Organizations increasingly govern work through ratings, rankings, recommendations, allocation systems, monitoring tools, and predictive assessments. Yet control may begin before an algorithm issues an explicit instruction. Workers can infer what an opaque system values, imagine how traces will be judged, and alter conduct in advance. This Original Conceptual Theory Article develops a bounded explanation of that pre-command process. It integrates scholarship on algorithmic control, work design, labor-process power, electronic performance monitoring, behavioral visibility, opaque evaluation, surveillance, and worker agency. The article proposes three linked constructs: imagined evaluation, metric anxiety, and pre-algorithmic behavioral control. These constructs organize Recursive Compliance Theory, a proposed feedback model in which inferred expectations shape anticipatory self-discipline; self-discipline changes visible behavioral traces; and perceived algorithmic responses revise subsequent beliefs and conduct. The model also incorporates resistance, collective sensemaking, contestability, and human review as potential interrupting pathways. Its propositions are conditional rather than universal: opacity, evaluative stakes, dependence, legitimacy, task design, labor alternatives, and meaningful appeal should influence whether anticipatory adjustment emerges, intensifies, stabilizes, or reverses. The contribution is conceptual, not empirically validated. Existing evidence supports components of the argument but does not establish the integrated sequence, causal strength, prevalence, welfare effects, or organizational performance consequences. Testing requires temporally ordered, multilevel designs that distinguish inferred rules from actual system logic, employee experience from organizational outcomes, and technological capability from legitimate authority.
