Post-digital discourse analysis and hybrid assemblages

Foundational work by Bauman and Briggs (Reference Bauman and Briggs1990) on entextualization establishes that any text is extracted from and shaped by its prior contexts of production and reshaped through subsequent acts of recontextualization. Silverstein and Urban (Reference Silverstein and Urban1996) extended this insight and foregrounded the “natural history of discourse” as a trajectory across social settings. Drawing on these traditions, Wortham and Reyes (Reference Wortham and Reyes2015) developed a framework for tracing how social identities and actions are accomplished across linked speech events. Within Critical Discourse Analysis (CDA), concepts of intertextuality and interdiscursivity have captured how texts embed and transform prior texts across social contexts (Fairclough, Reference Fairclough1992).

Concerns about the blurring of the online/offline distinction predate the post-digital label itself (Moore, Reference Moore2019). What has shifted in what Cramer (Reference Cramer2014) describes as the post-digital condition is not the logic of discourse circulation per se, but the specific sociotechnical conditions through which trajectories of discourse now operate at unprecedented scale and speed. The long-standing dichotomy between online and offline has become increasingly difficult to maintain, not simply because digital mediation has become “ordinary,” but because the effects of digitally mediated discourse now routinely extend to shape people’s offline social life.

The term “post-digital” is not, however, a neutral one. While originally coined to capture disenchantment with the ideology of inevitable digital advancement (Cramer, Reference Cramer2014), it has been taken up more critically to foreground how a set of conditions, including platform dominance, algorithmic gatekeeping, and uneven distributions of communicative reach, has real social consequences (see Blommaert, Reference Blommaert2019). In this paper, the post-digital lens serves as an analytical heuristic to examine how platform features actively mediate whose discourses get circulated and to what extent. For (socio)linguistics and discourse analysis, this makes the critical examination of technological affordances newly urgent, alongside the established tools for tracing discourse across space, time, and social context.

Specifically, digital discourse analysis in the post-digital age attends to how meaning emerges through hybrid assemblages where traditional boundaries, like physical and digital, human and machine, are constantly renegotiated (de Souza e Silva et al., Reference de Souza E Silva, Campbell and Ling2025; Lyons, Reference Lyons2019; Tagg et al., Reference Tagg, Eriksson and Vásquez2025). Importantly, users within these hybrid assemblages are not passive recipients of platform design; they bring their own practices and ideologies to negotiate platform affordances and exercise their user agency with high metapragmatic awareness of platform constraints, so as to achieve specific interactional goals (see, e.g., the widespread use of “algospeak” to evade content moderation algorithms (Steen et al., Reference Steen, Yurechko and Klug2023)). This awareness of platform design is central to the analysis of LIHKG’s user practices that follows.

Affordances-in-practice as discourse-mediating resources

The concept of affordances was originally developed to understand the relational potential between an actor and their environment (Gibson, Reference Gibson1979; Hutchby, Reference Hutchby2001). Within social media research, however, it has frequently been applied to describe relatively stable properties in platform infrastructure rather than the outcome of situated practices (boyd, Reference boyd and Papacharissi2010). Costa’s (Reference Costa2018) concept of “affordances-in-practice” reframes affordances as socially and relationally constructed rather than given. Costa’s ethnographic research on social media use in Turkey demonstrates how Facebook’s “public by default” design was reappropriated by users who created multiple accounts, fake and anonymous profiles, and private chats to maintain social norms and boundaries. This shows that affordances are perceived and situated within specific cultural contexts (see also Lee, Reference Lee2007). Affordances therefore “take shape only through specific material and social practices” rather than existing as predetermined potentials (Costa, Reference Costa2018, p. 3649).

Affordances-in-practice have specific implications for the study of verbal aggression, such as the misogynistic discourse discussed in this paper. As the case study in this article illustrates, each affordance acquires its ideological meaning from users’ repeated, collectively ratified deployment of both platform features and discursive resources within a platform-specific subcultural context. Central to the analysis is how users’ metapragmatic knowledge shapes their deployment of media resources (Androutsopoulos, Reference Androutsopoulos2021). Users recognize not only the technical possibilities of platform features but also the subcultural expectations governing their use. This community expectation enables platform features to sustain and amplify harmful discourse over time.

Online-offline nexus

While affordances-in-practice capture how discourse is enacted within the forum, they do not explain why forum-based discourse contributes to consequences far beyond the platform, in mainstream news, in offline reputational damage, and in the target’s own life. Tracing these cross-platform trajectories requires conceptual and methodological tools that can map how discourse transforms from “small things” (Blommaert, Reference Blommaert2019) online into macro-level social phenomena with tangible real-world effects. Blommaert defines small things as “actual moments of interaction taking the shape of meaningful social conduct” that have “powerful ordering effects” in communities where they become “normatively ratified” (p. 486). These seemingly trivial micro moments of interaction, such as emojis, upvoting/downvoting, or coded language, allow users to do identity and social work beyond their literal meanings (see also Darvin, Reference Darvin2022).

This study draws on Androutsopoulos’ (Reference Androutsopoulos, Georgakopoulou and Spiliotiin press) scalar model of the OON to trace how LIHKG misogynistic discourse achieves consequences beyond the forum. OON analysis maps three scales of discursive trajectory across online and offline landscapes. At the micro level, LIHKG users initiate online discussions with reference to real-time offline incidents (simultaneous nexus action). Over time, forum threads and mainstream news reports feed into each other to trigger new rounds of discursive circulation (stepwise interlocking nexus action). At the macro level, discursive practices on the forum become social currency and shape how targets are publicly constructed both online and offline. As will be demonstrated in the case study, misogynistic labels that emerge in LIHKG threads can circulate across platforms, attract mainstream media attention, and ultimately shape offline reputation and social phenomena.

LIHKG: platform affordances and subculture

Launched in 2016, LIHKG is Hong Kong’s most popular local online platform, with over 24.6 million visits monthly as of September 2025 (Similarweb, n.d.). The platform operates as a traditional forum structure where membership is restricted to those with Hong Kong university accounts or local internet service providers, creating a predominantly young and male user base (approximately 64% male, according to Similarweb, n.d.). Unlike global social media platforms dominated by automated algorithmic curation, the visibility of LIHKG posts is largely determined by users’ active interaction with its interface design.

In terms of platform interface and affordances, LIHKG adopts chronological threading (instead of reverse chronological arrangement like most social media feeds), upvote/downvote systems, nested quoting, custom emojis, platform-specific slang, trending (熱門) feeds on the sidebar, and a maximum of 1,000 replies per thread. While these features are not unique to LIHKG, they acquire culturally specific meanings through users’ situated practices. The trending feed, designed for highlighting popular content, becomes a convenient tool for users to collectively mobilize harmful content or attacks. The 1,000-reply thread limit, which is intended as an organizational constraint of the forum, transforms into a powerful resource for sustaining serial discussions on the same topic, as users immediately create successor threads upon reaching the 1,000-reply capacity. In addition, custom emojis originally serving as forum mascots are often reappropriated as culturally coded visual resources that reinforce users’ stance in their verbal message. While comparable affordances exist on platforms like Reddit and 4chan, many global platforms are primarily English-based and attract geographically dispersed user bases. LIHKG, by contrast, is Cantonese-based, largely shaped by its strong localist subculture with specific platform slang. The result is a form of targeted discourse directed at recognizable local actors and events, which requires analysis grounded in the specific sociolinguistic context rather than generalizations from global platforms.

LIHKG embraces “hard plastic” subculture (Zuser, Reference Zuser2022), a trolling subculture marked by nonsensical humor, deliberate absurdity, coded language, and politically incorrect discourse. According to Zuser (Reference Zuser2022, p. 274), this subculture is characterized by a “celebratory use of Cantonese profanity” and a trolling writing style that negotiates between activism and local identity. As one study participant noted in an interview, “[b]eing plastic is just for everyone to have a laugh […] it’s purely because we know it’s meant to be nonsensical.” However, with a predominantly male user base, this normalization of absurdity simultaneously creates an environment where harmful discourse can be easily trivialized and legitimized through collective endorsement via upvoting and serial thread creation. In addition, this hard plastic culture operates through “vernacular masculinity,” a curiosity toward all things bodily and sexual, which “speaks the sexual language of the curious, obscene and vulgar” (Wong, Reference Wong2024, p. 116). As Wong (Reference Wong2024) observes, vernacular masculinity manifests in the casual use of dehumanizing labels for women, sexual objectification, and clickbait threads that presume heterosexual male audiences through sexually explicit language. The convergence of hard plastic subculture and vernacular masculinity creates a unique masculine subcultural environment where misogynistic discourses become simultaneously playful and harmful, creating what Wong (Reference Wong2024) identifies as a space where the “luminous attention to the convergence of misogynistic cultural products eventually obscures its embedded gender hate” (p. 115).

Analytical approach: Critical discourse-centred online-offline nexus ethnography

The overall analytical approach adopts what I refer to as “Critical Discourse-Centred Online-Offline Nexus Ethnography (CD-OONE),” integrating insights from CDA, digital ethnography, and OON. It draws on and extends Androutsopoulos’ (Reference Androutsopoulos2008) discourse-centred online ethnography, which combines systematic observation of digitally mediated interaction with direct engagement with social actors. The term “ethnography” here is used in the tradition of digital ethnographic and technographic practice (Georgakopoulou, Reference Georgakopoulou2024; Hine, Reference Hine2020), encompassing three interconnected components of data collection and analysis:

1. (i) Platform observation: This involved the research team’s sustained observation of LIHKG’s subcultural norms, interactional conventions, and affordances as both insider users and researchers during the project period. Observed elements include how features such as upvoting, GIFs, and custom emojis are creatively appropriated as discursive resources. Fieldnotes were made throughout to document evolving platform practices and subcultural norms.

2. (ii) Discourse data collection and analysis: The “Calling Out Crazy Women” series was imported into MAXQDA for thematic and discourse-level coding, focusing on misogynistic discourse strategies, multimodal assemblage, and affordance deployment. Focal episodes for close analysis in this article were selected on the basis of recurrence of discourse strategies across threads and evidence of cross-platform or offline uptake.

3. (iii) In-depth interviews: Semi-structured interviews were conducted with eight active LIHKG users and/or bystanders of online aggression. Interview data are drawn on selectively in the present paper to illuminate participants’ awareness of platform affordances and subcultural norms, and to triangulate interpretations derived from the discourse data. The interviews were conducted in Cantonese and transcribed verbatim. Selected quotes for this article are translated into English and participants are fully anonymized.

While the broader study collected data across all three components, the analysis presented in this article draws primarily on the discourse data, with selective interview insights used to corroborate and enrich the analysis where relevant.

OON provides the conceptual tools for tracing how discourse moves across scales, from individual platform interactions to mainstream media coverage and offline social consequences, enabling the analysis of both the micro-linguistic features of aggression and their macro-level discourse trajectories. This study adopts a qualitative, case-based tracing of the online-offline trajectory of Kisabbb’s case, following the tradition of hybrid ethnography, which tracks subjects who naturally inhabit social media and constantly move in, out, and between digital and physical spaces (Liu, Reference Liu2022). As an influencer exists primarily across digital platforms, “offline consequences” are defined in this study as embodied effects on the target(s) as well as broader uptake of discursive practices beyond the original digitally mediated interaction that can be triangulated from multiple sources (e.g., forum posts, the influencer’s own social media posts, and mainstream news coverage). This article documents plausible discursive trajectories on the basis of these converging contexts, without assuming a single causal nexus from forum discourse to any given offline outcome.

Nomination strategies and upvoting

Misogynistic discourse on LIHKG manifests through the strategic deployment of culturally coded linguistic resources that reduce women to stereotypical categories. In critical discourse studies, nomination or referential strategies are used to negatively categorize out-groups. Hart (Reference Hart2010, p. 56) defines referentials as the “dichotomous conceptualization of an out-group defined with respect to a coalitional in-group which both text-producer and text-consumer are assumed to belong to.” In the study context, these referential strategies work to delegitimize targeted women by categorizing them as fundamentally “Other” to the predominantly male user base of the forum, constructing what Reisigl and Wodak (Reference Reisigl, Wodak, Wodak and Meyer2015) refer to as “explicit dissimilation,” the process of discriminating social actors by placing them in categories distinct from the in-group.

What distinguishes this from conventional CDA nomination analysis, however, is that the current analysis foregrounds how these naming practices on LIHKG acquire their delegitimizing power through affordances-in-practice. Naming in this context is a discursive practice that emerges through users’ repeated and situated engagement across serial threads, upvoting and downvoting, and increased visibility on trending feeds. Costa (Reference Costa2018, p. 3651) emphasizes that affordances are “never specific to a platform only, but are always specific to the relation between the platform and the situated users.” The analysis reveals that platform features act as a discursive resource for categorizing women into stereotypes, notably through four frequently occurring derogatory labels and naming practices in the forum data.

Example 1

Example 2

Example 3

Example 4

At the linguistic level, these labels function as naturalizing and depersonalizing metaphors or assign specific negative qualities that frame social actors as deficient (Wodak & Reisigl, Reference Wodak, Reisigl, Tannen, Hamilton and Schiffrin2015). The use of “chicken” in Example 1 not only undermines the target’s social capital by tying her achievements to sexualized labor but also perpetuates the idea that visible signs of wealth or celebration cannot redeem her from the group’s predetermined stigma. The use of 豬 (“pig”) in Example 2 works to dehumanize, reducing a woman’s physicality to something unworthy of empathy. When the label “MK” is deployed in Example 3, it enacts negative social judgment through local class stigma, suggesting that appearance and perceived vibe signal low cultural status. In Example 4, 大媽 (“big mom”) mobilizes ageist ideologies to delegitimize the target based on perceived lack of youth or style. Each example shows how referential expressions serve as social boundary markers to delegitimize and even silence women within LIHKG.

The significance of these examples lies in the high upvote counts (157, 106, 87, and 52, respectively), which represent affordance-mediated collective endorsement of each evaluative stance toward the target women. In Blommaert’s (Reference Blommaert2019) terms, each upvote is a “small thing” that acquires its ordering force through aggregation; that is, when a nomination is endorsed through collective upvoting, the naming practice achieves normatively ratified status within the LIHKG community. Crucially, this endorsement is publicly quantified and visible to every user through the platform’s interface. This visibility extends the reach of each derogatory label beyond the originating post. The upvote system transforms a single user’s derogatory comment into an “acceptable” truth. This is why the analysis should treat the upvote as an integral part of the nomination strategy itself.

Interview data confirm that users recognize misogynistic naming as a platform-level culture. As one female LIHKG member, Brenda, reflected:

Brenda’s account suggests that these nomination practices are not isolated incidents but part of a platform-level misogynistic culture that users recognize and anticipate. Her observation that sexual offenses are often reframed as “labour disputes” illustrates a recontextualization where violence against women is neutralized and where victims are downplayed as service providers or paid workers. This practice exemplifies the affordances-in-practice argument, as the delegitimizing function of these labels relies on users’ shared subcultural norms within a discursive environment in which vernacular masculinity and linguistic practice are so integrated that misogynistic nomination no longer requires justification.

Discourse amplification through nested quoting

Nested quotation, the practice of “quoting previous body texts” to establish interactional coherence in asynchronous environments (Fu et al., Reference Fu, Abbasi and Chen2008), contributes to reinforcing aggression on LIHKG. This affordance enables users to construct visually complex structures that sustain misogynistic language by making the process of discourse layering explicitly visible. Once quoted and nested in multiple layers, the derogatory referentials discussed in the previous section can be repeatedly recycled and remixed. This practice demonstrates that affordances-in-practice function through accumulation, aligning with research on verbal aggression, which argues that online bullying relies fundamentally on repetition (Law et al., Reference Law, Shapka, Hymel, Olson and Waterhouse2012). Figure 1 illustrates this nested quoting in practice.

Figure 1.

Nested quoting reinforcing visibility of negative naming practice 雞 “chicken” (highlighted).

Figure 1 is extracted from a thread discussing a female influencer, Yan, collectively referred to as “chicken Yan” here. The excerpt begins with a user noting the rapid deletion of her Instagram stories, followed by a chain of replies that scrutinize and shame her relationships, financial status, and moral character, using the quoting function to link their replies to specific previous comments. The quoting function creates a visual stack marked by vertical indentation lines, where the derogatory label 雞 (“chicken”) is repeatedly embedded and visibly displayed across multiple turns in the conversation.

While the primary function of such nesting is to maintain interactional coherence (Fu et al., Reference Fu, Abbasi and Chen2008), Eklundh (Reference Eklundh2010, p. 29) notes that quoting also serves to “draw attention to the exact form of the message… often as part of a negative stance.” In this example, the quoting affordance enables users to build visually layered sub-threads where each reply amplifies and remixes the initial insult. This recursive practice effectively turns individual instances of verbal aggression into a communal practice, making it highly visible to all users. As a result, the nested quoting structure not only intensifies the misogynistic label, but also signals to newcomers that such language is accepted and expected, further embedding these patterns into the forum’s everyday discourse.

Custom emojis as ratified aggression

LIHKG provides users with a platform-specific emoji system that serves as an embedded visual affordance. These emojis originated from HKGolden, LIHKG’s predecessor, where users developed custom icons because early web forums lacked standardized emoji support (Yu et al., Reference Yu, Tay, Jin and Yuan2023). Today, LIHKG users continue to use these forum-specific icons over standard Unicode emojis to mark subcultural authenticity and in-group membership, serving as markers of local identity. These distinct cultural meanings are acquired through years of collective use in specific discursive contexts. A famous local example is the LIHKG pig emoji (連豬). As it frequently appeared on protest sites as a mascot during the 2019–2020 Hong Kong protests, the government eventually banned its appearance after the pig was featured in a public festive display, claiming that it would “cause discomfort” to the public (Hong Kong Free Press, 2020).

Another notable emoji is the clown emoji, which symbolizes “clownish” absurdity central to the hard plastic (硬膠) subculture as discussed before (Zuser, Reference Zuser2022). In the “crazy women” thread series, the clown is often deployed in a serial manner, together with the “throw-up” emoji to signal ridicule and disgust. This combination is particularly evident in body- and appearance-shaming comments targeting female influencers, as illustrated in Figure 2 below.

Figure 2.

LIHKG post combining derogatory nomination and custom emojis to target a female influencer.

(“Bao Zuan/glittery gel nails = Bao Zhuan/diamond (Putonghua) (three clown emojis) So fucking locust-like (four clown emojis followed by two throw-up emojis)”)

In Figure 2, a user, after posting a screenshot of a female influencer’s video footage, shames her for using a pun based on Mandarin pronunciation (“Bao zuan” = “Bao Zhuan”), labeling this linguistic choice as 蝗 (“locust”), a dehumanizing label used by some local Hongkongers to stigmatize mainlanders as resource-consuming pests. This derogatory referential is visually amplified by a string of four clown emojis followed by two throw-up emojis. Here, the clown emoji functions as a specific marker of the hard plastic subculture discussed before. It potentially indicates that the influencer’s adoption of Mandarin slang is incongruous with LIHKG’s grassroots linguistic “policy” of Hong Kong Cantonese. By repeatedly displaying the clown emoji, the user frames the target’s behavior as deserving mockery. The throw-up emojis then escalate this ridicule into disgust. Together, these visual resources perform a dual act of Othering. The clown excludes the target from the local in-group, while the throw-up emoji indexes a rejection of perceived “mainlandized” appearance and fashion sense (glittery gel nails). This demonstrates how LIHKG’s custom emojis serve to reinforce cultural boundaries and deepen the ideological separation between Hong Kong and mainland Chinese women.

Interview participants confirm this pattern of escalation through the use of platform-specific emojis. As Samuel observed, LIHKG emojis generate “有種團體嘅感覺…比香港人有一種歸屬感” (“a sense of community... a sense of belonging for Hong Kongers”), while simultaneously encoding “諷刺…幸災樂禍” (“sarcasm and schadenfreude”) that “大部份嘅網民都係好容易就可以理解到佢嘅背後嘅意思” (“most netizens can easily understand the meaning behind it”) (Samuel, interview). Serena similarly noted that the mocking tone of emojis on LIHKG is “好明顯” (“very obvious”). The participants’ normalization of the negative connotations of LIHKG emojis suggests that the aggressive undertone of these emojis is no longer contained within LIHKG alone, but has spread into a wider Hong Kong digital vernacular. This illustrates the effect of the hybrid assemblage in which platform symbols, subcultural norms, and shared local knowledge converge to make aggressive intent self-evident.

Forum-specific slang as manual algorithmic discourse

For many years, online forum culture in Hong Kong has developed distinctive linguistic practices where users innovatively adapt language to navigate platform constraints and reinforce forum identity. Notably, the use of “LM” (shorthand for 留名 “leave name”) illustrates a long-standing creative practice originating in HKGolden. On LIHKG, users continue to adopt this practice to engage with LIHKG’s platform affordances. It is one of the interactional “small things” that contribute to the perpetuation of aggressive discourse on the forum. Figure 3 shows the opening messages from the 240th thread of the “Crazy Women” series. It shows a sequence of replies consisting almost entirely of one-word “LM” posts, occasionally accompanied by custom emojis like the LIHKG dog.

Figure 3.

Opening posts of the 240th “Calling Out Crazy Women” thread with “LM”.

Visually, this chain of “LM” comments produces a dense block of near-identical turns, each from a different user. This immediately creates an impression of high activity and popularity without contributing any substantive content. It illustrates users strategically exercising their agency in response to specific platform constraints. As noted, the 1,000 replies per thread limit means that to continue discussions on the same topic, users must manually migrate to a new thread. In Figure 3, the series of “LM” posts at the start of the 240th thread thus signals continuity from the previous 239th thread, allowing users to re-establish their presence in the new interactional space. Unlike social media platforms that push notifications algorithmically, LIHKG relies on users’ ongoing activity to keep a thread visible and alive. “LM” therefore serves the dual function of a personal bookmark to leave a digital footprint for future notification updates, while simultaneously “bumping” the thread to the top of the trending board to sustain its popularity. By repeatedly posting “LM,” users manually overcome platform constraints, ensuring that their chosen topic remains active. This metapragmatic awareness is reflected in interviewees’ accounts: several independently described their LIHKG engagement using the Cantonese idiom 食花生 (“eat peanuts,” meaning spectating from the sidelines without intervening, similar to “grab the popcorn” in English). “LM” is the textual equivalent of “eating peanuts” behavior, keeping the thread alive and the audience visible.

The analysis suggests that LIHKG users are not passive subjects of algorithmic mediation but active agents who intervene to determine post visibility and engagement. In the context of the data on misogynistic discourse, this manual reinforcement of aggressive labels contributes to users collectively endorsing and normalizing the hostile language, signals tacit community support, and encourages further participation by other users (Bruns, Reference Bruns2019). The following section examines the case of one Hong Kong female influencer to trace how these routine affordances and micro-practices turn local forum practices into wider, post-digital cross-media memetic aggression, through the lens of OON.

The case of Kisabbb

This section applies CD-OONE to a single case to trace how the affordances-in-practice and discourse strategies analyzed in the previous sections play out in a cross-media discursive cycle involving a Hong Kong female influencer known online as Kisabbb. This case is selected because it clearly demonstrates how a target-specific discourse event unfolds into a sustained long-term aggression cycle through the interplay of naming practice and platform affordances across media spaces. It foregrounds the paper’s central argument that post-digital misogyny operates not through isolated incidents but through ongoing appropriation of platform affordances that produce memetic effects across online and offline communicative contexts.

Posting her first videos on YouTube in 2014, Kisabbb gradually gained popularity through creating comedic and lifestyle content. Her growing visibility across social media drew the attention of LIHKG users from around 2017, whose early threads discussed being attracted to her body shape. Although these posts centered on the body of Kisabbb, they were framed with a largely positive and admiring undertone (e.g., one post has the subject line “Have you noticed that KisaBBB is actually pretty hot?”). In 2020, however, unedited beach photos of Kisabbb were taken without her consent and circulated online, which triggered immediate body-shaming discourse on LIHKG. Some users expressed shock at the perceived discrepancy between her filtered social media photos and the unedited images. One commenter remarked: “點解真係咁嘅樣..我寧願我以前從來無鍾意過你” (“Why do you really look like that… I’d rather I had never liked you before”). It was in this incident that users established the dehumanizing referential 坦克車 (“tank”), a target-specific label whose meaning transforms as it moves across contexts, from a body-shaming referential on LIHKG, to memetic slang, to a news headline, and back to the forum.

Across the LIHKG data, users repeatedly invoke 沙灘坦克 (“beach tank”) as Kisabbb’s primary identifier, with the label appearing in thread titles, comments, and nested quotations that ensure its continuous visibility. The phrase “按F進入坦克” (“press F to enter the tank”), a gaming reference to entering vehicles in first-person shooter games, emerged as a persistent mocking meme applied specifically to Kisabbb’s body. One user wrote: “有勇氣咪按f進入坦克囉” (“If you have the courage, press F to enter the tank”). This shows how a neutral gaming slang expression is transformed into a sexualized insult that frames Kisabbb’s body as an object to be occupied or rejected. This memetic aggression demonstrates affordances-in-practice operating through cumulative small things: every time the “tank” label is employed, alongside upvotes/downvotes and nested quotation, the term is redeployed and amplified collectively, leading to its memetic spread and mutation (Shifman, Reference Shifman2014). Initially exclusive to Kisabbb, the “tank” label has since generalized to describe other women perceived as overweight. The case shows, similar to the “#Karen” label, how a target-specific attack becomes a convenient tool for targeting specific out-groups.

In early 2025, Kisabbb directly addressed the sustained online aggression toward her over the years through an Instagram post. She wrote: “而家2025年仍然有人翻嗰啲沙灘相…我真係夠晒你哋呢班鍾意網絡霸凌嘅人” (“It’s 2025 and people are still digging up those beach photos… I’m so done with you lot who love online bullying”). In naming the experience explicitly as “online bullying,” Kisabbb asserts her own discursive position from being the passive object of entertainment on LIHKG to being the victim of online harm actively scrutinizing misogynistic discourse. In her Instagram post, she characterizes the experience as “俾人偷拍已經慘, 仲要俾人網絡欺凌咁多年” (“It was bad enough being secretly photographed, but having to endure cyberbullying for so many years on top of that is just terrible.”), explicitly naming the sustained nature of the incident. Rather than ending the abuse, her counter-discourse triggered what one interviewee, Adrian, described as “越反越紅” (“the more you oppose, the more popular it gets”) phenomenon on LIHKG, where collective opposition to a public figure tends to sustain rather than diminish their visibility.

On the day following her Instagram post, the Hong Kong mainstream news outlet HK01 covered Kisabbb’s Instagram story with the headline “網紅 Kisabbb 重提當年被偷拍事件稱被欺凌多年” (“Internet celebrity Kisabbb revisits being secretly photographed years ago, says she’s been bullied for years”). While the headline foregrounds Kisabbb’s experience of being bullied, the visual and textual framing in the news text reconstructs her counter-narrative through the very body-shaming lens that motivated her response. Specifically, HK01 published multiple screenshots and photographs of Kisabbb’s bikini photos, presented under the “即時娛樂” (“real-time entertainment”) section. Through strategic placement of multimodal texts, alongside selected quotes from her Instagram narratives, the news article recontextualizes Kisabbb’s voice as entertainment rather than a testimony of abuse. Notably, the article directly references the tank nomination practice on LIHKG, pointing out “有不客氣的網民取笑現實中的 Kisabbb 是『坦克車』” (“some unkind netizens make fun of the real-life Kisabbb as a ‘tank’”). This reframing uses “取笑” (“make fun of”) that repackages this misogynistic labeling as playful ridicule or, at most, a jocular insult (Dynel, Reference Dynel2021). The entertainment and jocular framing together undergo resemiotization (Iedema, Reference Iedema2003) where Kisabbb’s counter-discourse is removed from its original Instagram context and reinserted into the consumer-entertainment context for further online circulation.

This recontextualization by HK01 is not an isolated incident but reflects a broader journalistic practice in Hong Kong’s media ecology. Koo (Reference Koo2024), drawing on interviews with journalists from HK01 and other outlets, documents how click-based journalism has increasingly marginalized political reporting in favor of soft news that can be sensationalized without being politically sensitive. As such, LIHKG content targeting female influencers represents readily available soft content that is affectively charged, already viral, and requiring minimal editorial resources to repackage as “entertainment” news. As of April 2026, a search of HK01’s archive returns over 4,500 articles referencing LIHKG, while LIHKG returns over 48,000 posts citing HK01. Together, these figures suggest a feedback loop where digital discourse becomes news and news becomes digital discourse.

This news article was reposted to LIHKG on the following day. Many users responded by repeatedly recycling the 坦克 (“tank”) label alongside additional delegitimizing discourse. One user responded: “出返原圖仍然係架坦克車” (“Even if you (Kisabbb) post the original photo, you’re still a tank”). Another user’s meta-commentary on the news article captures an OON effect, stating “冇坦克相直頭唔記得呢個『網紅』” (“Without the tank photos, I’d completely forget this ‘influencer’”). This comment reveals how the news photographs and associated derogatory label initiated online have become Kisabbb’s defining public identity. The discursive circulation and trajectory of this case can be illustrated as follows:

The trajectory exemplifies two scales of OON identified by Androutsopoulos (Reference Androutsopoulos, Georgakopoulou and Spiliotiin press) as stepwise interlocking nexus action in which each event (non-consensual photographs, forum discussion, counter-discourse, and media coverage) feeds the next; and massive memetic uptake, where the target-specific “tank” label acquires generic meaning as a body-shaming resource applicable to any woman deemed overweight. These scalar effects would not have been possible without the LIHKG platform affordances. It is the combination of serial threading, nested quoting, upvoting, and custom slang and emojis that enables the label to sustain over a 5-year period and beyond across media contexts.

Interview data corroborate the conditions that make misogynistic discourse possible on digital media. Amy noted:

Another interviewee, Samuel, similarly observed:

Together, these accounts suggest that visibility, for female influencers, creates the conditions for misogynistic attack, in that it invites normative expectations about femininity, and when these expectations are not met, further collective aggression toward the female targets follows (Duffy et al., Reference Duffy, Miltner and Wahlstedt2022). Kisabbb’s case thus illustrates how sustained online presence as an influencer, rather than any specific act, becomes the trigger for memetic aggression within LIHKG’s vernacular masculinity culture.