| Description: |
Some scientific domains require distributional descriptions not merely as summaries of ignorance, but because the stable regularities to be explained attach to outcome distributions rather than to fully specified trajectories. This paper argues that such cases reveal a structural constraint on any adequate account of how one outcome is fixed from among multiple admissible possibilities. I introduce the Boundary Adequacy Constraint (BAC), a necessary condition on any adequate account of outcome fixation. The BAC requires a law-governed admissible space, boundary-defined outcome distinctions, non-invertibility at the outcome level, explicit scale-relativity, and resistance to reinterpretation as mere ignorance. The framework clarifies when probabilities represent objective constraint-structure rather than epistemic limitation, reposes irreversibility as boundary-localized representational information loss, and provides a criterion for assessing proposed accounts of selection without committing to particular mechanisms or interpretations. |