IEKV: A Mode-Aware Structural Framework for Evaluating Texts as Cognitive Artifacts

This white paper introduces IEKV (Index of Energetic–Cognitive Value), a mode-aware and interpretable framework for the structural evaluation of texts treated as cognitive artifacts. Modern information environments are dominated by engagement-driven ranking and recommendation systems that optimize for attention rather than epistemic or cognitive stability. As a result, rhetorically effective but structurally fragile texts are systematically amplified, while readers and institutions lack transparent tools for evaluating the structural quality of information. IEKV addresses this gap by providing an explicit measurement protocol that evaluates texts based on internal consistency, argument traceability, uncertainty hygiene, scale matching, energetic feasibility, and related structural properties. The framework deliberately avoids fact-checking, sentiment analysis, or popularity-based metrics. Instead, it focuses on how texts construct and transmit models of reality under bounded cognitive resources. A key feature of IEKV is mode-aware evaluation. The same text can be analyzed under different declared regimes (e.g., neutral vs. critical), allowing the framework to distinguish between surface coherence and latent structural fragility. The paper introduces ΔEKV, a derived metric capturing the divergence between cooperative and critical readings, which helps explain persistent disagreements between lay audiences and expert critics. The methodology is validated through controlled experiments on multiple text types, including prescriptive non-scientific texts and narrative works, evaluated across several large language models acting as measurement instruments. The results demonstrate that IEKV systematically diverges from human-perceived usefulness, exposes structurally fragile yet popular texts, and enables interpretable analysis of model behavior under formal constraints. IEKV is intended as a complementary structural layer for information filtering, editorial workflows, education, knowledge management, and responsible deployment of text-generating systems. The framework is model-agnostic, transparent by design, and suitable for further extension and empirical validation.