Reference Drift in Organizations: Shared Reference as a Maintained State, and What Speed Does to It

Organizational transformations routinely fail despite documented agreement. This paper proposes an explanation located not in commitment or communication skill but in the erosion of shared reference: the sufficiently common frame through which people across an organization interpret priorities, reality, meaning, trade-offs, and next sensible actions. The paper makes three claims. First, shared reference is best understood as a maintained operating state rather than an asset: it decays without recurring maintenance, and organizational acceleration removes precisely the occasions on which that maintenance used to occur as a by-product of slower work. Second, the resulting erosion — reference drift — proceeds through six nameable mechanisms: semantic inflation, interpretive sedimentation, cascade attenuation, novelty overload, social disconnection, and AI-amplified divergence. Third, because people systematically overestimate how well their frame is shared, assessment cannot rely on agreement items; it must compare revealed classification behavior with espoused aspiration. The paper positions these claims against common ground theory, sensemaking, strategic consensus research, and boundary objects; presents the framework at concept level; illustrates it with four documented reference failures; and closes with a research agenda that names the open validation questions.