Mapping textual horror: a cognitive linguistic exploration of the Backrooms discourse

Abstract The present study investigates the linguistic mechanisms underlying textual narratives that evoke horror in the Backrooms discourse. Originating on 4chan (an anonymous image-based internet forum) in 2019 and rapidly evolving through participatory storytelling, the Backrooms discourse depicts unsettling urban spaces marked by shifting architectural forms and ambiguous spatial configurations. Drawing on cognitive linguistic frameworks, including frame semantics and Cognitive Grammar, this exploratory analysis examines how language creates unease by systematically disrupting prototypical schemas. A corpus of 5,676 words from 10 Backrooms descriptions (Levels 0–9) is analyzed using an annotation scheme that identifies semantic incongruities, agentless constructions, and non-prototypical framing structures. The findings reveal that textual horror is generated by contradictions in spatial frames, dehumanization via passivation, and deliberate hedging that shifts attention from human agency to the environment. These linguistic features undermine conventional image schemas such as container and source-path-goal , contributing to cognitive disorientation. The study proposes a way of analyzing digital horror narratives and suggests that further quantitatively oriented research may expand understanding of the role of language in constructing textual horror.