Contribution
The paper sorts the structuralist turn in consciousness science into three programmes and argues that only one of them is in trouble. The metaphysical claim, that “the phenomenal character of an experience is fully determined by its relational properties” (abstract), is attacked by three arguments. The methodological programme (structural constraints on neural correlates) and the epistemological one (inferring others’ experience from aligned similarity structures) are said not to need it. Its most useful tool is the distinction between being in a structure and being the structure.
Key insight
Saying that experiences stand in relations is cheap and widely accepted. Saying that an experience is nothing but its place in those relations makes the relations prior to the experiences. It then follows that adding an experience changes all the others, and that a system with one experience is impossible. Both consequences look false, and a structuralist who accepts them must explain why.
Assumptions
- Local experiences. “Experience” means a local experience within a conscious scene, not the total state, but “my arguments would equally apply” to a global version (pp. 4–5).
- The target. Lyre 2022 is quoted as the explicit statement: “any phenomenal experience is fully individuated by its place in a Q-structure. Phenomenal content and character are relational, not intrinsic properties of sentient subjects” (p. 5, citing Lyre p. 6). Clark 2000 and Tsuchiya & Saigo 2021 are the others.
- Structure built on appearance. The colour structure is “built upon how similar colours appear to human subjects”, not on objective features of colour. Metaphysical structuralism “should be framed in a way that is neutral between, say, representationalism and sense-data theory” (p. 8). Rosenthal’s objectively built quality spaces are set aside as “not a direct target” (fn. 4).
- Metaphysical means necessary. The view is read as true in every possible world, so a single possible counterexample defeats it (p. 11).
- Physicalism is not at issue. The intrinsic property Relational Mary learns “does not need to be a quale in Jackson’s sense, since it can be a property that is fully reducible to the physical” (fn. 5).
Key results
- Three types of structuralism (§2). Methodological: search for neural correlates under the structural similarity constraint of Fink et al. 2021, “Neural substrates of phenomenal types must share the structure governing the phenomenal types they are associated with” (pp. 3–4), or build theories with structural explanantia, as IIT does. Epistemological: use structure to know what others experience, as Kawakita et al. 2025 aligned colour similarity structures across subjects without supervision and failed to align neurotypical with colour-blind ones (p. 4). Metaphysical: “the fact that an experience is the way it is entirely depends on the relational properties that experience has with other experiences” (p. 4). The truth of the epistemological view does not entail the metaphysical one; support would be “at best abductive” (p. 5).
- Three senses of “structural” (§2.3, pp. 5–7). An experience has an internal structure (like an ammonia molecule); it is in a structure (it has relational properties, to actual or possible experiences); or it “just is the structure”, so that “it has only relational properties”. The first two are “fairly uncontroversial”. Only the third informs metaphysical structuralism, and it follows from neither of the others.
- Relational Mary (§3.1, pp. 7–12). Mary has seen every colour but
orange and knows the full intersubjective similarity structure, including
orange’s place in it. P1: if metaphysical structuralism is right she knows
everything about seeing orange without seeing it. P2: she does not know
what it is like until she sees it. The argument is “a specific
application of Newman’s problem” (p. 7). Two replies are answered.
- She lacks phenomenal knowledge of the orange-relations. Then the relations are activated only by “coming into phenomenal contact” with orange, which makes “the relata … metaphysically prior to the relations”. That licenses orange being in a structure, not being one (p. 10).
- She can imagine orange, so learns nothing about orange itself. Metaphysical structuralism needs her to be able to do so necessarily, and an aphantasic Mary cannot. “Relational Charles”, told the similarity profile of a new species, could not picture it (pp. 11–12).
- The novel experience argument (§3.2, pp. 12–15). A system that has
seen red, yellow and green sees blue; its structure changes from R to R’,
so “seeing red when R’ is instantiated should not feel like seeing red
when R was instantiated”. P2: “A novel experience does not change other
experiences in a subject’s phenomenal repertoire.” The same instability
across subjects would undercut the intersubjective comparisons
epistemological structuralism relies on (pp. 12–13). Replies:
- Build the structure on possible experiences. Then modify the system’s sensorium so that it can register new frequencies; its possible repertoire changes and nothing suggests its old colours do (p. 13).
- Experts do report changed experiences. Finer discrimination need not mean changed content; after Dennett 1988 the change may be in judgement. The structuralist then has two options. Either “perceptual judgments, together with other cognitive capacities like attention and memory, are fully constitutive of consciousness” (p. 14), which “aligns metaphysical structuralism with the eliminativist claim that there is nothing more to experience than cognitive capacities (Frankish, 2016)” (p. 15). Or a clean line between perceptual content and judgement must be drawn, which is “both conceptually and methodologically unclear” (p. 15).
- The lonely experience argument (§3.3, pp. 15–17). If relations are prior, a system needs at least one phenomenal relation and so at least two experiences; identity of an experience with itself will not do, since the structures are built on relations sensitive to difference. P2: “It is possible to conceive of a system having only one experience.” It is said to be nomologically possible too: a global-workspace system with one smell detector is conscious of that smell when the sensor fires and of nothing otherwise (pp. 16–17). Footnote 10 offers brain organoids as a further case; footnote 11 puts the burden on the structuralist.
- What survives (§4, pp. 17–18). Methodological and epistemological structuralism need only that experiences have relational properties, “not that they have only relational properties”. “I could correctly infer the travel time between two cities just by looking at their relational properties … without assuming that those cities have no intrinsic properties whatsoever” (p. 18). The cost to the methodological programme is that a theory explaining consciousness only relationally “cannot be comprehensive” (p. 18).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Three structuralisms (methodological, epistemological, metaphysical) are distinct, and the first two do not entail the third | strong as a conceptual distinction | §2, §4 |
| C2 | ”Having”, “being in” and “being” a structure are three senses of “structural”, and only the third supports metaphysical structuralism | strong as a distinction | §2.3 |
| C3 | Relational Mary learns something on first seeing orange, so relational knowledge does not fix character | weak to moderate: valid, with an intuition premise (P2) that structuralists may deny | §3.1 |
| C4 | Metaphysical structuralism implies that a new experience changes all others, which does not happen | moderate on P1; P2 rests on phenomenology and on a judgement/content distinction the paper concedes is hard to draw | §3.2 |
| C5 | Making judgements and cognitive capacities constitutive saves the view only by aligning it with eliminativism | weak: an alignment asserted, not argued in detail | §3.2, pp. 14–15 |
| C6 | A system with a single experience is conceivable and nomologically possible, so metaphysical structuralism is false | weak: conceivability, and a sketched workspace system whose consciousness rests on assuming the workspace theory | §3.3 |
| C7 | Methodological and epistemological structuralism survive without the metaphysical thesis | moderate | §4 |
Concepts
- Methodological structuralism: the structural similarity constraint on neural correlates, or structural explanantia in theory building.
- Epistemological structuralism: using similarity structures to know what other systems experience.
- Metaphysical structuralism: an experience is fully determined by its external relational properties; it “is the structure”.
- Local and global experience: a content within a scene, or the total state of consciousness. A global experience can be related only to other possible global experiences (p. 6).
- Relational Mary, Relational Charles: knowledge-argument variants in which all relational facts and all other experiences are known.
- Lonely experience system: a system able to have just one experience.
Connections
The paper builds on Jackson’s knowledge argument and its replies (Lewis, Nemirow, Conee, Dennett), on Dennett’s “Quining qualia” for the judgement/content point, and on Fodor & Lepore on holism. Its targets are Lyre 2022 (LIT-1344, NOTE-1060), Clark 2000 and Tsuchiya & Saigo 2021. It credits Kob 2023 with the point that structuralist claims about the nature of consciousness are “largely orthogonal” to the method’s fruitfulness (p. 17). Aldé 2026 (LIT-1348, NOTE-1055) cites this paper as a critique that rests on intuition, which his own arguments are meant to avoid. Isaac 2025 (LIT-1361) is not cited by Negro.
Bearing on the record
- CLAIM-412 (constitution by interpretive organization). This is the
objection the claim already lists, and the reading sharpens it in two
places.
- Relational Mary is the realization problem A4 drew from it. The claim’s reply, realized and not described organization, meets it only in part. Negro’s Mary does not merely possess a description: she has all the other colour experiences realized. His argument is that the relations to orange can be realized only once orange is, which makes the relatum prior to its relations. A constitutive thesis that makes the experience the organization owes an answer to that ordering.
- Making judgement constitutive. The exchange’s note said Negro “considers making the relational organization of judgments constitutive”. He does: it is the first option in reply to the novel experience argument, and he says it aligns the view with eliminativism (Frankish), “at the cost of denying the existence of phenomenal properties as independent of cognitive capacities” (p. 15). The claim differs from that option in its terms (TERM-073: interpretation is organization, not judgement, and mostly not reportable) and in its stated realism (“It does not say that experience is unreal”). But the claim does make phenomenal properties not independent of the organization, and Negro would read that as his first option. The claim must either accept that experience is not independent of interpretive organization and argue that this is not eliminativism, or say why its organization is not a cognitive capacity in Negro’s sense. The pressure is the same as THEORY-633’s, from the other side.
- The lonely experience argument asks whether interpretation, which requires “a distinction” (TERM-073, part 1), can occur with one experience. Negro’s smell system has only on and off. Whether distinguishing presence from absence suffices is a question the claim has not asked.
- Novel experience. The claim’s relations are causal and temporal as well as similarities, so it need not say every new experience changes all others. It need not say they cannot either. Whether a change in the organization changes old experiences is a prediction the claim could make and test.
- CLAIM-411 and QUESTION-254. Negro’s lonely-experience system is a sufficiency case in reverse: a minimal organization that he takes to be conscious only because a workspace theory says so. It shows how a verdict on a minimal system can be borrowed from a theory, which QUESTION-254 says an answer must not do.
- CLAIM-136 (relata without primitive identity). Negro’s arguments are about phenomenal character, not about the identity of relata, and the claim is about communicative acts. Two transfer by analogy. His three-sense distinction maps onto the claim’s options: non-eliminative realism says that relata are in a structure; the strict version says their identity is given only by it; neither says that the relata are the structure. And his novel experience argument is the holism objection to the strict version: if a token’s identity is only what its relations differentiate, adding a participant re-individuates every other token. CLAIM-136 does not yet say whether it accepts that. The weak version, numerical plurality without differentiation, is not touched, and Negro’s reply to the first objection to Relational Mary, that relata are then “metaphysically prior to the relations”, is close to CLAIM-134’s charge against the weak version.
- THEORY-577. Negro’s appeal to Dennett, that a reported change in an experience may be a change in judgement, is the same separation of judgements from what they are about that THEORY-577 uses.
- THEORY warranted. THEORY-868 states the paper’s thesis: the structuralist evidence supports “in a structure”, not “is the structure”. It is the objection to THEORY-880.
- No instruction for machine-learning practice; nothing for the anthology.
Limitations
- Each argument turns on an intuition or conceivability premise: what Mary learns, that old experiences do not change, that a lonely experience is conceivable. Aldé (2026) says this is why such critiques “have failed to convince numerous scholars”. Lyre accepts the holism consequence outright (dichromats experience all colours differently), so the novel experience argument’s P2 is exactly what a structuralist denies.
- The nomological case for a lonely experience assumes the global workspace theory, which the author himself thinks is probably incorrect (fn. 9).
- The eliminativism charge against making judgement constitutive is one paragraph long. It is not shown that every such view is eliminativist.
- The arguments target the view that an experience is its external relations. A view on which character is fixed by internal and external organization together, or by causal and not similarity relations, is not addressed.
- The paper does not consider whether relations to merely possible experiences avoid the novel-experience consequence beyond the one reply it answers.
Open questions
- Is there any test of P2 of the novel experience argument? Similarity judgements among old colours before and after a new colour becomes available (as in the laser-stimulation “olo” studies Aldé cites) would bear on it.
- Can the line between perceptual content and judgement be drawn, as the second option needs?
- What would a minimal experiencing system look like on a theory not chosen to make it conscious?