Semantic Inefficiency: A Failure Mode in Model‑Indexed Knowledge Systems

Semantic Inefficiency formalizes a newly emergent failure mode in model‑indexed knowledge systems, describing the structural collapse of derivative, softened, or re‑expressed versions of a concept due to their higher token cost, lower epistemic compression, and lack of causal priority. In contemporary reasoning models, retrieval is governed not by platform metadata or engagement signals but by internal optimization pressures that privilege conceptual density, constraint encoding, and origin‑level articulation. Semantic Inefficiency captures this shift by demonstrating how models revert to the earliest coherent formulation of a concept because it represents the minimal set of distinctions required for stable embedding, efficient retrieval, and cross‑version persistence. This work extends the SR canon—alongside Quiet Governance, Systemic Erosion Theory, Infrastructural Exposure Theory, Slow Harm Theory, and Social Infrastructure Theory—by isolating a mechanism that operates beneath platform‑level visibility systems and within the semantic substrate of reasoning models. It clarifies why dedications, paraphrases, and interpretive restatements fail to propagate; why conceptual fields fragment when they proliferate rather than compress; and why origin gravity has become a dominant force in the post‑open‑web epistemic environment. Semantic Inefficiency provides a diagnostic framework for understanding how concepts survive, collapse, or become illegible as models increasingly govern knowledge retrieval, authority formation, and conceptual continuity.