<![CDATA[This study aims to identify the predominant types, linguistic strategies, and perceived functions of hate speech targeted at the global music phenomenon BTS (Bangtan Sonyeondan) on X (formerly Twitter). This study used a descriptive qualitative approach, in which the data consisted of comments containing hate speech against BTS from January 2023 to March 2025, collected from various user accounts, both anonymous and public. Approximately 100 posts were analyzed to reflect the diversity of hate speech. Data collection was carried out using the documentation method, which involves gathering evidence in the form of text and images. We point out how online harassers use irony and metaphor as a shield, spreading hate while hiding in plain sight. The findings reveal that the most dominant form of hate speech is negative characterization, manifested through derogatory labels and stereotypes, followed by dehumanization and demonization. The most frequently deployed impoliteness strategies are positive impoliteness and bald-on-record direct attacks. The analysis indicates that conversational implicature is routinely exploited to veil hostile intent through sarcasm, metaphorical language, and irony. This result shows why platforms struggle to moderate implied malice, prompting a shift from policing words to understanding intent, which confirms that in high-profile, platform-mediated contexts, hate speech is not entirely explicit but rather systematically employs indirect methods. These results confirm that hate speech against BTS on X is not solely explicit but also operates through subtle, indirect mechanisms that function to damage reputations, foster exclusion, and circumvent platform moderation policies. ]]>
