Christian List (2016), Noûs 52(2), 295–319 (2018 issue; online 28 July 2016) — DOI-10.1111/nous.12162
Key takeaways
- Agency composes and experience, on this argument, does not. A suitably organised collective meets a thin belief–desire definition of agency and has “consciousness as awareness” (access, attention, even sleep and waking), so all of Schwitzgebel’s functional marks can be granted. Whether it also has phenomenal consciousness is a separate question, settled only by a psycho-physical bridge principle.
- Conditional on IIT, the answer is “(close to) nothing”. The argument is that group agents are partitionable at low cost in Φ, not that their parts are minded. Exclusion is offered as a second, independent route.
- The verdict is asserted from three structural reasons. No group’s Φ is computed or even estimated, and List allows that a “biomorphic” science-fiction collective might score higher.
Standing in the record
Filed on 2026-10-02 at the owner’s request, as one of the works bridging Minsky’s Society of Mind (LIT-429, with the memos LIT-399 and LIT-402 and Singh’s exposition LIT-444) to Schwitzgebel’s “If materialism is true, the United States is probably conscious” (LIT-159, read in NOTE-131). NOTE-340 is the close reading of the same day, and it placed the work: Active.
Where it sits on the seam. It sits on the composition side, at the exact point where agency is separated from experience. It accepts that agency is built from minded parts and denies that experience comes with it. Against Schwitzgebel’s claims as NOTE-131 numbers them:
- C3 (a whole of conscious parts can be conscious): contests, for actual group agents only. It does not deny the possibility in principle. Its main argument is about coupling structure, not about the parts being minded.
- C4 (Putnam’s anti-nesting principle is unjustified): neutral. Putnam is not discussed.
- C5 (IIT’s exclusion postulate has absurd consequences): contests by reliance, not by argument. §6.5 invokes exclusion as an extra reason and notes that Schwitzgebel rejects it, but does not answer the one-ballot reductio. The main low-Φ argument does not need exclusion.
Schwitzgebel answered it in his book: The Weirdness of the World (LIT-216; see NOTE-188) adds List 2018 to its IIT footnote and gives the “Max goes to sleep” case against exclusion. No stand-alone reply to List was found.
What it gives the bridge, and why it is filed for it:
- It turns Minsky’s architecture into evidence against experience. List’s first reason is that a group agent is “decomposable into functionally relatively self-contained units” (p. 32). That is the society-of-agents picture. Under IIT, decomposability itself lowers Φ, so the more a whole looks like a society of nearly independent specialists, the less experience the theory gives it. The argument applies to a society of mindless agents too, so it would bear on a Minsky-style mind exactly as on a nation unless its agents are densely recurrently coupled. The paper does not draw that consequence.
- Its third reason is Chalmers’s objection from Schwitzgebel’s §6 (most capacities are the members’ own, not their relations’), quoted from LIT-159. NOTE-131 records Schwitzgebel’s reply, that global-workspace and “fame” models of human consciousness are also subsystem-driven. List does not engage it.
- It is the IIT application that NOTE-131’s open question asks for (“does a formal IIT calculation actually put a polity’s Φ above a brain’s?”), but only in heuristic form, and with the simplified Aaronson Φ rather than IIT 3.0 (n. 17). It does not settle that question. LIT-185 and its reading NOTE-122 are where the record holds IIT’s current formal problems: Barrett et al. argue the measure is undefined for real systems, which cuts against any confident low-Φ verdict for a group as much as against a high one.
Other neighbours. Dennett’s “Real Patterns” (LIT-220) is the stance List’s thermostat remark leans on (p. 6) without endorsing interpretationism. The indicator approach of Butlin et al. (LIT-056) is exposed to List’s worry about global workspace: GWT “simply asserts functionalist conditions for access consciousness and then stipulates” phenomenal ones (p. 24). List grants groups a global workspace in principle, which is the case NOTE-131 says neither LIT-056 nor LIT-159 addresses. Pilgrim et al.’s collective intelligence framework (LIT-046) treats the same groups computationally and says nothing about experience; List’s agency/awareness/experience split is the vocabulary for keeping those apart. The paper restates and extends the normative claim of List and Pettit’s Group Agency (LIT-391, chs 7–8). List says the book asserted an asymmetry between individual and corporate rights without fully defending it, and this paper’s consciousness argument is offered as the missing ground.
It carries no instruction for machine-learning practice, so nothing belongs in the Anthology of the SOTA.