Against the Ontologization of Probability: Records, Events, and the Limits of Structural Realism

This paper argues that probabilistic structure should not be treated as ontologically primitive. Instead, probability is defined relative to conditions of recordability: it organizes spaces of possible records rather than properties of events as such. The analysis distinguishes between events, records, and probabilistic structures, and shows that probability acquires meaning only when stable, distinguishable, and reproducible records are available. In this view, Kolmogorovian probability theory provides a formally consistent framework, but does not specify how probabilistic domains arise from empirical observation. As a result, probabilistic formalism underdetermines ontology. Different ontological models can generate identical probabilistic structures, which implies that probability cannot serve as a direct guide to the nature of reality. The paper critically examines Ontic Structural Realism (OSR), arguing that it risks conflating representational structure with ontological structure. It concludes that any ontological interpretation of probabilistic or correlational structures requires independent justification beyond formalism.