Skip to content

9. The critical mechanism: types before numerical comparisons

Question. What precisely licenses the move from the multiradial representation in IUT III, Theorem 3.11, to the log-volume bound in Corollary 3.12's Step (xi)? This is the IUT-N1 -> IUT-N2 edge in the claim ledger. The source-paired dispute guide reports both interpretations; the source-named object/arrow crosswalk compares their diagrams. This file records smaller mathematical tests, not a new proof or a verdict on the published argument. The later Project LANA interim report proposes a more specific compatibility test; it explicitly does not prove it.

Graduate-level walkthrough (intuition only)

The issue is an ordered comparison between values from different constructions, not just an isomorphism of underlying objects. Let \(r=-|\log q|\) denote the fixed native \(q\)-value and \(s=-|\log\Theta|\) the proposed bound in their respective normalizations. Suppose a paper-licensed choice \(\beta\) yields a normalized output log-volume \(t_\beta\) from a hull region, with \(t_\beta\le s\). To infer \(r\le s\), we would still need a compatible numerical law such as \(r\le t_\beta\) for that same choice and normalization. Then transitivity supplies the bound. This is an elementary conditional implication, not an assertion that IUT supplies the first inequality.

The ledger below asks how the \(\Theta\)-link, permutation transport, vertical log-Kummer maps, and hull/determinant construction might compose. The report's proposed equality of native and reconstructed maps is a candidate test, not an established equality; even such an equality would need a well-typed evaluation into the bounded output. A vertical hull-volume bijection is genuine information, but does not by itself transport the fixed native \(r\) horizontally.

Further preparation, not proof evidence: the object-identity ledger, dictionary, and Wikipedia on isomorphisms.

An initial object and transport ledger

The entries below are an audit index, not an inventory of all the objects in IUT. A citation to a definition or a participant's report does not by itself verify a map's type or its numerical effect. Page numbers refer to the hosted preprint PDFs listed in the source register; earlier versions may differ.

ID Distinct object or passage Source and reported relation What remains to establish
M1 One fixed collection of initial \(\Theta\)-data (field, curve, prime, local choices) IUT I, Definition 3.1, PDF p. 61; re-used for both columns in IUT III, Theorem 3.11, p. 153 Sharing initial data does not make later theaters or their log-volume functions identical
M2 \(\Theta\)-Hodge theaters at labels \((0,0)\) and \((1,0)\), each with its own arithmetic holomorphic structure IUT I, Definition 3.6, PDF p. 87; IUT III, Theorem 3.11, p. 153 Preserve the labels; a common name is not an identity between theaters
M3 \(\Theta\)-pilot at \((0,0)\) \(\leftrightarrow\) \(q\)-pilot at \((1,0)\) IUT I, Corollary 3.7(i), PDF p. 88, and IUT III, Definition 3.8(ii), pp. 112–113: prime-strip poly-isomorphism; IUT I, Remark 3.7.1, p. 89, does not distinguish one cross-theater isomorphism Not a ring/scheme isomorphism; no numerical log-volume identity follows solely from this correspondence
M4 Multiradial representations \({}^{0,\circ}\mathcal R_{\mathrm{LGP}}\) and \({}^{1,\circ}\mathcal R_{\mathrm{LGP}}\) IUT III, Theorem 3.11(i), pp. 153–155: a different, permutation-symmetry poly-isomorphism transports representations across columns; Step (xi-b), pp. 181–182, invokes it How this map combines with M3, and whether the same log-volume function survives horizontal transport, was not established by the passages independently checked
M5 Vertically varying log-Kummer correspondences IUT III, Theorem 3.11(ii), pp. 155–156: one specified component is precisely compatible with log-volumes; (Ind3) records only upper semi-compatibility for others Vertical compatibility is not by itself a theorem of horizontal M4 log-volume preservation
M6 Abstract, concrete (including the \(\ell^\star\)-member \(\Theta\) family), and arithmetic-degree real lines Scholze–Stix, section 2.2, PDF pp. 9–10 draws six nodes with a literal equality at the bottom; its text proposes inserting a \(j^2\) factor somewhere on the left, but the figure labels no such arrow; see the six-node redraw Decide whether IUT's actual path must factor through this proposed diagram of identifications or uses a separate hull/containment; the figure draws no hull arrow
M7 A possible region \(P\), its holomorphic hull \(\phi(P)\), and the arithmetic line obtained by \(\det^{\otimes M}(\phi(P))\) IUT III, Remark 3.9.5, pp. 127–128, (Ob1)–(Ob3), (Ob6), pp. 131–135, and (Ob8)/(Ob9), pp. 137–139; Step (xi-c), PDF p. 182, visibly writes \({}^{1,\circ}\overline{\mathcal U}\supseteq{}^{1,\circ}\mathcal U\), followed by determinant/log-volume in (xi-d), p. 183 Hull inclusion and (Ob9)'s vertical log-volume bijection do not themselves pick the native \(q\) value in the output bound; audit its admissible choice and numerical law
M8 Native $- \log(q) $ and the Step-(xi) output region $\mathbb R_{\leq-
M9 Value-group real line \(R_{\mathrm{val}}\) and volume-container real line \(R_{\mathrm{ss}}\) Project LANA report, section 9.2, PDF pp. 45–46: the latter uses equal-volume classes of measurable adelic regions; these are the report's two distinct pointed real-vector-space constructions Show that the native and reconstructed maps use compatible source and target copies, not merely isomorphic-looking lines; separately connect these lines to IUT III's numerical output region
M10 \(\eta_q\) from the native \(q\)-pilot and \(\eta^{\mathrm{anab}}_S\) reconstructed after a choice of integral-structure data \(S\) Same report, section 9.2, PDF p. 46: asks for a suitable \(S\) with \(\eta_q=\eta^{\mathrm{anab}}_S\) (9-1); p. 49 says the team does not have a proof Establish existence of that \(S\) and a compatible equality of maps; the report's proposed reduction is not itself a theorem of IUT III

To complete an edge, write source (theater, object, numerical copy) -> operation -> target (theater, object, numerical copy), with its exact domain, codomain, hypotheses, preserved operations, permitted choices, and source passage. Unknown is an acceptable cell; silently writing = between different copies is not.

flowchart TB
  A["Theta-pilot (0,0)"] -->|"Theta-link: pilot correspondence"| B["q-pilot (1,0)"]
  A -->|"Theorem 3.11: representation"| C["representation (0, circ)"]
  C -->|"permutation poly-isomorphism"| D["representation (1, circ)"]
  D -->|"hull / determinant"| H["output region"]
  B -->|"native log-volume"| Q["native q-value"]
  Q -. "membership asserted in Step (xi-f)" .-> H

This is a dependency sketch, not a commutative diagram: the dashed membership edge is what the argument must justify. In particular, the top correspondence and the middle poly-isomorphism are different maps.

Semantic dictionary for this edge

"Not licensed" means not established by the cited, checked passage alone, not a claim that no other IUT lemma supplies the fact.

Operation Structure it gives or retains What it does not identify by itself Next legitimate comparison to check
\(\Theta\)-link (M3) A poly-isomorphism of specified prime-strip/Frobenioid data and a pilot correspondence The theaters' ring/scheme structures, a distinguished cross-theater isomorphism, or equal log-volumes Type the pilot images in their respective theaters before assigning numbers
Permutation transport (M4) A poly-isomorphism between the two labeled multiradial representations The native \(q\)-pilot volume function with the transported \(\Theta\)-pilot volume function Check its numerical effect under (Ind1)–(Ind3), separately from the \(\Theta\)-link
Vertical log-Kummer (M5) Precise log-volume compatibility for a specified component, upper semi-compatibility for others The horizontal compatibility of M4 Use the stated component and quantifiers; do not extrapolate to the entire lattice
Hull and determinant (M7) An invariant container and an arithmetic line with a one-sided volume comparison An invertible ordered-real-line map or equality of the original region's volume Show that the native \(q\)-value lies in the correct output region (M8)

Remark 3.9.5(vii), (Ob8)/(Ob9), PDF pp. 137--139, supplies a vertical log-Kummer comparison and a natural bijection of hull log-volumes across a log-link; it is not an empty citation. The arrow audit in 09b records why this still leaves the particular native \(q\) value and the choice-compatible output bound as distinct obligations.

A concrete candidate for the missing comparison

Project LANA's interim analysis, sections 8.3 and 9.2 describes two proposed calculations of the right-hand \(q\)-pilot log-volume. It constructs \(R_{\mathrm{val}}\) from a BPS's value-group portion and \(R_{\mathrm{ss}}\) from its log-shell/volume-container data. The native calculation gives a map \(\eta_q\); the anabelian reconstruction gives \(\eta^{\mathrm{anab}}_S\) depending on an admissible choice \(S\). Its proposed main goal, equation (9-1) on PDF p. 46, is

\[ \exists S\ \text{admissible},\qquad \eta_q=\eta^{\mathrm{anab}}_S \quad\text{as suitably identified maps }R_{\mathrm{val}}\to R_{\mathrm{ss}}. \]

In the report's interpretation, compatibility would make the original \(q\)-pilot log-volume one of the admissible output-region values in the right-hand volume container, yielding the numerical comparison. The report has not proved this compatibility (section 10.5, PDF p. 49). Its section 8.2, PDF p. 42, warns that merely knowing two weakened objects are isomorphic is too weak to specify the necessary link. This is a candidate shared question for our audit, not an assertion that proving (9-1) alone checks every IUT dependency. Keep the pinned conditional Lean variant separate: it assumes a numerical inequality and does not establish (9-1).

The report sections 10.2–10.5, PDF pp. 47–49 agrees that the real-line hexagon drawn by Scholze–Stix does not commute, and says that making a proof factor through that diagram would lose too much precision. It proposes that the intended argument instead compares two constructions within the right-hand holomorphic structure. Its team neither proves that the original argument establishes (9-1) nor reaches complete consensus that it does not. Agreement about the hexagon therefore does not settle the distinct compatibility claim.

A typed conditional bridge, not a proof of Step (xi)

The dependent formalizer compared IUT III, Step (xi-c), PDF p. 182 with the Scholze–Stix hexagon reproduced in Project LANA's report, Figure 7, PDF p. 47. The figure has no containment arrow. The paper explicitly has \({}^{1,\circ}\overline{\mathcal U}\supseteq{}^{1,\circ}\mathcal U\): the overlined object is the holomorphic hull of the plain one. PDF text extraction drops this overline and can falsely print the two sides as \(U\supseteq U\); check the page image. This source-level distinction does not establish that the critics' loop is avoidable or that the native \(q\)-volume lands in the hull.

Here is a deliberately small ordinary-mathematics schema, not a theorem verified in IUT or Lean. For fixed initial data, let \(V=R_{\mathrm{val}}\) and \(W=R_{\mathrm{ss}}\) be the two different spaces in the report. After specifying compatible identifications, take maps \(\eta_q,\eta^{\mathrm{anab}}_S:V\to W\), a designated \(q\)-input \(x\in V\), an admissible reconstructed-output set \(\mathcal A_{T,S}\subseteq W\), and a shared, normalized log-volume evaluation \(\nu:W\to\mathbb R\). Suppose:

  1. A suitable admissible \(S\) makes \(\eta_q=\eta^{\mathrm{anab}}_S\) as maps \(V\to W\) — the report's still-unproved condition (9-1).
  2. The reconstruction actually places \(\eta^{\mathrm{anab}}_S(x)\in\mathcal A_{T,S}\); this must be checked using the hull, indeterminacies, and the same \(S\).
  3. The output bound \(\nu(z)\le -T\) holds for every \(z\in\mathcal A_{T,S}\), and the native normalization is \(\nu(\eta_q(x))=-Q\).

Then, by substitution rather than by identifying labels,

\[ -Q=\nu(\eta_q(x)) =\nu(\eta^{\mathrm{anab}}_S(x)) \le -T, \qquad\text{hence}\qquad T\le Q. \]

This does not assume the disputed native membership \(-Q\in\mathbb R_{\le -T}\): condition 2 concerns the independently constructed reconstructed output, and condition 3 bounds that output set. Neither source supplies this complete typed implication as a proved lemma. The report explicitly lacks a proof of condition 1 (PDF p. 49), and its section 9.3 outlines, rather than establishes, the needed passage from (9-1) to the output region. Conditions 2–3, including the common normalization and the actual quantifiers over IUT choices, remain separate source-level obligations. For full Corollary 3.12 they would have to hold for its paper-stated data, not merely one chosen \(x\).

To see why "the two spaces admit an isomorphism" cannot replace condition 1, take \(V=W=\mathbb R\), \(x=-1\), \(\eta^{\mathrm{anab}}_S(v)=v\), \(\eta_q(v)=v/2\), and \(\mathcal A_{T,S}=(-\infty,-1]\) with \(\nu\) the identity. Both maps are order-preserving isomorphisms and the reconstructed value \(-1\) lies in the bounded set, but the native value \(-1/2\) does not: \(T=1>Q=1/2\). This is a countermodel to replacing compatibility by mere isomorphism, not a model of IUT's actual maps.

A weaker numerical test, with its own unproved premise

For the designated inputs that the paper actually needs, equality of the two maps is stronger than the following one-sided compatibility:

\[ \nu(\eta_q(x))\leq\nu(\eta^{\mathrm{anab}}_S(x)). \tag{2} \]

If the reconstructed membership and output bound in conditions 2–3 above are independently established, (2) gives \(-Q\leq\nu(\eta^{\mathrm{anab}}_S(x))\leq-T\). This is another conditional ordinary-mathematics lemma, not a claim that IUT III or the Project LANA report proves (2). It does not assume the desired native output-region membership, but its direction, admissible choice \(S\), common evaluation \(\nu\), and quantified set of inputs would all require source-level justification. Reverse the sign in (2) and the conclusion need not follow: the countermodel just above has \(-1=\nu(\eta^{\mathrm{anab}}_S(x))< \nu(\eta_q(x))=-1/2\) while \(T>Q\).

The alternative is strictly weaker at one input: with \(V=W=\mathbb R\), \(x=-1\), \(\nu\) the identity, \(\eta^{\mathrm{anab}}_S(v)=v\), \(\eta_q(v)=2v\), and \(\mathcal A_{1,S}=(-\infty,-1]\), both maps are order-preserving linear isomorphisms and (2) holds at \(x\), but the maps are not equal. There is an important quantifier trap. If (2) were asserted for every \(v\in V\), and both evaluated maps \(\nu\circ\eta_q\) and \(\nu\circ\eta^{\mathrm{anab}}_S\) were additive, applying it to \(v\) and \(-v\) would force their evaluations to be equal everywhere. Even then, equality of the maps themselves would require an additional property such as injectivity of \(\nu\). Thus a one-sided test may only be genuinely weaker on a restricted input set; it is not a shortcut around the paper's quantifiers or the missing comparison.

A small algebraic reduction that can actually be checked

Let \(T\) and \(Q\) denote real numerical values corresponding to \(|\log(\Theta)|\) and \(|\log(q)|\) only if the two passages being compared use the same normalization and compatible real-number copies. If \(Q>0\), then the displayed form of IUT III's conclusion

\[ \forall C\in\mathbb R,\qquad -T\leq CQ\ \Longrightarrow\ C\geq-1 \tag{1} \]

is equivalent, as an ordinary real-number statement, to \(T\leq Q\). For the forward direction, choose \(C=-T/Q\) in (1); this forces \(-T/Q\geq-1\). Conversely, any \(C\) in (1) satisfies \(C\geq-T/Q\geq-1\) when \(T\leq Q\). The inequality quoted by Scholze–Stix as the intermediate Step-(xi) conclusion, \(-Q\leq-T\), has the same algebraic orientation as \(T\leq Q\). For \(Q=0\) and \(T\geq0\), the premise in (1) holds for every \(C\) and the conclusion is false. IUT III's proof itself explicitly invokes \(|\log(q)|>0\) at Corollary 3.12, PDF p. 173; we have not separately derived that assertion from the initial data. For example, \(T=3,Q=2,C=-3/2\) is a counterexample to (1), while \(T=2,Q=3\) satisfies it.

This removes one formal algebra ambiguity from 05, reproduction check 6. It does not establish that the two papers' \(T\) and \(Q\) are the same normalized quantities or that the Step-(xi) comparison producing \(T\leq Q\) is legitimate. The primary text states \(Q>0\); an independent check of its underlying derivation remains a separate task. These are source-level and mathematical obligations, not consequences of this lemma.

Two copies: a non-proof countermodel to an untyped inference

Consider two distinct ordered one-dimensional real vector spaces \(V_A=\mathbb R e_A\) and \(V_B=\mathbb R e_B\). The basis symbols are coordinates for this example, not distinguished parts of the structures. Let \(x=e_A\), \(y=e_B\), and measure coordinates in \(V_B\) by \(\mu_B(t e_B)=t\). Both \(f_1(e_A)=e_B\) and \(f_2(e_A)=2e_B\) are positive linear isomorphisms. Yet

\[ \mu_B(f_1(x))=1\leq\mu_B(y)=1,\qquad \mu_B(f_2(x))=2\nleq\mu_B(y)=1. \]

An inequality transported by one choice of isomorphism does not therefore hold for every choice. If an additional structure distinguishes the basis and requires its preservation, \(f_2\) is no longer admissible; if a theorem supplies a uniform estimate, that estimate must be stated and checked. This model tests the distinction between a correspondence, an allowed transport, and a numerical bound. It models neither IUT's indeterminacies nor Mochizuki's potentially nonlinear log-volume operation, and so refutes neither party's actual argument.

A source-grounded example of why a hull changes the question

Mochizuki's separate 2018 explanatory model, (LVEx), PDF pp. 24–26 takes a region \(R_{a,b}\subseteq\mathbb R^2\) and adds its reflection to form \(S_{a,b}\) for \(a,b>0\). A direct area calculation gives

\[ \mu_{\log}(R_{a,b})=\log(4a+2b) \quad<\quad \mu_{\log}(S_{a,b})=\log(4a+4b). \]

The \(4a\) is the central rectangle; the first region has two one-sided strips of area \(b\) each, and its symmetrization fills the reflected strips as well. For \(a=b=1\), this is simply \(\log 6<\log 8\). The example has a separate linear gluing map; its strict volume inequality comes from taking a larger region, not from multiplying a number by a scalar. The independent trial recomputed these areas, but neither this example nor the two-copy example above proves that IUT's actual Step-(xi) output contains the native \(q\)-pilot value. That is precisely the compatibility still to be checked. The exact parameter experiment in 05 §5.8.1 corrects this guide's numerical example, characterizes when the analogy's strict sign pattern can hold, and finds two admissible shapes with the same input log-area but different hull log-areas. The \(a=b=1\) example here shows only an area increase: its positive input log-area does not meet that stricter sign pattern.

A terminology-free hull experiment

Let a finite group \(G\) act by measure-preserving maps on a measure space, and set \(H(A)=\bigcup_{g\in G}gA\) for a measurable set with \(0<\mu(A)<\infty\). Then \(A\subseteq H(A)\), \(H(H(A))=H(A)\), and

\[ 0\leq\log\mu(H(A))-\log\mu(A)\leq\log|G|. \]

The first two assertions follow from the identity and group composition; monotonicity and finite subadditivity of measure give \(\mu(A)\leq\mu(H(A))\leq\sum_g\mu(gA)=|G|\mu(A)\). Nevertheless the hull's output does not depend on input volume alone. For a four-point set with counting measure, let \(G=\{1,\sigma\}\) where \(\sigma=(1\ 2)(3\ 4)\). The sets \(A=\{1,3\}\) and \(B=\{1,2\}\) each have volume \(2\), but \(H(A)=\{1,2,3,4\}\) has volume \(4\) while \(H(B)=B\) has volume \(2\). This finite example isolates the difference between an invertible map of numbers and a set operation that needs more than a number as input. It is a candidate for a tiny machine-checked toy lemma; it does not identify IUT's holomorphic hull with this orbit union or give any of IUT's required comparisons.

The next exact obligations

  1. Type every real-number copy and comparison in Step (xi), including which normalizations and \(j^2\) factors each uses. Match the hosted IUT III passage to the version cited in the 2018 objection.
  2. Try to justify the report's \(\exists S,\eta_q=\eta^{\mathrm{anab}}_S\) using the paper's actual input/output maps and Ind1–3. If it cannot be typed or established, record the first failing comparison rather than writing a success-shaped substitute.
  3. Note IUT III's explicit \(Q>0\) in the Corollary 3.12 proof (p. 173); if auditing its derivation from the initial data, supply that premise separately. Decide whether the two cited inequalities concern the same normalized \(T,Q\).
  4. Spell out the allowed (Ind1)–(Ind3) choices and the precise quantifier over transports: a chosen map, all maps, or an invariant statement on a quotient. Check which preservation or bound Step (xi) supplies; do not import one from the toy model.
  5. Give both sides this same typed statement. A genuine reconciliation needs a reviewed derivation under agreed hypotheses; a disagreement about types or admissible maps is an explicitly recorded stopping point, not a vote.

The independent-reading procedure and its results belong in 09a. The provisional ledger above should be refined when exact domain/codomain passages have been checked.