Semantic Fidelity: Lexicon for AI Meaning Preservation

This component contains the canonical lexicon for the Semantic Fidelity Project within the broader Reality Drift framework. The lexicon defines a working vocabulary for analyzing how meaning is preserved, altered, or lost as information moves through AI systems and other representational environments. It distinguishes foundational conditions such as Semantic Fidelity and Semantic Drift from mechanisms including Recursive Compression, Recursive Mediation, Context Loss, Hierarchy Flattening, Constraint Collapse, and Ground Erosion. It also defines recurring failures in AI workflows, including Retrieval Drift, Memory Drift, Agent Drift, Verification Failure, and Verification Substitution, along with optimization and ecosystem effects such as Proxy Optimization, Evaluation Blindness, Synthetic Coherence, Provenance Failure, Citation Laundering, Evidence Laundering, and Synthetic Realness. The lexicon is intended as a reference and diagnostic layer for identifying where representational loss enters a system, how it accumulates, and why correction may fail to restore contact with the underlying source, intent, evidence, or real-world condition. This component should be treated as the current canonical vocabulary for the Semantic Fidelity Project unless superseded by a later version.