NOTE-1060: Lyre — Neurophenomenal structuralism

Contribution

The paper states neurophenomenal structuralism as an ontology, where its companion paper (Fink, Kob & Lyre 2021) had stated it as a method. Every specific phenomenal character is said to be fixed by its place in the structure of a quality space, and that structure is said to be mirrored in neural maps. It then follows the thesis out: to the Newman problem, to reduction, to structural representation, to holism and to qualia inversion. It is an agenda. The author says most of its topics “deserve a much more detailed consideration” (Conclusion).

Key insight

Only differences cross the boundary between the world and the nervous system, so the brain works on relations, and so does what it makes experienced. An orange experience is orange because of where it sits among all possible colour experiences, between yellow and red, and nothing intrinsic is added. Whether the state is conscious at all is a separate question, answered by some general mechanism the paper does not choose.

Assumptions

• Change detection. “Neural sensory systems work by change detection, and only relational properties can cross the 3N boundary” (Conclusion). The 3N boundary is the neural/non-neural interface, inward transduction and outward motor transformation. Footnote 3 likens it to a Markov blanket.
• Two assumptions of the programme (Introduction; restated in “Structure, homomorphisms, and structural individuation”): “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.” And: “Q-structure is mirrored in N-structure. The mirroring has to be understood in terms of structural similarity, more precisely, in terms of a surjective homomorphism from N to Q.”
• Specific, not generic. It “does not tell us when a mental state is conscious in general” (Introduction). Phenomenal character is “experienced content”, a composite of the neural correlate of content and a general neural correlate of consciousness, after Marvan and Polák. The general mechanism is “at best system-intrinsic, not state-intrinsic”, so character is “as much relational as content” (“Phenomenality, content, and characters”).
• Structure is a set of relations over relata that “are not specified otherwise than by the relations in which they stand” (Introduction). “Leibnizian individuality and intrinsicality must be given up.”
• Metaphysically necessary supervenience. Asserted, not argued: “There is, I admit, no argument for upholding Q–N supervenience with metaphysical necessity … other than the request to adhere to the overall scientific, more properly physicalist, spirit” (“Reductive physicalism”).
• Colour as the case. Only colour is worked through. Interoception and the body are set aside “for reasons of space” (“The neural/non-neural interface”).

Key results

All results are arguments. By section:

• The Newman problem of the brain. If the brain had relations only, it could know nothing of the world but cardinality, which is Newman’s objection to Russell. Lyre takes Russell’s reply to Newman: we are “directly acquainted” with some spatiotemporal relations. Spatial separation on the skin or retina and temporal order are carried over into the separation and order of neural activity, giving “spatiotemporal grounding of N in W via perception”. Sensory intersections (seeing a surface and feeling it warm) spread that grounding through the whole system. Footnote 6: “it is the nature of spatiotemporal relations, and only their nature, that can pass the 3N boundary.”
• Structural individuation. The nature of the relations is fixed and “the relata are determined up to isomorphy only”. A structure with no symmetries, such as three relata at 3, 4 and 5 m from each other, makes them “maximally structurally individuated”, and still “the structural individuation does not amount to the attribution of intrinsic properties to the relata” (“Structure, homomorphisms, and structural individuation”).
• Quality spaces. Discriminability is the root relation, and similarity is its dual, after Clark: “similarity is naught but a failure to discriminate”. Just-noticeable differences give units and similarity ratings give order, from which the colour solid is built (“Quality spaces, Q-structure, and psychophysics”).
• Neural maps. Self-organizing (Kohonen) maps place neighbours in feature space next to each other in the wiring. A map from cone opponency onto cortex “mirrors the structure of color space”; Brouwer & Heeger 2009 found colour-space similarity structure in V4, which FKL take as a candidate structural correlate. Plasticity and multiple realizability make the map many-to-one, hence “surjective homomorphism” (“N-structure”; “Empirical evidence and consciousness studies”).
• Method. Ask “how a given stimulus relates to other stimuli”, not whether it is conscious. Similarity ratings and representational similarity analysis are the tools (“Empirical evidence and consciousness studies”).
• Not structural realism. “In particular, neurophenomenal structuralism is no realism about the phenomenal.” Q-structure is mirrored in N-structure, the neural maps bear the content, and an exclusion argument gives all causal work to N (“Structural realism”).
• Reductive physicalism. Shapiro’s dilemma removes the multiple realizability argument: realizers that mirror Q-structure share the causally relevant relational properties, so “structurally construed Q-types can a fortiori be reduced to N-types”. With necessary supervenience, a “tightened exclusion argument” follows: “the Q-structure plays no causal role at all; all causal work is taken over by the N-structure” (“Reductive physicalism”).
• Structural representation. Neural maps are a third type of structural representation, of feature structure, beside spatial and temporal ones. They mirror “abstract (rather than concrete spatiotemporal) W-structure” by partial homomorphism, so they refer to the world only indirectly. Exploitability makes the programme “pragmatic or instrumentalist”, and content is “instrumental at best” (“Structural representation”; “Reductive physicalism”).
• Holism. “Think of the Q-structure as an elastic web.” Any change can change the whole. Its predictions: red-green dichromats “will experience all colors differently”; ageing listeners do not notice that what they now hear as very high was once only high, because “what changes are not single experiences, but the web” (“Neurophenomenal holism”).
• Qualia inversion. Inversion needs a non-trivial automorphism of colour space. Actual inversion is ruled out because human colour space is “highly asymmetric”. Nomological inversion is ruled out by sensory intersections: “dark-colored objects absorb heat better than light-colored objects”. Metaphysical inversion with quiddistic qualia “cannot be ruled out”, and “certainly cannot be captured by any structuralist conception”. The programme “upholds the identity of Q-space isomorphs” (“Qualia inversion and (finally) the hard problem”).
• The hard problem. “what-it’s-like character reduces to holistic phenomenal content (which in turn reduces to neural vehicle representations)”. Since it addresses only specific consciousness, “it can never fully solve the hard problem, but may contribute to its solution” (Conclusion).
• Footnote 21, IIT. IIT “is intrinsicalism about phenomenality, and thus the opposite of structuralism”, despite its qualia space.

Claims

id	claim	strength	support
C1	Only relational properties of stimuli cross into the nervous system; neural codes are difference codes	moderate as a general principle; stated, not argued from data	”Change detection and relational coding”
C2	Spatiotemporal relations, and only they, carry their nature across the boundary, which answers the Newman problem for the brain	weak: an argument from Russell’s reply, with examples	”The Newman problem of the brain”, fn. 6
C3	Each specific phenomenal character is fully individuated by its place in a quality space	an assumption of the programme, motivated, not shown	Introduction; “Quality spaces”
C4	Q-structure is mirrored by a surjective homomorphism from neural maps	weak: one cited imaging result for colour (V4)	“N-structure”; Brouwer & Heeger 2009
C5	With necessary supervenience, structural mirroring yields type-reduction and causal exclusion of Q	conditional on a premise the author says he cannot argue for	”Reductive physicalism”
C6	Phenomenal structuralism is holistic: a dichromat experiences every colour differently from a trichromat	a prediction, untested here	”Neurophenomenal holism”
C7	Spectrum inversion is actually and nomologically impossible, though not metaphysically	weak to moderate: asymmetry of colour space for the first; an argument from sensory intersections for the second	”Qualia inversion”
C8	Neurophenomenal structuralism is not a structural realism about the phenomenal	definitional	”Structural realism”

Concepts

• W-, Q-, N-domain: worldly stimuli, phenomenal states, neural states. A B-domain, the body, sits between N and W and is set aside.
• 3N boundary: the neural/non-neural interface, “boundary of neural transduction” inward and “boundary of neuromotor transformation” outward.
• Q-structure, Q-space: the structure of a quality space, the totality of discriminations a sensory system can make.
• Structural individuation: individuating relata only by the relations they stand in, with the nature of the relations fixed. Relata are then fixed up to isomorphism.
• Surjective homomorphism (epimorphism): a structure-preserving, many-to-one map, here from N onto Q.
• Type-1, -2, -3 structural representations: of spatial, temporal and feature structure. Neural feature maps are type 3.
• Neurophenomenal holism: each experience’s identity depends on the whole Q-structure, so local change propagates.

Connections

It builds on Fink, Kob & Lyre 2021 and, most of all, on Austen Clark’s quality-space theory (1993, 2000). It takes Isaac (2013, 2014) for structural representation and for structural realism about colour qualities, which it calls “otherwise quite similar” but about the world–experience relation where this paper is about the neural–experience relation. It takes the Newman objection from Newman 1928 and the reply from Russell’s letter. It puts itself beside Malach 2021 and Tsuchiya & Saigo 2021 among structuralist neuroscience. Negro 2025 (LIT-1346) and Aldé 2026 (LIT-1348) both quote its first assumption as the explicit statement of the view they attack. Isaac 2025 (LIT-1361, read 2026-10-25 in NOTE-1163) cites this paper among the structuralist programmes it joins, and denies that quality-space geometry alone generates meaning. One cross-reference slips: the qualia section places neurophenomenal holism in the section “Structural realism”, where the section “Neurophenomenal holism” is meant.

Bearing on the record

• CLAIM-412 (constitution by interpretive organization). Lyre’s thesis and the claim share a shape: character is fixed by organization, not by an intrinsic quale. They differ in three ways that matter. Lyre’s relations are discriminations, a similarity structure; the claim’s are interpretive, with context and consequence (TERM-073). Lyre makes structure fix only which experience a conscious state is, and leaves whether it is conscious to a separate general mechanism; the claim makes organization constitute occurrence as well (TERM-072, A11: “occurrence, character, and unity”). And Lyre takes the reductive and instrumentalist route, with content “instrumental at best”. So Lyre is the nearest held statement of the claim’s character half and no support for its occurrence half. The comparison is mine.
• CLAIM-411 and QUESTION-254. Lyre’s split between neural correlates of content and a general correlate of consciousness is the same gap these entries name, conceded from the structuralist side: structure alone does not say when a state is conscious. He declines to fill it, listing workspace, recurrent processing, higher-order theories, synchrony and critical-brain views as compatible.
• CLAIM-136 (communicative relata without primitive identity). Lyre explicitly denies being a structural realist, and his relata are experiences, not communicative acts. But his account of structural individuation is the strict version of CLAIM-136 in another domain: relata “are not specified otherwise than by the relations in which they stand”, are fixed “up to isomorphy only”, and gain no intrinsic properties even when maximally individuated. His treatment of inversion shows what the strict version costs. Where colour space had a symmetry, the structuralist must identify the symmetric experiences (“the identity of Q-space isomorphs”), and Lyre avoids that only by appeal to the empirical asymmetry of colour space. CLAIM-136’s weak version, numerical plurality without differentiation, is not available to him, since two indiscernible experiences would be one experience. Lyre also restates non-eliminative moderate OSR as “there are relations and relata, but … there is nothing more to the relata than the relations in which they stand”, which matches the claim’s own description of it.
• CLAIM-138 (Newman and independently specified relations). Lyre’s reply to the Newman problem is that “we must specify the nature of the relations”, spatiotemporal ones, by acquaintance. That is the remedy CLAIM-138 says concedes non-structural input. Lyre does not see a cost; he makes spatiotemporal relations primitive and carried across. A further instance for CLAIM-138’s parallel, outside the communicative domain.
• THEORY-095 (Inverted Earth). Lyre names Block’s inverted earth as an “outlandish” case and replies only to inversion within colour space. His structuralism is internal in THEORY-095’s sense, Q mirrored in N, so it is one of the accounts THEORY-095 leaves open.
• THEORY-633 (illusionism). Lyre is a realist about specific character and reductive about it. He calls intrinsic qualia “mysterious”, not illusory, so he is not an illusionist, but his “Qualitative properties seem to be intrinsic properties, but they are not” (quoting Clark) is the move THEORY-633 generalizes.
• THEORY warranted. The record holds no statement of phenomenal structuralism as a positive thesis. THEORY-880 states Lyre’s, with its holism and inversion consequences, so that the two objections read in the same batch (THEORY-868, THEORY-867) have a target.
• No instruction for machine-learning practice; nothing for the anthology.

Limitations

• Programmatic. The author calls it “largely of programmatic nature” and offers “neither a clear proof for structuralism nor a knockout argument against intrinsicalism”.
• The reduction rests on metaphysically necessary supervenience, which the author says he has no argument for.
• The empirical support for the neural half is one imaging study of V4, cited, and representational similarity analysis in general, which footnote 19 says gives only “proxy vehicles”.
• Only colour is worked through. Interoception, the body and other modalities are named as future work.
• The holism predictions (dichromats, ageing hearing) are offered without data, and Negro’s novel experience argument (NOTE-1054) turns exactly this consequence against the view.
• The nomological case against inversion turns on a heat-absorption correlation between colour and temperature experience. That ties colour experience to other experiences; it does not show that no symmetry of the whole cross-modal structure exists.

Open questions

• Which general mechanism makes a structurally individuated content experienced? The paper leaves it open and says it is compatible with many.
• Does the holism prediction hold? A test would compare similarity structures for colours both groups discriminate, in dichromats and trichromats, and ask whether the shared part differs.
• If spatiotemporal relations are carried across the boundary by their nature, is the view still structuralist about those relations, or does it rest on something non-structural, as CLAIM-138 would say?
• Can the unique hues and the opponent structure of colour be read off the geometry of the colour space? Aldé (NOTE-1055) argues they cannot.