This theoretical paper explores collective responsibility and accountability in embedded groups, suggesting new frameworks for evaluation.
Paper 25, Responsibility Without a Global Judge, weighted a typed responsibility basis into a per-agent verdict profile and explicitly forwarded the collective case: the non-additive allocation of responsibility over a group's role, authority, contribution, and provenance graphs — the problem of many hands. This paper pays that allocation debt without positing a global allocator: a subsystem that jointly possesses complete access to an institution's structure graphs; an authoritative apportionment rule; infallible counterfactual foresight over every member's options; a uniquely correct share-decomposition; and binding authority to assign it. Collective and institutional responsibility is reconstructed as a specification-relative, profile-valued, challenge-open allocation over a group's organizational structure. A declared allocation specification Sigma_A maps the organizational structure, the field of members' individual verdicts, the applied norm, and the group's corporate-agency status to a typed allocation profile — bearer-indexed allocations together with a case-indexed open-residue register. The absence of a global allocator is made load-bearing: it forces assessor-relativity, a warrant discipline separating attributed from warranted allocation, mandatory corrigibility, public declarability of Sigma_A, and a standing-to-allocate condition distinct from the bare power to issue an allocation. Two results organize the account. The Non-Additivity result (R-A) establishes that, for any admissible Sigma_A responsive at an organizational-interaction matched pair, the warranted collective allocation is not a function of the members' individual-verdict multiset — collective responsibility reduces neither to the sum of individual verdicts (against individualist reduction) nor to nothing (against eliminativist dissolution). The Many-Hands Residue result (R-B), R-A together with a threshold-gap existence premise grounded in distributed foreseeability, establishes that warranted collective answerability can obtain while no individual member is over threshold; every such residue carries an explicit disposition under a trichotomy — allocated, admissibly open, or (only when silently dropped) a genuine void. The culpability-void reading of the artificial-intelligence "responsibility gap" is recast as an admissibility failure, not a metaphysical inevitability. Corporate agency is a derived, non-mysterian standing; blame is non-conserved, so it resists the additive decomposition that conserved liability legitimately admits. Collective responsibility is neither group guilt nor a sum of private verdicts: it is corrigible, typed allocation from inside the history. ALAC Paper 26. First panel of the seventh triptych (Collective Responsibility 26 -> Value/Normativity 27 -> Persons/Personhood 28). Part of the Consistency-Constrained Information History (Akashic Ledger Architecture Cosmology) series.
No takes yet. Share an insight, caveat, or question.
Tomoyuki Uchida (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: