Instagram-based Art as Feminist Phenomenological Practice

In this piece, I compare Husserl’s classic phenomenology with my own conception of algorithmic phenomenology or phenomenology of the masses, in which a false universalism of digital culture can be seen in the images that are recirculated by a mass-appeal algorithm. I use Bonnie Mann’s definition of feminist phenomenology, as derived from Beauvoir after her readings of Husserl and Merleau-Ponty, to explore how art on Instagram can subvert this algorithmic phenomenology. I use the internet artist Arvida Byström as a case study for how the algorithm prevents work with diverse particularities and a strong political message from arising, while still enabling infiltration for one’s own interests as a feminine artist.

In The Myth of Sisyphus (1942), Albert Camus famously accuses Edmund Husserl, founder of phenomenology, of philosophical suicide. Husserl’s “original” phenomenology is indeed an attempt to describe the world through experience rather than explain it in terms of a pre-existing grand principle. However, he also wishes to use these experiences to reach some sort of essence of each object of knowledge, which returns Husserlian phenomenology back into the trap of an abstract idealism that Camus, as an absurdist, detests: “he sounds like Plato” (2013 [1942], 34). However, since Husserl, the phenomenological notion of essence has shifted and even, in some areas, faded away. Merleau Ponty, for instance, speaks of a post-corporeal essence that can only be deduced when the object is historicized (Fielding, 2011, 524) (Butler, 1988, 521). Due to this sympathy for the “bodily subject in its concrete existence” (Heinämaa, 1999, 118), Merleau-Ponty’s re-elaboration on phenomenology was favored as a starting point by Simone de Beauvoir in her encyclopedic 2-volume account of historicized womanhood, The Second Sex (1949).

Several feminist scholars of philosophy question whether the feminist phenomenology employed in Beauvoir’s work was simply an application of a ready-made traditional phenomenology towards the new and uncommon topic of woman, or whether it was a fundamental rupture of what phenomenology was as a method and as a set of aims (Mann, 2018, 43) (Fielding, 2011, 518). I follow Bonnie Mann in her groundbreaking essay “The Difference of Feminist Phenomenology: The Case of Shame”, in positing that feminist phenomenology is a ruptured modification that defies the abstraction present in post-Husserlian phenomenology. For Mann, Beauvoir’s diverse and rich account is made possible through an interest in material specificity that would be rejected by a traditional phenomenologist’s reverence of a “generic subject” (2018, 61). As well, feminist phenomenology pushes past Merleau-Ponty’s corporealization of phenomenology in its insistence on material application to social injustice. Mann states that feminist. phenomenology. employs an “affective modality” which overrides the disinterested, catalogue-like nature of previous phenomenologists – The Cambridge History of Literary Criticism describes Husserl as aspiring to “a pure and rigorous enterprise that would eschew all prejudgements and dogmas” (Holub, 1995, 290) – with a reflexive embrace of the irritation and humor that accompanies the recognition of injustice (2018, 57). Mann concludes that the “difference” of feminist phen. is that, unlike traditional phenomenology, it is not a method but is inherently a practice: “a kind of doing that must be undertaken, a kind of thinking that must be thought, in order to be understood” (2018, 55).

Reading off of Bonnie Mann’s differentiation of feminist phenomenology, I seek new circumstances in which the “cannon to the canon” (2018, 44) rupture has occurred like the blow Beauvoir once delivered to phenomenological abstraction. In the post-digital context, some algorithms can be said to be practicing a sort of unreflexive phenomenology. For instance, when prompted with the word “woman”, the dall-E image creator combines all of the images online that have been labeled or identified with “woman” in order to create the most generic – or supposedly so – depiction of a “woman”. This algorithmically defined essence of “woman” favors appearance over a priori judgment, just as prescribed by Husserl. It also concocts the false universal that feminist phenomenology rejects – it erases the visual conceptions of “woman” that are rarely encountered online due to factors such as lack of commodification, lack of fulfillment of beauty standards, ambiguity in gender presentation, technical complications (ie: private domains) and so on. Judith Butler illustrates this shortcoming through her assertion that phenomenology fails to account for phenomena that cannot be easily categorized or recognized as such: “the bodies that are excluded fail to materialize according to intelligible norms, and hence are not seen as bodies that count, as bodies that matter” (qtd. in Fielding, 2011, 523).

This erasure of non-normative phenomena online has been examined in more of a technical manner by social critics of technology. Cyberfeminism attempted to raise concerns about social biases embedded in technological innovation. They ask, if men are the dominant designers, manufacturers, and professional users of a piece of technology, how can its function be representative of or useful for women (Wacjman, 2006)? And is there a way to “hack into the big daddy mainframe” of male-dominated tech spaces (VNS Matrix, 2001)? For cyberfeminists, this project most often culminated in unconventional forms of art and writing that used the hypertextual nature of the internet to explore alternative narratives. It also revitalized the psychoanalytic language of 20th century French feminists in order to describe the qualities, particularly communal creation and hypertextuality, that these new art forms embraced. More recently, James Bridle offered a critique of what he calls “computational thinking”, which has been designed into computers from their very origins as war machines. Computational thinking is “the belief that any given problem can be solved by the application of computation” despite living in an “incomputable world” (Bridle, 2018, 10). He exposes the mainstream assumption that algorithms can reach some sort of transcendental answer for social and political issues, and presses us to reach for “new language” and “new metaphors” rather than maximizing our algorithmic knowledge (Gat, 2018). Both the cyberfeminist lineage and Bridle strive to problematize technological design, and favor an action-based approach of playful re-description rather than a detached cataloguing of algorithmic knowledge. This emphasis on a metaphor rather than visibility and practical openness rather than the search for truth is reminiscent of what Mann identified as “oscillation” within Beauvoir’s writing: the key ability to shift between the concrete, personal circumstance of a schoolgirl or a housewife to the general “features of a human experience” for women (2018, 57). This oscillation is reliant on metaphor, using storytelling to make comprehensible the sublime suffering of many individuals while also not reducing the particularities of each woman.

Like DALL-e, several social media networks, notably Instagram, have adopted an algorithmic approach to content in which the most attended-to content is rewarded with more attention. Thus, what is general is rewarded with more generality, and thus a dominant post seems to transcend towards ubiquitous objectivity. I will call this method a phenomenology of the online masses, in which the total sum of appearances is defined by those that appear the most often because of their general quality. To curate a feminist phenomenology in response – or rather within – such a platform would require a practice of infiltration that balances mass appeal with radical messaging. In other words, can an Instagram user create a reflexive, interested, and oscillatory application of the algorithm itself, therefore “hacking” its aims, just as Beauvoir “hacked” the original abstract nature of phenomenology? In order to answer this question, I will use the Instagram feed of internet artist, musician, performer, and model Arvida Byström as a case study. Can Byström, through her algorithm-aware content-making, use the algorithm and its rewards in a manner that is not intended to be used?

In their article “Shedding More Light on How Instagram Works”, company representative Adam Mosseri outlined the “variety of algorithms, classifiers, and processes, each with its own purpose” that decide which content gets prioritized on Instagram (2021). If I am to read “the algorithm” as a phenomenology of the masses, this notion of several algorithms can appear to disrupt my characterization of the algorithm as a pure re-inscription of generality: how can I claim that popular Instagram content simply reflects the interest of the majority on a single scale, when its specific prioritizations are configured by a minority in a multi-step manner? However, I think this discrepancy is remedied by the innate goal of platforms like Instagram, and thus their complex concoction of algorithmic processes: to gain as much ad revenue as possible. If an application intends to gain as much ad revenue as possible, it will attempt to keep its users on it as long as possible. Thus, content that is successfully consumed (liked, commented, and reshared) by its very first users is pushed onto the explore page to be further consumed, since its chances of consumption are highest. Mosseri writes several times that they are primarily using interaction as a measure of “how likely the post is to be interacted with”, creating a tautology of interaction in which what can essentially be boiled down to hype is the only driving factor (2021). Hype is a peculiar form of self-advertisement which carries with it the baggage of mass appeal: the most base and common interests become the most likely to succeed, thus once again sidelining experiences that are particular and complex. In this manner, the tautology of an ad-revenue-based platform becomes a phenomenology of the masses. Bent on maximum profitability, the top-performing content on Instagram is intuitive to an imaginary “generic subject” (Mann, 2018, 70).

One of the most common sub-poisons of mass appeal is sexual appeal, in which content performs very well when it broadcasts an individual commonly classified as attractive and/or said individual doing something that commonly arouses sexual interest. Many users have noticed their engagement significantly increase through the posting of their face or through their body. Internet artist Yoona Bang writes on a bikini selfie on her profile:

“GOOD GIRLS LISTEN TO THEIR ALGORITHM

MY ALGORITHM LOVES ME

BECAUSE I POSTED IN A BIKINI

FEEDING MY ALGORITHM

LIKE THE GOOD WOMAN I AM”

(@yo0n__a, 2023)

This post signifies the awareness amongst users of how sexual appeal affects engagement, and thus their visibility as phenomena in the algorithmic scheme of the app. Users also note that political posting tends to decrease their engagement, as proven recently by the new default setting that limits a user’s exposure to “social and political content” (2024). This user awareness is what makes it possible to even try to subvert the algorithm artistically, by creating appearances that do not align with user intention; for example, how Yoona Bang uses a mirror selfie in a bikini to get algorithmic attention for what is actually an art piece and/or cultural commentary. This notion of algo-awareness is a starting point for a potential feminist phenomenology on Instagram. Yoona participates in but does not “venerate” this algorithmic phenomenology, just as Beauvoir infiltrates traditional phenomenology without an earnest intent to uncover the abstract essence of “woman” (Mann, 2018, 71).

Arvida Byström has mastered algo-awareness: through her performance art, she has garnered 190k followers (@arvidabystrom). She features her (tall, slim, white) body in almost every post, “feeding her algorithm” in girly outfits, while also being recognized as a fine artist and even a theorist. Her visual images are highly conceptual, intended both as commentary on technologically mediated conceptions of beauty and eroticism as well as appeals to those conceptions. A long-running project involves Arvida interacting with her hyperrealistic, human-sized AI sex doll Harmony in various matching lingerie outfits (@arvidabystrom, 2023). Another common theme is photos of Arvida’s body, which have been edited with DALL-e’s Out Paint feature to abjectly tamper cute objects like fawns, butterflies, and dolls into her skin (@arvidabystrom, 2023). Through her collation with a phenomenology of the masses, Byström appears aloof at first sight, like a mere phenomenon with sexual appeal rewarded by the algo. But her reward brings with it her threat. Male users, particularly, might engage with her appealing physique and cute outfits before they notice an intense uncanniness in the blurring between body and tech, natural and edited. Or, the uncanniness may be reincorporated into the users’ internal configuration of what is attractive or worthy of being considered a successful performance of womanhood – indeed, Byström runs a subscription-only page on Sunroom in which users pay to see more explicit variations of this uncanny content (@arvidabystrom, 2024). Byström has successfully performed a sort of Trojan horse into the phenomenology of the masses, as advocated by Mann (2018, 50) – but is it one that is oscillatory, like Beauvoir’s? And does the story she tells have an interested affect towards social change?

I would claim that Byström’s page lacks the oscillation necessary to make a feminist prioritization of embodied specificity clear. Despite being a particular body herself, her body is very normative: white, tall, blonde, and conventionally attractive and frequently present in pink and promiscuous clothing, she is a common replica from the perspective of the algorithm. She also rarely features other female bodies aside from that of her sex doll, and thus, instead of invoking the “polyvocality” present in Beauvoir’s citation-rich approach, her appearance is just a trope of the false universality that Beauvoir sought to dissect (Mann, 2018, 56). It is almost as if she is an empty metaphor of conventional femininity, relatable to many girls’ realities or desires in terms of appearance, without any shared story. With no particular struggle or strain, Byström seems to reproduce the phenomenology of masses, albeit with the uncanny imagery that makes her additionally marketable to the fine art world.

As well, Byström does not seem to take an affective stance towards the topics of gender performance, eroticism, and technology in relation to the experiences of women. While she does explore these topics artistically, she does not come to a political message demonstrated through her work. In the caption of an Instagram post, Byström writes: “The feminine and the cute can be fun, enjoyable, dark and complex all at the same time.” (@arvidabystrom, 2023) In this way, Byström’s work perhaps fulfills Sara Heinamaa’s much more philosophically conservative analysis of Beauvoir’s The Second Sex, in which she claims that the book is not intended as a “thesis about women’s socialization, but a phenomenological inquiry into the constitution of the meaning of sexual difference” (Mann, 2018, 51). If feminist phenomenology were to stop at mere explorative inquiry, Byström would very much align with her visual meandering of mainstream femininity. However, Bonn goes further and displays that Beauvoir very much does have a thesis on women’s socialization. Particularly, Beauvoir rejects traditional femininity very clearly throughout the entire treatise, and encourages women not to resign themselves to such qualities: “To be feminine,” Beauvoir says, “is to show oneself as weak, futile, passive and docile.” (348, qtd. In Mann, 2018, 52 ) This political thesis is what delimits the definition of feminist phenomenology and prohibits a mere playful inquiry from entering into its bounds.

Returning to Camus’ critique of Husserl in The Myth of Sisyphus, he further laments that “instead of encountering here a taste for the concrete, the meaning of the human condition, I find an intellectualism sufficiently unbridled to generalize the concrete itself” (2013[1942], 36). If I follow in Mann’s formulation of feminist phenomenology, its founding text appears to answer the call – isn’t a 1000-page treatise on the history, science, literature, archetypes, and case studies that surround the idea of womanhood a most profound and laborious “taste toward the concrete”? When the corporeal specificity of Beauvoir’s work is outlined, it becomes clear that Camus’ well-known remark that The Second Sex “humiliated the French male” is most likely a sexist one rather than one caused by philosophical disagreement (Gloag, 2020). An equivalent figure to Beauvoir is yet to rupture the phenomenology of the masses we see present on Instagram, other social media, and other algorithmic methods of knowledge-making. Of course, the claim could be made that Arvida Byström is one of many art influencers who carry an algo-awareness, and perhaps there are others who have more oscillation and political affect present in their work. However, I selected Byström for her particularly large following – she is one of few who are able to doubly articulate themselves as such to both the masses and the intellectual scene, instead of resigning themselves as a “micro-influencer” to a niche audience. Thus, it is no coincidence that she inhabits a conventionally feminine appearance that is neutral in political tone and easy to commodify. As digital spaces become the primary method for finding new ideas and choosing which to prioritize, Byström’s case is a reminder that we must remain intentional and laborious in our knowledge-making for feminist practices and beyond. To call for change, we must not just trick – we must also convert.