Abstract
This article provides a multiplatform systematic analysis of contextual factors in what we term memetic misogyny, a form of implicit, polysemous and ephemeral visual gender-based hate. We draw on a feminist ethnographic approach with a semiotic analysis applied across three case studies of meme-based misogyny (Greta Thunberg, Karens and anti-feminist memes related to protest #SisterIDoBelieveYou) on Twitter/X, Facebook, Reddit and YouTube to argue that memetic misogyny is created, maintained and co-shaped by (1) misogynistic contents (depictions, aesthetics and narrative framing), (2) platform affordances (real and imagined) – above all meme affordances and (3) the hegemonic discourses of the cultural communities where these are distributed. In combination, these three create what we call mythologies of memetic misogyny, that is, gendered socio-cultural and socio-political narratives that present misogynistic ideas as if they were natural, universal and timeless gendered truths, often going unnoticed because of the volatile, polysemous and polycontextual nature of memes.
Introduction
Memes have been a significant point of interest for research on gendered hate (Andreasen, 2021; Massanari and Chess, 2018; Tanner and Gillardin, 2025), wider practices of online hate and violence (Askanius, 2021; Koch, 2024; Thorleifsson, 2022), including what has been described as mediated, online or networked misogynies (Banet-Weiser and Miltner, 2016; Bulut and Can, 2023; Vickery and Everbach, 2018). This article extends this work by examining the misogynistic potential of memes and conceptualising memetic misogyny. We have previously written that visual misogyny draws on the technological affordances of networked visuals for purposes of (a) gender-based online violence (e.g. revenge and fake porn, online flashing, non-consensual dissemination of sexual, explicit or intimate imagery) and (b) stereotyping, mocking, ridiculing or otherwise discrediting women, non-binary and queer individuals, as well as different aspects of womanhood, femininity and lifestyles associated with them (Özkula et al., 2024a). We therefore see visual misogyny as the wide array of visual and multimodal practices expressing, solidifying, or enforcing both online gender-based violence or hierarchical gender relations based on heterosexual cisgender male dominance. In this article, we explore what we call memetic misogyny, a form of visual misogyny that draws on memetic affordances that are produced through memetic contents, meme practices, and meme-related affordances of social media platforms (above all where these are anchored in visual and memetic logics, e.g. YouTube or Instagram).
Memes significantly draw on affordances of visual communication, but unlike some of the examples mentioned above, they tend to be less explicit in their articulations and visualisations of hate and violence. Their prevalent ‘mash-up’ character often infuses socio-cultural and socio-political narratives with sarcasm, insider jokes, puns/double-speak and other forms of humour, but also reframes them through editing processes that adapt or mimic certain narratives and styles. The conventional use of humour in memes also masks hate and/or violence, a dynamic that has been said to need investigation in more contextualised ways (see Schmid, 2023: 15). The aim of this article is therefore to begin a systematic exploration of the diverse polysemous and contextual nuances of memetic misogyny through a semiotic analysis of memes grounded in three ethnographic case studies on Twitter/X, Facebook, Reddit and YouTube: memes about Greta Thunberg, Karens and under hashtag #SisterIDoBelieveYou. We do so by asking (1) what kinds of contexts are referenced in misogynistic memes; (2) how is this reference established (i.e. content, aesthetic and narrative) and (3) in what ways do these creative memetic processes reflect the specific contexts in which they have been shared (i.e. what’s the role of platform affordances and communities of practice in this process). In doing so, we aim to generate an inductive overview of memetic misogyny that captures memes in diverse contexts. We use this conceptualisation to illustrate the implicit, polysemous, polycontextual and ephemeral nature of meme-based misogyny that makes these practices difficult to identify and stymie.
In what follows, we outline the relationship between (social) media platforms, memetic content and gender-based hate, as well as different articulations of visual misogyny. We then summarise our ethnographically grounded semiotic analysis of memes. Finally, we lay out the findings of our contextual semiotic analysis to argue that memetic misogyny is generated, maintained and co-shaped through (1) misogynistic contents (content, aesthetic and narrative), (2) platform and meme affordances (real and imagined) and (3) the hegemonic discourses of the cultural communities where these are distributed. In combination, these create what we call mythologies of memetic misogyny, a visual form of implicit, polysemous, polycontextual, ephemeral and therefore difficult-to-identify gender-based hate. The article consequently extends existing knowledge in the field by defining the scope of memetic misogyny in its polycontextual co-creation that makes it an elusive and often underdetected practice.
Memetic misogyny in context
(Social) media and visual misogyny
Misogyny long predates memes, the Internet and social media. A classic example is the work of Laura Mulvey (1975) on the male gaze, which illustrated how classical Hollywood films positioned women as objects of desire. Today, mothers on TV shows in the Global North are largely portrayed as white, young, effortlessly slender and attractive objects of desire (Geena Davis Institute, 2024); in advertising, female characters ‘are shown more often in the kitchen, shopping and cleaning’ (Davis, 2020), and women of colour are subjected to a Eurocentric form of violence rendering them into exotic objects of desire (Schacht, 2019; Schutte, 2017). These representations mirror the supposed desires and perspectives of heterosexual cisgender males (often white and Western) and therefore (rein)force hierarchical gender relations that benefit them, and they extend to online spaces too. A pervasive and hostile misogynistic rhetoric on the Internet, defined as e-bile (Jane quoted in Vickery and Everbach, 2018: 13), is used to ‘injure, discipline, and subjugate women’ (Bárcenas Barajas, 2023; Brooke, 2022: 372; Esposito, 2023). Images, image-sharing practices and platform affordances play a fundamental role in this process since social networking sites (e.g. X, Instagram, TikTok and Facebook) are anchored in the notion that users engage with each other through the production, circulation and various forms of engagement with mutimedial and multimodal materials. These materials include visuals such as photographs, GIFs, cartoons and memes, especially on ‘visual social media’ platforms and apps that rely more strongly on visual forms of communication and therefore strongly facilitate, if not promote, the formation of visual practices and cultures (e.g. YouTube, Instagram and TikTok; see Leaver et al., 2020). However, thus far, little work conceptualises misogyny and its contextualised applications in reference to visual social media cultures.
In part, this gap stems from the difficulty of defining and disentangling misogynistic practices. Misogyny generally involves contempt, hatred, toxicity or hostility towards women based on a value system that assumes certain gender roles and hierarchies (i.e. male superiority in patriarchal regimes), and punishes women for moving beyond these boundaries. Communities of the manosphere, such as incels, pickup artists and red-pillers, are emblematic of these value systems. However, as part of a wider belief system, misogyny also involves the establishment or maintenance of this unequal power distribution and asymmetric gender roles through the policing and subordination of women (Manne, 2017; Weiss et al., 2024). It serves to ‘police and enforce these social roles’ by punishing women for not conforming to them, for example when they do not adhere to the notion of a woman’s role as necessarily caring and loving (Manne, 2017: XIV). This punishment may manifest through (both benevolent and hostile) sexist or anti-feminist movements and practices, which may or may not be explicitly hateful (Manne, 2017; Weiss et al., 2024) but seeks to reproduce patriarchal and heteronormative hierarchies/discourses. In visuals, this may involve explicitly formulated hate such as visual insults or threats and the exploitative use of images and videos (e.g. image-based sexual abuse such as revenge porn) and gender identities, as well as subtler forms of exploitation that reinforce gender hierarchies by subjugating women, queer and non-binary individuals (i.e. groups outside of heteronormative cisgender white men), for example through stereotyping, diminishing and shaming them (Özkula et al., 2024a). Visual forms of hate speech, symbolic violence, sexist commentary and movement-hijacking can therefore be said to constitute ‘visual-misogynistic repertoires’.
Another difficulty in identifying misogyny is that it also dovetails with other political ideologies and wider value systems, in particular those pertaining to neo-fascist and far-right groupings, including: anti-multiculturalism, racism, islamophobia, anti-globalisation and antisemitism, both as a precursor and potentially a predictor of shifts to far-right contents and the ideologies associated with these (see Askanius, 2021; Weiss et al., 2024). These communities of practice are rooted in wider notions of how gender and sex align with hierarchies of privilege, power and entitlement. Thus, although a wide array of literature stresses the persistence of misogyny, the term is difficult to apply in analytical terms. This reticence may be attributed to the difficulty of identifying toxic communication as a structural logic, the at times implied nature of hate and violence and seemingly harmless expressions of misogyny (see Manne, 2017). Even so, according to Manne (2017: 12), ‘the term misogyny is (. . .) both a word we need as feminists, and one we are in danger of losing’. In line with Manne’s work, we therefore explicitly use the term misogyny to draw attention to the structural and systemic conditions that foster, replicate and solidify asymmetric gender roles, which (we argue) are facilitated through platforms’ memetic affordances and communities of practice.
Misogyny in memes
While memes may have the appearance of being relatively harmless, hateful memes have garnered significant interest in recent years. Internet memes are primarily visual digital artefacts ‘that: (a) share common characteristics of content, form, and/or stance; (b) are created with awareness of each other; and (c) are circulated, imitated, and transformed via the Internet by multiple users’ (Shifman, 2014: 341). Memes often narrate a joke or statement made in relation to a common narrative, piece of insider knowledge or (pop-)cultural reference, and often imitated and repeated through aesthetic choices and mash-ups (see Bulatovic, 2019; Davison, 2012; Milner, 2016; Rogers and Giorgi, 2024; Shifman, 2013, 2014). Although memes may be circulated and applied transnationally, they are also often situated in particular socio-political events and geopolitical or cultural norms (McSwiney et al., 2021; Paz et al., 2021); they therefore both reflect specific cultural contexts and constitute cultural units in themselves (Situngkit, 2004). A particular challenge of interpreting memes (for both users and researchers) therefore lies in the different meanings that can be inferred from these multiple layers, that is, polysemy.
Despite their whimsical and humorous form of entertainment, memes often integrate social or political commentary. This is reflected in practices described as the ‘memefication of political discourse’ (Bulatovic, 2019), meme-based/memetic activism (Pilipets and Winter, 2017) and ‘playful activism’ (Cervi and Divon, 2023). These concepts stress the nuanced ways in which humour and play in memes are used to construct political narratives in a more digestible, attractive, seemingly harmless and at times ambiguous way. Through irony, satire and ridicule, memes may signal, foster or solidify group affiliation or ideological belonging, or, conversely, create distance from the targets by dehumanising them (see Askanius, 2021). For example, a meme mocking Greta Thunberg’s expertise on climate change may indicate scepticism towards climate change being anthropogenic, and the use of group-specific vocabularies, such as describing women as Stacies (vapid superficial women), may indicate incel affiliation. As such, identity politics may be both omnipresent and difficult to disentangle from meme cultures (Massanari and Chess, 2018). At the same time, their playful, ambiguous and generally ‘unserious’ character transforms otherwise aggressive or problematic contents into more socially acceptable socio-political commentary (Massanari and Chess, 2018).
Notions of ‘memetic hate’ (e.g. Koch, 2024) or ‘memetic violence’ (Halliday, 2018; Thorleifsson, 2022) are, in that regard, not new. Memes may seem harmless on a literal level as ‘not everything is said’ and their meaning is fixed through context, that is, the discourses and communities surrounding them (Askanius, 2021: 160). Memes therefore lend themselves particularly well to certain forms of hate speech 1 as ideas are articulated more playfully and ambiguously, which makes it difficult to identify hate, but may also normalise it in the process (Aranda Serna, 2024; Bárcenas Barajas, 2023; Massanari and Chess, 2018; Matamoros-Fernández, 2023; Matamoros-Fernández et al., 2023; Schmid, 2023; Tanner and Gillardin, 2025). Explicitly hateful memes have at times even achieved virality, such as the far-right white-nationalist slogan ‘your body, my choice’ in response to the Trump regime’s victory in the 2024 US election. The humorous orientation of memes suggests that hateful memes are more often ambient in nature, that is, toxicity lies in the background and reflects normalised views rather than being explicitly targeted and organised (Siapera, 2019). And yet, the playful and seemingly harmless nature of memes may seem ambient, when it may, in fact, be targeted (see the example of the Nordic Resistance Movement in Askanius, 2021). Either way, the playful character of memes reinforces power and naturalises misogyny through seemingly harmless content and the manipulation of users’ beliefs (Bárcenas Barajas, 2023; Recuero, 2024). Thus, memes may reflect a range of socio-cultural and socio-political narratives and construct collective identities around these issues (Gal et al., 2016; Hakoköngäs et al., 2020) – whether based on solidarity or hate, a difference that may be difficult to ascertain.
Following Halliday (2018), we therefore attribute ‘memetic power’ to meme practices as they reflect, spread and instil ideas around gender, race, sexuality, ability, ethnicity, religion and nationality, and therefore constitute a form of symbolic power and violence that expresses misogyny as a systemic logic. However, as we will argue, memetic hate is difficult to identify and mitigate as the evolution of memes through content, affordances and cultural practices often renders them implicit, polysemous, polycontextual and ephemeral.
Towards a conceptual framework of memetic misogyny
We have chosen memes here because, compared to other artefacts circulating on social media platforms, they are highly dynamic; that is, their meanings and connotations are not static. Since ‘memes adapt and evolve as they spread across environments’ (Peeters et al., 2021: 3), ‘memetics’ are subject to a range of meme-specific attributes. This view has also been subject to criticisms as the concept neglects human agency and the medium where memes are shared (Peeters et al., 2021). Medium specificity matters, as platforms that strongly adopt meme logics facilitate their production and circulation through possibilities for curation such as remixing, overlays, stitching, filters, templates and other creative features, as well as through digital tagging and reposting mechanisms that enable the rapid adaptation, modification and recirculation of such contents. We have therefore considered memetic attributes here on the level of content alongside medium specificity (= platform affordances) and community-specific meme practices. This means that misogynistic memes do not exist in isolation, but that they are created and shaped simultaneously by (1) what the meme depicts (= content), (2) by the technical infrastructures that make certain forms of circulation possible or restrict others (= platform affordances) and (3) by the cultural practices and shared imaginaries of the communities that spread and interpret them (‘realised imagined’ platform affordances, see Gibson, 1977). 2 In what follows, we outline what these three factors entail.
Content
Extant research has explored several layers of visual-misogynistic content. Studies have, for example, variably considered visual depictions (i.e. what is in an image) alongside narrative and aesthetic frames (i.e. how depictions are stylised and narrated in an image), as follows:
(a) Trivialisation, infantilisation and ‘cutification’ of women and femininity, for example the framing of gendered practices such as vain or trivial (Burns, 2015) or rendering women childlike or vulnerable through cute aesthetics (Jacobs, 2022).
(b) Stereotyping and nostalgia rooted in traditional gender roles, for example through the reinforcement of traditional gender roles for public figures and assumed ‘deviants’ (see Özkula et al., 2024a), reinforcement of heteronormativity (Andreasen, 2023b), etc.
(c) Downplay and/or normalisation of or harm/violence towards women, for example ‘just a joke’ memes about sexual violence or the rapeability of women (Andreasen, 2021, 2023a) and memes ‘poking fun at violence’ (Gusic and Lundqvist, 2023).
(d) Vilification of women and female identities, for example the depiction of female political figures as villains (Özkula, 2023), the construction of a ‘monstrous feminine’ (Massanari and Chess, 2018), the vilification of black women’s bodies, experiences and aesthetic features (Halliday, 2018) and ‘bestialisation’, that is, animal comparisons of women (Prieto-Blanco et al., 2022).
(e) Victim-blaming and/or sympathy for the (alleged) perpetrators (Andreasen, 2021, 2023a; see also #SisterIdoBelieveYou in Özkula et al., 2024a).
(f) Wider practices of sexualisation and commodification of women.
Not all of the above-listed works have considered the role of platform specificities in how these contents are generated, narrated and framed.
Platform affordances
This part of the conceptual framework is therefore concerned with the technical and infrastructural features of platforms that facilitate, stymie or otherwise shape how misogynistic memes circulate. Given that misogyny is shaped by the specificities of the sites where visuals are constructed and/or circulated, we have previously described visual misogyny as ‘platformed’ (Özkula et al., 2024a, as derived from ‘platformed racism’, see Matamoros-Fernandez, 2017), the notion that platform affordances play a role in if and how misogyny is created, circulated, made visible or invisible. For example, anonymity paired with practices of lighter/absent moderation (see Thorleifsson, 2022) produces very different regimes of visibility of misogynistic content than closely moderated and curated ones (see Özkula et al., 2024a). The proliferation of digital hate may consequently be influenced by platform-environmental transparency (e.g. WhatsApp’s opacity facilitating extremism, see Chagas, 2023), platforms’ moderation practices (Thorleifsson, 2022) and options for users’ anonymity (Massanari and Chess, 2018). Scholarship on visual or multimodal misogyny has already covered a wide range of platforms, including Instagram, Twitter, Facebook, 4chan, 9gag, Reddit, Imgur and TikTok (Andreasen, 2021; Halliday, 2018; Özkula et al., 2024a; Tanner and Gillardin, 2025; Thorleifsson, 2022). Even so, most work to date does not consider platform affordances per se or does not address how memetic contents are influenced by platform logics.
This is even though memes have been shown to ‘acquire specificities depending on the site where they are constructed’ (Rogers and Giorgi, 2024). Platform affordances consequently bear particular relevance for memes as platforms not only affect how memes circulate and gain or lose visibility but also how they are modified in content. Thus, memes as a format in themselves afford ephemeral and polysemous forms of communication, but these communications are also facilitated by platforms that afford memetic practices. Platforms may provide different features that facilitate memetic engagement and distribution, for example by offering ‘memetic templates’ such as on TikTok through lip-syncing, duets and point-of-view (Cervi and Divon, 2023) or sound templates (Tanner and Gillardin, 2025). Individual platforms have also been shown to foster more or less the emergence of memetic cultures based on the dominance of the visual over other communicative forms (i.e. visual social media; Lee and Abidin, 2023: 4–5; Tanner and Gillardin, 2025). These patterns suggest that misogynistic memes may potentially be more common on meme-oriented and ‘meme-literate’ platforms and subspaces such as TikTok (see Matamoros-Fernández, 2023; Tanner and Gillardin, 2025), in particular where lax and context-unaware moderation practices allow for them to linger. It also means that the meaning of a meme is perpetually unstable, as it is by its very nature a product of its sites of circulation and adaptation, as well as the communities that inhabit these.
Communities of practice
The third cornerstone of our conceptual framework concerns user practices and cultures, highlighting how the meaning and visibility of memes are negotiated and sustained through collective engagement (as opposed to infrastructural aspects). We aggregated the earlier listed misogynistic narratives specifically in relation to visual misogyny here, but extant research has also shown some of these narratives to be rooted more widely in meme cultures, for example trivialisation, concealment under humour and the subsequent normalisation of hate speech. 3 Key to this development are communities of practice in that human agency is not only exercised in producing memes, but also adapting and transforming them for joint cultural experiences (e.g. see Miltner, 2018; Shifman, 2013), as well as retaining and making them (in)visible. Communities of practice also play a significant role in how a given meme is applied and understood (see, e.g. the example of the ‘Social Justice Warrior’ in Massanari and Chess, 2018). Despite platforms partially managing content visibility by, for example promoting or shadow-banning content, users have shown to engage in practices of resistance, that is, strategies for circumventing and negotiating these effects. Contexts of toxicity therefore refer to the conditions and configurations of how toxic/misogynistic content is ‘produced, published, and interpreted’ (Tanner and Gillardin, 2025: 3). This suggests that the visibility of visual hate and violence is not merely dictated by platform providers’ activities, but a hegemonic result of the visibility negotiations between platform providers and communities of practice. Platform affordances and communities of practice are, as such, difficult and at times impossible to disentangle from each other as affordances are relational, that is, platforms’ infrastructural uses and cultural practices co-constitute each other.
In what follows, we will consequently argue that memes are not only polysemous but that their meaning is continuously (re-)constructed and (re-)negotiated in hegemonic, polycontextual and ephemeral resignification processes, which may render memes unreadable beyond the communities in which they have been distributed.
Data and methods
This article investigates these contexts through a context-sensitive, ethnographically informed visual analysis of memes from three case studies of visual misogyny: (1) Greta Thunberg memes in climate-sceptic groups, (2) memes shared on Twitter under #SisterIDoBelieve, a movement in response to a gang rape case in Spain and (3) Karen memes on YouTube, Twitter and Reddit. These memes were posted in different contexts, reflecting the political environments in which they were produced and the platform spaces through which they were distributed. The resulting polysemy (the co-existence of multiple meanings within the same artefact) is typically challenging for visual analysis as interpretations may vary across spaces and audiences due to differing understandings of symbols and contextual references (Boxman-Shabtai and Shifman, 2014; Özkula et al., 2024b). In this case, these varying contexts have been taken as the subject of analysis as visual misogyny persists regardless of platform and national contexts, but the cultural and political references and their creative applications vary accordingly. An analysis of visual misogyny was conducted towards identifying (as per the research questions) (1) the kinds of contexts referenced in polysemic misogynistic memes, (2) how contextual references are established and (3) how these creative processes reflect the specific contexts in which they have been shared.
These questions were explored through three ethnographic case studies. Case 1 (Greta memes, DENY Facebook group: May 2019–June 2021; wider Greta memes cross-platform: May 2023–June 2024) focuses on static memetic images of Swedish climate activist Greta Thunberg distributed within climate-sceptic groups online (more detailed research design in methods paper: Özkula, 2023). Case 2 (#SisterIDoBelieveYou, May 2018–August 2020) relates to multimodal communication under the Twitter (now X) hashtag #SisterIDoBelieveYou, which emerged in response to a controversial court decision on a gang rape case in Spain (full research design in García-Mingo and Prieto Blanco, 2021; Prieto-Blanco et al., 2022). Case 3 (Karens, May 2023–June 2024) consists of a multi-platform observation of Karen memes (both video and static memes) on Reddit, YouTube and Twitter. Karen memes originally criticised middle-class, middle-aged privileged white women displaying racist or otherwise rude, vitriolic or entitled behaviour, expressed in phrases such as ‘I would like to talk to the manager’ (see Brady et al., 2023). Karens are similar to the ‘Social Justice Warrior’ meme analysed by Massanari and Chess (2018) in that they started as a humorous trope but evolved into a pejorative commentary through their use across various online communities. In aggregation, these case studies span several years of observation and several hundred memes, alongside other visual, textual and multimodal social media contents (> 1000 visual artefacts, as posted under one Twitter/X hashtag, one dedicated Facebook group and in multiple platforms such as YouTube channels and subreddits). In aggregation, they therefore lend themselves particularly well for capturing diverse contextual and polysemous gendered memes.
These cases were chosen based on their different iterations of memetic misogyny. #SisterIDoBelieveYou was anchored to a specific rape case with an identified victim/target and perpetrators. This resulted in targeted actions, both feminist and anti-feminist. For instance, images featuring the faces and names of the five rapists circulated on Twitter as part of an escrache (the Spanish term for organised in-your-face rallies outside of private residences of public figures). In comparison, the Greta case was not a gendered or rights-based case. Gender was decentred here, and any gendered commentary was a by-product of the group’s wider political leanings and values, which were reinforced as part of the specific Facebook subspace in which they were active. Finally, the Karens became a mainstream trope that invited different applications and interpretations depending on the spaces and communities in which these memes were shared. Both the Greta and (especially) Karen memes were therefore more ambient in nature than #SisterIDoBelieveYou and not situated in legal or formal-political contexts. The Karen memes in particular were often situated in seemingly depoliticised contexts. The different case studies therefore allowed for a capture of different types and attributes of misogyny, such as implicit and explicit, targeted and ambient forms, directed at named or unnamed individuals. Observing the images linked to these three cases provided insights into comprehensive multi-platform practices (Twitter, Facebook, Reddit and YouTube) and into the platform features that determined their use (e.g. hashtags and video titles). For example, the analysis of a given visual also entailed the coding of how it evolved, such as how the meaning of an image was altered and subverted, and a consideration of how platform affordances were utilised for these purposes (e.g. the mixing of hashtags for hijacking).
While visual content analyses (both thematic and discourse) are a popular method in the field, we chose ethnography to capture contextual aspects of these memes, including how their meaning is created and how it evolves through communities of practice. Memes and the myths they encapsulate are subject to constant change, depending on communities’ practices of belonging and identity formation. We focus on memes as they are particularly complex to read contextually in that they do not have static polysemous meanings. It is through this back-and-forth process that ‘mythologies’, that is, cultural narratives or systems of meaning that present socially constructed ideas as if they were natural, universal and timeless truths (Barthes, 1964), are created. The Barthesian approach allows to account for the socio-cultural context of audiences and the communities of practice in the signification process. A Barthesian semiotic reading of the memes was conducted in light of the depictions and narratives identified in existing scholarship (i.e. a priori themes, e.g. women as villains, puppets, monsters and beasts – cf. p. 6 in the literature review) as well as how these contents were constructed through aesthetic choices (e.g. colours and stylisation). This involved routine practices of joining, observing and following the case-specific hashtags, groups, channels and collections, and exploring, archiving and analysing visuals shared in these.
This analysis was rooted in the development of an embodied feminist ethnographic sensibility towards the corpus that focused on capturing contextual aspects across content, platforms and practices. This means our research process featured dialogue, self-reflection on the researchers' positionalities and embodiments, iterations of analysis from context to specificity, an acceptance of uncertainty, ambiguity and polysemy in the research process and being lost in the data (Özkula et al., 2024a; Prieto-Blanco et al., 2022). This included a detailed consideration of research ethics. Given the problematic character of these data, we opted to anonymise the targets’/victims’ identities in text and visuals except for the public figures, but tried to retain the violent and affective character of these contents. The public channels, hashtags, subreddits and groups circulating misogynistic content are not anonymised here.
The visual analysis was conducted in tandem with a reading of contextual aspects such as group practices, group and moderator descriptions of the community identity and rules, framing devices as offered by the platform spaces (e.g. hashtags and video titles) and the wider use of platform affordances. Existing misogynistic narratives in literature (cf. p. 6) were used to establish exploratory a priori themes for analysis. These categories were used as provisional starting points and informed our ethnographic observation throughout. They were used as flexible sensitising concepts. Relevant visual aspects (e.g. aesthetics) were interpreted alongside tags (e.g. hashtags, tags and subreddits), group framing elements (e.g. Facebook group and subreddit rules and self-descriptions of collective values and identities) and visual-semiotic narration (e.g. symbols, metaphors and other rhetorical devices). The relationship between the multimodal elements was interpreted in light of users’ and communities’ practices observed (e.g. deletion/moderation of content and expulsion of members). Then, themes were generated based on the interaction of these categories across all three case studies and in light of both (adapted) a priori and later established codes for identifying how different contextual elements co-constructed misogynistic narratives. Codes were grouped in a reiterative process under flexible categories that involved content, affordances and user practices, and then refined into the final categorisation.
In what follows, we present an ethnographically produced inductive exploration of memetic misogynies across these three case studies that shows the different entanglements of memetic content and context, and serves as a basis for future systematic testing on larger samples (i.e. deductive research).
Entanglements of contents and contexts in memetic misogyny
The findings showed a polysemic make-up of memes that resulted from content-based factors (what is being depicted, and how, e.g. colour, text), community-based ones (insider knowledge, regional/ cultural knowledge), cohesive community values and hegemonic uses of platform affordances.
Content and framing factors
Memes in the dataset featured narratives of women’s exploitation and discreditation as previously identified (cf. list of p. 6) alongside other visual-misogynistic content. Images of individuals featured prominently in the dataset (e.g. Greta Thunberg, perceived Karens and the rapists). These images would often be photographic, but there were also cartoons and mash-ups. Fictional depictions of violence against women included drawings, image manipulation, mash-up and/or use of text, and this featured prominently in the dataset. Another common occurrence was that text was used to shape the meaning of the visuals, that is to direct viewers to a preferred meaning (see ‘anchorage’ in Barthes, 1977). For example, the overlaid text “Karen” on an image (e.g. see figure 4) encourages viewers to interpret loud women as overreacting and therefore as Karens, and by extension, to associate overreaction with a female quality, thereby contributing to the reinforcement of heteropatriarchal stereotypes of femininity and womanhood.
Over time, the ethnographic observations also showed that the visuals were not static in meaning but evolved (they were resignified), that is, meanings were ephemeral, dynamic and subject to change based on the spaces where they circulated (= resignification, an affordance of both memes as a format and the platforms that facilitated these uses). For example, a particular subset of memes was used to resignify symbols otherwise associated with feminist protest. In the case of #SisterIDoBelieveYou, the presence of such memes was not numerous, but they reflected a consistent effort over time to hijack well-known feminist content and infuse it with anti-feminism. Such memes (see Figure 1) depicted, for example what seems to be a home-made cardboard placard with a crowd on the street in the background. In the meme, the image is repeated, but the text on the placard changes: on top it says ‘Sister, I Do Believe You’ (the chant that gave rise to the hashtag and that expressed trust in the victim of the gang rape), while on the bottom the text is an interpellation to the victim to present evidence – ‘Sister, [we need] evidence, not faith’, thereby connoting disbelief in the victim. The bottom text also includes the handle @feminisciencia: a Twitter account created in 2017 to ‘provide a vision of gender issues based on science’ and that supports the Men’s Rights Movement (feminismocientific.wixsite.com). Other times, contrarian hashtags such as #LiberatingFeminism (#FeminismoDeLaLiberacion) 4 were added to images and anchored them to an anti-feminist meaning.
Resignification of hashtag #SisterIDoBelieveYou.
Memetic strategies of resignification also included the use of colour and symbols, as showcased in Figure 2. The colour purple and the feminist fist, both originally associated with feminist protests under the hashtag #SisterIDoBelieveYou, are featured on a new text, namely the slur ‘Puta Zorra 5 Feminazi’ (‘Feminazi Fucking Bitch’), which anchors these two images to anti-feminist meanings. In the image on the right, this text is paired with a photographic image of a fox. In the Spanish cultural context, foxes symbolise cleverness and resourcefulness, but also dishonesty, cunning and deviance. The image therefore connotes feminists as deceptive and manipulative, thereby vilifying and de-victimising them. This frame is underscored by the label of the ‘feminazi’, a pejorative portmanteau of ‘feminist’ and ‘Nazi’ to criticise feminism as radical/extreme, militant and/or authoritarian, as well as the ultimate enemy against whom the use of violence is justified. Both these cases (Figures 1 and 2) show how memetic strategies of resignification were used to reinscribe feminist messages with misogynistic meanings, portraying feminists and victims of sexual violence as lying, scheming and oppressive (similar to women/feminists as villains or beasts).
Resignification of colour purple in case #SisterIDoBelieveYou.
Other common aesthetic-memetic adaptations included edited fictional representations that either highlighted women as aggressors or narrated violence towards them. This included exaggerated aggressive stylisations of targeted women such as fictional or non-fictional angry Karens (images of real private persons were omitted here) including public figures such as Greta Thunberg and Kamala Harris (see Figure 3). In line with the notion of the monstrous feminine (= a priori category, cf. literature review) that acts as an aggressor or threat to patriarchal regimes, these aesthetic choices portrayed women in politics as aggressive, villainous and authoritarian. They therefore constituted a critique of political women’s chosen social roles by depicting them as deviant and enemies. Other fictional representations, such as cartoons, depicted harm towards these women in response to their alleged deviance, such as Greta being beaten or sexual-predatory behaviour towards her (omitted). While both kinds of depictions were often humorous as part of memetic frames, they narrated the role of these women by either demonstrating their threat potential and sometimes, consequently what they ‘deserved’, that is, the physical and rhetoric punishment for their refusal of their assigned social role as caring, loving, agreeable, silent and/or ‘obedient’ (misogyny).
Aggressively stylised depictions of women.
Similar to the textual frames in #SisterIDoBelieveYou, in Karen videos, these narratives were at times underlined by hashtags such as #karenbeaten, #karengetsbeaten, #karengetskarma, #deserved or #justice on Twitter, and video and compilation titles such as ‘Karens getting what they deserve’ or ‘Karens getting arrested’ on YouTube. Compared to textual anchoring in these images, these tags framed the visuals according to the cultural communities that produced and shared them (an entanglement of meme contents and collective practices). In doing so, these visual representations and their captions trivialised physical harm towards women by displaying it as mere jokes, as well as normalising or banalising it by indicating that it was justified given their deviance. They were rooted in a tacit moral agreement that loud and complaining women deserved violence – whether the context of their actions being rightful or bigoted was included or not. Although some of these captions implied (if not explicitly stated) the violent character of these memes, the generic description of these being Karens (i.e. not a specific or even real person) alongside the implied humour (i.e. Karen memes being ‘only a joke’) downplayed and minimised their aggressive potential. These spaces also combined memes depicting more serious harms, such as genuine physical violence (e.g. slapping and beating), with comparatively less harmful or generic commentary on Karens – fictional or real (see Figure 4). Without active long-term participation in these groups, their harmful contents were therefore at times ‘buried’ and masked alongside more generic commentary memes that did not reference specific women.
Generic de-individualised Karen memes.
In aggregation, these memes highlighted several memetic patterns. Images and memetic templates were used for adaptation and mimicking, making them ephemeral. More than that, curation techniques were used to subvert meanings or infuse them with misogynistic frames. These techniques included visual resignification (e.g. stylisation and colour) and textual resignification, both in-meme and as part of platform features such as tagging mechanisms. This use of platform affordances consequently both reflected and placed memes within specific group values.
Context: platform affordances and communities of practice
These multimodal memetic narratives were shaped by an interplay where users not only harnessed platform-specific logics such as framing and tagging mechanisms but also adapted these to suit their communities of practice, a factor also apparent in moderation practices. While tweets posted under #SisterIDoBelieveYou were public and largely moderated by the platform (though also subject to user-driven reporting), posts in closed and user-administered groups were subject to their cultural communities – an effect reinforced by platform specificities that allowed for constrained visibility based on the targeted cultural community. In the case of Greta memes, exclusionary subspace moderation processes meant that group cohesion was expressed through continuous misogynistic commentary and violence towards Greta Thunberg implied in memes, and maintained by trolling or expelling users that questioned these practices – a hegemonic process that evaded platform rules based on a group’s moral agreement (e.g. that depictions of violence towards these women was acceptable). Across all three cases, memes expressing gender-based violence were also overshadowed or masked (and therefore overlooked by platforms) by practices that distracted from the underlying misogynistic narratives: the use of humour, political counter-narratives, movement-hijacking and de- or re-victimisation processes – not only in meme content but also in cultural communities’ uses and negotiations of platform affordances.
In the case of the Karen memes, cultural communities were targeted through video and collection titles on YouTube, subreddit descriptions and tapped into through tags and hashtags, which varied in their level of aggression. Subreddits at times specified the thematic scope of the group, for example, by banning the display of the supposed Karens’ real identities, pictures of real people/Karens, ‘witch hunt’ posts, or the wider remit of hate – a practice not applicable to YouTube’s infrastructure and applied more liberally on Twitter. As such, platform norms and cultures still mattered in how content was and could be regulated. These descriptions, rules and user-led moderation processes reflected what was expected, desirable or acceptable within a given cultural community, expressing their collective moralities and identities. Even so, these rules were subject to actual uses (i.e. a hegemonic process). Hateful comments, including verbal insults, persisted (e.g. c*** and b**ch) and some Karen subreddits have since been reframed or closed – in 2024, a Karen meme subreddit closed due to difficulties in moderating the increasing amount of doxing, misogyny and racism, and has since referred users to other subreddit communities. Thus, user communities still achieved memetic resignification and hijacking within the platform features.
Cultural homogeneity was therefore subject to preservation processes, often through appropriation and resignification techniques. Although memes draw on known logics and tropes (e.g. the Karen), they are also appropriated in (re-)distribution to produce new meanings, for example, through oppositional targeting and developments in format. Anchorage therefore happened not only via text embedded in the memes (as with the placards in Figure 1), but also via contextualising text added in meme (re)distribution. These developments show how users harnessed platform affordances in targeted ways towards reframing and (re)creating meaning that suited their communities’ value systems. Although Karens were originally meant to criticise a generic imaginary of a racist, privileged white middle-aged woman, meme communities at times applied these tropes for public shaming, labelling, mocking and attacking women (of all ages and ethnicities) – irrespective of the contextual information supplied with it (which may be false, accurate, or missing). Thus, while Karen memes may be humorous critiques, they have also been appropriated for call-out practices re-attributing blame to so-called Karens as entitled and/or bigoted individuals. To illustrate, videos shared on YouTube under titles such as ‘Karens getting what they deserve’ or ‘Crazy Karen getting arrested/ tased/ beaten’ were often based on decontextualised videos showing violence against angry women (see Figure 5), a format that utilised the humorous notion of Karen memes, but has developed beyond its original meaning. A Karen meme may in that regard be both racist (by the Karen) and misogynistic (towards the Karen), and a woman both a perpetrator and victim (e.g. expressed in group descriptions such as ‘FuckYouKaren’) – a polysemous result of the complex hyper-contextual/multi-contextual character of memes, but also their appropriation and resignification by user communities. These entanglements of memetic contents, real and imagined affordances, rendered the memes polysemous, ambiguous and ephemeral.
Video frames of Karens getting beaten, tased or arrested (unidentifiable frames chosen).
As such, memes may become subversive practices that create oppositional, conflicting and at times unstable meanings. This equally applies to the meme collections that comprised violent memes alongside more generic ones. For example, where the description of a video showing a woman being handcuffed or beaten as ‘Karen getting what she deserves’ was buried alongside generic Karen memes, it inadvertently legitimised violence based on the community’s moral agreement. That meme in isolation therefore does not necessarily or clearly articulate whether this commentary is targeted at a specific Karen case or constitutes part of wider misogynistic and violent narratives. Thus, beyond memetic cultures and content moderation processes, group homogeneity and collective identity played a significant role in how a given meme was read and whether and how memetic misogyny was maintained – an issue that is not necessarily visible without a closer reading of the given communities of practice.
A particular hindrance in this context was the polysemy of memes, that is, their multiple and sometimes multilayered meanings and narratives. A contextual complexity of these images therefore lies in identifying their primary narrative. While posts against the feminist movement behind #SisterIDoBelieveYou were more targeted and specifically related to gender violence (i.e. the gang rape case), neither climate change nor Karen narratives were, in principle, rooted in issues of gender equality. Misogyny existed, in that sense, sometimes as a primary narrative and was targeted, or as a secondary narrative among interlinked or potentially competing narratives. How far a given case expressed gender-based violence depended, as such, on the community’s framing and wider practices, particularly where multiple narratives converged (example ‘Greta as Karen’ in Figure 6). Tacit moral agreement within a given group therefore dictated how primary narratives were determined and understood, and the higher-level meaning or ‘meta-language’ that is inferred from them (i.e. the myth). Polysemy therefore required a closer knowledge of community values, which emerged from longer-term participation.
Greta Karens.
Longer-term engagement also allowed for an understanding of cultural, regional and pop-cultural references that determined how a given meme was read. Cultural references were coded into memes both textually and visually through specific vocabularies. Well-known examples include self-identification labels such as red-, blue- or black-pillers, which sometimes appeared in subreddits using Karen memes, as well as visual stereotypes such as the Karen. A reading of these cases as misogynistic therefore required a knowledge of group-specific codes and references.
This pattern was also reflected in a meme shared under #SisterIDoBelieveYou (see Figure 7) featuring a woman looking at the camera lens in the foreground and a man looking at a mobile phone in the background. Our regional knowledge allowed us to recognise these two individuals as Pablo Iglesias (political leader of the new left in Spain) and Dina Bousselham (Iglesias’ political advisor at the time). It is also insider knowledge that allowed us to identify this image (a) as a manipulation of an original campaign against gender violence by the Spanish Ministry for Health, Social Services, and Equality; (b) as related to a the theft of Bousselham’s sim card in 2016 which was later linked to a corrupt police commissioner (Villarejo) and (c) as a hijacking of a feminist issue – an attack on the law against gender violence popularly known as ‘Only Yes is Yes’. The meme depicts Pablo checking a mobile phone. It is implied that he is checking Dina’s phone to control her, an action associated with domestic violence by the governmental campaign designed by the very new left Pablo Iglesias led at the time. The meme obscures the socio-political context that the pictured woman is Iglesias’ political advisor, whose phone was stolen and its contents filtered to a Spanish magazine and to the political party Podemos. Without the relevant geopolitical knowledge, this myth of political hypocrisy would therefore remain occluded. Thus, although the meme did not seem overtly misogynistic, the regional and insider knowledge allowed for its anti-feminist and contrarian readings.
A memetic critique of the Spanish party Podemos and the feminism they and #SisterIDoBelieveYou represent.
In accumulation, these examples highlighted that mythologies of memetic misogyny were (beyond merely elements depicted within a given meme) a product of complex entanglements of meme contents, platform affordances and communities’ memetic practices (cultural, geo-political, linguistic and symbolic codes and references) through which meaning was appropriated, resignified and negotiated. Memetic meaning (here: misogyny) was consequently rendered implicit, polysemous, polycontextual and ephemeral. It was therefore difficult to identify, as its meaning was not static but unstable.
Discussion: contextual complexities in memetic misogyny
These findings show memetic misogyny as a product of the complex entanglements of memetic contents, platform affordances that allowed for hiding, distributing and framing these memes, as well as the construction of moral hierarchies in communities of practice. On the level of content, misogyny was created through common depictions, the targeted use of aesthetics, and common narrations of women based on existing social roles that express gendered hierarchies. Despite the mainstream applicability of many of these narratives, platform specificities mattered, above all in terms of the digital mechanisms available to users. They were harnessed for visibility markers or as framing devices, for example, hashtags for visibility and framing in the case of Twitter, subgroup descriptions and administration for limiting visibility on Facebook, title descriptions and compilations for framing on YouTube and subreddits for the creation of Reddit meme collections. This entanglement of users’ interpreted and subversive uses of platform features allowed them to subvert and appropriate memetic meanings towards anti-feminist and misogynistic messages.
Depending on the specific cultural communities these were subject to, networked features were used for organised, targeted, coordinated and amplified misogyny (#SisterIDoBelieveYou on Twitter) or for ambient forms of misogyny that resulted from carefully curated homogeneous communities with moral agreement on gendered social roles and/or gendered violence (e.g. subreddits and Facebook groups). As such, how misogyny was expressed extended beyond the memetic content to how affordances were (re)imagined and practised by the cultural communities in which they were shared, including how they were circulated, (re)contextualised, resignified and (in)visibilised. These contextual complexities demonstrate that, while visual depictions and their textual framing play a significant role in how memetic misogyny is generated, there are further influences that impact their reading: platform specificities (i.e. ‘real’ affordances), user practices (i.e. ‘imagined’ affordances) and the cultural communities surrounding these (overview in Table 1, combining themes from the literature review with findings from the ethnographic project).
Content and context in visual misogyny.
We have termed these instances misogyny here in line with Manne (2017) as they reflect punishments (both physical and rhetorical) based on gendered expectations of both the targeted individuals’ and more generic gendered social roles. As part of the latter, many of the memes discussed here were ambient in nature, and oftentimes commented more generically on non-specific women’s attributes, actions and the deserved consequences. For example, in both the Karen case and the gang rape victim, violence was geared towards the generic corpus of either ‘feminists’, ‘female activists’ or women deemed ‘Karens’. As such, these memes created mythologies of misogyny that could be applied to any and all women, regardless of who they are or what they did in a given context. These visual and multimodal practices therefore expressed, solidified or enforced gender-based online violence or hierarchical gender relations based on heterosexual cisgender male dominance.
Conclusion: From Memetic Misogyny to mythologies of hate
This article provided insights into contextual factors in a form of digital hate: memetic misogyny. We did so through an ethnographic exploration of three case studies that build on extant scholarship documenting wider practices of visual misogyny. The results illustrate that while memetic misogyny is a global and cross-platform phenomenon, it is situated within platform-specific subcultures of use – both practices emerging from imagined affordances and the cultural communities that attribute meaning to these. Memetic misogyny is created, spread, co-shaped and maintained through a range of (1) content (i.e. common depictions, narratives and aesthetics), (2) platform affordances – both real and imagined and (3) the cultural communities that (re)inscribe memes with meaning. In combination, these contextual factors create what we term mythologies of misogyny as a higher-order reading that requires complex contextual knowledge. They are gendered socio-cultural and socio-political narratives that present misogynistic ideas as if they were natural, universal, timeless gendered truths, and often go unnoticed because of the volatile, polysemous and polycontextual nature of memes.
The cases highlighted above show that the role of platform specificities is somewhat limited as cultural communities steer how these visuals are read, spread and (mis)understood, and how memetic practices are established on that basis. The memes therefore moved beyond shared/global interpretations as their reading requires insider knowledge only acquired via group inclusion in a concrete community of practice. Outsiders/viewers may only be able to interpret these memes on a superficial level, thereby missing their hateful/misogynistic character (i.e. interpreting the meme with lacking tacit/insider knowledge). For non-community members such as policymakers, the scientific community, platform providers and the general public, this means that misogynistic memes bypass voluntary and hired human content moderation, as well as automated ones, and identification processes conducted through artificial intelligence. Meme-based videos constitute a particular challenge herein as they consist of multiple frames that are continuously (and often invisibly) cut, edited and narrated. As a result, the mythical qualities of misogynistic memes constitute a significant challenge for policymakers and industry in this field. We therefore join scholars such as Matamoros-Fernández et al. (2023) with a call to policymakers to take humour seriously, to differentiate between harmful and harmless humour and to recognise that memetic harm may be difficult to ascertain and identify without an understanding of the individual cultural communities that inscribe these with meaning. This also suggests that memes in themselves may be socially harmful visual artefacts as they are permanently unstable and elusive.
The contribution of this article is fourfold in that regard. First, it provides a more nuanced contextual understanding of how memetic misogyny is generated, circulated, co-shaped and maintained, that is, the four-part model of content, real and imagined platform affordances and cultural communities these are embedded in (see Table 1). Second, it draws out the specific role meme cultures inhabit in these processes in that their polysemous, polycontextual and ephemeral make-up allows for multiple readings that trivialise and mask misogyny. Third, it demonstrates how meme practices allow for the proliferation of mythologies of hate that keep moving beyond shared interpretations and therefore bypass policymaking and moderation processes. Finally, the article develops a definition of memetic misogyny, a sub-practice of platformed visual misogyny, as: the wide array of memetic visual and multimodal practices expressing, solidifying or enforcing both online gender-based violence or hierarchical gender relations based on heterosexual cisgender male dominance (often white and Western) emerging from fluctuating memetic misogynistic meanings in contents (depictions, aesthetics and framing), platform affordances (real and imagined) and the meaning-making processes of the cultural communities where these are shared and appropriated.
These contributions carry relevance for wider contextual complexities of digital violence. Practices of memetic misogyny suggest that meme cultures provide a creative and liberal playground for both ambient and targeted violence. This applies to wider practices of digital gender violence, as well as varying forms of networked visual violence (due to the implicit nature of visual communication). It also applies to the wide array of digital violence practices, as memetic trivialisation equally applies to other forms of violence, such as racism, xenophobia, homophobia, islamophobia, antisemitism, ageism, ableism and intersectional cases of digital hate and violence. Going forward, this means that scientific communities will need to further unpack the mythologies that mask and simultaneously normalise misogyny. Given that this was an inductive exploration of contextual factors in a form of visual misogyny, further systematic and especially deductively oriented analysis is needed to test this theory on larger samples of data and for understanding user readings and responses to these. On a policy level, this means that there remain unmitigated issues in digital violence that will carry significance for how platforms regulate content and, consequently, themselves.
Footnotes
Acknowledgements
The authors would like to thank special issue editors for their efforts, the collective surrounding our wider work on visual misogyny, including the Communication and Information Studies (CAIS) research group, for their insights, the European Communication Research and Education Association (ECREA) Visual Cultures section for the networks they enabled and the proyecto Digital Visualities and Sexual Rights (DiVISAR) (
). Patricia and Suay co-led the CAIS 2025 working group Critical Interventions into Platformed Visual Misogyny.
Funding
The authors received no financial support for the research, authorship and/or publication of this article.
Ethical approval and informed consent statements
Research conducted here was ethnographic in nature. Given the controversial practices reported here, efforts have been made to make users unidentifiable.
Data availability statement
Given the sensitive nature of the data, the raw data have not been made available.