise birthday party. Capturing the reactions of animals and babies on camera, where innocence (and by implication, authenticity) is guaranteed, is one way of producing such content. But the quest for reaction content can also involve ‘cloughtlighting’ (a hybrid of ‘clout’ and ‘gaslighting’) in which tricks are pulled on the unsuspecting, with the aim of creating content that will go viral on social media (Jarrar et al. 2020). Gender imbalances are common here, as men stage surprises and pranks, that catch women in the street (or their own partners) unawares, often so as to capture anger, shock or confusion as a form of ‘content’ for others’ enjoyment, even if they purport to be demonstrating ‘kindness’ (Hunt 2023).

Some pieces of reaction content become so widely shared, as to become memes in their own right, being converted into gifs that can also be used as feedback. The endless production of new facial reactions that appear on YouTube and TikTok in particular is a limitless source of content, from which affective expressions can be mined and deployed online. Alternatively, some social media users will use a single photograph of their baffled or disgusted or horrified (etc) face, as their reaction to a piece of news or content. On Instagram or TikTok, the term ‘reaction video’ tends to refer to videos that users make of themselves, reacting to something else they’ve seen on the same platform, using features such as TikTok’s ‘Duet’ that enables the reaction to appear alongside the original video. These inevitably prompt subsequent reaction videos – which brings us to the phenomenon of reaction chains.

Reaction chains

Social media platforms have a ‘synoptical’ structure, that allows large groups of users to all be visible to one another simultaneously. This is in contrast to the structure of print or broadcast media, in which the audience is invisible to itself, or to a traditional theatre design, in which audience attention is focused primarily upon the stage. Audience research and consumer surveys pre-dated digital media, achieving great popular interest when they were first devised (Igo 2009). But the real-time audience reaction to a radio broadcast (say) would not intrude into the experience of it. As Luhmann observed, ‘mass media’ enable ‘action’ but not ‘interaction’, which is what grants them a ‘societal’ status (Luhmann 1995). Only in situations where live events are arranged so as to blur audience and performer (such as the positioning of the Greek chorus alongside the actors, where it can react to events) is the one-to-many structure of the ‘broadcast’ or the ‘performance’ cast aside.

A fundamental feature of the cultural norms and valuation of digital content creation, in the context of social media platforms, is that users are (in dramaturgical terms) simultaneously ‘audience members’ and ‘actors’. There is no formal distinction between what counts as a performance, and what counts as a reaction to the performance. Indeed, the person notionally on the ‘stage’ may use their platform to offer their own reactions to another event; meanwhile, the person who is notionally in the ‘audience’ may react in ways that are so engaging, that they become the main focus. For this reason, while it is heuristically helpful to distinguish between reaction as feedback and as content, in practice the two phenomena are very blurred, in some of the ways already detailed.

A recognition that organisms (and potentially machines) are all constantly reacting to one another, without any single actor as the original instigator or overall governor, is a further feature of the cybernetic imagination. Cybernetics famously originated in efforts to aid the accuracy of anti-aircraft guns, starting from the premise that the gunner and the aircraft were in a communication loop with one another, each reacting to the movements of the other, in potentially predictable ways (Galison 1994; Kline 2015). Such an idea of a self-organizing system, without any central authority, helped Friedrich Hayek to hone his defence of markets, whose genius, he argued, lay in their provision of a real-time indicator (prices) which everybody could watch simultaneously, without anyone needing to have a view of the whole or conscious responsibility for it (Hayek 1945). Hayek imagined a ‘catallaxy’, in which millions of decisions were coordinated with one another, without any central coordinator.

These kinds of synoptical situations, where all actors are visible to and (in a cybernetic sense) in communication with one another, are ripe for what I term ‘reaction chains’ of various kinds. In the case of the anti-aircraft gun, each party constantly reacts to the others’ reactions; in the case of markets, especially financial markets, the calculations of each actor must also attempt to factor in how ‘the market’ (i.e. all other actors) is likely to react to a piece of information. In financial markets, the distinction between empirically verified fact and rumour becomes irrelevant; it only matters how these pieces of information are likely to affect prices, which in turn depends on how all parties expect them to affect prices (Vogl 2022). For the same reason, financial markets are liable to suffer from herd behaviour driven by collective sentiment, in which pessimism (or optimism) of others becomes grounds for further pessimism (or optimism). Asset values are reputational in nature, depending on collective expectations of their future value, which is enabled by the market synopticon (Adkins, Cooper, and Konings 2020; Feher 2018). Just as the panic on the face of the other in a crowd represents useful information, so the rapid responses of others in the market are something for investors and speculators to anticipate and learn from. Markets, like social media platforms, are spaces in which observations are constantly observed (Esposito 2011; White 2002).

One way of understanding the neoliberal policy paradigm that emerged in the 1980s was as a deliberate relinquishing of efforts to plan economic systems (which were deemed to have failed due to unanticipated reactions, manifest in inflation), and instead an opening up to the emergent forces of financial reaction-chains (Krippner 2012). Monetary policy, which became an increasingly potent tool of government over this period, became designed – right down to the precise choice of words used in press conferences – to anticipate emergent financial responses (Braun 2015). Thus the policy-maker no longer acts towards a goal of their own (as under a plan such as price-setting), but seeks to provoke and orient reaction chains in various discursive and non-discursive ways. But it is in the nature of financial markets that appeals to empirical objectivity (by any party) do not necessarily outweigh the surges of positive and negative sentiment, when sentiment is itself a form of information.

The synoptical structures made possible by the internet, and social media platforms in particular, have witnessed some similar forms of herd behaviour, in which reputations become over-inflated, or else trashed, as mass sentiment rises and falls (Arvidsson 2011; Davies 2021b; Rosamond 2019). Underlying these dynamics are mechanisms whereby individuals react to the reactions of others. As we have seen, reaction content (videos of entertaining reactions) seem to have a high propensity to attract attention and reaction feedback, and thus achieve virality. Few things seem to prompt reactions like reactions. But leaving the issue of reaction videos (i.e content) to one side, the decentralized and synoptical structure of social media yields some peculiar political and cultural effects, in enabling reactions to rapidly become the ‘main event’, thanks to the reactions that they themselves provoke.

One clear model of this is that of trolling. The basic format of trolling, deriving from a situation in which everyone is simultaneously ‘audience member’ and ‘performer’, is triadic: it creates a feedback loop connecting the troll, the troll’s target and the troll’s peer group. Exploiting the fact that boundaries and contextual distinctions are suspended in online situations, a troll will intervene in the typically non-humorous (even grief-laden) situation of others with a prank of some kind, so as to provoke laughter in a different group altogether, and win acclaim for themselves (Phillips 2015). The troll deliberately seeks to set off a reaction chain, whereby their own intervention provokes first of all bewilderment, offence or anger on the part of their target, leading to enjoyment on the part of the peer group. An accusation that someone is ‘trolling’ implies that their actions are purely aimed at generating a reaction, which will raise their own profile in the process. Abstaining from a reaction (even when there is a grave insult or injustice being waged) therefore becomes seen as a way of rebuffing such attacks online, breaking the reaction chain not through forgiveness (as Arendt argued) but through simple ignorance of the aggressor.

Between behaviour and critique

Whether as ‘feedback’ or as ‘content’, or as some kind of ‘chain’ that links these up, much of the fascination with and utility of reactions is that they appear to circumscribe deliberation of one kind or another. Neoliberal scholars praised markets precisely because they rescue society from reliance on moral deliberation (e.g. Friedman 1953). Sociologists of critique have emphasized how rival ‘orders of worth’ (assumed moral metaphysics) become mobilized by actors in situations of uncertainty, which can be translated into ‘tests’ of worth in the form of methodologies, measures, audits and disciplinary frameworks (Boltanski and Thévenot 2006). But as already noted, this assumes classically modern rhythms of discipline and liberal governance that may no longer hold in the context of platform capitalism. Where the ideals of critique and judgement assume a liberal subject, capable of critical distance, the ideal of reaction assumes an embodied cyborg (or what Deleuze terms a ‘dividual’), which communicates impulsively and non-verbally. Facial, neural, embodied and pre-conscious behaviour attain a kind of truth, where they come to appear unmediated by discourse or reflexivity (Andrejevic 2013).

Nevertheless, we are accustomed to treat reactions as important valuations of one kind or another, even while (or because) they have been untethered from the autonomy of critique. Live comedy becomes a paradigmatic cultural form in this respect, inasmuch as the immediate laughter of the audience would seem to count as a better measure of value than any subsequent review, written the next day, by a critic (Friedman 2014). An embodied reaction (such as laughter) potentially tells us something about both what prompts it, and the person responding – perhaps even about humans in general. It offers psychological, even anthropological, insight.

For the sociology of value and valuation, the turn from critique to reaction suggests a significant transformation in the implicit organization of everyday cultural and moral life. One way of grasping this, following Goffman, is in a shift from theatrical formats of observation to laboratory formats, and related transformations in the status of observers (Goffman 1999). The modern ideal of the critic is as an observer who achieves sufficient distance from the observed, as to leave no trace of their own private identity or prejudice in their assessment, with the expectation that their judgement will count as generally valid (Daston and Galison 2010; Kant 2007). Numerical reports and scores add to a judgement’s sense of impersonality (Porter 1995). The reactive person or body, by contrast, is one that has been deliberately immersed in a situation, so that their responses can be observed by others, by way of an experiment. The moral or cultural authority of reaction lies precisely in the absence of self-conscious reflection or critical distance. A degree of distance is necessary for the figure who plans and arranges the prompting of behavioural impulse is necessary, but this is the distance of the scientist, not of the critic. The epistemological assumption here is that, where data is generated by an unwitting experimental subject, their lack of reflexivity and the absence of any validity claims for their behaviour offers greater opportunity for a scientific, causal analysis.

Reactions have indeed played an important role in the history of scientific psychology, and its development as a discipline separate from philosophy. Early experiments by Wilhelm Wundt in Leipzig sought to measure ‘reaction time’, while one of Wundt’s crucial innovations was the spatial reorganization of his office, to include a separate area for the practice of experiments (Nicolas and Ferrand 1999; Rieber 1980; Robinson 2001). Psychophysicists, such as Gustav Fechner, sought to measure how different physical objects produced different quantities of ‘sensation’, as a bodily response (Crary 2001; Heidelberger 2004). Psychology experiments necessarily involve some spatial and optical reorganization of everyday situations, and an implied power imbalance between observer and observed. These artificial arrangements have been presented as vehicles to achieve deep insights into the nature of human beings, including their spontaneous and pre-conscious behaviour. ‘Reality’ must be ‘provoked’ into revealing itself, especially at historical junctures where the question of the ‘human’ seems most pressing (Lezaun, Muniesa, and Vikkelsø 2013).

The template, epistemology and politics of behavioural experimentation have exited formally designated laboratories in various ways. ‘In vivo’ experiments and randomized control trials treat ‘real world’ situations as opportunities for experimental learning, so long as feedback loops can be established (Kelly and McGoey 2018; Millo and Lezaun 2006; Muniesa and Callon 2007). ‘Pranks’ can be used against the unsuspecting public, not so much to learn from them, as to teach them something (McLeod 2011), or just to extract entertainment from them. Hidden cameras and reality television add a further example of this, which take the affective atmosphere of high-surveillance societies, and intensify them for entertainment value.

The digital platformization of everyday life in the early twenty-first century means that all manner of social, economic, political and cultural situations can be manipulated experimentally, for purposes of learning on the part of platform-owners – something that has intensified paranoid epistemologies and conspiracy theories, weakening the credibility of liberal institutions (such as ‘mainstream media’ and electoral processes) at the same time (Andrejevic 2013). Meanwhile, the conventions of everyday interaction are thrown into some uncertainty as the metaphor of the ‘platform’ becomes ubiquitous, and any activity can potentially become elevated as an object of observation and evaluation, not only by the operators of surveillance infrastructure, but by other smartphone users (Gillespie 2010). Doubt is cast as to when and whether one is (in a formal sense) on a ‘stage’, in an ‘audience’ or in a ‘laboratory’, once these situations (and the roles that accompany them) become swiftly interchangeable.

Returning now to the example of ‘reaction videos’, we can see these as a cultural effect of the blurring of ‘theatrical’ and ‘experimental’ situations, and of critical judgement and behaviour. The capacity of smartphones to extract ‘content’ from one situation, and to share it in another, has contributed to a kind of everyday ‘context collapse’, whereby relationships between observers and observed are manipulated unexpectedly. One manifestation of this is simply how public spaces or beauty spots become converted into backdrops, for those trying to generate engaging content for their Instagram followers. Another is the rise of pranks and filmed surprises of various kinds, which (whether benign or not) combine theatricality with experimentation. The unsuspecting subject of the prank is deliberately given no time to develop a critical relationship to it, as it is their affective and impulsive one – be it positive or negative – that is being provoked and captured.

This turn from critical valuation to behavioural valuation becomes more explicit in one prominent genre of music reaction videos, where inter-generational cultural differences are a means to witness a song being heard for the first time. There are now various ‘Mom reacts’ YouTube channels, based around the premise of sharing songs which are well-known to younger generations, with the YouTuber’s mother. One famous music reaction video features teenage twin brothers listening to Phil Collins’s 1981 song In The Air Tonight, ostensibly for the first time. The entertainment value lies in witnessing the surprise, pleasure or shock of these listeners. YouTube features hundreds of videos of babies hearing their mothers for the first time, thanks to new cochlear implants.

This kind of ‘reaction content’ deploys an experimental epistemology and method, to capture the impact of novel experiences for the very first time, often where that experience (i.e. the song, film, sound or sight) is far from novel to most of those watching the video and witnessing the reaction. By artificially engineering such a situation, with a view to capturing the reaction as content, behavioural impulse is being deliberately provoked as a means of valuation. The true value of (say) In The Air Tonight or of a mother’s voice becomes supposedly expressed on the face of the naïve listener. The implication here is, firstly, that value can no longer be grasped via conventions of critical judgement, perhaps because content is now so over-abundant, old songs now so instantly available, that we are no longer able to rely on critics or our own experience to measure cultural value. By contrast, the face of the naïve listener being ‘triggered’, ‘provoked’ or ‘lit up’ becomes a barometer of some real underlying value. The reaction thereby reflects back upon the artefact or experience that is used to provoke the behavioural response. The behaviour of babies and animals have a communicative potential precisely because they are least acculturated – another symptom of the decline of symbolic efficiency.

This turn towards the face of the naïve witness as a measure indicates an ambivalent concern with the ‘human’ element of valuation, that falls short of a complete post-humanist turn to cybernetics. The fascination with impulsive behaviour embeds the study of human psychology very firmly within the broader terrain of animal psychology (Mills 1998). In the context of the reaction video, the reflexes of the face and body possess an ‘authenticity’ to the extent that they seem involuntary, and rooted in biology and the pre-conscious. Yet, the allure of these responses (and the effort that is made to provoke them in the first place) derives from a cultural and moral preoccupation with the nature of value, and the capacity of the human animal to measure and agree on it. Economists and psychologists have periodically entertained the Benthamite hope, that the human mind or brain may contain some ‘ultimate’ barometer of value (Colander 2007; Wade Hands 2009). This kind of neural or cyborg subjectivity attains its influence and profile, not because it demolishes humanist problems of value and valuation, but because on some level it promises to settle them peacefully – the hope that we might be ‘wired’ to achieve aesthetic consensus, without needing to deliberate first. Fields such as neuro-economics and neuro-marketing may start from the premise of circumventing what consumers say they value, but then shift the measure of value into the body instead (Schull and Zaloom, 2011).

Lurking in an ambivalent space between critique and behaviour, reactions (as captured in reaction videos) have an ambivalent form of predictability, which arguably accounts for their charisma and virality. On the one hand, the reactions which are sought and shared online are eminently predictable and unsurprising. The child filmed unwrapping a present gets happy and excited when they see it; the woman interrupted in the street with a bunch of flowers looks confused and slightly creeped out; the teenager watching Silence of the Lambs for the first time looks scared. The familiarity and predictability of these responses are a grounds for empathy and feelings of common humanity, grounded in something that is deemed ‘authentic’ (as opposed to ‘fake’). On the other, the precise split-second moment of reaction, the exact motion of the face and hands, is spontaneous and unforeseeable, subtly different from every other. There is still a possibility of going in a different path altogether (what if the child opening the present bursts into tears?). Suspense and drama is not eradicated altogether. The reactive (and surveilled) subject hovers in an ambivalent evaluative space between critical autonomy and animal behaviour, that signals a zone of compromise between the biological and human sciences (Rose 2013).

As in the paradigmatic case of laughter erupting spontaneously from the live audience, affective and embodied reaction is something that has long been recognized as a capacity of crowds (Borch 2012). Populist leaders and critics of liberal democracy, such as Schmitt, have long believed that instinctive, immediate crowd responses have an authenticity about them that representation (whether via parliament or media) lacks. Crowds express their valuations, not through judgement but (Schmitt argued) through acclamation, praising the figure they value, denouncing the one that they don’t. Acclaim is one way of framing the ambiguous valuation mechanism, that sits between behavioural impulse with critique, and may indeed be seen as the chief mode of expression facilitated by ‘reaction’ buttons such as Facebook’s (Dean 2017). This turn towards non-representational valuation, hovering between predictable impulse and autonomous judgement, is therefore how the logic of the social platform combines with that of populism, arguably benefiting the Right more than the Left. But it may also open up space for ethical and political consideration of what is deserving of ‘personhood’, in the grey area between ‘human’ and ‘post-human’ (Osborne and Rose 2023).

Between judgement and artefact

The concepts of ‘platforms’, ‘platforming’ and ‘platformisation’ have grown ubiquitous in the twenty-first century, not only due to the rise of large digital platforms (such as Facebook et al) but as metaphors for how certain ideas are elevated and acted upon (Gillespie 2010). A society organized around the logic and metaphor of the ‘platform’ is one in which there are constant cultural, moral and political choices as to where attention should be paid, what deserves to be promoted in the public sphere (and conversely demoted), and who one should share one’s ‘platform’ with. Digital platformization, combined with new metaphors of ‘platforming’ and ‘platformisation’, occasion a breakdown of the dramaturgical organization of everyday life as theorized by Goffman, in which there are implied boundaries between ‘stage’, ‘back-stage’ and ‘audience’ (Goffman 1999). In an architectural sense, a ‘platform’ is a far more ambiguous entity than a ‘stage’. As a result, in a society of platforms, boundaries between public and private, actor and audience, actor and critic, are far more fluid. In the case of social media platforms, for example, every user is simultaneously watcher and watched, valuer and valued.

These are the conditions in which the surge of reaction content has occurred, whereby watching and listening become performances in their own right. But broadcast formats already existed, which turned valuation into a ‘spectacle’ (Muniesa and Helgesson 2013). The BBC TV show, Antiques Roadshow, first aired in 1979, sees members of the public bring their antiques and heirlooms for the inspection by experts. While the experts remark on all manner of details of the objects concerned, the inspection isn’t over until the crucial moment when they ask, ‘and have you had it insured?’. This is the prelude to offering their own expert monetary valuation. In some cases, these valuations are far higher (or lower) than expected, producing a frisson of excitement on the part of the owner and the live audience. This elevation of valuation into a kind of public performance has since become familiar in a range of reality TV formats, such as talent shows, where attention is heavily focused on the opinions and reactions of a panel of judges, who develop distinctive personae in how they deliver their valuations. Formats such as The Apprentice and Dragon’s Den take aspects of contemporary capitalism (hiring processes and venture capital investment respectively), gamify them, and convert the resulting decision-making and justifications into entertainment. Documentary and history formats, in which presenters are filmed gazing lovingly at a monument or painting, operate in a similar fashion, turning valuation into a performance.

Common to all these media and genres is something which deserves the attention of sociologists of critique and valuation: the conversion of judgement into an artefact or type of content, to be judged. Once the critic or viewer enters the frame of observation and surveillance, so it becomes possible to watch, assess and enjoy how others watch, assess and enjoy things. Philosophically, this speaks of an anxiety that is core to modernity of how to establish some foundations for value and valuation, in the absence of any transcendent or extra-human values (Foucault 2005; Habermas 1987). Kant’s answer to this was to turn critique upon itself, to use human reason to establish basic principles as to what reason can and can’t establish. But this is perilously close to Nietzsche’s response, which is to question the very value of values as such, once humanity is its own measure. Disputes between ‘critical sociology’ and ‘sociology of critique’ centre around whether the priority of sociology should be to provide the grounds of value (through a theory of justice) or to map existing spheres of value and valuation (through a pragmatic reconstruction of everyday theorizations of justice).

The conversion of judgement into a form of embodied, affective performance, to be observed, enjoyed and judged by others, represents another mutation in the reflexivity of valuation and critique. Reaction videos take critical reaction, and aestheticize and somatize it, such that it becomes a spectacle. Thanks particularly to platform infrastructures, that allow for dramaturgical divisions between ‘stage’ and ‘audience’ to be suspended, these videos produce what might be termed vicarious valuation practices, in which a particularly expressive or extrovert individual comes to emote, judge and react on behalf of others, who watch this performance in the manner of a more conventionally passive ‘audience’. As already noted, something like a Greek chorus represents a precedent for this vicarious valuation, in which the boundary between the play and the audience reaction is blurred. But how are we to make sense of the phenomenon of vicarious valuation in today’s digital public sphere? Three logics are discernible.

Firstly, there is a simple question of critical authority in play. The ideal of art criticism, at least in the modern era, assumes that certain individuals are better qualified than others to judge a novel or opera (Bourdieu 1984). Some of these are sufficiently qualified as to become professional critics, who are paid for their valuations, which may then be read or watched as valuable cultural genres in their own right (literary critics can end up more celebrated than the authors they review). Processes of judgement and evaluation have long been vested with literary and theatrical value in their own right. What appears to change, with the rise of talent show formats and subsequently ‘reaction videos’, is that the source of value shifts towards affective impulse and the body. Certain faces and bodies (being especially expressive or sensitive to aesthetic impact) serve as an authoritative measure of value, thus representing or confirming the value of whatever song, film, performance or game they are currently ‘consuming’. There is thus a shift from a form of liberal credentialism, in which the authority of the critic derives heavily from their humanistic education, reputation and publication record, to one of networked virality, in which critical authority lies in how ‘engaging’ and shareable a reactive performance is.

Secondly, vicarious valuation offers a route through which dispersed, privately-held feelings and values can acquire a public status and prominence. In a relatively mundane sense, this is how many opinions and judgements travel mimetically, especially through social media: a reaction or valuation will often be shared or liked precisely because others share it, and find it to be a better expression than they themselves are capable. But this is also the logic of populist demagoguery, in which a given leader claims to be expressing what their followers feel (perhaps secretly), offering a more direct form of representation than parliamentary democracy (Müller 2017). Thus a figure such as Donald Trump (with a background as a reality television ‘judge’ on The Apprentice) masters the art of reacting with disgust and outrage to the 24/7 news cycle, emoting spectacularly, and relieving his followers of the pent up urge to do so (Wahl-Jorgensen 2018). In the digital public sphere, the high-profile political reactor (and reactionary) ensures that certain feelings of anger acquire a mass presence.

Finally, vicarious valuation is a means of accessing cultural value empathically and indirectly. Thus, somebody reporting very fluently on their appreciation of The Beatles, say, may enable the listener to share in some of that pleasure, through learning of the effect of the music on the original listener. But in the case of reaction videos, this vicarious relation once again becomes more somatic and affective, such that the emotions of the viewer, listener or gamer are transmitted via the face, body and tone of voice, unmediated by discourse – holding out the faint hope of a ‘pure’ or ‘authentic’ empathic connection. Insights from ‘social neuroscience’ suggest that witnessing another person in obvious pain or ecstasy affects similar parts of the brain to those of the person originally affected (so-called ‘mirror neurons’) (Cacioppo and Patrick 2009; Rose and Abi-Rached 2013). This sort of finding adds scientific authority to an already existing cultural circuit, in which affectivity is generated and shared, with the aim of setting off emotional mimesis. Thus, the sight of a highly expressive and positive reaction on the face of the other (thanks to some happy surprise or enjoyment of a song) provides a small ‘hit’ of positive affect, which leads it to become shared with others as a kind of emotional gift.

Conclusion: a sociology of reaction?

Sociology, as Weber first proposed it, is the study of ‘action’, in the sense of practices which have mutually understood social meaning, and which take account of others (1978, 4). For action to be understood by the sociologist, it is necessary for its immanent meaning and reasons to be brought to light, through some hermeneutic process. As Habermas argued, building on Weber, actions are reasonable in terms of the different types of validity claims that they implicitly make, be these concerning some objective outcome that they seek, some norm they seek to uphold, or some subjective feeling or desire they express (Habermas 2015). The pragmatist ‘sociology of critique’ proceeds along the same lines, on the assumption that actors are mutually intelligible and equipped to evaluate one another (Basaure 2011; Boltanski 2011). ‘Action’ has also been privileged philosophically and politically, as the uniquely human capability that gives birth to novel political associations and institutions, and moves history forwards (Arendt 1958). Action is contrasted with mere ‘behaviour’ (or, for Arendt, ‘social’ life), in which individuals obey autonomic and biological impulses, losing their distinction from the rest of nature. Thus ‘action’ carries a heavy burden, not only methodologically for the sociologist, but in the hopes of modernity and of democracy.

Action, of course, rarely occurs without some kind of audience or observer, which may or may not be a sociologist. One question is whether action is intentionally aimed at making an impression on the observer, that is, as interaction or what Habermas termed ‘dramaturgical action’, whose validity consists in convincing the observer of its authenticity (2015, 86). What, then, might be added to the sociology of action and of critique, by turning our attention to reaction? There is both a pessimistic and a more optimistic answer to this.

Pessimistically, it is worth recognizing the extent to which the infrastructures of neoliberal society trap us into cycles of reactivity, due to the synoptical arrangements whereby everyone observes the observations and responses of everybody else. Thus, social media platforms render people so alert to the possible responses of others, that they internalize online feedback into their very sense of self, with known harms to young people in particular (Haidt 2024), while public institutions and space are utilized and designed around maximizing digital feedback. Democracy becomes dominated by cycles of trolling and trap-setting, in which politicians select their words and their policies in ways that are aimed only to provoke reactions in their opponents, and not at action. Economic policy becomes geared around reassuring ‘the markets’ and hoping that emergencies such as climate change can be alleviated through emergent, spontaneous responses, rather than through deliberate planning. One might argue, following Luhmann, that these are simply examples of the ‘self-referentiality’ of all social systems, which ensure a degree of both closure and openness to new meaning and knowledge (Luhmann 1995). On this basis, Luhmann believed that ‘planning’ was ultimately impossible, seeing as it immediately generated resistance to itself. But neoliberalism was explicitly conceived as an alternative to coordinated political action, which was assumed to have failed across multiple fronts by the 1970s, and thus a deliberate block on planning, deliberation and action. The architecture of both financial markets and giant platforms intends to trap us into cycles of reaction, and not simply interaction, which may not equate to reactionary politics in the conventional sense, but certainly obstructs action as Arendt conceived it. Sociologists needn’t become behaviourists to recognize that society has been governed and formatted around behaviourist and post-humanist assumptions, which cannot be grasped on the assumptions of communicative rationality.

On the other hand, and more optimistically, this paper’s own examination of ‘reactions’ in the digital public sphere has hopefully demonstrated that hermeneutic understanding is not defunct, and that the status of the ‘human’ is not expunged altogether from these self-referential cycles. On the contrary, some vestige of critique and a search for validity remains at work, albeit in an ambivalent form. I’ve emphasized two key ambivalences surrounding the place of reaction. Firstly, reactions are provoked and observed in search of some kind of universal, trans-cultural critical faculty, as if the pre-conscious reflexes of the face and brain might betray the truth of cultural value in some way. Secondly, reactions are constantly morphing into performances (or actions), and back again, as the question of what to ‘platform’ or promote is radically opened up. In neither case is it clear what validity claims are in play, and the obstacles to a hermeneutic sociology of critique or of action are profound. But we can find elements of deliberation, criticism and valuation in the affective reflexes of the digital public sphere. The question, politically, if forms of deliberate, coordinated action are to be resuscitated is how to move beyond this ambivalence, break out of ‘reaction chains’, and move purposefully towards the future, rather than emergently and reactively.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

William Davies

William Davies is Professor of Political Economy at Goldsmiths, University of London.