Modal logic analysis distinguishes human and AI text geometry

Analyzing semantic neighborhood graphs through modal logic reveals structural differences between human and machine-generated text. Transformer embeddings exhibit higher frequencies of axioms B, 4, and 5, allowing for a statistical differentiation between natural and synthetic corpora based on graph symmetry and transitivity.