1.1 Digital Vigilantism and Doxxing

Doxxing is closely linked to digital vigilantism, or digilantism, defined as “a process where citizens are collectively offended by other citizen activity, and coordinate retaliation on mobile devices and social platforms” (Trottier, Reference Trottier2017, p. 55). This “weaponised visibility” involves “naming and shaming” to expose, shame, and punish individuals, often resulting in enduring psychological and physical harm to targets. Those engaging in digilantism take justice into their own hands to address perceived wrongdoings.

The affordances of digital media – anonymity, global reach, and ease of sharing – have empowered ordinary web users to act as digilantes. Crucially, as society becomes increasingly mediatized and connected, one does not need to be a hacker to retrieve and share others’ information publicly online. As soon as people post content to their social media profiles, whether it is intended to be shared privately or publicly, they are already exposing themselves to an unknown and unintended audience. For example, a “private” Facebook post intended for friends only can easily be reshared by a friend to an unexpected audience. This phenomenon is called the “context collapse” of social media, that is, the “collapse of infinite possible contexts” into one, with an “infinitely ambiguous audience” (Wesch, Reference Wesch2009, p. 23). The diminishing of contextual boundaries enables digilantes to weaponize everyday personal data. The widespread availability of personal data and the ease of sharing information have made doxxing a powerful tool for digilantes to achieve various purposes. In recent years, doxxing has become a common tactic for grassroots activism. For example, in the Hong Kong 2019–2020 protests, targets of political doxxing included politicians, activists, police officers, and journalists (Lee, Reference Lee2020).

The impact of doxxing extends beyond privacy invasion to psychological distress, reputation damage, and even physical harm to the victims. Digital affordances, characterized by “spreadability” (ease of sharing and circulation), “searchability” (ability to search and find content), and “visibility” (ability to make content visible to specific audiences) (boyd, Reference boyd and Papacharissi2010, p. 43), make “forgetness” almost impossible, that is, the shared information remains accessible and visible indefinitely online (Garcés-Conejos Blitvich, Reference Garcés-Conejos Blitvich2022). The repeated resharing and persistent disclosure of personal information online can endanger victims and subject them to ongoing scrutiny.

1.2 Anti-Doxxing Laws

In response to the surge of doxxing cases in recent years, legal measures have been implemented in various countries and jurisdictions to regulate doxxing. While some jurisdictions rely on existing laws, others introduce new or amended legislation to prohibit doxxing offences.

In Australia, the Parliament passed the Privacy and Other Legislation Amendment Act 2024 under the Criminal Code Act 1995 to formally criminalize doxxing. This Act covers two new doxxing-related offences: one is for general doxxing, with a penalty of up to six years of imprisonment, and another is for doxxing based on protected characteristics such as race and religion, with a penalty of up to seven years in prison (Parliament of Australia, 2024).

The legal implications of doxxing in the United States are more complex and continue to be contested. Unlike some jurisdictions, there is currently no comprehensive federal anti-doxxing law, and the act of publishing someone’s personal information is generally protected under the First Amendment. However, doxxing can lose constitutional protection when it constitutes “true threats” or involves “incitement to imminent lawless action” (Cremins, Reference Cremins2024). As of 2025, about twenty states have enacted specific anti-doxxing legislation with varying scope and definitions. However, there remain constitutional tensions as many state laws may conflict with First Amendment protections due to their broad scope. The US case demonstrates the typical and ongoing tension between protecting individuals’ rights to privacy and preserving freedom of expression (for details, see Cremins, Reference Cremins2024).

In Hong Kong, the context of this study, an anti-doxxing law was introduced under the Personal Data (Privacy) Ordinance (PDPO) in response to the rise in doxxing cases during the 2019 protests. The law is enacted under a two-tier structure: (i) a summary offence for disclosing personal data without consent with intent to cause harm, liable to a fine of HK$100,000 and up to two years in prison (Section 64(3A)); and (ii) an indictable offence, whereby, in addition to what is stated in the first-tier offence, the disclosure causes harm. The offence is punishable by a maximum fine of HK$1,000,000 and up to five years of imprisonment (Section 64(3C)). The Privacy Commissioner for Personal Data (PCPD) has also been granted more power to investigate and prosecute doxxing offenders (PCPD, n.d.).

Many jurisdictions are yet to implement a specific anti-doxxing law. In some countries, existing laws have been updated to accommodate the changing landscape of technology that shapes online offences. The UK does not have any specific anti-doxxing law to date. Doxxing-related practices such as harassment and unauthorized personal data disclosure are prosecuted under existing laws, including the Protection from Harassment Act 1997, the Malicious Communications Act 1988, and the Data Protection Act 2018 (Crown Prosecution Service, 2024).

Even with these legal measures, applying the doxxing law in real cases remains a challenge. First, it is difficult to strike a balance between protecting individuals’ privacy and people’s right to free speech. In addition, the affordances of digital media, such as users communicating anonymously and sharing information across multiple platforms, have made it difficult to track down those truly responsible for malicious doxxing. A further complication involves the varying interpretations of legal terms like “intent” and “harm” between institutions and lay actors, sometimes even leading to redefinitions of doxxing that reframe and legitimize harmful actions, as will be demonstrated in Section 5.

1.3 Doxxing as a Social Practice: A Turn to Language and Discourse

The challenge in defining doxxing stems from its multifaceted nature, which is best understood through the lens of social practice. As Trottier (Reference Trottier2017, p. 68) argues, doxxing as a tool of digilantism “should not be regarded as an aberration from other digital media practices, but instead located on a continuum of forms of user-led policing and citizenship.” The leakage of Tesla employees’ data by Dogequest described at the beginning of this Section exemplifies how data sharing is weaponized as a tool of protest and retribution under the name of public interest and social justice. In other words, doxxing is not an isolated phenomenon; rather, it is situated in a continuum of practices enacted and amplified by digital media affordances and digital cultural practices that encourage massive sharing and bottom-up policing.

A limited body of academic research has conceptualized and investigated doxxing as a behavioral, psychological, and ethical matter. Scholars from such disciplines as sociology, communication studies, law, psychology, and information science have examined the motives behind doxxing, the methods employed by doxxers, and the consequences and harm caused (Douglas, Reference Douglas2016; Trottier, Reference Trottier2017; Chen et al., Reference Chen, Chan and Cheung2019; Anderson & Wood, Reference Anderson and Wood2021; Cheung, Reference Cheung, Bailey, Flynn and Henry2021; Huey et al., Reference Huey, Ferguson and Towns2025). Despite much of online doxxing being enacted through written comments or messages posted on digital platforms, little attention is paid to the role of language that shapes the phenomenon.

This Element is amongst the first to contribute a language and discourse perspective to the growing body of doxxing scholarship. It conceptualizes doxxing as a social practice primarily enacted, represented, and sustained through discourse (Lee, Reference Lee2020). Drawing on insights from multiple analytical tools such as netnography, critical discourse analysis (CDA), and pragmatics, this Element aims to unpack the discursive construction of doxxing practices. It offers a comprehensive analysis of a range of discourse features surrounding doxxing. It also reveals how ordinary people and institutions construct, negotiate, and contest its meaning.

This Element is organized as follows: Section 2 outlines the theoretical underpinnings and methodological approaches for researching doxxing as language and discourse. Drawing on the case of doxxing in Hong Kong, Section 3 examines the linguistic and discourse strategies employed in doxxing-related online discussions. Sections 4 and 5 continue to explore attitudes and perceptions toward doxxing, as well as how it is interpreted by participants, based on survey and interview data. This Element also examines selected legal documents of doxxing to reveal the limitations of top-down legal and institutional discourses surrounding doxxing. Section 6 concludes the Element by reflecting on the theoretical, methodological, and policy implications of the study. By analyzing the discursive strategies and public perceptions of doxxing, the study seeks to offer insights into the intersection of language, power, and policy in digital landscapes.

2.1 Doxxing Discourse: Insights from Critical Discourse Studies

Having established doxxing as a social practice in Section 1, the central argument of this section, and in this Element as a whole, is that doxxing should be understood as not just an online behavior, but also a discursive action. This is because doxxing and its related practices are primarily produced, circulated, and sustained through language and discourse. Michel Foucault’s conceptualization of discourse as a system of thought that shapes social reality (Foucault, Reference Foucault1972) provides a framework for understanding doxxing as discourse. Foucault argues that discourse is not merely language, but it is also a system of power that produces and regulates knowledge, identities, and (un)acceptable behaviors. In view of this, doxxing can be understood as operating within such discursive systems, where the act of disclosing personal information becomes a “technology of power” (Foucault, Reference Foucault1971) to punish targets through public shaming, resistance, or silencing. For example, during the US Capitol riot in 2021, doxxing of rioters was used as a form of counter-discourse against far-right extremism, which was often framed through discourses of resistance. Foucault’s conceptualization of power and discourse aligns with doxxing’s dual role in both exposing those who participated in the riot and constructing collective identities (e.g., “defenders of democracy” vs. “threats to democracy”) as the events unfolded.

The field of Critical Discourse Studies (CDS), or what is commonly referred to as Critical Discourse Analysis (CDA),Footnote ³ provides a suitable theoretical framework for examining doxxing because it views discourse as a social practice that both shapes and is shaped by the social actors, contexts, institutions, and social structures in which it occurs (Wodak, Reference Wodak, Leung and Street2014, p. 303). CDS is also “problem-oriented and interdisciplinary” in approach (Wodak & Meyer, Reference Wodak, Meyer, Wodak and Meyer2009, p. 2), which makes it particularly relevant in understanding a complex social practice like doxxing. This Element understands discourses of and about doxxing through the lens of CDS, in that people’s actions and attitudes toward doxxing and its related activities are shaped by the everyday discourse that we routinely encounter in online and offline communicative landscapes (Teo, Reference Teo2000). As a method of analyzing language and discourse, CDA is a contextualized approach that examines how language is used to justify, legitimize, and normalize acts of publicly exposing personal information, thereby revealing the complex social dynamics that underpin the practice of doxxing.

Although to date there has been limited discourse analysis of doxxing (but see Lee, Reference Lee2020), there is an established body of language-based research analyzing various linguistic dimensions of conflicts, aggression, and abusive behavior in digital media. For example, linguistic devices such as irony, metaphors, and nonstandard orthography are commonly employed in enacting flaming and hate speech online (Herring, Reference Herring1999; Assimakopoulos et al., Reference Assimakopoulos, Baider and Millar2017). In addition to overtly aggressive linguistic markers such as profanity and racial slurs, scholars have also observed an emerging pattern of “covert hate speech,” that is, utterances that do not carry explicitly offensive language while causing harm (Baider & Constantinou, Reference Baider and Constantinou2020), such as dehumanizing metaphors like referring to refugees or immigrants as “mice” or “worms.” At the macro-level of discourse practices, researchers adopt CDA to understand the strategic ways in which ideologies are represented through verbal aggression online. For example, a delegitimation strategy, such as rejecting the professionalism of certain social groups, is often employed to justify verbal abuse online (Kagan et al., Reference Kagan, Kagan, Pinson and Schler2019). Focusing on trolling, Hardaker (Reference Hardaker2013) identifies a range of covert discourse strategies, such as the strategy of “digress,” that is, “straying away from the purpose of the discussion.” These covert strategies may seem harmless on the surface, but can still cause emotional discomfort to targets and possibly lead to offline physical abuse.

A closely relevant body of work that informs this Element’s focus on doxxing discourse is Garcés-Conejos Blitvich’s (Reference Garcés-Conejos Blitvich, Johansson, Tanskanen and Chovanec2021, Reference Garcés-Conejos Blitvich2022) extensive research on online public shaming (OPS). Through digital discourse analysis and netnographic approaches, Garcés-Conejos Blitvich (Reference Garcés-Conejos Blitvich2022) discusses six cases of OPS in the United States involving the exposure of perceived racism against visible minorities, where doxxing is a key tactic. Informed by pragmatic theories of (im)politeness, Garcés-Conejos Blitvich argues that OPS reflects emotional and coercive impoliteness, where heightened emotions like anger and indignation are used to punish targets’ perceived wrongdoings, which is racism in this case. The research reveals the important role of language for bottom-up social regulation and resistance. It also informs the understanding of the dynamics of doxxing for the research reported in this Element, which aims to examine the complex interplay between discourse, morality, and social justice in online environments.

In the context of digital media, CDA offers a toolkit for digital discourse researchers to scrutinize the complex power relationships among social actors in online public spaces. It also helps unpack the role of digital activism and counter-discourses online in shaping offline socio-political events. Notably, Social Media Critical Discourse Studies (SMCDS) is an emerging approach to discourse analysis that addresses the changing dynamics of discursive power in social media (Bouvier & Machin, Reference Bouvier, Machin, Guillem and Toula2020; KhosraviNik, Reference KhosraviNik2023). SMCDS argues that social media has transformed the traditional power relationship between authors and readers, allowing for increased reader agency in engaging with and responding to political content through online commenting. SMCDS also examines how social media users actively consume and produce political discourse to influence public opinion and political discussions. This study also applies CDA to analyze online doxxing discourses, so as to uncover how netizens and bystanders use language to justify or normalize doxxing.

2.2 Intertextuality and Recontextualization

Doxxing is primarily constructed through intertextuality, as the act of publicly exposing an individual’s personal information depends on doxxers’ sourcing, organizing, disseminating, and resharing a range of texts from both online (e.g., targets’ social media profiles) and offline (e.g., public records). Intertextuality, the ways in which texts echo, incorporate, and respond to previous texts and voices, is a core concept in CDS (Fairclough, Reference Fairclough2003). The notion has been applied in a range of data types and contexts, including the discourse of online hate speech and racism, where speakers mobilize intertextual references to increase the authority of their own stance (Hodsdon-Champeon, Reference Hodsdon-Champeon2010) or to perform what has been called “intertextual impoliteness” (Badarneh, Reference Badarneh2020). Intertextuality also demonstrates the ways in which texts are inherently historical, in that a text always “responds to, reaccentuates, and reworks past texts” to “make history” (Fairclough, Reference Fairclough1992, p. 270). The historicity and temporality of texts are particularly relevant to the current study of doxxing. To doxx someone is to create and recreate the target’s personal history through the assemblage of prior discourses that may be publicly or privately available online or offline.

In digital media, intertextuality is amplified by social media affordances. Users become “intertextual operators” (Androutsopoulos, Reference Androutsopoulos and Coupland2010) who effortlessly modify and appropriate texts across media platforms through copy-pasting, taking screenshots, forwarding, and resharing. Personal and other private information that was intended for private sharing on one platform can now be easily reappropriated and repurposed, reinforced by context collapse, as discussed in Section 1. This repurposing of texts is enabled by recontextualization, a process where social actions are represented and transformed by discourses in new contexts (van Leeuwen & Wodak, Reference Van Leeuwen and Wodak1999; van Leeuwen, Reference Van Leeuwen, Wodak and Meyer2009). In digital media, recontextualization occurs when content is shared across platforms and adapted to fresh discursive settings (Androutsopoulos, Reference Androutsopoulos2014). Recontextualization plays a crucial role in digitally mediated socio-political protests, where semiotic resources are rapidly recontextualized and distributed across physical and online spaces (Khosravinik & Unger, Reference KhosraviNik and Unger2016). In addition, social media’s real name policy and users’ limited privacy awareness further facilitate the unexpected circulation of personal data, contributing to the proliferation of doxxing (Wauters et al., Reference Wauters, Lievens and Valcke2014).

2.3 Discourse Strategies of Legitimation, Justification, and Argumentation

Another concept from CDS that informs the current study is legitimation. As previously noted, doxxing is a criminal offence in some regions, including Hong Kong, and may fall under data privacy laws elsewhere that do not have specific anti-doxxing legislation. Given this background, it is reasonable to expect that in online forums, those engaged in doxxing or doxxing-related practices, including active users and bystanders, may manipulate or strategically deploy their language to continue exposing others’ personal information without appearing to engage in illegal activities. A CDA approach allows researchers to uncover covert discourse strategies that justify, legitimize, or even normalize the controversial, and at times legally ambiguous, behavior of doxxing.

Legitimation discourse, as defined by Reyes (Reference Reyes2011, p. 782), refers to “the process by which speakers accredit or license a type of social behavior.” Van Leeuwen (Reference Van Leeuwen2007, p. 92) outlines four primary categories of legitimation discourse: authorization (reference to authority, figures, or tradition), moral evaluation (references to value systems), rationalization (references to goals and uses of institutionalized social action), and mythopoesis (narratives that reward legitimate actions). These categories were initially developed to understand top-down legitimation processes of institutional actions, based on the assumption that “[l]egitimation is always the legitimation of the practices of specific institutional orders,” such as policy documents that justify compulsory education (Van Leeuwen, Reference Van Leeuwen2007, p. 92). However, legitimation strategies are equally applicable in grassroots contexts, where counter-discourses seek to resist or subvert mainstream institutional narratives (Feltwell et al., Reference Feltwell, Vines and Salt2017).

Legitimation is closely tied to the discourse process of justification, which refers to the linguistic realization of “defensive reactions to implied or inferred accusations” (Wodak, Reference Wodak1990, p. 132). Both legitimation and justification rely on argumentation strategies that allow the speaker or writer to position themselves as unbiased actors, “free of prejudice or even as a victim” (Wodak, Reference Wodak, Graddol and Thomas1995, p. 8; see also Reyes, Reference Reyes2011). In CDA, argumentation involves providing reasons that serve to explain social actions in order to (de)legitimize them. Legitimation discourse is also goal-oriented. Reyes (Reference Reyes2011) observes that one of the most common goals of legitimation is to seek approval or acceptance from others by presenting potentially controversial action as serving a wider group or community. This Element demonstrates how the legitimation of doxxing emerges as a meaningful strategy employed by both ordinary people and institutions.

Against this conceptual and theoretical backdrop, this Element examines the discursive nature of doxxing and its implications for forensic linguistics research. Using the 2019–2020 Hong Kong protests as a case study, the Element explores how doxxing practices are discursively represented and constructed in both online discussions and institutional documents. The overarching research questions guiding this inquiry are: How do online participants, bystanders, and policymakers discursively construct, represent, and perceive doxxing? What are the implications of these discursive constructions for understanding and addressing doxxing?

2.4 Context: The 2019–2020 Hong Kong Protests and Doxxing of Police Officers

In 2019–2020, shortly before the global spread of the COVID-19 pandemic, Hong Kong witnessed one of its largest socio-political movements in history. The anti-extradition law amendment bill protests (also known as the Anti-ELAB movement) were a year-long pro-democracy movement against the proposed amendment of an extradition bill that would allow transfer of fugitives to any jurisdiction that Hong Kong lacks a formal agreement with, including mainland China. This sparked widespread concerns over potential threats to Hong Kong’s legal autonomy as a special administrative region of China. The movement is believed to be leaderless and self-organized mainly on digital communication via social media and mobile apps such as Telegram and traditional online forums like LIHKG (Lee, Reference Lee2020).

Despite the formal withdrawal of the bill in October 2019, tensions between the protesters, the government, and the police continued to escalate. Riot police were criticized for using excessive force, such as tear gas and rubber bullets, to disperse the protesters.Footnote ⁴ The protesters urged the government to meet their five demands, which included establishing an independent commission of inquiry into police use of violence. Media reports also revealed certain police officers concealing their identification numbers and badges and covering up their faces while on duty on the protest sites.Footnote ⁵ When the police became unidentifiable, this lack of transparency made it challenging for protesters to file complaints against them.

As the tension between the activists and the police surged, numerous doxxing incidents emerged on online forums and social media, notably the Hong Kong-based forum LIHKG, where the personal information of numerous police officers and their families was exposed. As of the end of 2019, the Office of the Privacy Commissioner for Personal Data (PCPD) of Hong Kong received over 1,500 complaints that involved the doxxing of police officers and their families (PCPD, 2020). Information being disclosed includes their full names, phone numbers, home addresses, spouses’ and children’s names, and schools attended, and so on, which was published along with pictures of the targets extracted from their personal social media profiles (Forsyth, Reference Forsyth2020). What followed was a series of public shaming and harassment of these officers and their families online and offline (SCMP, 2019). As a result, the High Court granted an interim injunction order to ban all unauthorized publishing or disclosure of the personal data of police officers and their families. For a while, a handful of doxxing-related posts were removed from LIHKG.

Doxxing in Hong Kong predates the rise of social media. However, it became more widespread during the 2019 protests. As a result, Hong Kong’s legal framework surrounding doxxing has evolved significantly. Based on a preliminary review of 124 media reports on doxxing between 2015 and 2022, four key phases of doxxing in Hong Kong have been identified:

1. (i) Early concerns (2015–2016): In 2015, the government called for restricting access to public registers to protect privacy. As the Legislative Council Election approached in 2016, there were increasing cases of doxxing identified. Hong Kong Privacy Commissioner Stephen Wong urged internet users to respect the privacy of individuals.

2. (ii) Legal measures and academic research (2017–2018): In November 2017, a financial consultant was charged with using others’ personal data in direct marketing without consent. Personal data obtained from public domains were also protected under the Personal Data (Privacy) Ordinance (PDPO). Academics also called for legal responses against cyberbullying. For example, Chen et al. (Reference Chen, Chan and Cheung2018) noted that 30 percent of secondary school students had experienced doxxing.

3. (iii) Protests (2019–2020): During the protests, several arrests were made for disclosing personal data of police officers and their family members. Notably, a thirty-two-year-old man was arrested for exposing the data of over 2,300 police officers on the Telegram channel “DadFindBoy” (Maragkou, Reference Maragkou2019). In October 2019, an interim injunction was granted to ban doxxing against police officers and their families.

4. (iv) Criminalization (2021 to date): The Personal Data (Privacy) (Amendment) Ordinance 2021 was passed on October 8, 2021, to criminalize doxxing. The first arrest under this law was made in December 2021 in which a thirty-one-year-old man was accused of posting the personal details of the victim with whom he had monetary feud.

2.5 Data Collection and Analysis

The data and analysis discussed in this Element are based on a research project funded by the Hong Kong government’s Public Policy Research Funding Scheme (PPR project code: 2021.A4.075.21A). The original study examines how doxxing is defined, perceived, and justified online and among young people in Hong Kong, comparing these perspectives with legal and government definitions to inform education policy development that enhances critical awareness of doxxing. The research design for this study adopts a discourse-centred online ethnographic (DCOE) approach (Androutsopoulos, Reference Androutsopoulos2008), combining systematic observation of online discursive practices and insider perspectives from interviews with internet users. The study adopts a multi-staged design with four main phases of research activities:

2.6 Data Processing and Coding

The study adopted an iterative research process, where the research activities were constantly refined as the project progressed. Multiple data sources were triangulated to understand a particular theme from multiple perspectives. The online questionnaire results were generated through descriptive statistics. The purpose of the questionnaire findings is (i) to gain general insights into the participants’ experience, perceptions, and attitudes of doxxing; (ii) to identify suitable participants for Phase 2 of the study; and (iii) to complement and contextualize the results of the qualitative discourse analysis in subsequent phases of the study. All the LIHKG posts, interview transcripts, and policy documents were imported to the qualitative analysis software MAXQDA to create a database of analyzable texts. Recordings of the interviews were transcribed verbatim by a student assistant. Codes were developed inductively from the data.

Coding and analysis were performed in two stages: The first was a thematic analysis. Following van Dijk’s (Reference van Dijk2009) semantic macro-analysis, the research team performed both open and axial coding to identify salient themes and subcategories in the data. The second stage of coding paid close attention to the linguistic and discursive strategies using CDA tools and concepts. This phase of coding identified argumentation strategies related to the dissemination, legitimation, and naturalization of doxxing. This was achieved through repeated close and critical readings of the linguistic and content choices made in the LIHKG data, the interviews, and policy documents. The CDA was primarily informed by groundwork related to doxxing by the researcher (Lee, Reference Lee2020) and previous research on online aggressive behavior by other researchers (e.g., Assimakopoulos et al., Reference Assimakopoulos, Baider and Millar2017; Baider & Constantinou, Reference Baider and Constantinou2020). The interpretation of discourse practices was also situated within the broader socio-historical background of Hong Kong to uncover the ideologies that shape the participants’ perceptions. The findings from analyses of forum posts and interviews were compared with institutional discourses. The purpose of this comparative analysis is to reveal possible mismatches between young people and the official understanding of doxxing.

To ensure rigor and transparency, the data were coded and repeatedly reviewed by both the author and a full-time research assistant. Regular consultation sessions were also held with trained student assistants, who are also regular users of LIHKG, to interpret context-specific language found on the forum. This collaborative approach helped validate interpretations in the identification of doxxing incidents and discourse patterns. Coding memos were also produced to document major decisions made in the coding, analysis, and interpretation of data. The memos also noted theoretical or conceptual connections that emerged from the data. Patterns of digital practices and discourse strategies, and similarities and differences among social actors (i.e., whether they take up similar or different strategies when representing the same instance of doxxing) were noted. The coding matrix was drafted and constantly reviewed and revised by all members of the project team (see Appendix for the full coding scheme).

3.1 Doxxing as Intertextual and Recontextualized Discourse

To illustrate what a doxxing incident may look like on LIHKG, consider a thread with the subject line, “There’s a dog that really wants to be famous. Will you help it or not?!!” This thread exemplifies the structure and rhetorical strategies often used to expose personal information for public scrutiny. The various components of this post will first be unpacked to reveal how doxxing operates intertextually within the forum’s discursive norms. Doing so also helps to contextualize the data discussed in the subsequent sections.

The euphemistic title hints at doxxing behavior. First, the metaphor of a “dog” serves as a derogatory reference to the target – a police officer – a common dehumanizing label used by activists and protesters. The playful and ironic tone downplays the seriousness of the content, which involves disclosing personal information for the purpose of naming and shaming, a common tactic in online shaming and digilantism (Murumaa-Mengel & Muuli, Reference Murumaa-Mengel and Muuli2021). The post also calls to action, “Will you help it or not?!!,” which actively invites participation from other forum users.

The majority of the initial posts in this thread are represented by the most direct and explicit form of intertextuality, or what Fairclough (Reference Fairclough1992) calls “manifest intertextuality.” The post is a lengthy one, comprising nine screenshots taken from a police officer’s Facebook profile. The thread comprises three intertextual components:

1. – Part 1 of the thread includes a screenshot of the officer’s photo from his Facebook account, captioned with his full name and family members’ names (wife, son) in Chinese, which reads, “Name of dog: [full name of the officer]; Family history of dog: A mother, a wife called [name of wife], a son called [name of son].”

2. – Part 2 of the post features a collage of the officer’s Facebook screenshots and online comment exchanges, including a political debate with a student and a meme mocking the police commissioner.

3. – Part 3 of the post contains five private photos of the officer’s family and friends from Facebook.

3.2 Linguistic Re-Appropriation of Doxxing

This section examines how doxxing-related practices are linguistically reframed by LIHKG users through strategic language use. Drawing on concepts of CDA, the analysis focuses on how forum participants creatively re-appropriate language to construct doxxing as morally acceptable. Three recurring linguistic devices are identified across the dataset – metaphor, euphemism, and irony – which work together to downplay the perceived harm of doxxing.

3.3 Legitimation Discourse Strategies

Legitimation strategies, as theorized within CDA, aim to address the fundamental questions of “Why should we do this?” or “Why should we do this in this way?” (Van Leeuwen, Reference Van Leeuwen2007). There are four primary conceptual categories in Van Leeuwen’s (Reference Van Leeuwen2007) legitimation framework: authorization (reference to authority), moral evaluation (reference to value systems), rationalization (reference to goals and uses), and mythopoesis (legitimation through storytelling). These strategies operate through complex discursive manipulations in order to transform potentially controversial topics or actions into socially acceptable practices. When discursively framing the doxxing of police officers on LIHKG, participants provide arguments to answer “Why do we doxx?” or “Why is doxxing the right thing to do?”

3.4 Legitimation and the Moral Economy of Doxxing

In summary, this section has shown that the legitimation of doxxing is achieved through a complex interplay of linguistic and discursive strategies, including dehumanization metaphors, irony, euphemism, rationalization, (re)definition, negative-other construction, and victimization. Table 1 and Table 2 summarize the linguistic and discursive strategies employed by LIHKG users in the data to legitimize doxxing and related practices.

The analysis of legitimation clearly demonstrates the discursive nature of doxxing, rather than treating it as just a behavioral issue. As argued throughout the Element, language and discourse play a central role in shaping what counts as doxxing and why it is the way it is. LIHKG users often rely on linguistic features such as metaphors, rhetorical questions, and irony to rationalize or redefine doxxing practices. From a critical perspective, these linguistic choices are not random; rather, they are strategically deployed by social actors to serve wider discourse functions, such as moral disengagement, subversion, and self-defense. These legitimization and discourse processes are best understood through the lens of what Thompson (Reference Thompson1971) refers to as the “moral economy,” which describes a legitimizing notion that informs crowd actions that are based on defending perceived rights or customs supported by a wider consensus of the community. By drawing on a shared set of linguistic and discourse processes that legitimize doxxing, LIHKG users collectively negotiate the moral norms of online practices. It is worth noting that these strategies are not mutually exclusive but are often used in combination. For example, in the data, the strategy of (re)definition often co-occurs with blame shifting, whereas dehumanization is evident in almost all cases of negative-Other construction.

The analysis of legitimation discourse also sheds light on understanding the limitations of legal frameworks to regulate digital behaviors such as doxxing. As the discussion in this section reveals, online users can skillfully and strategically manipulate their language to downplay and legitimize what might be controversial practices. In legislation, it is important to address the subtle yet recurring discursive processes in digital communication that often lead to the perpetuation of online harm.

4.1 A Spectrum of Attitudes

The interview data reveal the multifaceted attitudes toward doxxing among university students in Hong Kong, who constitute the core group of LIHKG users. The findings show how discourse actively shapes users’ ethical norms, legitimizes digilantism, and negotiates power dynamics in online spaces. A central theme in the interviews is the pervasive discourse of ambivalence surrounding doxxing, which often emerges from a perceived failure of the legal system to uphold justice or ensure accountability. The examples come from both focus group and in-depth individual interviews with participants who are either active LIHKG users, bystanders, or those who had first-hand experience of doxxing incidents, as well as from respondents’ open-ended comments in the online survey. Attitudes toward doxxing in Hong Kong span a complex spectrum ranging from positive evaluation to normalization, conditional acceptance, and even disapproval.

4.2 Ambivalence and Moral Tension

In the interviews, the participants were asked to reflect on the acceptability of notable doxxing cases in Hong Kong. A number of interviewees showed signs of moral ambivalence (Carr, Reference Carr2009), in which they simultaneously endorsed doxxing practices and expressed their own ethical uncertainty. This moral ambivalence is not simply a sign of confusion or inconsistency, but, as Carr (Reference Carr2009) argues, reflects complex emotional and moral conflicts, and is an inevitable part of human experience. In the data, such moral ambivalence also allows participants to negotiate the moral conflicts between personal values and perceived social justice achieved by doxxing, as illustrated in Respondent K’s words:

The phrase 理性上 (“Rationally”) illustrates what Reisigl and Wodak (Reference Reisigl, Wodak, Wodak and Meyer2009) call perspectivization, that is, the expression of a speaker’s involvement that positions their perspective. Here, the stance marker “rationally” signals Respondent K’s perspective and frames his viewpoint as being based on reason or logic. This logical perspective is contrasted with the more subjective judgment that follows, that some “deserve” private sanction outside the law. This contrastive framing allows Respondent K to acknowledge legal norms (protecting privacy) while simultaneously positioning certain individuals as morally exempt from them. Also using the term “rationally,” Respondent S remarked:

The repetition of the modal of uncertainty “可能” (“maybe”) and the self-reflection “滲透埋自己嘅一啲 ideology” (“some of my ideology influences this … ”) reveals the respondent’s internal discursive struggle, where competing evaluations of doxxing coexist – one that sees doxxing as “understandable” in some situations, and another that is “wrong.” By using hedging and self-reflection, Respondent S constructs a discourse that enables holding both positions to manage and negotiate ethical boundaries in digital contexts.

Ambivalence is often expressed through conflicted discourse. Respondent F noted:

This juxtaposition reveals the respondent’s internal contradiction between empathy with victims and rational endorsement of doxxing. The emotional complexity is encapsulated in Respondent R’s reflection:

Both statements reveal the affective nature of ambivalence (Carr, Reference Carr2009). Carr argues that expressions of moral ambivalence are often shaped by emotion, as divided feelings are crucial for how people experience and negotiate moral dilemmas. Respondent R’s admission of happiness coexists with moral doubts, and Respondent T’s fear of harm is coupled with the pragmatic function of doxxing. Respondent R’s rhetorical questions (“Is it wrong? Is this humanity’s dark side?”) parallel Respondent T’s explicit acknowledgment of ambivalence (“I have different views, both sides”). Structurally, the two statements follow a pattern of first presenting an ethical concern (“it may cause harm”), followed by an adversative conjunction (“but”) [adversative], and ending with a competing justification (“warn society”). This discursive structure serves to balance ethics with subjective experience.

4.3 Morality and Legality in Doxxing

The analysis reveals a spectrum of attitudinal stances toward doxxing. These attitudes are often shaped by a competing discourse between morality and legality, where participants’ shared notions of justice and accountability frequently override formal legal definitions of doxxing. In Section 3, the analysis of LIHKG posts reveals relatively homogenous legitimation discourse strategies to doxxing as a justified action to punish perceived wrongdoing. The interview data in this section demonstrate more diverse and ambivalent stances, where participants struggle with competing legal, moral, and affective discourses. On one hand, some participants legitimize doxxing through moral appeals to social justice (“伸張正義”) and public interest (“公眾利益”); on the other hand, they openly acknowledge its potential to push moral boundaries. This ambivalence exemplifies dialogical tension (Bakhtin, Reference Bakhtin1981), where institutional legal discourse (“doxxing is illegal”) clashes with grassroots discourse (“doxxing upholds justice”). This multiplicity and complexity of attitudes suggests that doxxing cannot be understood or interpreted through purely legal terms. As the next section demonstrates, institutional discourses often criminalize doxxing in terms of privacy violation and harm, overlooking the cultural pragmatics of shame and grassroots discourses of moral logics that sustain present-day digital vigilantism.

5.1 Awareness and Accessibility of the Doxxing Law

When it comes to the 2021 amendments to the PDPO that criminalized doxxing in Hong Kong, there is a significant gap between institutional efforts to define and legislate against doxxing and the public’s understanding of what the law entails. Only 33 percent of the survey participants reported being aware of the doxxing law and the criminalization of doxxing. Participants were then shown the PCPD’s definition of doxxing, and details of the scope of two-tier doxxing offences in the doxxing law, as quoted:

Regarding the general definition of doxxing provided by the PCPD, 71 percent of the respondents find the definition to be clear or somewhat clear. Among the respondents who found the PCPD’s definition of doxxing “unclear” or “somewhat unclear,” some found certain key terms, such as “personal data” ambiguous. Others were unsure whether simply sharing publicly accessible information could still be considered illegal doxxing. Some expressed concerns regarding the lack of clarity about what constitutes intent. A number of participants suggested that the PCPD’s definition is too broad or too subjective, leaving much room for misinterpretation.

When it comes to the two-tier offences, only 45 percent of the respondents found the two-tier law to be “clear” or “somewhat clear.” However, this means over half of the respondents found the language of the law unclear and hard to understand. The respondents were asked to identify key reasons that might hinder their comprehension of the doxxing law (Table 3).

As indicated in Table 3, over 44 percent of the respondents considered the legal wording too technical, and 31.2 percent found some terms difficult, especially “specified harm” (指明傷害), “intent” (意圖), and “personal data” (個人資料), which seemed loaded and lacked clarity. Respondents who selected “Others” provided written explanations of why they found the law difficult to interpret. One respondent asked, 「指明傷害」指的是身體傷害還是精神傷害 (“Does ‘specified harm’ mean physical or mental harm?”); another stated, 傷害一詞過於含糊 (“Harm is too vague a term”); while another remarked, 意圖太廣泛, 人人有不同理解 (“Intent is too broad – everyone understands it differently”). These comments reveal semantic and lexical ambiguities of the law’s language (see Section 5.2).

The perceived complex structure of the two-tier offence also hindered comprehension. 61 percent of the respondents said the law was “too long,” and many struggled to differentiate first- and second-tier offences. One respondent wrote, “I don’t know the difference between the first- and second-tier offences” (不知第一級和第二級的分別) ; another said, “It’s hard to distinguish between tier one and tier two offences” (難以分別一級同二級罪行). Although a tiered structure is common in legislative drafting to enhance clarity and precision, in the case of the two-tier doxxing offences, the distinction between offences based on intent, recklessness, and actual harm caused was not obvious to readers. The language used in both tiers is highly similar and may fail to achieve its intended clarity. If these relatively educated university students struggle to understand the distinction, it is likely that less educated members of the public would find it even more challenging to determine what actions fall under which tier.

Some respondents also noted grey areas in the law, especially concerning re-sharing already publicly available personal data. In the open-ended comments in the survey, one respondent wrote:

Similar uncertainties were also expressed by multiple interviewees. The law does not clearly and explicitly define what kind of personal data is protected once such information enters the public domain. Others mentioned the challenge of proving intent or recklessness. They also noted the lack of an objective standard for assessing intent and harm:

1. (1) 如何證實人是否有意圖進行一件事 (“How do you prove someone intended to do something?”)

2. (2) 受傷害的客觀標準 (“What’s the objective standard for harm?”)

3. (3) 難以定義某些概念, 如實際指明傷害 (“Some concepts like ‘actual specified harm’ are hard to define.”)

Overall, these concerns emphasize the challenge in identifying and assessing subjective mental states and experiences like harm in the legal process. There is a broader disjunction between the technical and rather stringent requirements of the law and the public’s understanding of these standards. With this background in mind, we turned to the interview data to probe more deeply into how participants made sense of legal definitions in practice. The discussion focuses on their interpretations of four key expressions found in official definitions and legal documents. The recurring keywords that were variably interpreted in the interviews are: “personal data,” “intent,” “harm,” and “public interest.”

5.2 Diverging Interpretations of Key Legal Terms in the Doxxing Law

5.3 Discursive Tensions and the Limits of Legal Language

To sum up, this section has demonstrated the rather significant discursive mismatch between legal representations and public interpretations of doxxing. The findings suggest how legal discourse is shaped by and also shapes broader social norms and practices. While the law as communicated by the PCPD and the PDPO may appear precise through strategic lexicalization, such language use often leads to confusion, uncertainty, and resistance among the public. The analysis of four key legal expressions – personal data, intent, specified harm, and public interest – reveals the complex discursive space in which Hong Kong’s doxxing law is interpreted, contested, and applied. Interview data suggest that these key terms are ambiguous and far from self-explanatory. It is clear from the data that the participants interpreted these expressions through every day, contextual, and pragmatic perspectives, rather than rigid legal definitions. This more fluid perspective results in varying and contested standards of what constitutes personal data, harm, and intent in real-life situations. The persistent discursive tension between legal and lay meanings is evident in the study. To address this, it requires not only better top-down communication and education of the law per se; policymakers should also be better informed of how legal discourse interacts with moral reasoning in everyday digitally-mediated communication.

6.1 Summary of Findings

This study has demonstrated that doxxing is not merely a technological or legal phenomenon, but also a discursive practice shaped by language, power, and ideologies. Through critical discourse analysis (CDA) of online forum posts, interviews, and legal texts, three key findings emerge:

• – Doxxing is discursively legitimized: Doxxing is enacted and perpetuated by a moral economy (Thompson, Reference Thompson1971), a concept that is still guiding present-day global civil society and global social justice like digital vigilantism (Calabrese, Reference Calabrese2005), where grassroots, communal notions of informal justice and morality may override formal legal frameworks. Both the forum discourse and interview participants consistently legitimize doxxing through discursive strategies like irony, (re)definition, blame attribution, and victimization. These strategies reflect a grassroots moral logic where privacy violations are seen as necessary resistance to perceived institutional insufficiency.

• – Ambivalence and ideological negotiation: Attitudes toward doxxing reveal ambivalence, where participants simultaneously condemn doxxing as an immoral act and justify vigilantism. This ambivalence is not mere uncertainty but a dialogical tension (Bakhtin, Reference Bakhtin1981) between institutional voices and informal justice seeking from the bottom up.

• – Everyday vs. legal meaning: Keywords in the doxxing law, like “personal data,” “intent,” “harm,” and “public interest,” are reinterpreted through the lens of everyday, contextual, pragmatic meanings. When discussing doxxing motives, participants prioritize context over legal definitions, such as viewing car license plates as public data despite legal identifiability, or framing ironic praise as legitimate opinion. This mismatch suggests a fundamental tension between legal language that seeks static clarity and everyday discourse that is often more fluid and context-sensitive.

6.2 Doxxing and the Online-Offline Nexus

Doxxing illustrates what Blommaert (Reference Blommaert2019) describes as the “online-offline nexus,” the idea that our social actions and identities are shaped across both digital and physical spaces (or what Lyons [Reference Lyons2019] calls “phygital” spaces), which are now so closely intertwined that the online-offline dichotomy has become increasingly irrelevant. In this interconnected space, discursive actions online, such as doxxing and public shaming, can lead to consequences in the physical space, including harassment, threats, or even physical violence. As this study has shown, weaponized visibility (Trottier, Reference Trottier2017) in digital contexts can drive collective action and escalate harm beyond the screen. Context collapse (Wesch, Reference Wesch2009) allows content to be quickly reshared and reinterpreted by unintended audiences, who may recontextualize it from its original context and change its meaning and impact offline. For example, the doxxing of Hong Kong police officers’ families, including the release of home addresses and details of children’s schools, resulted in offline harassment, such as their children being bullied at schools.

The intertextual nature of doxxing, which repurposes personal data into oppositional narratives, creates what Androutsopoulos (Reference Androutsopoulos2024) refers to as a “feedback loop.” In this cycle, doxxing is often triggered by offline events, which then lead to physical world aggression, enacting and intensifying further verbal abuse and instances of doxxing online. For example, during the Hong Kong protests, the “DadFindBoy” Telegram channel, which gathered and exposed police officers’ personal information, became a space for coordinating online doxxing, which allegedly led to offline confrontations (Lee, Reference Lee2020). This case reveals that doxxing, as with many acts of online abuse, is rarely confined to one single online platform; instead, it spreads across multiple networked communities, gaining new meanings and reaching broader audiences as information gets circulated.

For forensic linguistics and discourse-based research of online abuses, the online-offline nexus presents some methodological challenges – How can harmful intent be identified when discourse is scattered across digital and physical contexts? and How should harm be assessed when online discourse leads to offline consequences and vice versa? The intertextual and recontextualized nature of doxxing (see Section 3.1), where the meaning of doxxing and its related practices can be easily reinterpreted to legitimize harmful actions, necessitates an analytical approach that takes into consideration the fluidity of context and meaning. In assessing harmful intent, researchers and investigators must focus not only on linguistic and discourse content within one single platform, but they should also trace the trajectories of meaning and intent as discourses are reinterpreted across media by different audiences.

6.3 Theoretical and Methodological Implications

This study brings to fore the crucial role of Critical Discourse Analysis (CDA) and pragmatics in understanding and responding to the complexities of doxxing in digital contexts. Legal drafting and communication often rely on static legal definitions and may overlook the fluid, context-dependent nature of digital communication. As this Element demonstrates, the anti-doxxing law in Hong Kong may be insufficient to uncover the covert discourse and pragmatic strategies used to legitimize doxxing, such as irony, euphemism, or “jocular insults” (Dynel, Reference Dynel2021), where intent is co-constructed by the writer and the intended audience rather than explicitly stated in the utterances.

By drawing on insights from CDA and pragmatics, the study reveals the limitations of legal entextualization, that is, extracting language from its original context and recontextualizing it within legal settings. This process may overlook the pragmatic complexity essential for interpreting discourse, and can distort or suppress the original meaning and voices of lay participants (Rock et al., Reference Rock, Heffer, Conley, Rock, Heffer and Conley2013). For example, as shown in the data, discourses surrounding “public interest” are highly contested between institutions and lay members of the public, with the law using the expression to justify data protection, while lay actors reframe the same concept to defend doxxing and shaming. These discursive tensions reflect participants’ ideological struggles over the legitimacy of doxxing, which is often shaped by culturally embedded norms such as “shame penalty” in Chinese societies, where public shaming functions as a mechanism of moral accountability (Chen, 2022).

Language is inherently ambiguous. Durant and Leung (Reference Durant, Leung, Barron, Gu and Steen2017) emphasize the importance of pragmatics in addressing semantic indeterminacy, implied and indirect meaning, and ambiguity of reference in legal interpretations. Pragmatic concepts like inference and co-text help resolve meaning uncertainty in linguistic evidence. For example, a case discussed in Durant and Leung involves the court inferring meaning in the defamation cases Lord McAlpine v Sally Bercow, where the interpretation of ironic utterances (e.g., “Why is Lord McAlpine trending? *Innocent face*”) was central to determining whether a tweet was defamatory. The judge eventually ruled that the rhetorical question and the phrase “innocent face” conveyed a defamatory accusation. In the ruling, the court considered the contextual cues, audience’s familiarity with the public figure concerned, and the news story on Twitter. Similarly, in the context of online doxxing discourse, the intended meaning of utterances can be easily manipulated or misinterpreted when taken out of their original context, especially with frequent use of covert discourse strategies. Coupled with CDA, pragmatic analysis becomes especially relevant in understanding digital communication, where context collapse can result in misinterpretation of speaker’s intended meaning. In the actual investigation of online harmful discourse, audience reactions and interactive comments should also be analyzed alongside the questioned utterances. Such reactive comments often provide important contextual cues regarding how a message is perceived, amplified, or reinterpreted by readers. Pragmatic concepts such as speech acts, face, relevance, and (im)politeness can shed some light on investigating ambiguous cases, which could potentially be interpreted either as light-hearted remarks or covertly harmful comments, potentially triggering doxxing.

Finally, a key methodological finding that emerged unexpectedly from the research process itself was when multiple interview participants began to reflect ethically on their own stance and online actions. This reveals an unexpected and unintended dimension of participant engagement that warrants further theoretical consideration. As Respondents S and K reflect:

These reflexive comments reveal the values of focus-group interviews as both a research tool and a discursive site for self-examination. The multi-phased research design, from online survey, to focus group and one-on-one interviews – created unprecedented opportunities for the participants to examine their own assumptions and develop their critical awareness about their own and others’ behavior as well as their language used online and offline. This also constitutes a form of research impact, where these young people, through engaging in the research project, enhance their awareness of the ethical and social consequences of doxxing, which fosters greater sensitivity to the harm it can cause.

6.4 Implications for Policy Communication

This study reveals significant gaps between official legal definitions and public understanding of doxxing in Hong Kong. Over half of the survey respondents found the PCPD’s official definition unclear, with participants expressing confusion about key legal terms. Additionally, only 33 percent of the participants were aware that a specific doxxing law exists in Hong Kong and that doxxing could be criminalized. There is clearly a misalignment between policy intent and public understanding of the law. The findings also suggest that the participants hold ambivalent attitudes toward doxxing, seeing it as both a potential tool for justice-seeking and an immoral practice. This complexity suggests that existing policies need to enhance both legal language communication and public education on the issue.

1. (i) Enhancing legal language clarity and education: In promoting the implementation of anti-doxxing law in Hong Kong, the government should collaborate with the PCPD and adopt a more proactive approach to clarify potentially ambiguous legal expressions. This includes: (a) providing an annotated list of common types of personal data protected under the law; (b) explaining how “specified harm” is assessed; (c) clearly defining “intent” and explaining how it is interpreted in the legal process. Policy communication must move beyond the assumption that precision equals clarity. Future legislative drafting should incorporate plain language principles without compromising legal precision (Assy, Reference Assy2011). This can include using concrete examples and accessible terminology to bridge the gap between institutional definitions and public understanding. Legal language education should also be integrated into the formal school curriculum, such as in Liberal Studies in Hong Kong. Beginning at the secondary or even primary level, common legal expressions and youth-friendly legal education should be introduced progressively. More broadly, comprehensive and regularly updated digital legal literacy programmes should be made compulsory from an early age. In Hong Kong, the Education Bureau should integrate legal literacy modules into existing curricula and establish partnerships between the PCPD and educators to co-develop age-appropriate legal education materials. A successful model in this regard is Singapore’s National Digital Literacy Programme,Footnote ⁸ which includes “Digital Safety & Security” and “Digital Responsibility” competencies as key strands of digital literacy education. The Programme aims to address important areas such as understanding of online behavior, protection strategies, and ethical digital conduct. While education alone may not eliminate harmful discourse online, it plays a role in enhancing the public’s critical thinking in both legal and media domains.

2. (ii) Improving policy communication: Currently, public education on doxxing and its legal implications in Hong Kong is primarily provided through the PCPD’s website, press releases, radio APIs, and TV APIs. These relatively traditional channels, while informative, may not effectively reach the most relevant audiences, that is, active users of social media and local discussion forums where doxxing discourse is frequently produced. More direct engagement with digital media users is necessary for meaningful dialogues with those who are frequently exposed to or even engaged in doxxing discourses. The PCPD should also collaborate with schools and universities to develop educational programmes and awareness enhancement activities to encourage ethical reflections among young people, who are frequent users of these digital media and are particularly vulnerable to online harm.

   The limited awareness of the doxxing law among participants indicates discursive gaps between policy communication and public perceptions. There is an urgent need to thoroughly examine public perceptions of the issue. Rather than relying solely on one-way, instructional communication channels such as websites, policymakers should create interactive platforms for public engagement with the ethical and legal dimensions surrounding doxxing and other harmful online practices. This could involve public forums and consultations, and collaborative definition-making processes that incorporate grassroots perspectives into legal discourse. The participants’ ethical self-reflection in the present study suggests that dialogical policy processes could potentially be more effective than top-down communication campaigns.

3. (iii) Integrating language-based research into policy development: The value of language and discourse perspectives is clearly evident in this study, which has shown that doxxing is largely discursively constructed. The government should establish a task force with representatives from the PCPD and the Department of Justice, who would work closely with language and communication scholars. Further discourse-based research should be conducted within this task force to investigate multiple manifestations of online doxxing across digital platforms and the range of underlying motives and social conditions that shape doxxing practices. Drawing on tools and insights from CDA, this task force should systematically analyze known doxxing cases to identify recurring linguistic patterns and discourse strategies that legitimize or normalize unauthorized data disclosure.

4. (iv) Support interdisciplinary research initiatives: Research funding agencies should establish dedicated funding schemes to support interdisciplinary research on digital behavior. Bringing together legal scholars, linguists, psychologists, and media researchers, such interdisciplinary initiatives should explicitly recognize the role of language in constructing, sustaining, and justifying doxxing behavior. A good example in point is the EU-funded C.O.N.T.A.C.T. project (Assimakopoulos et al., Reference Assimakopoulos, Baider and Millar2017), which brings together scholars, NGOs, and stakeholders from various EU countries to collaboratively address discriminatory discourse and hate speech. The discourse analysts on the project team use CDA to systematically analyze harmful discourse in online news comments, with a particular focus on covert hate speech and implicit language such as irony. This kind of interdisciplinary work also contributes to the growing field of forensic linguistics, as it shows how systematic analysis of linguistic evidence in digitally mediated interactions can generate insights to inform legal interpretation and the investigation of online crime.

6.5 Final Thoughts and Future Research

This Element does not suggest that discourse or language analysis can offer a definitive method for identifying doxxing cases, nor does it claim to use language analysis to resolve the complexity of legal-ethical tensions surrounding doxxing. Rather, it contributes to conceptualizing doxxing practices as discursively constructed and sustained through strategic linguistic and discourse practices, a perspective that is often overlooked in the legal process. Insights from CDA reveal ongoing negotiation of ideologies surrounding doxxing, where the boundaries of public accountability and moral acceptability are constantly redefined. The critical analysis of discourse in this study also raises broader questions regarding power relations in digital governance, such as who has the right to define justice and whose interests should be protected in digital spaces.

This study, of course, has limitations, which in turn present opportunities for future research.

First, the study context is limited to Hong Kong only, and the data were collected from one platform (LIHKG). Target participants were exclusively university students in Hong Kong. The current study clearly serves as an exploratory case study to lay the groundwork for future research on doxxing discourses. A more comparative and comprehensive approach is recommended to examine doxxing discourse across cultural contexts with a more diverse range of participants. More broadly, the discourse of harmful research spreads across polymedia (Madianou, Reference Madianou2016). A polymedia environment is characterized as “the constant availability of a range of mediational tools for interpersonal communication” (Androutsopoulos & Stæhr, Reference Androutsopoulos, Stæhr, Creese and Blackledge2018, p. 119). For linguistics research, this concept is useful in tracing how people navigate and make meaning across multiple communication media. By documenting the discourse trajectories of doxxing within polymedia environments, researchers can better understand how platform affordances, user practices, and cultural norms interact to shape digital practices.

In terms of analytical approach, the study relies heavily on qualitative analyses of forum discourse and interview data. Future research can integrate computational methods to effectively complement qualitative discourse analysis. Researchers can employ corpus tools to examine much larger datasets of how discourses of doxxing circulate and evolve across multiple online platforms (e.g., in Hong Kong, from LIHKG to Telegram). Findings from corpus linguistic analysis could potentially inform AI model training that aims to detect patterns of covert harmful discourse.

Future research would also benefit from longitudinal research to trace how both public and legal discourses surrounding doxxing change over time. This could involve analyzing media reports, social media comments, or legal documents before and after major policy reforms to see how discursive strategies shift in response to new policy measures. Wodak’s Discourse Historical Approach to CDA (Wodak, Reference Wodak2015) offers a valuable framework for such analyses, as it situates media texts within their broader historical contexts. Analyzing the discursive histories of doxxing would also shed light on its online-offline nexus (Androutsopoulos, Reference Androutsopoulos2024), as it situates online doxxing discourses within previous offline events, which in turn shape subsequent online doxxing practices.

Ultimately, the study of doxxing discourse is not just about identifying harmful language per se, but also about understanding how norms of ethics and justice are continuously negotiated in an increasingly mediated social world.