nucleation

TERM-044: agent (at a level of description)

2 min read

Properties6
StatusActive
Version3
Tagsagency
Linedistributed-agency
Used byCLAIM-155 — Willed agency is a kind between agency and selfhood: a control system whose behaviour is governed by a will-organization over several regimes is more than an agent and closer to a self, in the self-governing sense, without thereby being a subject, CLAIM-168 — Identifying agents by their will-organization is circular: will-organization is defined for an agent already given, and the agent-versus-device comparison presupposes an identified system, so neither can produce the factorization it is meant to find, CLAIM-170 — The agents in a system are a factorization of it relative to a level of description and a target; agents in general are identified by adaptation, and willed agents are individuated by their will-organization: a willed agent is the jurisdiction of a will-organization, CLAIM-174 — A candidate factor is an agent at a level when it is one system there, when a goal-directed description beats a device description by more than the Occam cost, and when its policy would adapt if its actions' consequences changed; it is a willed agent when, further, a will-organization over several regimes describes it better than none, and what passes that is the jurisdiction of a will-organization, CLAIM-157 — Relative to a stated target, a combined form is one system while its joint dynamics are nearly decomposable with the form as a block whose gap beats the Occam factor, and it lasts as long as that block persists; the unit is the gap-isolated block, not any member, CLAIM-167 — One-way drive rightly passes the threshold criterion: a metronome and its musician, or a nerve and its muscle, form one system relative to a target that needs both, and the direction of coupling gives the system's internal architecture, not its membership, CLAIM-418 — Possessing, enacting or being shaped by an organizational perspective does not establish an independent agent, and perspectival, agential and institutional boundaries need not coincide: several agents can share a perspective, one agent can act through several, and a perspective can be reproduced with no agent governing it; where a perspective contributes to an agent's activity it can enable and constrain the exercise of agency, and that exercise can reproduce or transform it, while whether agency universally requires such a perspective stays open
RivalsTERM-063 — agent, as an individual that is the active source of modulations of its coupling with its environment and regulates them against norms its own organization sets (Barandiaran, Di Paolo and Rohde)

Graph View

Definition

Fix a level of description and a target. A factor of a system is an agent there when it passes three tests:

  1. System. It is one system (TERM-049, CLAIM-157).
  2. Description. A goal-directed description of it beats a device description by more than the Occam cost of the richer model (LIT-911, priced as in CLAIM-154).
  3. Adaptation. Its policy would change if its actions influenced the world differently (LIT-909).

The test is CLAIM-174. Agency so defined is relative to the level and target, and can be wrong given them.

What it is not

  • Not every control system. A bare thermostat is a control system and fails test 3. With its designers it passes (CASE-044).
  • Not a willed agent. An agent need not have governance over several regimes. A trained policy’s training process is an agent on Kenton et al.’s account, and need not be willed. That is TERM-048.
  • Not a self (TERM-045).
  • Not the will paper’s “agent”. The will paper and the authenticity line use “agent” for an agent supposed, not tested (CLAIM-080: “the account asks how a given agent’s practical efficacy is organized, not what makes something an agent”). On the tiers those agents are willed agents (TERM-048). The older claims are not rewritten. Their sense is recorded here. CLAIM-104’s “an encompassing agent” and “agents at one level” already make agency relative to a level, as this term does.

Note of 2026-10-21: rivals proposed, and a sense this term refines

A2 of the exchange of 2026-10-20 proposed this term with CLAIM-170 as a paper, without citing either (CLAIM-170, note of 2026-10-21).

A11, A12 and A14 proposed filing competing definitions of agent beside this one. One is THEORY-071’s organizational sense, filed as a rival, TERM-063. The other is “agent, as an adaptive policy mechanism”, which is this term’s test 3 taken alone, as THEORY-801 defines it. This term refines that sense and does not rival it. THEORY-583’s cybernetic Self falls at the control-system tier, which this term places below agency (“Not every control system”) and which has no TERM of its own (TERM-048, CLAIM-155). A14 proposed a comparison section on homeostats, trained policies, training processes, organisms and collectives, marking unestablished verdicts as unresolved. That would be a new version.

Backlinks