Why Reference Must Be Fixed Semantic Infrastructure as a Condition of Empirical Theory

This paper establishes reference fixation as a necessary precondition for empirical theory. It shows that empirical claims, predictions, and confirmations presuppose a stable semantic infrastructure in which terms continue to refer to the same targets across admissible redescriptions. When reference is allowed to drift, empirical success becomes uninformative and theoretical content collapses. Within a closed admissibility framework, reference is not determined by usage, success, or pragmatic convenience. It is licensed only where standing can be preserved across changes in description, scale, or representation. Attempts to treat reference as flexible, negotiable, or retrospectively adjustable introduce silent redescription, allowing failures of fit to be masked as interpretive updates rather than recognized as standing failure. The analysis is eliminative. No semantic theory is proposed, and no method for fixing reference is offered. Instead, the paper establishes a boundary condition: reference must be fixed for empirical discourse to be admissible at all. Where fixation fails, claims do not become approximate, theory-laden, or context-sensitive—they become undefined. By enforcing reference fixation as infrastructure rather than interpretation, the paper dissolves debates over semantic holism, theory-ladenness, and reference under paradigm change. It closes these debates by showing that their persistence depends on allowing reference to drift in ways that undermine the possibility of empirical theory itself.