The Institutional Production of Causation: Predictive Classification and the Governance of Human Behaviour

Causation is increasingly produced prospectively through institutional systems rather than reconstructed retrospectively from events. Contemporary predictive systems generate causal environments through algorithmic classification, risk scoring, and procedural governance that structure behaviour, allocate resources, and determine outcomes before events occur. This article demonstrates that classification functions as causal infrastructure by enabling governance through probabilistic association without mechanistic explanation. Historical evidence reveals that classification-based governance has long operated through procedural consistency rather than causal understanding, with algorithmic systems automating and scaling this existing institutional logic. The analysis examines how predictive classification reshapes human behaviour by displacing decision-making, restructuring agency, diffusing responsibility, and producing stratified opportunity structures. These findings require reconceptualising causation itself: institutional systems do not measure pre-existing causal relationships but generate causal environments through systematic classification that operates prospectively to structure the conditions under which behaviour emerges. This preprint is part of an ongoing peer review process. This version is released as a stable preprint and is not intended for incremental updates. Substantive revisions will be incorporated into the peer-reviewed publication. Please cite the peer-reviewed version when available. Part of the TLIM Research Program (Tri-Layer Integrated Model). For the foundational definition of the TLIM framework see:https://doi.org/10.5281/zenodo.18667835 Program website:https://tlimresearch.org Landing page:https://hirokitamba-ops.github.io/paper2.htmlhttps://hirokitamba-ops.github.io/