AI & ML Paradigm Challenge

The gray area of logical reasoning is actually built into the rigid math of classical logic.

April 29, 2026

Original Paper

Approximate Entailment from Classical Analytic Decomposition

MARIO PIAZZA

SSRN · 6663716

The Takeaway

Logicians usually assume that fuzzy logic or external similarity scores are needed to handle statements that are mostly true. This paper shows that graded logical tolerance emerges naturally from classical analytic decomposition. The mathematical structure itself allows for a concept of approximate entailment without adding extra layers of complexity. This means that rigorous, hard-coded logic is more flexible than we previously gave it credit for. Understanding this helps bridge the gap between strict symbolic reasoning and the messy, probabilistic world of real-world data.

From the abstract

Fractional semantics extracts a rational invariant from classical analytic decomposition bycounting tautological leaves in stable, terminating proof search. This paper shows thatthe same invariant generates a canonical theory of approximate entailment within classicallogic. For each ε ∈ Q ∩ [0, 1], the ε-slack relation Γ |∼ε Δ accepts a classically derivableinference exactly when the loss of fractional value from premises to conclusion is boundedby ε. Strict fractional consequence is the zero-bu