We examine the continuing use of subjective workload responses to index an operator's state, either by themselves or as part of a collective suite of measurements. Lack of convergence of subjective scales with physiological and performance-based measures calls into question whether there is any unitary workload construct that underpins conscious experience, physiological state and the individual's profile of task-related performance. We examine philosophical and measurement perspectives on the divergence problem, and we consider three possible solutions. First, difficulties in reliable and valid measurement of workload may contribute to divergence but do not fully explain it. Second, workload may be treated operationally: use of specific measures is justified by demonstrating their pragmatic utility in predicting important outcomes. Third, further efforts may be made to develop representational workload measurements that correspond to real empirical phenomena. Application of formal standards for test validity can identify multiple latent constructs supporting subjective workload, including those defining self-regulation in performance contexts. Physiological and performance-based assessments may define additional, distinct constructs. A resolution of the diversity issue is crucial for ergonomics since the invalid application of workload measurement will threaten exposed operators as well as many others who are served by the complex technological systems they control. Relevance to human factors/Relevance to ergonomics theory Workload assessment is important for multiple applications of ergonomics, but subjective and objective workload measures often diverge. Current ergonomics theory fails to explain this divergence. This article provides two possible solutions to this challenge. First, specific workload measures may be treated as operational variables that predict important real-world outcomes. This approach eschews theory in favor of pragmatic utility. It is compatible with emerging "big data" approaches to performance prediction. Second, subjective measures ARTICLE HISTORY