The question
Take interpretive contexts (vision, touch, interoception, memory, action readiness) as a cover. Each context carries local interpretive states, which may or may not agree on overlaps (A7 §10). One answer: an experience is unified when the local states glue into a global section. A7 rejected it: “a system can have unified experience while maintaining incompatible interpretations. A person might simultaneously want to approach and avoid something”, and “the conflict itself can be a unified experience”. Its alternative: “Unity = A realized organization of local interpretations and their interactions”, “a process that maintains, negotiates, and responds to incompatible interpretations.”
A7’s example does not show what it was taken to show. Wanting to approach and avoid the same thing is a conflict, but a conflict is not by itself a gluing obstruction. A global state z = (desire to approach, desire to avoid) holds both desires with no inconsistency, since the two are not incompatible assignments to one common observable. A richer global state space can represent a conflict whose demands cannot be jointly met. So three things must be kept apart:
- Conflicts that glue. Opposed interpretations or demands that one global state represents together. Most felt conflicts may be of this kind.
- Genuine non-gluing. An empirical model with no global section under a cover and a correspondence between contexts that are fixed independently of the verdict, as CLAIM-197 and CLAIM-143 require.
- Phenomenal unity. Whatever makes an experience one.
The question is whether (2) can take part in (3): whether a process whose local interpretations genuinely fail to glue can still be one unified experience. A unified experience of conflict does not answer it, unless the conflict is shown to be of kind (2). The two readings come apart both ways. A7: “A globally consistent assignment could exist in a merely descriptive model without any corresponding unified process. Conversely, a genuinely unified process could contain locally incompatible states.” A6, of Lee-Youngzie et al. (LIT-1352): “compatible local sections in a specified mathematical model need not establish that there exists a conscious subject corresponding to their union. Nor does failure of a global section automatically imply the absence of experience.”
Why it is a question
The record already uses the global-section reading for a self. CLAIM-153 (Active): “A global section, one coherent self, exists exactly when the monodromy Φ around a cycle vanishes.” CLAIM-196, CLAIM-197 and CLAIM-198 place disunity on overlaps: preference reversal “is not yet disunity” until a self-prediction creates an overlap. Those claims are about whether a self holds together across contexts, not about the unity of one experience. If a genuine gluing failure on an overlap, kind (2), can be part of one unified experience, then CLAIM-153’s coherent self and phenomenal unity are different unities. The tag vocabulary already keeps four unities apart (ADR-024), but no claim says where a gluing failure falls among them.
It also bears on CLAIM-412, whose unity is a property of a process, and on Metzinger’s “single functional cluster” (THEORY-063), which is a dynamic criterion, not a sheaf condition.
A7’s decomposition of binding names what is at stake. Functional binding: how distributed processes coordinate. Perspectival binding: when coordinated processes constitute one frame. Phenomenal binding: “Why, or whether, does that organization constitute the unity of conscious experience?” A gluing condition is a candidate for the second, and the question is whether it can be the third. CLAIM-355 already records the exchange of 2026-10-17’s five bindings and says that it “does not solve phenomenal binding”.
What would answer it
Either a showing that the unified experiences offered as cases of non-gluing are, under the right cover, of kind (1) after all, as the approach–avoid case may be; or an empirical model of kind (2), with its cover and correspondence fixed independently, shown to belong to one unified process, together with a formal object, beside the sheaf, for a process that holds non-gluing sections together, with a reason to call it the unity of an experience and not only a functional coupling. Isaac’s dynamical structures (LIT-1361) and the record’s frustration of maintained demands (CLAIM-368) are candidate materials for the second.
Note of 2026-10-22: a candidate answer, and two constraints
A8 proposed the second alternative and gave it a dynamic form: “Unity consists in jointly consequential interpretation, not necessarily in agreement among interpretations.” Incompatibilities δ_ij enter the update of a common frame, z_{t+1} = H(z_t, {j_i}, {δ_ij}, v_t). A system keeps a unified perspective “by preserving unresolved tensions, revising interpretations, reallocating attention, or shifting its evaluative priorities”. A9 found in the record what any formal version of that must respect:
- CLAIM-197 and CLAIM-143. A disagreement is a gluing failure only on an overlap, and only once the correspondence between the contexts’ observations is fixed independently. “We cannot simply assign each neural module a local interpretation and treat disagreements as evidence of contextuality.”
- CLAIM-153’s own scope. It models a self’s consistency across contexts, and A9 noted that it “does not establish a general theory of psychological dissociation”. A9: use sheaf tools “to study the structure of contextual compatibility, not identify global sections with conscious subjects”.
A9 also tied the conflicts to the record’s frustration apparatus. Jointly unrealizable demands form a compatibility complex (CLAIM-347), and its minimal non-faces are the minimal conflicts. That is a candidate for the formal object this question’s second answer lacks, beside the sheaf: an organization of incompatibility, not an absence of it. But infeasible demands and failure of a global section are different mathematical facts. A complex with a non-face says that some demands cannot be met together. It does not say that their local descriptions fail to glue, and a global state can describe an infeasible set of demands consistently.
Note of 2026-10-22: the readings
Lee-Youngzie et al. (NOTE-1059, THEORY-875) formalize binding, not unity, and assume an experience is always unified. They read a sheaf failure two ways, as disunity in binocular rivalry and as context dependence inside one unified experience in the Ebbinghaus illusion, and give no rule for telling the two apart. The reading checked their formal content.
- The worked Alexandroff topology is wrong as printed.
- By the paper’s own gluing definition, rivalry is not a sheaf failure.
- Restriction maps that match trials by label make gluing always succeed, so the test has content only as an empirical consistency radius on measured data.
- “Global section” there means measured data, not an experience or a subject.
The paper gives no support for “unity is a global section”.
A candidate for kind (2). The paper’s Penrose tribar is a presheaf of local interpretations, compatible at each junction, with no global section. Yet the figure is seen as one object. If the cover (the junctions) and the correspondence are accepted as fixed independently, that is a genuine non-gluing model taking part in one unified percept, which is what this question asks about. The inference is the record’s, not the paper’s. “Seen as one” is report, and whether the percept is unified phenomenally or only categorized as one object is open. Its obstruction needs a cycle of three junctions, as CLAIM-153’s does.
Coherence views. Laukkonen et al. (THEORY-882) and IWMT (THEORY-876) take the coherence side. Laukkonen et al. treat incoherence as a loss of awareness. Neither addresses a felt conflict.
Isaac (LIT-1361), whose dynamical structures this question named as material, was not reached that day. It was read on 2026-10-25; see the Note of 2026-10-25 below.
Note of 2026-10-25: Isaac’s dynamical structure
Isaac (LIT-1361), read in NOTE-1163, gives the dynamical structure as a vector field of competing valences and affordances over one experienced space (the hot handle’s invitation and the remembered heat’s repulsion felt together). It has no cover, local sections or gluing condition, and forces in a field superpose, which is globally consistent. So it is material for a kind-(1) reading of felt conflict, one global state holding both sides, not a model of kind (2). Its two co-present levels of colour experience (THEORY-959) are another kind-(1) case. The inference is the record’s. The question is unchanged.
Note of 2026-11-01: Metzinger’s minimal phenomenal experience
The question names Metzinger’s “single functional cluster” (THEORY-063) as a dynamic criterion of unity. His later work, read on 2026-11-01, changes what that criterion can cover. Metzinger 2020 (LIT-1946, NOTE-1498) offers as the simplest consciousness a state with “neither centre nor periphery”. It “lacks the phenomenal quality of singularity or numerical identity”, and he adds “non-oneness” to “non-duality” (p. 13). Metzinger 2024 (LIT-1943, NOTE-1499) models that state as inference at the top level of a hierarchy “without any perceptual inference at the levels below it” (p. 477).
If that account is right, the minimal case has no local interpretations to glue or fail to glue. It is neither kind (1) nor kind (2). The question of phenomenal unity, kind (3), may not arise for it in phenomenal terms. It would be unified only in the thin sense that it is one state of one system. That marks a limit of the sheaf framing: the framing presupposes contexts, and the minimal case has none. It supplies no case of kind (2). The nondual “modes”, in which a world is present without subject/object structure (2024, pp. 470–471), are where kind (2) could be looked for. The pages read describe them only by report. The inference is the record’s. No status, title or warrant changes.