# Statistical Compression **Domain:** Probability / Modeling / Epistemology **Doc Type:** Canonical Concept Node **Maturity:** Developed ## Definition **Statistical compression** is the representation of causally dense trajectories through distributions, classes or summary variables that discard much of the path history producing them. Compression is necessary for finite observers and useful models. The danger begins when a compressed category is mistaken for the complete ontology of the subject. ## Predictive Governance A score or bin can make a population administratively legible while erasing the environmental and institutional causes embedded in individual trajectories. [[wiki/Score Separability|Score Separability]] limits migration of the compressed judgment, while [[wiki/Epistemic Status Separation|Epistemic Status Separation]] prevents a forecast from masquerading as identity. [[wiki/Peak Person|Peak Person]] is a harmful compression when expected yield replaces unfinished continuity. [[wiki/Operational Ontology|Operational Ontology]] determines which compressed representations acquire institutional force. ## Principal Source [[articles/Peak Person and the Predicaments of Prediction|Peak Person and the Predicaments of Prediction]] develops the Galton board as an allegory for civilizational sorting. ## Key Insight **A useful summary becomes an injustice when the institution governs the summary as though it were the whole person.**