Contribution
The authors’ own clarifications of Group Agency (LIT-391): which of the book’s claims are modest possibility claims, which conditions can be relaxed, and what the book holds about the responsibility and rights of group agents. For this record it is the nearest lawful view of the book’s Parts II and III.
Key insight
A group’s beliefs need its members’ acceptance, but they cannot be read off it. Knowing which members’ acceptance counts, and that the structure lets the putative belief play the role of belief, already requires the group-level view, so “it is the group belief that explains the role-based member acceptance, not the role-based member acceptance that explains the group belief” (p. 13).
Assumptions
- Functionalism about agency and attitudes, including group emotions if there are any (p. 5).
- An acceptable account of joint intention is presupposed, not supplied (p. 13).
- Normative individualism, adopted in the book “as just a plausible assumption” (p. 17).
- The existence of group agents: the book’s argument (GA 75, quoted p. 14) omits the empirical premise that collections do form and act on robustly rational bodies of attitude, which it “mostly take[s] for granted”.
Key results
- Gaus (pp. 1–5). The premise-based procedure is “merely a particularly simple – but also rather inflexible – instance” (p. 1). Ecological rationality is classical rationality relative to environment (p. 2). Robust group rationality may be weakened: domain restrictions where members’ views are cohesive, incompleteness where tolerable (pp. 3–4). Proposition-wise (“independence”) rules remain “extremely limited” even then (p. 4). Only a shared logic, not a shared structure of premises, is assumed (pp. 4–5). Group emotions need not be simple functions of members’ emotions; some corporations “may exhibit certain behavioural patterns usually associated with psychopathy” (p. 5, citing Bakan).
- Cariani (pp. 5–10). Democratization, decomposition and decentralization are ideal types; chapter 4 gives possibility results (quoting GA 102–103). Units of epistemic comparison are whole group agents with a given member reliability profile. Reliability is reported as positive and negative pairs because an always-safe medical panel could score high on an average (p. 8). An expert panel may do best with a supermajority rule that leaves many propositions undecided (p. 8).
- Sylvan (pp. 10–14). Non-redundancy is a matter of degree, and the theory lies towards the non-redundant end (citing GA 77–78). Workable group agents “allow any proposition, however long endorsed, to be rejected on grounds of inconsistency” and assign subgroups to issues (pp. 11–12). Executive members must act on the group’s propositions but “need not personally assent to it” (p. 12).
- Briggs (pp. 14–17). The book’s verdict, quoted from GA 182: hold group agents “responsible in their own right”, with “restricted rights as compared with individuals”. The book “do[es] not pay much attention to the range of responsibilities” (p. 15). Its answer to the objection that members, not groups, have control relies on Jackson and Pettit’s programme explanation; Briggs proposes List and Menzies’ difference-making instead, which the authors say they were aware of and “didn’t develop” (p. 15). Personhood is performative: a group agent can be “party to a system of accepted convention, such as a system of law” (GA 173, quoted p. 16). They are agnostic about hedonic group states (p. 16) and defend normative individualism tentatively: “individual agents are the producers of social arrangements, corporate agents the precipitates” (p. 17).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | The premise-based procedure is illustrative, not the book’s recommended design | strong | authors’ statement of their own intent, with the book’s pages cited |
| C2 | Relaxing robust group rationality leaves proposition-wise rules extremely limited | moderate | asserted with reference to the aggregation literature; not proved here |
| C3 | Group attitudes cannot be readily reduced to role-based member acceptance | moderate | argued against Sylvan |
| C4 | Group agents are fit to be held responsible in their own right | moderate | the book’s argument, quoted; not restated here |
| C5 | Group agents should have restricted rights, on normative individualism | weak | normative individualism is an assumption; the circularity argument is offered tentatively |
| C6 | Group agents exist | weak | the empirical premise is taken for granted, with anecdotal support |
Concepts
- robust group rationality — universal domain plus collective rationality (consistency, completeness); here its strength may be “adjusted from case to case” (p. 4).
- ecological rationality — rationality of a procedure in the environments to which it is adapted (p. 2).
- non-redundant realism — group attitudes are real, and no accessible recipe from individual facts makes the group-level posit dispensable (pp. 10–14).
- role-based acceptance — members treating a proposition as true for group purposes, by assenting in the group’s name or acting on it in an executive role (p. 12).
- normative individualism — “something is good only if it is good for individual human or, more generally, sentient beings” (GA 182, quoted p. 16).
Connections
It defends Group Agency (LIT-391, unread). The formal core it presupposes is read in LIT-705, and its point that robustness may be relaxed qualifies the conditions stated there. Briggs’s proposal draws on List and Menzies (LIT-722; THEORY-133). Pettit’s chapter LIT-718 states the fitness-to-be-held-responsible argument that the book’s chapter 7, unread, presumably contains; these replies quote only its verdict.
Bearing on the record
It is the source of no account here; it adds to two.
- THEORY-127. The authors say that weakening universal domain or
completeness complicates the formal arguments but that rules satisfying
proposition-wise supervenience remain “extremely limited” and exclude
“some of the most natural methods of aggregation” (p. 4). That supports
the claim beyond the strict conditions its title states, but on the
authors’ word, not by proof. The account’s
promote_whenis unchanged. The replies also confirm that the premise-based procedure is one illustration, so THEORY-127 should not be read as recommending it. - THEORY-139. These replies give the book’s own verdict, quoted from GA 182: group agents are fit to be held responsible “in their own right”. They add three things the account does not yet say. Fitness is assigned “only in the domain where it is equipped to form and act on attitudes” (p. 16). The book does not address the range of a group’s responsibilities (p. 15). And the control objection has two answers in the literature, programme explanation in the book and List and Menzies’ realization-insensitivity in Briggs’s proposal (THEORY-133). The concession that existence is taken for granted, and the call for empirical study of “the actual dynamics of group reasoning” (p. 15, quoting Briggs), match that account’s note that its empirical premise is asserted, not shown.
For the owner’s essay:
- §5.3. The quoted claim that the agential perspective lets us “see higher-level regularities” and respond “in abstraction from much of what happens at the lower level” (GA 6, quoted p. 10) is the authors’ version of what an agentive explanation adds.
- §7.1. Executive members may act on a group belief they do not hold (p. 12), so corporate endorsement is not member endorsement.
- §8.3. Workable group agents can reject “any proposition, however long endorsed” on grounds of inconsistency (p. 12), a structural condition for recursive revision.
- §8, on psychopathy (the record’s link). The authors grant that some corporations show patterns “usually associated with psychopathy” (p. 5). Wolf’s lecture (THEORY-147) separates accountability from fitness for reactive attitudes in exactly the psychopath’s case. Neither text mentions the other.
Limitations
Replies, not a treatise: arguments are compressed and often point to the book. The commentaries are known only through the replies. The book’s pages are quoted, not read. Several claims rest on the aggregation literature, cited and not reproduced.
Open questions
What range of responsibilities a group agent should bear, which the authors say the book leaves open. Whether the “extremely limited” possibilities under relaxed conditions include any rule a real organisation uses.