Answers are grounded in this paper and cite their sources.
A formal model predicts comprehension time in cognitive systems, indicating balance in processing knowledge and capacity.
Current models of cognitive processing quantify individual constraints — working memory capacity (Miller, 1956; Cowan, 2001), memory scanning time (Sternberg, 1966), forgetting dynamics (Ebbinghaus, 1885) — but do not integrate them into a prediction of comprehension time: the time a system requires to process material to the point where it can detect internal inconsistencies. In artificial intelligence research, evaluation focuses on production quality rather than comprehension, leaving the distinction between generating a plausible response and understanding the input largely unformalized. We propose a phenomenological equation that models comprehension time T as a function of information load I, verified prior knowledge K, thinking space C, and a system-dependent search cost coefficient α(C), conditional on a verification gate V: Given V = 1: T = α(C) · (I / K) · exp(I / (K · C)) Given V = 0: T is undefined. The model yields three results. First, the optimal thinking space is not “as large as possible” but C = I/K (human) or C = I/(2K) (transformer), determined by the balance between combinatorial load and search cost. Second, minimum comprehension time scales as (I/K)² for humans and (I/K)³ for transformers — a difference traceable to linear versus quadratic search cost scaling. Third, comprehension requires a verification process: without it (V = 0), the system produces without comprehending, and contradictions are systematically overlooked regardless of capacity or knowledge. Three falsifiable predictions are formulated. A concrete experiment using contradiction detection as a comprehension proxy is proposed, applicable to both human participants and transformer architectures.
No takes yet. Share an insight, caveat, or question.
Michael Schmid (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: