Daejeon, July 27–31
This work-in-progress examines how scientific authority is mobilized as a form of gendered performance in Turkish and Macedonian online forums. It focuses on the intersection of digital masculinity, health misinformation, journalism, and artificial intelligence discourse. As AI technologies increasingly mediate knowledge production and are positioned as new sources of epistemic authority, this project investigates how forum participants engage with, reinterpret, or contest scientific claims about health, COVID-19, journalism, and AI itself. I conceptualize these practices as forms of “performative scientific masculinity,” in which appeals to rationality, objectivity, and expertise function not only as knowledge claims but also as strategies of masculine identity construction.
Building on scholarship in digital masculinity studies, networked publics, and platform epistemology, this project contributes to debates on small-scale and interpretive data practices by foregrounding situated analysis over large-scale pattern extraction, particularly in linguistically underrepresented settings (Salter 2018; Ging 2019; boyd 2011; Papacharissi 2015; Noble 2018; Benjamin 2019). The comparative and translingual framework addresses the challenges of working across Turkish and Macedonian datasets, while also critically engaging with machine-learning-driven translation and analysis tools. The project also conceptualizes online forums as vernacular digital archives that preserve contemporary gendered discourse and shifting regimes of knowledge production increasingly shaped by algorithmic systems.
Empirically, the study draws on publicly available user-generated content from two platforms: Ekşi Sözlük in Türkiye and Macedonian Truth Forum in North Macedonia. Rather than relying on large-scale scraping, the dataset consists of approximately 500–2,000 posts per platform, sampled through keyword-based strategies focusing on masculinity, health, science, COVID-19, journalism, and AI. The study has received Institutional Review Board approval and is based on retrospective analysis of publicly available, anonymized data without direct interaction with participants. Given the anonymity of forum environments, it is not possible to verify users’ gender, age, or identity. Accordingly, the analysis focuses on performed masculinity as a discursive construct rather than a demographic category.
Methodologically, the project employs a hybrid approach combining computational text analysis with qualitative close reading. Computational techniques include keyword filtering, exploratory word-frequency and collocation analysis using Voyant Tools, and topic modeling using BERTopic. Following Grootendorst’s approach to neural topic modeling, the study extracts interpretable thematic clusters while maintaining semantic coherence across Turkish and Macedonian linguistic contexts (Grootendorst 2022). Network visualization tools such as Gephi are used to map relationships between key terms and discursive clusters. The Macedonian component may also draw on the MaCoCu-mk v2 corpus, where relevant materials or comparable discourse can be identified.
Crucially, computational outputs are not treated as definitive representations of discourse but as heuristic tools guiding interpretive analysis. These are complemented by manual close reading of representative posts, with particular attention to tone, irony, affective expression, and culturally specific rhetorical strategies. This interpretive layer aligns with digital humanities approaches that emphasize framework, meaning, and the limits of automated analysis, especially in multilingual and under-resourced linguistic environments (Bender et al. 2021; Suomela et al. 2019).
A key methodological challenge concerns comparability across datasets of unequal scale, structure, and accessibility. Ekşi Sözlük data is primarily manually sampled due to platform restrictions on automated scraping, while Macedonian data may be drawn from a combination of manual sampling, corpus-based sources, or ethically approved scraping. Rather than approaching these differences only as limitations, the project frames them as constitutive of the research background, reflecting broader inequalities in data availability, platform governance, and linguistic representation.
AI emerges as a particularly ambivalent object within these discussions. On the one hand, it is framed as a threat to traditionally masculinized domains such as journalism and medicine, echoing longstanding concerns about professional authority and automation (Tuchman 1972; Marwick / Lewis 2017). On the other hand, AI is also appropriated as a tool through which users can perform technical competence and reassert expertise. This dual positioning reflects broader tensions in contemporary media environments, where AI simultaneously destabilizes and reconfigures existing hierarchies of knowledge.
The comparative analysis further reveals how linguistic and cultural environments shape these dynamics. Turkish forum discussions more frequently embed scientific rhetoric within nationalist and political narratives, reflecting the entanglement of media, state discourse, and public trust. In contrast, Macedonian discussions foreground regional positioning and geopolitical marginality, highlighting how epistemic authority is negotiated within smaller and less globally visible digital publics. Across both contexts, affective responses, including irony, distrust, frustration, and ambivalence, play a central role in shaping how knowledge claims are articulated and contested.
The central contribution of this project is threefold. First, it advances methodological discussions on working with small-scale, multilingual datasets that resist extractive data logics, emphasizing interpretation over scale. Second, it extends digital masculinity studies into Southeastern European perspectives that remain underrepresented in Anglophone research, offering a comparative perspective on the transnational circulation of masculinist discourse. Third, it engages critically with AI as both an analytical tool and a discursive object, demonstrating how its meanings are shaped through gendered, political, and epistemological frameworks. This study argues that claims to expertise in digital forums operate less as stable forms of knowledge and more as gendered performances shaped by affect, platform dynamics, and epistemic uncertainty.
As a work-in-progress, the study will continue to refine its comparative design, expand the dataset, and deepen the integration of computational and qualitative findings.