Contribution
Before this chapter, the extended-mind thesis had no official statement from its authors. LIT-097 “never quite state[s] a single official thesis” (§1). After it there is one, and it is stronger than the reading most critics attacked: external entities partly constitute mental states “in virtue of the subject’s sensorimotor interaction with these entities”. The chapter also turns the foreword’s doubt (LIT-421) into a definite position on consciousness: perception and action are not a boundary for cognition, but they are one for consciousness.
Key insight
What made Otto controversial was never the skin. It was that the notebook reaches him only through seeing and writing. Make that the thesis, and the controversy, the strongest objection and the limit on consciousness all fall out of one feature: belief can tolerate a perceptual step between store and control, and consciousness, on this account, cannot.
Assumptions
- The 1998 thesis, as reconstructed. “A subject’s cognitive processes and mental states can be partly constituted by entities that are external to the subject, when those entities play the right role in driving cognitive processes” (§1). “Partly constituted” is read as partial grounding, or as nomological dependence for a property dualist. The body is taken as the key boundary, to separate the thesis from embodied-mind theses.
- Extended-circuit cases are accepted by both sides. Clark’s Diva, whose damaged division circuit is replaced by an external silicon circuit, is extended cognition by the 1998 statement. The chapter takes it that Adams and Aizawa (2008) and Rupert (2007, as the text has it; 2009 in the references) accept such cases. It does not quote them saying so.
- Consciousness correlates with direct availability for global control. This is imported from The Conscious Mind and “On the Search for a Neural Correlate of Consciousness” (§3), with Tye’s “poised” states as a parallel. It is not argued here.
- There are no plausible cases of extended consciousness. The chapter takes this as independent evidence and offers directness to explain it.
- Neutrality on the mind-body problem. Extended physical states need only “nomologically ground (or support)” consciousness for consciousness to count as extended, so a dualist can pose the question (§3).
Key results
- The 1998 thesis is too weak (§1). Following Farkas (2012), it is “near-obviously true in extended circuit cases” and is accepted by the thesis’s main opponents. Sticking with it “would be a little akin to the US declaring victory in Vietnam and going home.”
- Rejected reformulations (§1). Adams and Aizawa’s modal/nonmodal split and the science-fiction/non-science-fiction split both tie the thesis to dates (“the 2017 version of the thesis is false but the 2117 version… will be true”). Block’s quip is reported: false in 1995, true since smartphones. Farkas’s standing-state thesis misses mental rotation, which is not a standing state.
- The proposed thesis (§1). First: constitution “in virtue of the subject’s interacting with these entities via perception and action”. Then, to finesse doubt that Otto really perceives or acts: “in virtue of the subject’s sensorimotor interaction with these entities”. Chalmers calls it “extensionally correct” for paradigm cases and invites Clark to accept it “as the retrospectively official statement”. It counts Otto and Tetris in and Diva out. Brain-computer interfaces and intravenous stimulation go with Diva. Telepathy is a borderline case.
- Five objections answered in one line each (§2). (1) Different psychological profile: “implausible human chauvinism” (cited to Rupert 2009). (2) Derived content: biological memory’s content is derived from conscious states too (Adams and Aizawa 2008). (3) Coupling-constitution fallacy: “confuses the thesis with the argument for the thesis”, which turns on parity (Adams and Aizawa 2008). (4) Not conscious: as mental as standing beliefs is enough (Gertler 2007). (5) Cognitive bloat: reduce it with endorsement-on-retrieval and past-endorsement criteria, accept the rest.
- The perception-action objection (§2). “By far the strongest objection”: if the mental lies between perception and action, the thesis as now stated is false. It is “not a knockdown”, and the reply is the foreword’s: reject the boundary at “some cost of revisionism”. The thesis “also overthrows the hegemony of perception and action.”
- The bandwidth explanation is withdrawn (§3). Vold (2015) is “correct” that vision is very high-bandwidth, so low bandwidth cannot be why perception fails to carry consciousness.
- Directness (§3). Direct access is “not a matter of sheer speed or bandwidth” but “a relatively short and unmediated causal pathway”, and “crucially, direct access requires access that is not mediated by other mental states”. Standing belief needs two-step availability (to consciousness, then to control), which Otto’s perception supplies. Consciousness needs one-step availability, which perception cannot supply. Hence “even if there is extended cognition, there is no extended consciousness.”
- Limits stated by the author. Unconscious perception might escape the mediation argument. Then the pathway is still longer and parity fails, but a conclusive verdict “would require more detailed criteria for directness than I have given here”. The author has “no necessary and sufficient conditions for directness”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | LIT-097 states no single official thesis; its implicit one is constitution by entities playing “the right role in driving cognitive processes” | strong | textual reconstruction by a co-author, §1 |
| C2 | That thesis is too weak: it is satisfied by extended-circuit cases the opponents accept | moderate | Farkas 2012; the claim that Adams, Aizawa and Rupert accept Diva-style cases is asserted, not quoted |
| C3 | The interesting thesis is constitution in virtue of sensorimotor interaction | moderate as a proposal | fits Otto, Tetris and Diva; Clark’s reply to Fodor quoted as tacit support; no necessary-and-sufficient test |
| C4 | LIT-097’s claim that Otto’s access “is not perceptual at all” was a stretch | strong | concession by a co-author, §1, referring to the foreword |
| C5 | The common objections (profile, derived content, coupling-constitution, consciousness, bloat) are weak | weak | one sentence each, §2 |
| C6 | The perception-action boundary is the strongest objection, but not a knockdown | moderate | §2; the reply is revisionism, defended in the foreword |
| C7 | Low bandwidth does not explain the absence of extended consciousness | moderate | concedes Vold 2015: vision is high-bandwidth |
| C8 | Consciousness requires direct availability for global control | assertion here | imported from the author’s earlier work |
| C9 | Perceptually mediated availability is indirect, so there is no extended consciousness via perception and action | informal argument | C8 plus the definition of directness as unmediated by other mental states; the author says it is not a knockdown |
| C10 | Extended circuits could support consciousness | informal argument | directness poses no obstacle where access is not perceptual, §3 |
Concepts
- extended circuit — an external component wired or radio-linked into the brain so that it plays an internal circuit’s role (Diva). Extension here is subpersonal.
- extended circuit thesis — that such a circuit can partly constitute a mental process. The chapter says most opponents accept it.
- perception-action (or sensorimotor) extension — constitution in virtue of the subject’s seeing, reading, writing or pressing. This is the chapter’s “strong and interesting” thesis.
- boundary thesis — that the mental is constituted by processes between perception and action.
- direct availability — access along a short causal path not mediated by other mental states. It is not the same as speed or bandwidth.
Connections
- LIT-097 and NOTE-180. The chapter is the authors’ own revision of the 1998 paper. It answers two of NOTE-180’s limitations. On “is coupling enough for parthood?”, it now reads constitution as partial grounding and says the coupling-constitution objection “confuses the thesis with the argument”. On the perceptual-access reply, NOTE-180 called it “admittedly circular in form”, and the chapter withdraws it.
- The foreword (LIT-421, NOTE-369). The chapter is its sequel: the perception-action objection, the “Necker cube” revisionism and the consciousness question all begin there.
- The critics held here. Adams and Aizawa 2001 (LIT-438) and Rupert 2004 (LIT-422) are not cited; the chapter cites their later books (Adams and Aizawa 2008, Rupert 2009). Objections (1)–(3) of §2 are the later forms of the arguments in those two papers. Clark’s reply (LIT-403) answers the derived-content objection with the same Venn-diagram-and-Martian parity moves that §2 compresses into one line.
- Schwitzgebel (LIT-159, NOTE-131). NOTE-131 records that §6 of LIT-159 answers Andy Clark’s high-bandwidth-synchrony objection, inferred from Clark (2009), and a different objection Chalmers raised in correspondence. This chapter abandons the bandwidth idea (C7), which removes the support of that objection from its own side.
- Chalmers’s consciousness work. The directness criterion is the author’s earlier correlate of consciousness, from The Conscious Mind, which another reader is filing in this record now. It is a correlate. The chapter’s neutrality between grounding and nomological dependence keeps the hard problem out of the extended-mind question altogether.
Bearing on the record
- THEORY-043. The directness criterion is a published
architectural criterion for consciousness that a reader could apply to
both a society of mind and a nation. Because directness requires access
“not mediated by other mental states”, and a nation’s internal
information flow runs through its citizens’ perceptions and actions, the
criterion appears to exclude a nation and spare a mind built of mindless
agents. That is my application, not the chapter’s. It also shows that
the criterion is close to an anti-nesting principle: what disqualifies
the group is that its parts’ mental states sit in the path. The
criterion’s justification is the weakness THEORY-043 names: it is
offered to explain an independently held verdict. So it does not meet
that account’s
promote_when. It is the most promising candidate this record now holds for the computation the account asks for. - THEORY-023 and LIT-056. Direct availability for global control is the global-availability idea that global-workspace indicators formalise. The chapter says nothing about AI systems.
- LIT-138. A BCI falls with the extended circuits, outside the controversial thesis and possibly able to carry consciousness.
- No instruction for machine-learning practice; nothing for the anthology.
Limitations
- Both main theses are proposals. The sensorimotor statement is defended for fit with paradigm cases, and the consciousness argument rests on an imported correlate and on the absence of plausible cases.
- Directness is left undefined. The author says so.
- The opponents’ concession is asserted. The chapter says Adams, Aizawa and Rupert accept extended-circuit cases, but quotes none of them.
- The objections of §2 get one sentence each. In particular, the reply to the coupling-constitution charge assumes that parity, not coupling, carries the argument. NOTE-180 found that parity is stated in LIT-097, not argued.
- Short. About 5,700 words.
Open questions
- Does the directness criterion, applied to an explicit society-of-mind model and an explicit model of a nation, put them on different sides? If so, is “not mediated by other mental states” an anti-nesting principle?
- What becomes of extended belief through unconscious perception, which the author concedes his argument does not rule out?
- Does Clark accept the “retrospectively official statement”? Not checked here.
Corrections
- none to a seeded skim (there was no seed)
- Bibliography on the author’s site. consc.net lists the volume as “Wiley-Blackwell, 2018”. Crossref and the book’s DOI give Oxford University Press, 2019.
- Rupert’s date. The text says “Rupert (2007)” for the extended-circuit concession; the references list Rupert 2009.