This framework identifies the mandate gap in governance, highlighting its implications for authority and autonomy.
Modern governance of autonomous systems answers, increasingly well, the questions that can be settled at the act itself. It has almost no way to answer one that can't: whether the party who issued the instruction held legitimate authority to issue it, correctly scoped and valid at the moment it mattered. This paper names that missing property the mandate gap, and argues it is a distinct object of governance, separable from identity, provenance, and authorization, with its own failure mode. The paper proposes no mechanism. It separates four governance gaps the field has treated as one, shows that two are substantially addressed in current practice, that a third is now drawing serious work but is not yet closed, and that the fourth, the mandate gap, is addressed by none of that work, and it demonstrates that the mandate gap reduces to none of the adjacent properties it is routinely mistaken for. Each of the prior gaps addresses a derivative of the transaction: whether it occurred, whether it was permitted, whether it was correct, whether it can be proven. Each terminates at the act. The mandate gap does not. Authority resolves upward rather than inward, and every delegation introduces a further link that must be established independently and does not close at the agent. The relation is N+1 rather than a further derivative, which is why the gap has been tolerable at shallow delegation and will not remain so as agentic delegation deepens. Drawing on four independent bodies of work, in physiology, in the theory of promises, in institutional design, and in criminology, the paper reconstructs five conditions a legitimate agentic commitment has always had to meet and offers them as a test rather than a design. The argument is that the discipline skipped a step it already knew how to take, that the step is centuries old in other domains, and that naming it is the precondition for building it. The aim is a vocabulary and a research agenda, not a solution to defend.
No takes yet. Share an insight, caveat, or question.
Justin J. Kaplan (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: