Existential Cognition: A Computational Analysis of Existential Discourse in Online Communities

Questions concerning meaning, mortality, freedom, and the nature of existence have long been central to philosophical inquiry and, more recently, to empirical research in existential psychology. Despite this interest, much of the literature remains theoretical or limited to small-scale studies, leaving a gap in our understanding of how existential concerns are articulated in everyday, naturalistic discourse. This thesis addresses this gap by investigating the structure of existential cognition—the cognitive processes through which individuals engage with fundamental questions of being—through a large-scale computational analysis of over 330,000 Reddit comments. The study employs a multi-layered unsupervised learning pipeline, integrating topic modeling, semantic clustering, and a fine-tuned transformer-based model, Existential-BERT-2.0, to capture both lexical themes and deep semantic nuances. Across analytical methods, six recurring domains consistently emerged: meaning and purpose, epistemic doubt and uncertainty, metaphysical reflection, death and mortality, meta-cognitive reflection, and emotional coping. The first five represent the core conceptual structures of existential cognition, while the sixth reflects the psychological responses to such thoughts. By combining computational text analysis with humanistic interpretation, this study demonstrates how fine-tuned machine learning architectures can provide a data-driven, empirical perspective on how the fundamental questions of existence are navigated in contemporary digital discourse.