# Inference–Articulation Distinction **Domain:** Machine Intelligence / Governance / Communication **Doc Type:** Canonical Distinction **Maturity:** Developed ## Definition The **inference–articulation distinction** separates being modeled through behavior from being heard through an intentional semantic act. Telemetry can estimate preference, danger, loyalty or need, but it cannot by itself establish consent, testimony, refusal or political representation. Articulation allows a subject to name context, contradict a classification, introduce a hypothesis and explain why observed behavior does not mean what the model assumes. ## System Function [[wiki/Planetary Administrative Intelligence|Planetary Administrative Intelligence]] requires both modes. Inference supplies scale; articulation supplies semantic error correction. Without a [[wiki/Semantic Feedback Channel|Semantic Feedback Channel]], the system risks confusing its own intervention effects with independent facts about the population. The distinction grounds [[wiki/Model Contestability|Model Contestability]] and limits [[wiki/Operational Ontology|Operational Ontology]]: a model of a person is evidence about that person, not a substitute for the person's addressable voice. ## Corpus Context [[articles/Ambiguity Will Destroy Man and Machine|Ambiguity Will Destroy Man and Machine]] makes the distinction the surviving human interface within machine-mediated governance. [[wiki/The Machine|The Machine]] preserves an imperfect articulation layer; [[wiki/Samaritan|Samaritan]] treats inference as sufficient authority. ## Key Insight **Observation is not dialogue, prediction is not representation, and inference is not consent.**