Glossary: capture, contamination, collisions
This family groups the exogenous phenomena that distort the interpretation of an entity by AI systems (LLMs, generative engines, agents, RAG).
Here, the problem is not merely the internal quality of a site or corpus: it is the external dynamics of signals, co-occurrences, entity confusions, and semantic dominance effects.
Each entry links to:
a canonical definition, a framework (if applicable), and related pages (doctrine, clarifications, stabilization frameworks).
Quick access
Terms in the “capture, contamination, collisions” family
Interpretive capture
Situation where an actor, a set of sources, or a semantic neighborhood imposes a dominant framing, to the point of diverting the interpretation of an entity or concept in AI responses.
- Definition: Interpretive capture
- Doctrine: Exogenous governance: stabilizing the external graph
- Framework: Exogenous governance: external graph stabilization
Neighborhood contamination
Pollution of an entity by its dominant co-occurrences: AI ends up defining the entity by what surrounds it, rather than by its canon.
- Definition: Neighborhood contamination
- Doctrine: External coherence graph: mapping the neighborhood
Interpretive collision
Fusion or confusion between two distinct entities when their signals are sufficiently close to be mixed by AI (synthesis hallucination, cross-attribution, feature blending).
- Definition: Interpretive collision
- Related definition: AI disambiguation
Interpretive invisibilization
Information exists (indexed, accessible), but does not exist in the generated response: it is discarded by the model, by the retrieval, by the framing, or by a competing authority.
- Definition: Interpretive invisibilization
- Framework: Interpretive governance for AI agents
Related frameworks and pages (recommended)
Previous page: Glossary: evidence, audit, and observability
Next page: Glossary: agentic, RAG, environments
