IEKV: A Structural Viability Framework for Evaluating Knowledge Under Scaling

This white paper introduces IEKV (Integrated Evaluation of Knowledge Viability) — a structural framework for evaluating knowledge constructs beyond citation-based and attention-based metrics. IEKV addresses a systematic blind spot in contemporary research evaluation: the absence of tools for diagnosing structural viability and scaling risk of ideas. While existing metrics primarily measure social uptake (citations, visibility, impact), IEKV evaluates whether a knowledge construct remains coherent, operational, and context-aware when applied beyond its original scope. The framework treats scientific and strategic texts as systems, not signals, and assesses them along six dimensions: Positive dimensions:Cognitive Coherence, Structural Novelty, Realizability Penalty dimensions:Cognitive Overload, Methodological Noise, Contextual Mismatch These dimensions are aggregated using a fixed multiplicative rule designed to reflect fragility and failure dynamics in complex systems. The paper presents: a formal IEKV evaluation protocol (MVP), a multi-evaluator robustness experiment using heterogeneous AI models, comparative case studies across normative, conceptual, academic, and operational texts, and a dual-loop validation architecture combining AI-based diagnostics with human expert panels. Importantly, IEKV does not assess truth, correctness, or impact. Instead, it functions as a pre-scaling diagnostic instrument, identifying structural risks and failure modes before ideas are institutionalized, operationalized, or widely deployed. IEKV is intended as a complementary layer to peer review and bibliometrics, particularly suited for: grant and proposal pre-screening, interdisciplinary research evaluation, foresight and strategic planning, editorial and reviewer support, AI training data curation, and organizational knowledge governance. The framework emphasizes transparency, reproducibility, and explicit limitations, and is designed to support informed human decision-making rather than automate epistemic authority.