Why Do I Keep Crying?

Abstract

This article explores how my performance videos Googling Things in Hell 1 (Tammy) & 2 (Daniel) (2021) capture the contradictory nature of emotional performances in surveillance capitalism and convey them to the viewer. When we perform emotions online through social media posts and comments, or as in these videos, searching for content on Google, personalized content isolates us by using our emotions as data to provide or sell us what the algorithm thinks we want. But the generic nature of our search results shows that our feelings are never truly our own and are always structurally determined by capitalism.

Following Eve Sedgwick and Adam Frank's reading of the mid-century psychologist Silvan Tomkins, the article presents a ‘weak theory' of the emotions under investigation by staying close to the surface of the emotional performances captured in the videos. But, by thinking through the aesthetic and form of the screen image, and the textured feel of the activity taking place on screen, the article reveals the ‘strong' theory of emotions deployed by Google Search, which aims at profiting from those emotional performances by interpreting them as potential requests for information about products or services provided by Google’s advertising clients.

This reductive way of interpreting emotions is contrasted by the videos' interest in the specificity of the online performances of emotion captured on-screen. The article concludes by suggesting that these performances of emotion are necessarily interested in the conditions of their own possibility. Using ideas by the artist Andrea Buttner and the affect theorist Lauren Berlant, the article claims that my videos' maintenance of interest in their on-screen performances of emotion might help a viewer acknowledge, understand and reflect on the political conditions in which they take place.

This article explores how my performance videos Googling Things in Hell 1 (Tammy) & 2 (Daniel) (2021) capture the contradictory nature of emotional performances in surveillance capitalism and convey them to the viewer. When we perform emotions online through social media posts and comments, or as in these videos, searching for content on Google, personalized content isolates us by using our emotions as data to provide or sell us what the algorithm thinks we want. But the generic nature of our search results shows that our feelings are never truly our own and are always structurally determined by capitalism.

The question that underpins this investigation is, ‘How does it feel to have our feelings mediated by surveillance capitalism?’ However, following Eve Sedgwick and Adam Frank’s reading of the mid-century psychologist Silvan Tomkins, this article will avoid trying to answer this question directly because it risks homogenizing the emotions under examination. Instead, by examining emotions of sadness, shame and despair performed as Google searches, this article offers descriptions, reflections and tentative ideas about these emotions that cannot be universally applied to emotion in general. In some areas of theoretical work this would be considered a flaw, but as Sedgwick and Frank note, for Tomkins, an effective affect theory is a weak theory, relying on description, closeness to its object and being ‘expandable only through textured analogy’ (1995: 519, footnote 19). Strong theory tries to apply itself everywhere all the time, but weak theory is never universal because it is concerned with the particularities of its object rather than general applicability. This article stays close to the surface of the emotional performances captured in the videos, thinking through the aesthetic and form of the screen image, and the textured feel of the activity taking place on screen.

Googling Things in Hell 1 (Daniel), still from HD video, 10 minutes, 2021

This article produces a weak theory of the emotions performed in the videos. However, this weak theorizing will reveal the strong theory of emotions deployed by Google Search, which aims at interpreting and ultimately profiting from those emotional performances. The article concludes with a revaluation of the political meaning of ‘sadposting’ and ‘doomscrolling’, suggesting that the online performances of emotion taking place in the videos are necessarily interested in the conditions of their own possibility. Using ideas by the artist Andrea Buttner and the affect theorist Lauren Berlant, the article finishes by describing how my videos’ maintenance of interest in their on-screen performances of emotion might help a viewer acknowledge, understand and reflect on the political conditions in which they take place.

Googling Things in Hell 1 (Tammy), still from HD video, 7 minutes, 2021

GOOGLING THINGS IN HELL

Googling Things in Hell 1 (Tammy) & 2 (Daniel) (2021) are performance videos in which Tammy Reynolds or Daniel Oliver sit on a plastic chair against a backdrop of psychedelic visuals, using their phones to search for things online¹. While we never see their faces, we observe their search terms and results through close-up shots of their phone screens and open captions in square brackets. The videos are part of an installation called In Hell (2021), consisting of 118 performance videos in which Tammy and Daniel portray themselves engaging in various activities during a fictional artist’s residency in hell. All the performances were filmed in a green screen studio at Kingston School of Art in London over two weeks in 2021.

Each performer searches the same sequence of phrases in both videos. The viewer sees the performers entering search terms and receiving real-time results. The editing is consistent across both videos, alternating between an over-the-shoulder close-up of each performer’s phone screen and a mid-shot of their torso and legs while they sit on a chair, holding their phone. The close-ups allow the viewer to read the information displayed on the phone screen. These shots hint at the voyeuristic perspective of the viewer regarding the emotional performance. What the viewer watches is not a theatrical reenactment of Internet searches, but rather a making visible of the private details of personalized search results received by each performer during the performance².

The performers play themselves as characters. They used their own phones when we filmed, and they did not use incognito mode, an ad blocker or virtual private network (VPN). The results we see on-screen are their actual personalized results based on the information Google has about them³. Through the use of personalized search results as part of their content, these videos address their political and economic context in which commercial digital platforms owned by enormous companies have both expanded and disrupted the public sphere by incentivising online performances of emotions.

SURVEILLANCE CAPITALISM

The action in the videos takes place in the ‘emotion economy’ of what Shoshana Zuboff calls surveillance capitalism in which emotions are part of a ‘behavioural surplus’ that companies like Google collect and draw on to predict and modify behaviour for profitable ends (2019: 126, 273). This is visible in the videos in the performers’ Google search results based on personal data that Google holds about them. It is also formally represented through the videos’ use of over-the-shoulder close-ups to film each performer’s phone screens, as well as the open captions that display the performer’s search terms as they type. The viewer, through the voyeuristic camera angle and the on-screen text, becomes part of the surveillance operation, watching and assessing as new data points are created and triangulated with each search term.

Zuboff’s investigation of surveillance capitalism tends towards the sci-fi dystopian in its focus on the way that machine learning and artificial intelligence (AI) are able to conduct ‘emotion analytics’ on a wide range of inputs, including human facial expressions, and in her claims that surveillance capitalism is an attempt to impose ‘a new collective order based on total certainty’ (Zuboff 2019: 8, 267). This article will avoid interpreting the performances of emotion in the videos as though they are surveilled by the super intelligent/malevolent AI that stalks the imaginary of Zuboff’s book. The performances of emotion taking place in the videos are expressed in language that does not require the more advanced technology discussed by Zuboff in order to interpret it. The sentiment analysis applied to the search terms in the videos is much more straightforward: matching search terms with words found on websites or in the advertising copy of the companies who pay for Google Ads. Zuboff’s interest in presenting surveillance capitalism as being able to name what we feel more accurately than we can ourselves means that she draws on an idea of emotions as pre-linguistic, with their expression in words being a translation of their internal reality (Zuboff 2019: 267). The idea of pre-linguistic feelings is closer to the way many theorists think about affect, such as Rebecca Coleman (2017: 527) who describes affect as ‘elusive, excessive, non-rational and/ or difficult to articulate through language’⁴. This article contains a performative analysis of emotions, investigating the way emotions take form in language, and understanding these performances as one of many ways in which emotions are constituted rather than merely described.

Following J. L. Austin’s work on performative utterances, we can say that when we name or express our emotions, we perform actions (naming, expressing) rather than just describing an internal state (1962: 12). These discursive actions have outcomes, or can set off a chain of reactions. In these videos, each time the performers search for a phrase, Google returns results based on an emotion it calculates is expressed in the search term, and both the action and reaction (although not the inner workings of the algorithm that produces the reaction) are available for analysis. This article does not need to enter the realm of the ‘technological sublime’, in which technology is beyond comprehension, to analyse the emotions at play and begin to understand what is happening in their performance (Ames 2018: 2).

BIO-MORALITY

The search terms can be roughly separated into five categories: phrases and questions related to anxieties around bodily norms (for example, ‘normal sized penis’)⁵, questions suggesting a desire to understand or overcome sadness and other negative feelings (for example, ‘why do I keep crying?’)⁶, questions about the quality or status of things that seek impossibly objectiveanswers (for example, ‘who is the best person?’)⁷, existential questions (for example, ‘why do we die?’)⁸ and one question that seems to be more an expression of interest than a performance of emotion (‘why does my cat eat my earwax?’). Some of these were written by me, some were discovered by typing in the beginning of a phrase and allowing the Google Search autocomplete feature to suggest the ending and one was suggested by Tammy.

Putting aside the question of cats and earwax for now, the search terms depict a contemporary kind of sadness – a mix of anxiety, depression and shame, as well as a confusion regarding societal norms, quality and material success.

This is the felt reality of the capitalist equation between emotional valence and moral value that Alenka Zupančič calls bio-morality. Biomorality establishes happiness as a moral imperative under capitalism by equating feeling good with being a good person and feeling bad with being a bad person (Zupančič 2008: 5). Under bio-morality, feelings of sadness, anxiety or stress are never appropriate because they are bad feelings. If we express bad feelings this is understood as a moral failing, rather than as an appropriate response to our circumstances. Many people in the UK experience stressful and precarious working conditions, rising costs of living and real-terms pay cuts (Leaker 2023). These structural features of capitalism are very likely to produce bad feelings, and yet we are discouraged from attributing them to our circumstances (Mental Health Foundation 2023).

But is bio-morality at play in the search results we see in the videos? When Tammy googles, ‘Why do I keep crying?’, the top result is an advert for Better Help Online Therapy. The ad contains an extended version of the search term: Why do I keep crying for no reason?, answering itself with a nullifying tautology. It does the work of bio-morality through the privatized cognitive behavioural therapy (CBT) model of neo-liberal self-care in which negative feelings are located entirely in the minds of individuals, and are pure pathology that require fixing, rather than analysis. One example of this is the way in which large companies offer their staff access to mental health hotlines or meditation apps as a way to encourage them to seek individualized therapeutic solutions to discontent at work rather than taking part in collective political struggle for better conditions (Purser 2019: 16, 57).

Googling Things in Hell 1 (Tammy), still from HD video, 7 minutes, 2021

Under the conditions of surveillance capitalism, one of the outcomes of Tammy’s performance of sadness is economic activity, with Better Help paying Google for the right to show Tammy their advert. If bio-morality is functioning here, then it isn’t as a violent act of shaming that stops any expression of negative emotions. Bio-morality in surveillance capitalism doesn’t need to shame us so that we never express our sadness, but just enough so that we perform our sadness in order to find solutions to it. Conveniently enough for Google and its advertisers, those solutions come in the form of commercial services such as online therapy.

COLLABORATING WITH YOUR UNCARING ALGORITHM

Tammy and Daniel sit in swirling psychedelic isolation, mute and muted. They type their search terms into Google and they receive personalized search results in return. This is the outcome of a collaboration between each performer and an uncaring algorithm that uses their data to decide what they see. The performances of emotion captured in the Googling Things …  videos are collaborations in which the performers act out an expression of emotion online alongside the data that Google holds about them.

When the performers type their words into Google and press enter, those words become actions that result in pages and pages of links and images. The search terms are performative, in the sense proposed by J. L Austin, of being words that do things. Austin’s idea of performativity was taken on by later theorists such as Sara Ahmed (2004) who used it to think about the constitution of identity through linguistic performances. The collaboration between the performers and Google is performative in the specific way described by Sara Ahmed in her writing on the performativity of disgust in that it constitutes a group identity.

Ahmed writes that when we express our disgust in words we perform two related actions: designating both the objective status of the disgusting thing as disgusting, and constituting a community of the disgusted who agree on the status of the disgusting thing. In Ahmed’s examples of expressions of disgust posted on Internet message boards in the wake of the 11 September 2001 attacks on the World Trade Center,

The speech act, ‘That’s disgusting!’ generates more than simply a subject and an object; it also generates a community of those who are bound together through the shared condemnation of a disgusting object or event. (Ahmed 2004: 94)

When we express disgust we want other people to be disgusted alongside us. It is a way to reassure ourselves that we are part of a majority enforcing certain norms within a specific context. The formation of this community is one of the motivations behind expressing disgust but, crucially, it is also an essential outcome of its performance. The formation of a community is one of the necessary results of the action we perform when we publicly express disgust.

The performances of emotion in the videos also constitute an object in the form of a named emotion, and a community bound together by that emotion. The critical difference is for whom that community is constituted. For Ahmed, those who perform disgust constitute a community of the disgusted for themselves. But in the videos, when the performers express their emotions in the commercial, semi-private form of Google search, they collaborate with Google in constituting themselves as a particular demographic for Google’s advertisers.

STRONG THEORY AND EMOTIONAL HEURISTICS

In the videos, the performers are never satisfied by their searches. They scroll through their results, idly clicking on websites that catch their eye, skimming the pages, then clicking back. The end point of each Google search is another Google search. ‘Normal sized penis’ leads to ‘normal sized forehead’ leads to ‘normal sized Babybel’⁹. Why can’t Google give the performers the answers they need?

For most of the Google searches captured in the videos, the personalization of the results takes the form of a re-ordering of similar pages from big websites including the National Health Service, Quora, Healthline, Wikipedia, wikiHow and Reddit. The results are personalized, but only to the extent of rearranging content from big providers such as public health bodies, commercial companies or large databases of user-generated content. The idea that our search results are individual to each of us might lead us to believe that under surveillance capitalism our emotions are cut off from everyone else’s. But the generic nature of the web pages returned by the search engine suggests that our sadness is not our own, and is instead structurally determined by life under capitalism.

In the videos this contradiction is played for laughs. Some of the search terms express existential anxiety or despair but the top results are comically non-specific or upbeat. ‘Why do we die?’ gets the Wikipedia page on death. ‘When can we love again?’ gets a chirpy how-to guide from wikiHow. The disparity between the specific emotions performed in the search terms, and the general nature of the webpages in the results hints towards an economic fact: Google and other platforms that collect huge amounts of data about us in order to personalize our content are only really interested in our individual characteristics when they align us to a specific demographic that might be interested in paying for a product or service.

The introduction promised that this article would stay close to the emotions performed in the videos in order to construct a weak theory of the feelings at play. But what has become clear is that Google’s search results are for the most part the outcome of a very strong theory of emotion that aims to maximize ad revenue. Google’s strong theory of emotion processes all performances of emotion on the basis of their potential monetization. When a company is willing to pay for advertising in relation to a certain search term, Google categorizes emotional performances as symptoms of common conditions that require consumable goods or services in order to treat them (for example, crying is categorized as a symptom of treatable depression). However, if Google fails to find an interested advertiser, emotional performances are interpreted as enquiries for information (for example, ‘why do we die?’ becomes a request for information about death). What is notable is not the well-documented fact that Google and other large companies seek to monetize our emotions, but rather, that to do so, they construct a strong theory of emotion in which feelings are homogenized in a particular way. In Google’s strong theory of emotions, all emotions are understood as either problems that can be solved by purchasing a product or requests for information.

The comical failure of the search results to adequately respond to the performers’ desires to understand their own emotions indexes the strong theory of emotion that constitutes Google’s emotional-commercial demographic model. The insight here is aesthetic in the sense that it is inscribed in the video image. The viewer does not need to decode the mathematical logic of the algorithm in order to understand it, but rather can see it in the performer’s search results. The black comedy of the videos is an indexical mark of the economic process at play.

In the writing of the artist Andrea Büttner, shame is ‘an heuristic affect, which makes the norms of representation, along with the structures within which representation exists, visible’ (2020: 38). In Büttner’s theorization, shame is an emotion that can overflow the banks of its normative function to expose norms themselves to scrutiny. To recognize that we feel shame is to know something about the societal norms that are its basis. The performances of emotions in the Googling Things …  videos index an economic process and offer the viewer an opportunity to understand the wider political context of that process. While watching the performers google their emotions, we don’t learn much about the performers as private individuals, but we do learn something about the way selfhood is formed, described and categorized by the economics and politics of twenty-first-century surveillance capitalism.

MAINTAINING INTEREST

At first glance the politics of the Googling Things …  videos seem to take the form of a one note satire: ‘Aren’t we stupid to offer all of this data to a large company in the futile hope that we might find an answer to such personal questions?’

However, one search term is not like the others. ‘Why does my cat eat my earwax?’ seems to contain very little emotion – it can’t be read as a performance of depression, anxiety, shame or even disgust¹⁰. To return again to the work of Silvan Tomkins, we could categorize this question as a performance of interest. For Tomkins, interest is a fundamental affect in the way it appears alongside the beginnings of other feelings, pairing them with objects. We notice something, it interests us and interest connects that object to joy or fear or excitement or any other affect. As Sedgewick and Frank note, for Tomkins shame is also fundamental in the way it can be the end point to all other affects in a strong shame theory. For Tomkins, we can be shamed out of feeling interest. If interest causes us to look, then shame makes us look away. For this reason, in Tomkins’ framework, interest is at one end of a shame–interest polarity (Sedgwick and Frank 1995: 500). Shame shuts down interest in an object. Sustaining interest means we are not ashamed.

Googling Things in Hell 1 (Tammy), still from HD video, 7 minutes, 2021.

This search term is fuelled only by interest (we just have to know why our cat eats our earwax) and does not seem to turn into another emotional state. And it does not seem to produce a bio-moral shaming response from the algorithm. The search results allow us to sustain our interest further. The closest the search results get to a negative emotional response is a website that explains why it is fine for cats to engage in what might appear to be a disgusting activity (Traveling with Your Cat 2022).

But on reflection, sustained interest is also at play in all of the other search terms, even though they are also performances of other emotions and might invoke some bio-moral shaming content in the results. All the performances of emotion in the videos are simultaneously a performance of the emotion at hand and a performance of interest in that emotion via the on-screen acts of searching, scrolling, selecting and reading – and of course the creative acts of performing, filming, editing and viewing that produced these videos. This insight is crucial to understanding the politics of the videos. Googling an emotion is a performative expression of interest in that emotion, and the videos capturing these performances offer that interest as a potential response to the viewer. This self-reflexive emotional interest disrupts attempts to homogenize specific emotions into generalized valence by bio-moral shaming and the emotion economy of surveillance capitalism.

Performing emotions online is often derogatorily labelled as ‘sadposting’ and endlessly scrolling through negative news or alarming information as ‘doomscrolling’. And it may indeed be true that both of these activities can involve apolitical emotional wallowing in collaboration with our algorithms. However, acknowledging that all online performances of emotion can also be self-reflexive expressions of interest reveals that the totalitarian nature of bio-morality and the apparent omniscience of surveillance capitalism are not the whole story.

Lauren Berlant was a writer who engaged with public spheres as ‘affect worlds’, and thought about how negative emotions might be imagined as ‘forms of attachment’ rather than simply the effects of ‘bad power’. For example, Public Feelings, a ‘feel tank’ run by Berlant and their collaborators organized an International Day of the Politically Depressed in 2012, including a march where they distributed T-shirts and stickers that read, ‘Depressed? It might be political’ (2012: 340).

In their book Cruel Optimism, Berlant ends with a discussion of artworks that they call ‘ambient art in the recessive mode’ (2011: 231). This minor political form of art making is aligned to Melanie Klein and Eve Sedgwick’s understanding of the ‘depressive position’, in which a subject,

[a]cknowledges the broken circuit of reciprocity between herself and her world but who, refusing to see that cleavage as an end as such, takes it as an opportunity to repair both herself and the world. (Berlant 2011: 259)

If we maintain interest in the negative emotions we perform online we also perform this act of acknowledgement and stay open to the opportunity of repair. Repair here does not come in the form of a therapeutic solution, but as a recognition of the political nature of our feelings, and an openness to reconnecting with the political, despite the way it makes us feel.

This acknowledgement necessarily exceeds the comprehension of algorithms that reduce our emotions to commercial data points, and it disrupts the restrictive demographic groups within which companies such as Google want to locate us. Perhaps, by maintaining interest in emotions that run counter to the bio-moral imperative to be happy, there is a minor politics to be found in ambient artworks, such as these videos, and the communities of performers, artists, viewers and critics that form up around them. A politics in which that community can acknowledge the broken circuits, and keep asking questions that might lead to opportunities to repair them. Why are we depressed, anxious and exhausted? Why do we hate our jobs? Why does capitalism feel so bad? Why do our cats eat our earwax?

Notes

1 Tammy Reynolds is a performer, producer and drag artist who sometimes performs as Midgitte Bardot. On stage they ‘sing/dance/ speak/scream/shout/eat my trauma’. They make work with their disability, describing themselves as ‘always disabled … always a dwarf’ (Artsadmin 2020). Daniel Oliver (2015) is a performance artist, lecturer and researcher who makes ‘raucous, dyspraxic-led performance worlds’.

2 The personalized results each performer received on the day of filming would be different if they were to google the same search terms today. The results depend on ad spends by different companies, each webpage’s place in Google’s page ranking system and individual changes in the performer’s data over time.

3 Including search history, location data and other identifying information that they have willingly or unknowingly handed over to Google Search, Gmail or one of Google’s many other services

4 It is important to mention here that some theorists use affect and emotion interchangeably, including Silvan Tomkins, who referred to shame, anger and other nameable emotions as affects. This text uses affect when writing directly about Tomkins’s theories, but it is interested in emotions that take shape in language, rather than any pre-linguistic affects

5 Normal sized penis; normal sized forehead; normal sized babybel; normal sized hands; normal sized tongue; why does my vaginal discharge smell?; why do my testicles feel like a bag of worms?

6 Why am I a bad person?; Why do I keep crying?

7 Why is contemporary art so bad?; who is the best person?

8 Why do we die?; when can we love again?

9 If you know Babybel cheese but you are confused by the phrase ‘normal sized Babybel’ then I urge you to google it now. It will turn your world upside down. Babybel sold in UK shops is the mini version of the squidgy, wax bound cheese. Across mainland Europe there are also midi and maxi sizes of the cheese. I find it very helpful to show students images of midi and maxi size Babybel if I want

10 Tammy suggested this question as an example of something they googled in the middle of the night. This is why I enjoy collaborating with performers. to quickly problematize the idea of cultural norms having objective correlations.