Board status — ratified 2026-07-08 (supersedes the roll-up statuses below): DeepRoot — Granularity/cost-floor is CERTIFIED-IRREDUCIBLE · RESOLVED +0 and its sibling DeepRoot — Shape is DERIVED-GIVEN-anchor · RESOLVED +0; the combined Deep-Roots roll-up is closed (board: 33 RESOLVED +0 · 0 ANCHORED · 0 OPEN). The body below is the frozen 2026-06-29 mid-audit record, preserved verbatim; its “OPEN roll-up” language is historical and governed by this banner. Source of truth: the gate board.
Per-gate dossier (Quantum / TOE root). Serious candidate, NOT validated; no status was ever upgraded. STATUS-UPGRADES:0. Frozen 13D branch
dcc66f1b2685/a5b1e6f9d951READ-ONLY / UNMUTATED.Honest grade (matches the live popup chip): Honest grade (ratified board 2026-07-08): CERTIFIED-IRREDUCIBLE · RESOLVED +0 — the cost-floor root is certified irreducible at no new cost; its content is one named, value-free statement (a positive operational cell, Δ₀ > 0) with ℏ identified as its residue value. Direction: strengthened (formerly carried mid-audit as REDUCED-TO-AXIOM). This dossier expands the live 30-second popup spine into its deep, checkable version; it never upgrades the grade.
Headline. Reality having a smallest step is no longer a vague confession. We have compressed it to one named, value-free posit — a positive operational cell, Δ₀ > 0 — and shown that ℏ is just the size of that step, a residue value rather than an added knob.
What this dossier establishes — and what it does not. It establishes that the granularity root of the program (call it R1: "reality has a smallest operational step / a positive cost floor on action") has been driven from a fuzzy posit down to a single sharp, target-free statement, with a largely proved scaffolding around it:
So the root is named, not faked, and ℏ is pinned as its residue value. What this dossier does not establish: it does not derive the granularity floor from a strictly weaker principle (that is the open frontier, §6); it does not derive the value of ℏ, k_B, the Bekenstein constant, or Δ₀ (these are residues by construction, outside what any reconstruction delivers); and it does not claim strict irreducibility ("no deeper principle anywhere could be more fundamental"), which is a universal negative unprovable for any root in any field.
The status word, precisely. The granularity sub-root (R1) is REDUCED-TO-AXIOM / AXIOM-CLOSED: reduced to exactly one named posit — the Uniform Operational Cell Law (Δ₀ > 0) — plus one genuinely atomic measured anchor (ℏ, the measured action spacing). The combined Deep-Roots gate carries the muted OPEN roll-up, because its sibling sub-root SHAPE rests on a permanent open wall (absolute geometric minimality, an uncomputable Kolmogorov question, a different root and out of this dossier's scope). This dossier is about the granularity root, whose honest standing is REDUCED-TO-AXIOM. (Status sources: handoff "Honest status"; RESULT_FTC_FORK_B_CELL_LAW_2026-06-23.md footer; RESULT_A1_FTC_UNIFORM_CELL_PREQUANTUM_DERIVATION_ATTACK_2026-06-24.md §8; CONSOLIDATED_OPEN_GATES_HANDOFF_2026-06-25/DEEPROOTS.md.)
The edge, stated as a confident bet, not a hedge. The strongest honest endpoint is a falsifiable bet: granularity is co-fundamental — the floor sits at-or-below quantum mechanics, thermodynamics, and gravity in the implication order — and the one bet that could upgrade it to a theorem is a pre-quantum reconstruction that derives Hilbert structure, the Margolus–Levitin bound, Landauer's kT ln 2, and the Bekenstein bound all as theorems of a single cost floor provably weaker than any of them. The deeper version — "no possible principle anywhere is more fundamental than this" — is a universal negative, unprovable for any root in any field. That is a limit on all knowledge, not a defect of this program. The ceiling is therefore honestly drawn: a named root with ℏ as its residue, and a concrete falsifiable path to promote it.
Every physical theory bottoms out on a small number of brute assumptions that everything downstream inherits. For the structure of reality, one of those is granularity: the claim that there is a smallest operational step — a positive cost floor on action, a quantum of action. The open question is sharp:
Can the existence of reality's smallest operational step — a positive floor ε > 0 on the cost of a distinguishable transition — be derived from something strictly weaker, or must it be posited?
No accepted physics derives the existence of a minimum action quantum from a deeper principle. Across the field the existence of ℏ as a smallest action step is taken, not earned. The deeper, field-wide version of the same question — deriving the compactness / finiteness of the operational state space from purely operational-causal principles — is recognized as the open frontier of quantum foundations.
The granularity question has two parts that must never be conflated:
The uncertainty principle illustrates the split exactly: "a phase-space cell floor exists" is structure; "it equals ℏ" is value. The semiclassical state count behind statistical mechanics, Bohr–Sommerfeld quantization, and the Bekenstein bound,
$$ N_{\text{distinguishable states}} \;\approx\; \frac{\text{phase-space volume}}{\hbar^{\,n}}, $$
is the same statement realized symplectically with ℏ as the cell size (RECORDABILITY_RESEED_PACKET/03_THEOREM_ATTACK_FTC...md §3). The program is only ever after the existence of the floor; the value is conceded as a "just is" number on the same footing as the four frozen headline anchors {M_Pl, α_i(M_Z), y_t, |V_us|} (ch_01_root_axiom_cost_floor.md §1.9).
The established physics — the "floor exists" half — is textbook, in three mutually independent currencies (ch_01_root_axiom_cost_floor.md §1.3):
These three are mutually independent (action; energy/information; information/region), so the floor's existence does not hang on any one of them. But all three are theorems of quantum mechanics or quantum field theory — Margolus–Levitin and Landauer presuppose a Hilbert space and orthogonal distinguishability; Bekenstein presupposes ℏ in its very constant. So none of them derives granularity from below; each confirms the floor while sitting inside the framework that already contains it. That is the precise sense in which the field has a best confirmation but no derivation.
Within the operational / general-probabilistic-theory (GPT) reconstruction tradition (Hardy; Chiribella–D'Ariano–Perinotti, "CDP"), the analogous structural object — finite operational dimension (a bounded number of perfectly distinguishable states) — is assumed as an axiom, not derived. Adopting it is adopting the cell law. The single near-theorem that produces finiteness of local degrees of freedom from below is Buchholz–Wichmann nuclearity, and that result lives inside algebraic QFT — so invoking it to ground granularity would be a quantum import, the very circularity the program is trying to defeat (handoff open-hole T3; RESULT_T3...md §"Orchestrator addendum").
The certification route — "prove the floor irreducible by showing every distinguishable transition costs at least ε" — is circular inside QM: "cost per distinguishable transition" presupposes distinguishability, and distinguishability is usually defined through Hilbert-space orthogonality. That makes granularity look downstream of QM (QM ⇒ ML floor) rather than its root. Breaking this circle is the decisive obstruction the program identified and partially defeated (§3.5, §4.2).
This is the load-bearing section. The granularity root was driven down a chain of equivalences and reductions to a single named posit. We give each link in full, mark exactly what is proved, conditional, or posited, and state the guards that keep the chain honest.
The first move declines the spatial lattice while keeping its job. What is quantized is cost / action, not space. In the owner's clean form (ch_01...md §1.2):
(L1) Infinite subdivision is HARMLESS. Gliding through a continuum of non-orthogonal (overlapping) states is free — none is a distinguishable accomplishment; overlap is free. (L2) Infinite prerequisite chains with a nonzero lower cost are IMPOSSIBLE. Only transitions to orthogonal (distinguishable) states cost; an infinite chain of such, each costing ≥ a fixed floor, needs infinite total resource and cannot complete under finite resource; orthogonality is dear. ⟹ Any completable (finite-resource) process has finitely many costly (distinguishable) steps.
The L1/L2 line is the non-orthogonal/orthogonal line. The established floors of §2.3 apply precisely to the orthogonal milestones (L2) and to nothing in the non-orthogonal continuum (L1). This is what lets the route decline a smallest length (which would break Lorentz invariance, §3.2) without losing the regulator that stops infinite descent.
The Lorentz corollary. A length floor picks a preferred frame (length contracts), breaking Lorentz invariance — the standard expensive objection to discrete spacetime. Cost, action, and information are Lorentz scalars; a floor on a scalar picks no preferred frame (ch_01...md §1.4). So the cost-floor reframe pays spatial granularity's biggest bill for free. This is structural firewall content G3 (see §3.7): the program supports constraint and sector-label granularity but does not claim spacetime discreteness.
The ℏ > 0 reading (a reading, not a derivation). Realizability ⟹ ℏ > 0: the world admits completable processes only if there is a positive floor on distinguishable cost, and that floor is the nonzero quantum of action. Sending ℏ → 0 returns the Zeno/UV pathologies (UV catastrophe, infinite point-charge self-energy, the unstable classical atom — each an infinite descent ℏ > 0 truncates). The axiom names that logic: because reality is realizable, it is quantum. This explains that ℏ > 0, never which ℏ (ch_01...md §1.5).
The decisive conceptual move is to define distinguishability without an inner product, Born rule, or trace distance. The root metric is the operational total-variation distance
$$ d_{\rm op}(r,s) \;=\; \sup_{T,e} \big| P(e \mid r, T) - P(e \mid s, T) \big|, $$
the supremum over admissible tests T and outcomes e of the gap in outcome frequencies between records r and s. This uses only records, tests, and outcome frequencies — no Hilbert structure appears in its statement. It matches the operational distinguishability of the Hardy / CDP GPT framework; Hilbert orthogonality is a downstream representation of d_op, not its definition (handoff "Banked"; RESULT_A1...md §1 premise 1).
This breaks the surface distinguishability circle: distinguishability no longer rigorously requires QM. (It does not by itself close the deeper certification circle — that the cost floor sits at-or-below QM in the implication order — which remains the open co-fundamentality certificate, §6 hole 3.)
A classical witness confirms d_op is well-posed pre-quantumly: take Ω a measurable space, tests = measurable functions f : Ω → [0,1]; then d_op reduces to the classical total-variation distance, and for distinct points indicator tests give d_op = 1 with no Hilbert structure present (handoff first-pass plug, "Established" demand (1)).
The first fully rigorous link. Let D be the space of distinguishable records and c : D → ℝ≥0 a cost function.
T5. If D is compact, c is continuous, and c(x) = 0 ⇒ x ∉ D (no distinguishable record is free), then ε = min_D c > 0.
Proof. A continuous function on a compact set attains its minimum (extreme value theorem). The minimizer x ∈ D, so by pointwise positivity c(x) > 0; hence ε = c(x) = min_D c > 0. ∎ (handoff "Banked"; GATE_REGRADE...md line 1488.)*
T5 is clean and unconditional given its hypotheses. The entire weight of the granularity question is thereby transferred onto its lead hypothesis: is D (equivalently R_phys, the admissible record space) compact?
The compactness hypothesis is itself reduced to a single operational lemma, FTC (Finite-Resource Operational Test-Compression).
The target T3. For a bounded causal record-support system R (finite extent, finite duration τ, finite total resource budget B = action or energy·time), with R_phys the admissible stably-retrievable records under d_op, claim: R_phys compact ⇒ N_max(R, Δ, τ) < ∞ for every Δ > 0.
Theorem A — finite resource does NOT imply compact recordability (a banked loss). The naive hope fails by a rigorous countermodel (RESULT_T3...md §2). Take a classical pointer x ∈ [0,1] with records r_x; finite extent, finite τ, finite energy-time B (a pointer at rest needs no energy scaling with the number of positions). The admissible readout test T_{x,y} ("is the pointer near x, excluding y") gives P(e_x | r_x, T_{x,y}) = 1 and P(e_x | r_y, T_{x,y}) = 0, so
$$ d_{\rm op}(r_x, r_y) = 1 \quad \text{for all } x \neq y. $$
Then [0,1] under d_op is an uncountable discrete space: for any 0 < Δ ≤ 1 every pair is Δ-separated, so N_max = ∞, the set is not totally bounded, hence not compact. ∎ The stated primitives genuinely do not yield T3.
The precise obstruction. The missing link is not "compactness" abstractly. It is a resource → resolution law:
$$ B < \infty \;\overset{?}{\Longrightarrow}\; \forall\,\Delta > 0,\; N_{\max}(R, \Delta, \tau) < \infty. $$
Finite extent / duration / energy-time do not stop arbitrarily sharp readout tests unless a law connects the resource budget to readout resolution. That law is FTC.
The relabel guard fires on the easy "fix." Defining B = finite information capacity (bits) gives N_max ≤ 2^B trivially — but that assumes finite record capacity, the conclusion. So: B = action/energy-time ⇒ T3 OPEN; B = finite information capacity ⇒ RELABEL-FAIL. Finite information capacity must be derived from more primitive resource assumptions, not assumed (RESULT_T3...md §3, "relabel guard").
Lemma FTC (the named target).
For every bounded causal R (budget B, duration τ) and every tolerance η > 0, there is a finite family of admissible tests 𝒯_η = {(T_j, e_j)}_{j=1}^{M(B,η,τ)} such that for all admissible records r, s: $$ d_{\rm op}(r,s) \;\le\; \max_{1 \le j \le M} \big| P(e_j \mid r, T_j) - P(e_j \mid s, T_j) \big| \;+\; \eta. $$
Plain words: every operational distinction makeable with bounded causal resources is η-approximable by finitely many bounded-resource tests — the pre-quantum analogue of nuclearity / finite local degrees of freedom, stated with no Hilbert space, trace distance, Born rule, Bekenstein bound, or QFT. M may grow as η → 0; it need only be finite at each η (RESULT_T3...md §4; ...03_THEOREM_ATTACK_FTC...md §2).
Theorem B — T3 conditional on FTC (valid proof). If FTC holds and R_phys is closed under d_op-Cauchy limits (completeness), then R_phys is compact (RESULT_T3...md §5).
Proof. Fix η > 0; FTC gives a finite 𝒯_η. Define the operational coordinate map Φ_η(r) = (P(e_j | r, T_j))_{j=1}^M ∈ [0,1]^M. The cube [0,1]^M is totally bounded; cover it by finitely many sup-norm cells of radius η and pick one representative per nonempty cell, giving a finite set {r_1, …, r_K}. Any r lies in a cell with representative r_k, so max_j |Φ_η(r)_j − Φ_η(r_k)_j| ≤ 2η, and by FTC d_op(r, r_k) ≤ 3η. Thus there is a finite 3η-net for every η ⇒ totally bounded; with completeness ⇒ compact. ∎
The constant d_op ≤ 3η and the cube-cover construction are explicit and checked (GATE_REGRADE...md line 1489). The completeness hypothesis is a genuine extra assumption: total boundedness alone does not give compactness without it. This dossier tracks it explicitly on the residue ledger (§5.3, §6 hole 4) — total boundedness ⇒ compactness only modulo Cauchy-completion of the record space, itself posited.
The next attack asks whether FTC itself is derivable from strictly weaker premises. The answer, reported as a banked loss, is no — finite resources alone do not force FTC, and a fortiori do not force a uniform floor.
The decisive delta-test countermodel (RESULT_FTC_FORK_B_CELL_LAW_2026-06-23.md §1). Admissible binary tests f_a(r) = P(e_a | r, T_a) ∈ [0,1]; d_op(r,s) = sup_a |f_a(r) − f_a(s)|. Take R = [0,1], tests f_x(r) = 𝟙[r = x] (the indicator of a single record). For distinct r ≠ s, f_r separates them, so d_op(r,s) = 1. For any finite family {f_{x_1}, …, f_{x_M}}, pick r, s ∉ {x_j} (possible, [0,1] uncountable): every selected test returns 0 on both, so max_j |f_{x_j}(r) − f_{x_j}(s)| = 0, while d_op(r,s) = 1. So FTC ($d_{\rm op} \le \max + \eta$) fails for every η < 1. ∎
Therefore finite causal support + finite duration + finite action/energy-time do not imply FTC unless the theory already forbids infinitely sharp operational tests. FTC is a finite-resolution law, not a theorem of finite resources. Adopting it is naming the real primitive, not deriving granularity (this is "Fork B" — the honest irreducible bottom — as opposed to "Fork A," a derivation that the countermodel rules out from the stated premises).
The irreducible bottom is stated as a single value-free posit:
Uniform Operational Cell Law (granularity root, AXIOM-CLOSED, UNPROVEN). There exists a single, system-independent constant Δ₀ > 0 such that, for every bounded causal record-support system, no two stable independently-retrievable records are operationally closer than Δ₀ in d_op, and no stable record occupies an operational cell smaller than Δ₀. Equivalently: the record-keeping substrate of reality has one universal positive resolution quantum.
This statement names a universal quantum and its role (a uniform grain on d_op); it fixes no value and mentions ℏ nowhere. It could have been written by anyone who believed reality is granular at a single universal scale, before ever measuring ℏ. Its value (Δ₀ ≈ ℏ in spectral units; N ≈ V/ℏⁿ semiclassical count) is the residue, derived only as a downstream shadow, never as input (RESULT_A1...md §6).
Sufficiency lemma — basin-packing: cell law ⇒ FTC (sound conditional). Assume an operational stability landscape with finite total variation Var_op(R) ≤ B, and that each stable record requires a basin of robustness depth ≥ Δ to be resolvable. Pairwise-resolvable robust records have disjoint depth-≥Δ basin-certificates, so
$$ N \cdot \Delta \le B \;\Rightarrow\; N \le \lfloor B/\Delta \rfloor < \infty. $$
Finite ⇒ totally bounded; with completeness ⇒ compact; FTC then follows by finite-pair test extraction over a finite η/6-net (triangle inequality), with M(B,η,τ) ≤ ⌈B/Δ⌉², i.e. M ≲ (B/Δ)² (RESULT_FTC_FORK_B...md §3; GATE_REGRADE...md line 1490). This is conditional on the cell law Δ > 0 — exactly the irreducible posit — so it is honest, not circular: it grounds FTC in a stability picture and isolates the irreducible bit (Δ > 0), which it openly adds rather than derives. No QM, Bekenstein, or thermal import enters.
A subtle point separates what finite resources do buy from what they don't. The basin-packing route, run with an η-dependent Δ(η), buys total boundedness — enough for the cost-floor face (ε > 0 via T5). But the finiteness face (a discrete milestone count) and the exact ℏ-graining correspondence (N ≈ V/ℏⁿ, finite) require the uniform cell Δ₀ > 0. The attack that tests precisely the uniform step is the basin-shallowing countermodel (RESULT_A1...md §2):
Take a bounded operational landscape with finite total variation Var_op(R) ≤ B. Populate it with a countable family of disjoint stable basins of depths d_n = B·2^{−n−1}, so Σ d_n = B/2 < B (finite total resource respected). Each basin is a genuine stable, retrievable record (positive depth ⇒ robust against sub-d_n perturbations). The depths have infimum 0: there is no uniform Δ₀ > 0 with every depth ≥ Δ₀. Consequences, exactly:
Therefore finite causal support + finite duration + finite action/energy-time + a continuous stability landscape do NOT imply a uniform positive Δ₀. They imply at most total boundedness. The gap between {η-dependent floor} and {uniform Δ₀} is exactly the program's named two-face obstruction — and it is now a theorem: a counterexample to the uniform claim exists under the stated premises.
The both-ends meeting point (RESULT_A1...md §3). Backward (from "reality is granular with a uniform grain"): a system-independent Δ₀ cannot be set by any per-system resource (those vary system to system; basin-shallowing shows per-system finiteness is compatible with inf = 0). A uniform floor must be a single universal constant of the record-keeping substrate. Forward (from pre-quantum primitives): finite resources give per-system total boundedness, never a substrate-universal positive minimum. The two ends meet only at "there is one universal action/resolution quantum shared by all record systems" — which is "ℏ exists as structure," the Uniform Operational Cell Law, a new primitive, not a reduction of it. This is the structural signature of an irreducible root.
A standing duality sharpens what the posit is without removing it. Identifying a generating phase on a closed circle with a discrete, equally-spaced action spectrum makes existence + discreteness of the action grain inter-derivable with compactness of the underlying phase (a Pontryagin-type "iff"). But this RELOCATES the uniformity posit onto the compactness of the phase rather than eliminating it — and attempting to derive the uniformity from unitarity is circular by the work's own finding (it would put granularity downstream of QM, the exact distinguishability circle the program broke) (handoff "Banked"; CONSOLIDATED.../DEEPROOTS.md R1; DEEPROOTS_COMPLETION_HANDOFF/specialist_reply.txt R1). Hence the gate-specific granularity posit count is ONE, not zero. Claiming "zero posits" would overstate the floor as assumption-free; that is the live PROMOTION_RISK the audit flagged and this dossier refuses (§7).
Every link above is gated by hard guards; tripping any one is an auto-fail (...03_THEOREM_ATTACK_FTC...md §7; firewall in handoff "Banked").
The orchestrator's circularity test passes: FTC concerns resolution / test-count, not per-record cost, so it smuggles neither the floor ε (which emerges only at T5) nor Δ (FTC holds at every η) (RESULT_T3...md §"Orchestrator verification header").
These are the moves that made the progress believable and reproducible — now shareable at working depth.
4.1 Cost, not length. The single highest-leverage reframe is that what is quantized is action / cost (a Lorentz scalar), not space (a length). This buys the regulating effect of "a smallest something" without paying spatial granularity's Lorentz-breaking bill, and it routes the whole question through distinguishable milestones (L2) while letting non-orthogonal continua (L1) be free. Everything downstream — the Lorentz corollary, the ℏ > 0 reading, the operational metric — is a consequence of this choice (ch_01...md §1.2–§1.5).
4.2 The pre-Hilbert operational metric. Defining distinguishability as the total-variation distance over admissible tests, d_op = sup_{T,e} |P(e|r,T) − P(e|s,T)|, breaks the surface circle that distinguishability "lives only in Hilbert orthogonality." This is what makes the entire reconstruction target-blind: nothing in the root statement mentions QM, so the chain cannot secretly assume what it sets out to ground (RESULT_A1...md §1 premise 1).
4.3 Granularity ⇒ MDL / finite-resolution, made operational. The right physical object is not "compactness" abstractly but a resource → resolution law (FTC): bounded causal resources forbid infinite operational shattering. Naming that object precisely is what converted a topological wall into a concrete physical statement the countermodel demanded — and what exposed that the law is as hard as a slice of quantum reconstruction, because classical physics lacks such a law and QM is the framework that supplies one (RESULT_T3...md §"Orchestrator addendum").
4.4 Run the hard attack; report the loss. The program deliberately tried to derive FTC and the uniform floor from finite resources and reported the failure as a result: the delta-test and basin-shallowing countermodels are banked wins precisely because they are honest losses for the derivation. A precise Fork B (named irreducible bottom) beats a hand-wavy Fork A (claimed derivation). This is the discipline that keeps the grade at REDUCED-TO-AXIOM rather than letting it drift to "derived."
4.5 The κ³/π / no-target-fitting blade. A candidate parent principle counts as a derivation only if it would have been written without knowing the answer is ℏ. Applied as a falsification test to every candidate (basin-packing, pre-quantum Margolus–Levitin, finite information per region, finitism, causal-diamond, GPT finite dimension), each either relabels, splits to the wrong (finiteness-only) face, runs circular through an assumed per-transition cost, or pins its keystone constant only via ℏ (target-fitting). No row would have been written without already knowing the target — the decisive signature that the only premises reaching the uniform floor are the floor itself in disguise (RESULT_A1...md §4).
4.6 Structure vs. value, held on every line. The posit fixes existence and universality of a positive quantum, never its magnitude. Δ₀ ≈ ℏ is the residue shadow (a T7 confirmation), so there is no free coefficient to reverse-engineer. This is why deriving "the value of ℏ" is not a plug-able hole — it is outside what any reconstruction can deliver, by construction (ch_01...md §1.9; RESULT_A1...md §7).
| Result | Statement | Status | Source (read) |
|---|---|---|---|
| T5 | D compact + c continuous + (c(x)=0 ⇒ x∉D) ⇒ ε = min_D c > 0 | PROVED (EVT + pointwise positivity) | handoff "Banked"; GATE_REGRADE...md L1488 |
| Theorem B | FTC + completeness ⇒ R_phys compact, finite 3η-net via cube-cover (d_op ≤ 3η) | VALID PROOF | RESULT_T3...md §5; GATE_REGRADE...md L1489 |
| Basin-packing | cell law Δ>0 ⇒ N ≤ ⌊B/Δ⌋ < ∞ ⇒ FTC, M ≲ (B/Δ)² | SOUND CONDITIONAL (no QM/Bekenstein/thermal import) | RESULT_FTC_FORK_B...md §3; GATE_REGRADE...md L1490 |
| Theorem A (loss) | finite extent+τ+action/energy-time ⇏ FTC; delta-test f_x(r)=𝟙[r=x] on [0,1] gives d_op=1 ∀ distinct records | DECISIVE COUNTERMODEL (banked) | RESULT_T3...md §2; RESULT_FTC_FORK_B...md §1 |
| Basin-shallowing (loss) | finite resources ⇏ uniform Δ₀; depths d_n=B·2^{−n−1}, Σ=B/2<B, inf=0 | DECISIVE COUNTERMODEL (banked) | RESULT_A1...md §2 |
| Pre-Hilbert metric | d_op = sup_{T,e} | P(e | r,T)−P(e |
| Reduction to one posit | granularity ⇒ Uniform Operational Cell Law (Δ₀>0); ℏ = residue value | AXIOM-CLOSED (named, not proved) | RESULT_A1...md §6/§8 |
The specialist results were checked, not rubber-stamped, by the orchestrator (RESULT_T3...md, RESULT_FTC_FORK_B...md verification headers): Theorem A's countermodel verified VALID; Theorem B verified a VALID PROOF with constants checked; the packing bound verified finite (base constant loose, immaterial); all guards G1–G5 verified PASS; the circularity test verified passing. The Fork-B countermodel was verified VALID & decisive; the basin-packing conditional verified SOUND and honestly labeled.
One correction was logged and is carried here: an earlier RESULT_FTC_FORK_B...md footer claimed the uniform Δ₀ was "RESOLVED / verified." The A1 attack demoted that to "uniform Δ₀ POSITED (conditional sufficiency lemma verified; uniform existence NOT derived — §2 countermodel)," reconciling five sibling docs that already held it OPEN (RESULT_A1...md §0, §8). This dossier states the uniform law as a posit, exactly as the corrected disposition requires. The re-grade reviewer independently confirmed this draft's predecessor avoids that one overclaim (GATE_REGRADE...md line 1523, "PASS with one fix").
The reframe relocates rather than eliminates constants. The honest ledger is {ℏ, k_B, Bekenstein constant, Δ₀}, plus the program-wide floor values {c·Λ_YM, ℓ_min, ρ < 1} and the completeness (Cauchy-closure) hypothesis of Theorem B (ch_01...md §1.9; handoff hole 4; GATE_REGRADE...md line 1522 reviewer). Each is "just is" by construction (existence given, value not): the axiom buys floor existence and never floor magnitudes. Δ₀ is a new residue introduced alongside ℏ and k_B — the reframe trades {cost floor ε} for {resolution floor Δ + compactness}; that is a better-motivated relocation, not the removal of a constant, and §6 hole 4 tracks it as such.
A reader should not attempt to derive the uniform floor from unitarity (circular by the work's own finding) and should not construct a finite-K register countermodel to "prove Cell-Law does not imply QM" — that construction is net-new, not in the frozen corpus, and was caught as fabricated (§6 hole 3, traps).
Each open hole is a work-package: precise statement, why it is hard plus the specific traps the verifier already caught, exactly what closes it (target-blind, with the refuting result that also counts as a close), the machinery and corpus files to start from, and the cross-gate leverage. Physics only; the QC-chip and out-of-scope engineering applications are firewalled out entirely.
(a) Precise statement. Prove Lemma FTC pre-quantumly from a distinguishability test-space + additive-cost axiom set, with no Hilbert orthogonality, no trace distance, no Bekenstein bound, no QFT nuclearity. Concretely: derive (a) the inner-product / Hilbert structure so that Margolus–Levitin τ ≥ πℏ/2E falls out, (b) Landauer kT ln 2, and (c) the Bekenstein bound — all as theorems of the cost floor. FTC is currently itself the irreducible posit; this is the bet that flips the arrow.
(b) Why it's hard / prior-attempt lessons. FTC is "as hard as a slice of quantum reconstruction" — the dossier estimates ~15–20% odds to flip the arrow on a single framework (RESULT_T3...md addendum; handoff hole 1). The basin-packing route (the primary handhold) is a sufficiency lemma that takes Δ₀ as hypothesis (Fork-B §3 says so verbatim); it derives the count, never the grain. Traps to avoid: (i) do not import the orthogonality you must derive (G3 auto-fail); (ii) do not adopt an equal-strength parent principle (e.g. GPT finite operational dimension — adopting it is adopting the cell law, RELABEL-FAIL); (iii) do not re-inject ε > 0 by hand; (iv) do not use a pre-quantum Margolus–Levitin whose normalization τ ≥ πℏ/2 is pinned only by ℏ — its keystone constant is the target (target-fitting) (RESULT_A1...md §4 table).
(c) Exactly what closes it. A reconstruction theorem deriving (a)+(b)+(c) above from the additive-cost axiom set, passing all auto-fail guards. Success criterion: the derived structure reproduces N ≈ V/ℏⁿ and the uncertainty relation as outputs. A refuting result is also a valid close: a proof that no such axiom set can derive FTC without re-importing orthogonality/finite-dimension (a clean no-go) would terminally confirm Fork B and is equally publishable.
(d) Machinery & inputs. Start from RESULT_T3_RECORDABILITY_REDUCED_TO_FTC_2026-06-23.md, RESULT_FTC_FORK_B_CELL_LAW_2026-06-23.md, RECORDABILITY_RESEED_PACKET/03_THEOREM_ATTACK_FTC...md, and the Hardy / CDP GPT reconstruction literature for the test-space machinery. Tools: operational/GPT axiomatics, total-boundedness arguments, the d_op metric.
(e) Leverage. Closing this closes Hole 3 (the co-fundamentality certificate is the same reconstruction object) and Hole 2 (T3 follows), and upgrades the granularity grade from REDUCED-TO-AXIOM toward DERIVED — the single largest possible move on this root.
(a) Precise statement. Prove "bounded causal recordability ⇒ R_phys compact" as a standalone pre-quantum theorem (or prove the FTC lemma directly) from bounded-causal-resource axioms ALONE.
(b) Why it's hard / traps. In GPT frameworks compactness is an axiom; the only near-theorem (Buchholz–Wichmann nuclearity) lives inside QFT, so invoking it is a quantum import. The live circularity to defeat: the "finite resource budget" B must not secretly be the per-milestone floor. Trap: keeping B = finite information capacity is the illegitimate relabel that gives N ≤ 2^B trivially — keep B = action / energy·time (RESULT_T3...md §3 relabel guard).
(c) Exactly what closes it. Either FTC proved from resource axioms (then Theorem B gives compactness), or T3 proved directly. Success: a finite 3η-net constructed from resource bounds alone. Refuting close: a sharper countermodel showing resource bounds provably cannot bound resolution (strengthening Theorem A) terminally confirms FTC is the bottom.
(d) Machinery & inputs. RESULT_T3...md §§2–5; the causal-diamond / Lorentz-scalar leg (necessary, not sufficient — the one already-clean leg supplying "bounded" and frame-independence). Combine with the basin-packing route in RESULT_FTC_FORK_B...md §3.
(e) Leverage. Discharges the completeness-conditioned compactness step and hardens the whole T5 chain; feeds Hole 1.
(a) Precise statement. Supply the non-circular pre-Hilbert definition of distinguishability (already partly banked as d_op) AND show the cost floor sits at-or-below QM in the implication order — the co-fundamentality certificate. The surface circle is broken (d_op needs no QM); the deeper certification circle (the floor is downstream of QM via QM ⇒ ML floor) is the recognized-open object.
(b) Why it's hard / traps the verifier caught. This is the same reconstruction object as Hole 1. The first-pass plug attempted to deliver the certificate and three fabrications were caught — do NOT repeat them (handoff first-pass plug, "Traps the verifier caught"):
- Fabricated finite-register countermodel N2 (Ω = {0,…,K−1}, indicator tests, d_op = 1 for r ≠ s, uniform cell Δ₀ = 1, N_max = K < ∞): this construction appears nowhere in the frozen corpus. The corpus's only countermodels are the continuum-pointer [0,1] delta-test model (RESULT_T3) and the basin-shallowing model (RESULT_A1 §2); neither is a finite-K register, and neither is offered to prove "Cell-Law does NOT imply QM." Do not invent it.
- "QM ⇒ Cell-Law (ℏ-graining)" asserted as a load-bearing "standard fact": not an established implication anywhere in the corpus; worse, the reseed packet explicitly forbids using ℏ-graining/uncertainty as an input step (G3 quantum-smuggling — "puts granularity downstream of QM, the exact distinguishability circle the whole program broke").
- The strict implication order "Floor-existence ≤ Cell-Law < QM" and "QM is SUFFICIENT BUT NOT NECESSARY for the floor": this co-fundamentality ordering certificate is fabricated as delivered. The corpus lists it only as the undelivered closes-by target (GATE_REGRADE...md lines 1502/1511; RESULT_A1...md §3–§4), at ~15–20% odds to ever flip one framework — never as a proven result. Stay target-blind; do not assert the ordering as proven.
(c) Exactly what closes it. A genuine implication-order certificate showing the cost floor is derivable from a strictly weaker premise than QM (so QM ⇒ floor but not conversely), with no ℏ-graining input. Refuting close: a proof that floor and QM are equivalent (mutually derivable) would establish strict co-fundamentality — also a valid, publishable endpoint and consistent with the declared ceiling.
(d) Machinery & inputs. RESULT_A1...md §3–§4 (the both-ends meeting point and the falsification test table); the banked d_op definition; GPT operational-distinguishability literature (Hardy; CDP).
(e) Leverage. Same object as Hole 1; closing it defeats the named decisive obstruction (the distinguishability circle) and certifies the headline "co-fundamental" claim.
(a) Precise statement. Δ₀ is introduced as a NEW residue (a posited resolution margin) alongside ℏ and k_B; the reframe trades {cost floor ε} for {resolution floor Δ + compactness} rather than removing a constant. Derive Δ₀ from a deeper resource law instead of positing it.
(b) Why it's hard / traps. The basin-shallowing countermodel (RESULT_A1 §2) proves a uniform Δ₀ is not forced by finite resources, so any derivation must supply the universality from outside per-system resources. Trap: do not silently drop Δ₀ from the residue ledger by claiming the reframe "eliminated" a constant — it relocated one. The completeness/Cauchy-closure hypothesis of Theorem B belongs on the same ledger and was previously glossed (GATE_REGRADE...md line 1522 reviewer).
(c) Exactly what closes it. A deeper substrate law from which a single universal Δ₀ follows. If it cannot be derived, the close is documentary: state explicitly that one residue (Δ₀) replaced another (ε) — better-motivated, not eliminated — so the ledger {ℏ, k_B, Bekenstein constant, Δ₀, completeness} stays complete and honest.
(d) Machinery & inputs. RESULT_A1...md §2–§3; RESULT_FTC_FORK_B...md §3; the residue table in ch_01...md §1.9.
(e) Leverage. Keeps the residue ledger honest (guards against a false "constant eliminated" claim); a genuine derivation would feed Hole 1.
(a) Precise statement. Test the unified hypothesis "one granularity ⇒ both the mass gap and Λ." Compute ρ_vac = (½) Σ_n λ_n over the actual finite spectrum of the frozen operator with no free regulator, and check whether granularity can make Λ small, not merely finite. (This is a Λ-gate test surfaced by granularity, not the granularity root itself.)
(b) Why it's hard / what is established vs. fabricated. The hypothesis RELOCATES: all eight reframes relocate on magnitude — granularity makes Λ finite but not small. This is established by dimensional analysis (handoff first-pass plug; GATE_REGRADE...md line 1505): the frozen operator's native scale is M_cutoff = 1/R₀ = 6.28 × 10¹⁶ GeV (the GUT/compactification cell), so the naive vacuum density ~M_cutoff⁴ = 1.56 × 10⁶⁷ GeV⁴ misses (2.3 meV)⁴ = 2.80 × 10⁻⁴⁷ GeV⁴ by 10^113.7. One granularity scale supplies one transmutation exponent (~161 in density for the gap, ln(M_cut⁴/Λ_YM⁴) = 161.2); meV needs a second independent ~262-decade exponent (ln(M_cut⁴/meV⁴) = 261.9) the single scale cannot provide. Traps the verifier caught — do NOT repeat: the zero-weight (T²-invariant) multiplicity rule m₀(p,q) = min(p,q)+1 if (p−q)%3==0 else 0 was asserted as computed/"verified on 3 known irreps," but aK6_status.json lists degeneracies_known:false and sign_stable:false, and k6_spectrum.csv marks all 35 modes BLOCKED_MISSING_ZERO_WEIGHT_MULTIPLICITY ("m0 not supplied by reference"). The downstream numbers built on it — N = 2920, Σ deg·C2 = 76416, (½)Σ = 38208 — therefore reproduce the arithmetic but rest on the named open blocker, not a derived input; presenting them as "computable, not a wall" is an overclaim against frozen status.
(c) Exactly what closes it. A canonical regularized (½)Σ λ_n via the ζ_K6(−1) analytic continuation OR the d=6 scalar a₄ Seeley–DeWitt coefficient (the Λ-14 a_K6 blocker — the divergence makes this object mandatory, not optional), plus the zero-weight multiplicity rule supplied by a real reference (not asserted), plus the graded (spin-c / fermion) sector sign. Prediction (target-blind): ~Λ_YM⁴ ~ 1.6 × 10⁻³ GeV⁴, ~10⁴⁴ too large vs (meV)⁴ ⇒ confirms RELOCATES (a valid negative close). Only a knob-free built-in suppression (e.g. spin-c index −3 or a kernel dimension) driving toward 10⁻¹²² M_Pl⁴ with NO H₀ would upgrade to conjecture.
(d) Machinery & inputs. The frozen eigenvalue CSV (.../New folder/...k6_spectrum.csv) with C2(p,q) = (p²+q²+pq)/3 + (p+q), dim(p,q) = (p+1)(q+1)(p+q+2)/2; aK6_status.json; the Λ-14 a_K6 dossier. Confirm C2(1,1) = 3 is the lowest deg>0 eigenvalue; multiply by 1/R₀² with 1/R₀ = 6.28 × 10¹⁶ GeV.
(e) Leverage. Cross-gate into the Λ (cosmological constant) gate and the a_K6 spectral-coefficient blocker; a knob-free suppression would also bear on the mass-gap transmutation story.
(a) Precise statement. "Discrete ⇒ positive gap" is the EASY direction; the recognized open wall is a uniform-in-a lower bound on the gap that survives the continuum limit and lands on ordinary SU(3)_c. Build the finite positive operator H on the pure-glue (Z₆-center) projection of K₆ in the smallest R₀-fixed truncation and test it.
(b) Why it's hard / traps. The finite-floor gap is inherited-established (standard finite-spacing reflection-positive transfer matrix), but that is distinct from a polymer-DERIVED gap; the still-open object is the subcriticality inequality z_∗ < 1/(𝓔_conn·A_fluc) uniformly through the marginal band g(2ⁿa) = O(1) (ch_01...md §1.6–§1.8). Trap: F⁺ (the frozen geometry) is inert / no shortcut here — the bridge must be proved from ordinary 4D pure-glue SU(3) block data, not by invoking the 13D branch; and the certificate alphabet must be defined only from local gauge-invariant block data, never from energy eigenstates (else it assumes the gap it claims). Do not bank the retired coincidences (κ³/π, 5+3=8).
(c) Exactly what closes it. Verify (a) λ₁ > 0 and stays bounded below as the truncation dimension grows toward the cutoff (not a truncation artifact), and (b) λ₁ = M_cutoff · exp(−2π/(b₀ α)) lands in [0.1, 0.3] GeV using ONLY frozen α(M_U) and SU(3) b₀, no fitted prefactor. Refuting close: λ₁ → 0 in the continuum limit, or a value outside [0.1, 0.3] GeV, is a clean negative.
(d) Machinery & inputs. The pure-glue Z₆-center projection of K₆; the No-Infinite-Descent Coercivity Theorem and Finite Gauge-Invariant Certificate Alphabet Theorem (ch_01...md §1.6–§1.7); the boxed z_∗ inequality. This is Gap-02 work — coordinate with the Gap-02 dossier.
(e) Leverage. Closing it bears directly on Gap-02 (Yang–Mills mass gap) and on the granularity ⇒ gap dissolution route; it would also feed Hole 5's transmutation exponent.
The ceiling, stated as the strength it is. The granularity root is co-fundamental, not strictly irreducible. "Co-fundamental" means the cost floor sits at-or-below QM, thermodynamics, and gravity in the implication order — a bounded, defensible claim. The deeper version — "no possible principle anywhere is more fundamental than this cost floor" — is a universal negative over an open-ended domain, unprovable in principle for any root in any field; the program itself rates it ~0%. That is a limit on all knowledge, not a defect of this work. So "co-fundamental, named root with ℏ as its residue" is the ceiling, not a hedge (handoff "Dissolved unicorns"; GATE_REGRADE...md lines 1510–1513).
Dissolved ≠ solved. Reducing granularity to one named posit is a reduction in vagueness and standing, not a derivation. The axiom count does not drop (granularity stays one root); no physics gap closes. Naming the Uniform Operational Cell Law is adopting a sharper primitive, not proving it.
Given-E ≠ derivation of E; selection ≠ derivation. This dossier covers the granularity root only. Its sibling SHAPE root — why this 13D geometry — remains a permanent open wall (an uncomputable Kolmogorov / shortest-description question), correctly REFUSED-as-axiom and OPEN; that is why the combined Deep-Roots gate carries the OPEN roll-up even though granularity is REDUCED-TO-AXIOM.
What is explicitly NOT claimed.
- Not that granularity is derived from a weaker principle (Fork A is ruled out from the stated premises by the delta-test and basin-shallowing countermodels). It is named (Fork B).
- Not "zero posits." The gate-specific granularity posit count is ONE (the Uniform Operational Cell Law). The Pontryagin iff relocates but does not eliminate the uniformity posit; deriving uniformity from unitarity is circular. Claiming "zero" was the live PROMOTION_RISK and is refused (CONSOLIDATED.../DEEPROOTS.md; DEEPROOTS_COMPLETION_HANDOFF/specialist_reply.txt).
- Not the value of ℏ, k_B, the Bekenstein constant, or Δ₀ — these are residues by construction; the program is about structure, not magnitude.
- Not spacetime discreteness (G3): a smallest length is explicitly not claimed; only a floor on the Lorentz-scalar cost.
- Not strict irreducibility (the universal negative above).
The anchors paid. Granularity rests on exactly one named, value-free posit (the Uniform Operational Cell Law, Δ₀ > 0) plus one genuinely atomic measured anchor (ℏ, the measured action spacing — an anchor, stated without its size). The Theorem-B compactness step additionally carries the completeness (Cauchy-closure) hypothesis on the residue ledger. The headline residue set is {ℏ, k_B, Bekenstein constant, Δ₀}.
The bottom line. Reality is granular because stable records require positive operational cells; the cell's existence is the root, its value (ℏ) is a residue. Finite resources provably cannot supply the uniformity of that grain (only per-system total boundedness), so the uniform quantum is named, not derived — and there is a concrete, falsifiable reconstruction path (Holes 1/3) that would promote it. Serious candidate, frozen, reproducible, load-bearing — not validated and not proven unique. STATUS-UPGRADES:0.
Dossier built 2026-06-29 by synthesis + expansion of the frozen corpus: RESULT_T3_RECORDABILITY_REDUCED_TO_FTC_2026-06-23.md, RESULT_FTC_FORK_B_CELL_LAW_2026-06-23.md, RESULT_A1_FTC_UNIFORM_CELL_PREQUANTUM_DERIVATION_ATTACK_2026-06-24.md, RECORDABILITY_RESEED_PACKET/03_THEOREM_ATTACK_FTC...md, TOE_Self_close_chapters/ch_01_root_axiom_cost_floor.md, PER_GATE_DOSSIERS/CONSOLIDATED_OPEN_GATES_HANDOFF_2026-06-25/DEEPROOTS.md, PER_GATE_DOSSIERS/DEEPROOTS_COMPLETION_HANDOFF/specialist_reply.txt, GATE_REGRADE_BESTCASE_AND_HOLES_2026-06-29.md (DeepRoot-granularity block, lines 1476–1523), and the live brief articles/GATE_BRIEF_DEEPROOTS.html. No number is fabricated; every quantity traces to a cited file. No status was upgraded. Frozen branch dcc66f1b2685 / a5b1e6f9d951 READ-ONLY / UNMUTATED. Committed: NO. Deployed: NO.