A · Source corrections and concordance

0 NAG Tags

A.1 Corrections and qualifications of source statements

This section gives the mathematical evidence for the source qualifications used in the main text. It concerns the versions listed in Section 0.2.2, and it does not claim an audit of every page of the 608-page text. The counterexamples below address literal statements; the separate remarks about incomplete arguments do not assert that their conclusions are false. Return to Section 11.2, Section 13.1, Section 19.3 for the statements used in the lecture.

Weak positivity does not imply order zero

Example A.1.1. (A weakly positive current of order one). On R4\mathbb R^4, write a symmetric (2,2)(2,2)-test form in the matrix normalization α=−∑∣I∣=∣J∣=2HIJ(x) d′xI∧d′′xJ,HIJ=HJI.\alpha=-\sum_{|I|=|J|=2}H_{IJ}(x)\,d'x_I\wedge d''x_J, \qquad H_{IJ}=H_{JI}. Define the coefficient functional ℓ(H)=H12,34−H13,24+H14,23.\ell(H)=H_{12,34}-H_{13,24}+H_{14,23}. If H=aξξTH=a\xi\xi^{\mathsf T} is a strong generator, the two-covector ξ\xi is decomposable. Its Plücker relation is ξ12ξ34−ξ13ξ24+ξ14ξ23=0.\xi_{12}\xi_{34}-\xi_{13}\xi_{24} +\xi_{14}\xi_{23}=0. Thus ℓ(H)=0\ell(H)=0 for every strongly positive form, pointwise in xx. Define T(α)=∂∂x1ℓ(Hα(x))∣x=0,\begin{equation} T(\alpha)=\left.\frac{\partial}{\partial x_1} \ell(H_\alpha(x))\right|_{x=0}, \end{equation} and let TT vanish on antisymmetric forms. This is a continuous distribution of order at most one. It is symmetric, and it vanishes on every strongly positive test. Both TT and −T-T are therefore weakly positive.

Choose a constant symmetric form α0\alpha^0 with ℓ(Hα0)=1\ell(H_{\alpha^0})=1, and a smooth compactly supported function χ\chi equal to one near 00. The tests αj=j−1sin⁡(jx1)χ(x)α0\alpha_j=j^{-1}\sin(jx_1)\chi(x)\alpha^0 have one fixed compact support and coefficient suprema tending to zero, whereas T(αj)=1T(\alpha_j)=1 for every jj. Hence TT is not of order zero.

The same phenomenon is recorded in [BGJK] (Examples 2.3.6 and 5.2.8). It directly contradicts the weak-positive assertion in [CLD25] (Proposition 7.2.9, p. 162). The proof there uses the relative-interior statement for the strong cone in Proposition 1.2.13 as if that cone could dominate every symmetric form. Example 7.4.3 shows that the required domination is impossible.

The replacement used in this book is Proposition 9.1.7, for ordinary positive currents. In degrees (1,1)(1,1), (n−1,n−1)(n-1,n-1), and the extremal degrees, the coincidence of the relevant cones removes this particular middle-degree obstruction. The counterexample must not be cited as a failure of order zero for ordinary positive (1,1)(1,1)-currents.

Return to positivity and continuous multipliers.

A locally integrable distribution does not fix point values

Remark A.1.2 (The literal implication in Proposition 8.3.2). The statement in [CLD25] (Proposition 8.3.2, pp. 201–202) quantifies over functions on a convex real open set which are not identically −∞-\infty. Its third condition is local integrability and positivity of d′d′′[u]d'd''[u], and it concludes pointwise finite convexity. The function in Remark 11.2.9 satisfies the third condition and fails the conclusion. The convolution step in the printed proof also invokes uniform convergence to this arbitrary function, which does not follow from local integrability.

For continuous functions, Proposition 11.2.8 gives a complete repair. Theorem 11.3.2 supplies the full Lloc1L^1_{\rm loc} repair: there is a unique finite convex representative, and convolution converges locally uniformly to that representative. This distinguishes the representative from arbitrary modifications on sets of measure zero. Adding upper semicontinuity alone does not repair the printed statement, as the same upward-point-spike example shows.

Return to the corrected continuous comparison.

Multiplication need not preserve a closed current

Example A.1.3. (Failure of the scalar multiplication assertion). In the tropical line X=RX=\mathbb R, let T=[1]T=[1] be the integration current of bidegree (0,0)(0,0). It is positive and closed and has the Bedford–Taylor property, as also stated in [CLD25] (Example 8.6.4). Let u(x)=1+x2.u(x)=1+x^2. This is a smooth positive function, but d′(uT)=[2x d′x]≠0.d'(uT)=[2x\,d'x]\ne0. Thus uTuT is not closed and cannot be a Bedford–Taylor current under [CLD25] (Definition 8.6.2). This disproves Corollary 8.6.7, p. 208, literally.

Even deleting closedness from that definition would not fix the claimed preservation of potential independence. Indeed, a(x)=xa(x)=x and b(x)=0b(x)=0 are smooth psh functions with the same zero Hessian. Nevertheless, d′d′′(a uT)=[6x d′x∧d′′x]≠0=d′d′′(b uT).d'd''(a\,uT)=[6x\,d'x\wedge d''x]\ne0=d'd''(b\,uT). The correct smooth multiplication statement is Proposition 8.6.5 of the source: the factor is a positive closed smooth form. A general positive continuous scalar does not satisfy the closedness requirement.

Return to the continuous recursive construction.

The successor step in the generated-sheaf comparison

Remark A.1.4 (The missing implication in Proposition 8.1.11). CLD25, Proposition 8.1.11, pp. 197–198, claims C25,α=C0∩P25,αfor every ordinal α.\mathcal C_{25,\alpha}=C^0\cap\mathcal P_{25,\alpha} \quad\text{for every ordinal }\alpha. At a successor stage, the right side permits a continuous endpoint uu which is locally a decreasing limit of uj∈P25,αu_j\in\mathcal P_{25,\alpha}. The approximants uju_j are not assumed continuous. The printed proof instead compares decreasing and uniform limits of members of C25,α\mathcal C_{25,\alpha}. The induction hypothesis identifies these with the continuous members of P25,α\mathcal P_{25,\alpha}; it does not make every uju_j continuous.

Dini’s theorem for continuous approximants proves the first-stage equality in Proposition 11.2.3. It does not supply the missing continuous approximants at an arbitrary stage. Even a uniform convergence conclusion for the original upper semicontinuous approximants would not place each of them in the continuous stage. An additional replacement or approximation argument is required.

We have no counterexample to the stated equality. The usable conclusions include the first-stage equality, the inclusions in Proposition 11.2.4, and the same-ordinal forward inclusion of Proposition 11.6.6. Theorem 11.6.4 handles a hierarchy on one fixed compact set, but transferring sheaf-generated sections to that hierarchy requires the extra representation in Corollary 11.6.5. Proposition 11.6.7 verifies the neighbouring L1L^1 comparison independently of the continuous gap. The independent PL synthesis later in the text supplies an equality on that specified class without invoking the disputed successor step. Section 11.7.4 supplies a sufficient bounded-stability criterion, and Corollary 11.7.6 applies the comparison inputs of [Wu26] in the projective envelope category. This is a scoped source-dependent application; it does not repair the original proof on every analytic space.

Remark A.1.5 (The stationarity wording). Lemma 8.1.4 of CLD25, p. 196, says that the ordinal hierarchy is eventually stationary, and its proof invokes a cardinality argument. Remark 8.1.6 on the same page says that stationarity is not known. Read without a qualification on the ordinal, these sentences conflict. A hierarchy of subobjects inside a fixed set of local sections cannot grow strictly through all ordinals; the eventual set-theoretic stabilization and the question whether the first stage suffices are different matters. The lecture does not choose an unstated ordinal bound on the authors’ behalf. It uses only the minimal-sheaf definition and the explicitly proved first-stage comparison.

Return to the generated-sheaf definition.

Normalized residual sections.

The literal corrections to CLD25 Lemma 17.4.10, Corollaries 17.4.11–17.4.12 and Proposition 17.8.2 are recorded, with the corrected proof and a one-point test, in Lemma 19.18.13 and Remark 19.18.14. The semiample conclusion is retained with its weighted maximum.

A fixed neighbourhood is part of an approximation statement

Remark A.1.6 (Proposition 17.8.4 and a repair in a common category). The proof of [CLD25] (Proposition 17.8.4, pp. 497–498) perturbs a nef residual class by a rational multiple of an ample class. Applying Proposition 17.8.2 to each perturbation gives psh on a neighbourhood of the chosen point. The argument then takes a uniform limit on one fixed neighbourhood. The printed application alone does not show that the psh neighbourhood can be chosen independently of the perturbation.

The logical point is elementary. On a real interval, let vε(t)=εt2−t4.v_\varepsilon(t)=\varepsilon t^2-t^4. For every ε>0\varepsilon>0, this function is convex near 00, since vε′′=2ε−12t2v_\varepsilon''=2\varepsilon-12t^2. On a fixed compact interval it converges uniformly to −t4-t^4, which is not convex on any neighbourhood of 00. Thus positivity on varying neighbourhoods is insufficient for the claimed limiting step.

In the good strictly analytic, nontrivially valued setting at interior points, [GJR] (Proposition 7.10) supplies one fixed neighbourhood and a uniform family of locally Fubini–Study approximants. Theorem 19.18.17 expands its residue-model construction: a single reduction tube works for every perturbation. Proposition 19.18.16 isolates the exact compatibility used. The PL synthesis uses this fixed-domain argument. This is an independent repair under its stated hypotheses, not a counterexample to Proposition 17.8.4. The graded construction of Proposition 19.18.20 and Theorem 19.18.22 now gives the integral PL version over arbitrary complete fields.

Scope and indexing in the residual comparison

Remark A.1.7 (What is and is not established about Theorem 18.8.4). We did not obtain a counterexample to [CLD25] (Theorem 18.8.4) or a refutation of its entire proof. Its claim is stated in Remark 19.3.1. The following local problems are identifiable.

  1. On p. 539, the residual numerical criterion for a product of dd Chern classes is expressed as a zero-cycle degree. The testing subvarieties must therefore have dimension dd. The printed reference to codimension mm does not give that dimension in general; later in the proof, the subvarieties are correctly taken of dimension dd.

  2. In the same proof, the rational-domain inequalities formed from fiεif_i^{\varepsilon_i} require the right-hand side ∣fi(x)∣εi|f_i(x)|^{\varepsilon_i}. Omitting the exponent on that side fails to make the chosen point satisfy the equality when εi=−1\varepsilon_i=-1 and ∣fi(x)∣≠1|f_i(x)|\ne1.

  3. Corollary 18.8.8, p. 541, states its equivalence for integral PL metrics, whereas its cited implication from Proposition 17.8.4 has a strictly integral PL hypothesis. The distinction is made in the source’s definitions; a reduction and descent argument is needed to use that implication in the larger category.

The first two items are incorporated in the reconstruction of Theorem 19.18.12. For the third, Theorem 19.18.15 treats arbitrary integral PL metrics with semiample residue, and Corollary 19.18.18 treats nef strictly integral PL metrics over a nontrivially valued field. Theorem 19.18.22 completes the general integral PL implication using prescribed graded premodels over the original field; Corollary 19.18.23 gives the resulting pure-dimensional comparison. None of them authorizes replacing the theorem by a general statement about arbitrary continuous metrics.

Remark A.1.8 (A bidegree/bidimension slip in the CLN estimate). In [CLD25] (Proposition 8.4.1, p. 202), the current is called of bidegree (p,p)(p,p) while the test has degree (p−q,p−q)(p-q,p-q) after qq curvature factors have been inserted. The displayed pairing has the advertised degree when pp is the bidimension of the initial current: its resulting bidegree is (n−p+q,n−p+q)(n-p+q,n-p+q). This is an indexing correction; the lecture uses the explicit complementary test degrees throughout.

Return to the PL source comparison.

Corrections in the other sources

Remark A.1.9 (The 2012 source is not an error-free reference version). The earlier text also needs qualifications. Its commutativity exponents in §1.2.6 and §3.1.4 conflict with the explicit total-degree algebra; Remark 7.1.5 gives the adopted correction. The real strong-cone decomposition in Lemma 5.2.3 is disproved by Example 7.4.3. The proof of the analytic Lemma 5.3.3 relies on that decomposition and requires a separate repair; the real example alone is not a counterexample satisfying every analytic hypothesis. Its continuous facewise-convexity criterion in Lemma 5.5.3 misses the corner mass; [GK19] (Remark 3.2 and Example 3.3) supplies the correction. The general pullback sentence following Definition 5.5.1 must be distinguished from the proved pullback result on a specified approximation class.

Remark A.1.10 (Two cautions about the GK positivity text). The incorrect real strong-span statement is repeated in [GK19] (Remark 1.12(i)). The analytic argument of Lemma 2.16 uses the affected decomposition; this identifies a proof obligation, rather than by itself disproving the analytic conclusion. Also, the slope-only criterion in Example 8.6 needs the edgewise Hessian condition for a piecewise smooth function. The smooth function −t2-t^2 on an interval has zero sum of its opposite outgoing first derivatives at each interior point and negative second derivative. The full measure formula in Proposition 16.2.7 retains both the edge density and the vertex masses.

Remark A.1.11 (A substantive correction already acknowledged in GJR). GJR, Remark 16.4, explicitly says that the argument in [GK17] (Examples 8.15 and 9.17) used an incorrect smoothness assertion for canonical tropicalizations of abelian varieties. Its construction in §17 shows that these maps need not be original CLD-smooth. The argument that canonical metrics are delta-metrics in that original sense consequently does not follow. This corrects the argument; it does not prove that the metrics fail to be delta-metrics. The distinction is important when using canonical metrics as examples of an enlarged calculus.

Remark A.1.12 (Elementary corrections in GJR v3). In GJR §10.6, pp. 63–64, the commutativity exponent should be (p+q)(p′+q′)(p+q)(p'+q'), and the last operator in the d′d' Leibniz rule should be d′d'. The printed alternatives fail respectively on two independent (1,0)(1,0)-forms and by bidegree counting. In the proof of Proposition 10.17, the affine Taylor polynomial is g(ν)+Dgν(ω−ν)g(\nu)+Dg_\nu(\omega-\nu); Proposition 17.2.7 uses this corrected expression. These are local corrections to formulas, not counterexamples to the corresponding theorems.

A.2 Additional source checks and explicit repairs

The regularized maximum in Proposition 16.9.1

Status: local formula error, with a repair.  The displayed kernel in CLD25, p. 466, integrates max⁡i(si+ti)∏i=1mθ(ti/ε) dt1⋯dtm\max_i(s_i+t_i)\prod_{i=1}^m\theta(t_i/\varepsilon)\,dt_1\cdots dt_m with ∫θ=1\int\theta=1, but omits the factor ε−m\varepsilon^{-m}. Its total kernel mass is εm\varepsilon^m. Consequently, adding a constant aa to every sis_i changes the displayed expression by aεma\varepsilon^m, rather than by aa. It cannot have the claimed translation behaviour and approximation bounds in that form.

A corrected regularized maximum is Mε(s)=∫Rmmax⁡i(si+εti)∏i=1mθ(ti) dt,\begin{equation} M_\varepsilon(s)=\int_{\mathbb R^m}\max_i(s_i+\varepsilon t_i) \prod_{i=1}^m\theta(t_i)\,dt, \end{equation} where θ\theta is a smooth nonnegative even kernel supported in [−1,1][-1,1] and of integral one. Convolution makes the expression smooth; convexity and coordinatewise monotonicity are preserved. Jensen’s inequality and the zero mean give max⁡isi≤Mε(s)≤max⁡isi+ε.\max_i s_i\leq M_\varepsilon(s)\leq\max_i s_i+\varepsilon. If one coordinate lies more than 2ε2\varepsilon below the maximum, it never attains the perturbed maximum and the expression is independent of that coordinate. This last property is what permits logarithmic branches vanishing identically at some points to be discarded locally.

The domain and sublevel qualifications in the same proof

Status: local domain and statement qualifications.  The proof uses logarithms log⁡∣fi/g∣\log|f_i/g|. The condition that the fif_i and gg have no common zero does not prevent all fif_i from vanishing at a point where g≠0g\ne0. At such a point every logarithmic branch is −∞-\infty, so the real smoothing formula is not directly defined. Add a finite constant branch below the desired sublevel threshold. The maximum is then finite; branches sufficiently far below it can be discarded. On a neighbourhood where a remaining logarithmic branch is active, its analytic function is nonzero and supplies a legitimate unit coordinate.

The intended barrier conclusion also needs two neighbourhoods: a smooth psh function on a larger domain, negative on a smaller neighbourhood of the point, with a compact nonpositive sublevel contained in the larger domain. The literal wording “negative on UU and {ρ≤0}\{\rho\leq0\} compact” makes the sublevel equal to all of UU. The proof itself chooses a shifted function that is negative near the chosen point, not necessarily everywhere on its domain.

These corrections concern the local construction used to apply the polar-negligibility theorem. They do not constitute a counterexample to the existence of suitable local psh barriers. The normalized smoothing and constant branch supply the appropriate mechanism in the good analytic setting of the lecture.

The hypothesis transfer in Corollary 8.11.3

Status: false in the unrestricted abstract tropical category.  Proposition 8.11.1 assumes a skeletal current with the Bedford–Taylor property. Corollary 8.11.2 retains a skeletal current locally expressible as a difference of such currents. Corollary 8.11.3 is then stated for an arbitrary positive closed current TT and a convex function pulled back by a harmonic chart. Its proof cites the preceding corollary to identify the BT product with multiplication by a G-form.

The cited statement alone does not supply that identification for arbitrary TT. One needs an extension theorem, a decomposition preserving the necessary hypotheses, or an independent positivity argument for the harmonic pullback. The skeletal/BT assumptions cannot simply disappear between the two statements. The cross calculation in Section 14.2 gives a counterexample in an explicitly verified abstract tropical category. It does not claim a counterexample on a smooth Berkovich space.

The ordinary continuous-generated-potential construction in Chapter 13 has an independent proof by positive continuous multiplication and closure under uniform limits. It does not need this unrestricted harmonic-coordinate implication. When a theorem uses the latter implication specifically, its additional justification must be supplied.

Local corrections in the higher-codimension formulas

Status: formula and index corrections.  In the setup of §18.7.1, the complete intersection YY has dimension n−rn-r, so the dimension bound at its points must be compared with n−rn-r, not with the codimension rr. In the truncation formula, the correct identity is log⁡max⁡(ε,∣f1∣,…,∣fr∣)=max⁡(log⁡ε,u),u=log⁡max⁡i∣fi∣.\log\max(\varepsilon,|f_1|,\ldots,|f_r|)=\max(\log\varepsilon,u), \qquad u=\log\max_i|f_i|. Writing log⁡max⁡(ε,u)\log\max(\varepsilon,u) is not meaningful when uu is negative and is not equal to the left side. On a compact support away from YY, uu has a finite lower bound; it need not be strictly positive. Proposition 18.7.3 likewise uses relative dimension n−rn-r in its proof, rather than the printed rr in its introductory statement.

The Levine calculation also provides a simple independent sign check: on the complement of the cycle, d′d′′h=α−β\mathrm d'\mathrm d''h=\alpha-\beta, so d′d′′(hω)=αr−βr\mathrm d'\mathrm d''(h\omega)=\alpha^r-\beta^r. The sign of the singular term is then fixed by Poincaré–Lelong. The final identity is the one stated in Theorem 18.7.5 and displayed in Chapter 22. A contradictory intermediate sign must be corrected locally rather than propagated through the calculation.

Sections 22.4–22.7 give the extension and fibre-comparison arguments with their geometric inputs stated. There is also a separate commutative-algebra issue in the proof of Proposition 18.7.3: a regular sequence alone does not put a neighbourhood of its zero set in Serre’s (Sr)(S_r) locus. Example 22.7.2 disproves that unrestricted inference. The relevant finite-flat argument can instead be made at the Abhyankar points, where dimension and depth coincide. This repairs the needed mechanism in the stated category without asserting that the entire original Green identity is false.

A correction acknowledged in the canonical-measure literature

Status: author-acknowledged gap in an earlier auxiliary claim.  In the introduction to Non-archimedean canonical measures on abelian varieties, Gubler explicitly withdraws an earlier purity claim for the tropical image in the general abelian setting, after a referee identified a gap. The corrected theorem allows polyhedral pieces of different dimensions and retains the canonical-measure formula. The totally degenerate case has the expected pure dimension in that result.

This correction is useful because it shows exactly what changes. One must not use the earlier purity claim in a dimension argument for mixed reduction. One may still use the corrected measure theorem with its actual dimension range. The correction says nothing by itself about the present-day status of every related purity question.

Closedness in the radial mass formula

In the consulted arXiv:1008.2856v1, Proposition 3.4 of Lagerberg states radial trace-mass monotonicity for a weakly positive current without listing closedness. The Stokes calculation in its proof needs closedness. Example 9.13.3 gives a counterexample to the unrestricted wording: a nonnegative compactly supported smooth scalar on R\mathbb R has a constant ball mass for large radii, so the mass divided by the radius decreases. The corrected result, including its regularized proof, is Theorem 9.13.2.

Corollary 3.5 of the same source does include closedness and ordinary positivity. These assumptions have separate roles. Closedness supplies the ball-growth estimate; ordinary positivity supplies coefficient domination by the trace. The distinction between weak and ordinary positivity is already explicit in the source’s Remark 3.3. The correction here is pinned to the consulted preprint wording, without asserting that every later printed version has the same omission.

Local domains and degrees in the toric paper

Three local repairs concern the consulted arXiv:2102.07392v1. In Example 2.3.5, the expression T↦(T,T−1)T\mapsto(T,T-1) is used as a map from Gm\mathbb G_m to Gm2\mathbb G_m^2. Its second coordinate is not invertible at T=1T=1, so this map is defined only off that point. The invariant function in the example is nevertheless original smooth: use that chart away from T=1T=1, and use the constant zero expression near T=1T=1, where ∣T∣=1|T|=1. Section 21.17 gives the local calculation. Thus the example’s conclusion survives the domain correction.

In Proposition 4.3.2 and Remark 4.3.4, some displayed factors lack the d′d′′d'd'' operators required by their assigned degrees. The differentiated product and the companion product with one scalar factor must be written separately. Proposition 21.21.1 records both formulas and their target bidegrees. Applying d′d′′d'd'' changes the leading scalar factor into a curvature factor; it is not an operation that leaves the bidegree unchanged.

Finally, positivity of the differentiated product does not imply positivity of the companion product when its undifferentiated scalar can be negative. Taking that scalar to be −1-1 and the remaining positive product nonzero gives a negative current. The companion product is defined without a sign restriction, but its positivity requires an additional nonnegativity assumption on that scalar. These qualifications leave the correspondence and product constructions available with their correct statements.

Upper semicontinuity in the classical maximum proof

The proof of Gubler–Rabinoff, Theorem 6.4 in arXiv:2506.13548v2, invokes continuity to say that the maximum set of a classically psh function is closed. The definition only provides upper semicontinuity. That is sufficient: if M=max⁡XuM=\max_Xu, the set {u=M}={u≥M}\{u=M\}=\{u\geq M\} is closed for an upper semicontinuous function. The rest of the open-and-closed argument is unchanged. The component reduction in the same proof is a finite cover by irreducible components, whose intersections may be nonempty. The reconstructed proof in Theorem 19.5.2 uses precisely these two elementary formulations.

This is a hypothesis slip in an intermediate sentence, with an immediate repair. It has a different mathematical status from the weak-positive order-zero assertion disproved above. Recording the distinction prevents a local wording problem from being presented as a failure of the classical psh theory.

A.3 A compact register of conclusions and dependencies

Source location Assessment Usable replacement or remaining obligation
CLD12 §1.2.6, §3.1.4 Sign error Use total-degree graded commutativity.
CLD12 Lemma 5.2.3; proof of 5.3.3 False real statement; analytic proof obligation Use ordinary positive forms for coefficient domination; the Plücker example gives the obstruction.
CLD12 Lemma 5.5.3, continuous form False literally Retain corner mass; use the smooth version or an explicit Hessian-plus-corner criterion.
CLD12 after Definition 5.5.1 General pullback not justified by the definition Use smooth or specified approximable potentials, or a separate functorial theorem.
CLD25 Proposition 7.2.9, weak-positive part False literally Ordinary positivity gives order zero; weak positivity in middle degree need not.
CLD25 Proposition 8.3.2 False pointwise statement Use continuous functions or distinguish an L1L^1 class from its convex representative.
CLD25 Corollary 8.6.7 False literally Positive continuous scalars preserve positivity, but not closedness or the BT kernel property.
CLD25 Proposition 8.1.11 Successor-step gap Indexed inclusion and fixed-compact barrier theorem proved; sheaf globalization remains. The neighbouring L1L^1 equality is valid.
CLD25 Lemma 8.1.4 / Remark 8.1.6 Conflicting wording Separate eventual set-theoretic stabilization from first-stage stabilization.
CLD25 Proposition 8.4.1 Degree slip Use bidimension pp for the displayed test degree.
CLD25 Corollary 8.11.3 False in the stated abstract generality The cross example disproves the arbitrary-current assertion; retain the proved skeletal/BT hypotheses.
CLD25 Lemma 9.5.4 False in the stated abstract generality The cross-times-line example has mass off the graph; the affine case remains valid.
CLD25 Proposition 16.9.1 Kernel normalization and local-domain issues Use (A.2.1), a constant branch, and nested barrier neighbourhoods.
CLD25 Proposition 17.8.4 Fixed-neighbourhood step missing GJR Proposition 7.10 gives an independent repair in its specified analytic category.
CLD25 Theorem 18.8.4 and Corollary 18.8.8 Local corrections; graded PL bridge supplied Theorem 19.18.22 replaces the missing strictness transfer; arbitrary continuous weights remain separate.
CLD25 §18.7 Local formula/index/sign corrections Use the correct truncation and dimensions; retain the extension theorem and regular-sequence hypotheses.
GK positivity Remark 1.12 / proof of Lemma 2.16 Real error; analytic proof obligation Ordinary-positive replacement.
GK positivity Example 8.6 Incomplete for piecewise smooth functions Include edgewise Hessian densities as well as vertex slopes.
GK canonical examples; GJR Remark 16.4 Acknowledged invalid smoothness argument Canonical tropicalizations can fail original smoothness; original delta-metric membership is a separate question.
GJR §10.6, Proposition 10.17 proof Local algebra/Taylor slips Correct total degree, differential operator, and constant term.
Gubler canonical-measure introduction Acknowledged earlier purity gap Use the corrected mixed-dimensional theorem.
CLD25 Proposition 18.7.3 proof Regular-sequence inference insufficient Example 22.7.2; use the local dimension and Cohen–Macaulay inputs in Section 22.7.
GJR v3 §17.2 and Proposition 17.4 proof Cycle ordering and slope-list slips The consistent theta-graph calculation and residual obstruction are in Section 17.8.
Lagerberg v1 Proposition 3.4 Closedness missing from literal statement The counterexample and corrected radial mass proof are in Section 9.13.
BGJK25 v1 Example 2.3.5 Torus-chart domain too large Exclude T=1T=1 from the unit chart and use a constant expression near it; Section A.2.7.
BGJK25 v1 Proposition 4.3.2 and Remark 4.3.4 Differential degrees and positivity of a scalar factor Retain DDDD on the curvature factors; a companion product needs a nonnegative scalar to be positive. See Proposition 21.21.1.
GRPSH v2 Theorem 6.4 proof Continuity invoked unnecessarily Upper semicontinuity closes the maximum set; Section A.2.8.

The register is an audit of identified statements in fixed versions, not an assertion that every page of every source has been independently verified. The most consequential repairs have complete elementary evidence in this volume. The missing-step entries identify precise obligations for a proof reconstruction and do not assert the negation of the theorem.

Which arguments remain usable after a correction

The false strong-span lemma affects any proof that bounds all coefficients by a strongly positive control form in middle degree. Replacing that step by ordinary-positive domination repairs the continuous multiplier argument used here. It does not automatically repair every theorem involving a different positive cone.

The scalar-BT counterexample prevents treating the class of BT currents as a module under arbitrary positive continuous functions. It does not obstruct the definition of uTuT as an order-zero current or the construction of d′d′′(uT)\mathrm d'\mathrm d''(uT) for an admitted psh potential and a positive closed background.

The successor-stage gap prevents using the claimed equality of all continuous generated stages as a shortcut to smooth approximation. It does not obstruct direct arguments using the continuous generated sheaf, nor the independent PL comparison proved through numerical model approximation and curve tests.

The fixed-neighbourhood gap affects a particular limiting argument. A separate theorem producing approximants on one common neighbourhood repairs the implication in its own category. This is why GJR’s model construction is mathematically useful even if the intended CLD conclusion is correct.

A.4 Current estimates and curvature comparisons

Why “replace smooth by psh” is not a definition

A differential algebra has products, two first-order differentials, pullbacks, and relations between local representatives. A psh function is generally not differentiable in the ordinary sense. Even when its Hessian is a positive current, its first derivatives and products with singular forms require definitions. The class is also not closed under negation, so it cannot simply replace the coefficient ring of a differential algebra.

The BT construction supplies one carefully defined operation: an admitted potential acts on a positive closed current by d′d′′(uT)\mathrm d'\mathrm d''(uT). It does not immediately supply every product d′u∧αd'u\wedge\alpha, every pullback of a current, or an algebra of arbitrary singular forms. Extending GK by psh coefficients therefore requires a choice of domain for all operations and a proof that the relations between representatives are respected.

One coefficient estimate, written out

On a fixed real chart, identify a symmetric test of degree (q,q)(q,q) with its coefficient matrix H(x)H(x) in the chosen sign normalization. For ordinary positivity, choose a positive definite control matrix II on the finite coefficient space and a cutoff equal to one near the support. If ∥H(x)∥op≤M\|H(x)\|_{\mathrm{op}}\leq M, then MI+H(x)≥0,MI−H(x)≥0.MI+H(x)\geq0,\qquad MI-H(x)\geq0. Pairing with an ordinary positive current gives a bound on the coefficient functional by its value on the fixed positive control form. Localizing and summing over finitely many charts gives the corresponding bound for the actual test.

The matrix argument uses the ordinary cone, which has full-dimensional interior in the symmetric space. For the strong cone in middle degree, the Plücker functional in the audit vanishes on every generator, so no strong control form can dominate every symmetric direction. This is the exact point where the superficially similar proof fails.

How to pass from smooth scalars to continuous scalars

Fix a positive test form β\beta. The functional a↦T(aβ)a\mapsto T(a\beta) on smooth compactly supported scalar functions is positive and bounded by a sup norm on each fixed compact support. It therefore extends to a Radon measure. For a general symmetric test, write it as a difference of ordinary positive tests and use the resulting signed measures. Compatibility on overlaps makes uTuT well-defined for continuous uu.

This construction does not identify the current with one global matrix of measures in a fixed tropical coordinate system. It builds compatible scalar measures for actual tests. That distinction is what lets the definition survive chart refinement.

The two-error convergence estimate

Suppose Tj⇀TT_j\rightharpoonup T through positive currents and uj→uu_j\to u locally uniformly. For a fixed test η\eta, first split ujTj−uT=(uj−u)Tj+u(Tj−T).u_jT_j-uT=(u_j-u)T_j+u(T_j-T). The first term is small by the uniform scalar bound and local boundedness of the positive masses. For the second, approximate uu uniformly by a smooth scalar aa. The term a(Tj−T)a(T_j-T) tends to zero on the fixed test by weak convergence, while the two remaining errors are bounded by ∥u−a∥\|u-a\| times the same mass control. This proves convergence without requiring derivatives of uju_j or a common chart for all their approximants.

Total mixed mass and properness

For smooth semipositive metrics on fixed line bundles over a proper nn-fold, the mixed degree formula gives the total mass. For globally uniform limits, weak convergence preserves this mass because the constant test 11 is allowed on the compact space. Local uniform convergence on a noncompact domain only controls compactly supported tests. The sequence δj\delta_j on R\mathbb R has mass one but converges to zero on those tests.

Thus the global mass formula is a consequence of a global comparison and compactness. It is not evidence for an additional local regularization theorem. This distinction answers the recurring question whether a mass or total-variation lemma is a direct consequence of established results: once the measure construction and degree comparison are available, the elementary estimate often is direct; establishing those inputs is the substantial part.

Restriction to a subspace and ambient extension

If a function belongs to a functorial positive approximation class, restriction to an analytic subspace can be obtained by restricting the approximants, provided the morphism and test theory are in the admitted category. The converse problem is different: a positive function or metric on a subspace need not extend positively to an ambient neighbourhood by ordinary extension of continuous functions.

An ambient extension theorem must control curvature in transverse directions and maintain the line-bundle transition data. If it is obtained from global sections or an FS construction, the generation and positivity hypotheses do the work. A proof that only extends the underlying continuous function has not proved a positive extension theorem. Thus a positive extension theorem requires both an extension of the data and control of its curvature.

A.5 Further version-specific corrections

Remark A.5.1 (CLD25 Corollary 16.8.6). The printed tropical Hausdorff exponent is the number of curvature factors. The bidimension of their product instead gives the complementary exponent. Example 19.17.2 disproves the printed statement on Gman\mathbf G_m^{\mathrm{an}} using one Gauss point; Proposition 19.17.1 proves a replacement for admitted differences over a nontrivially valued field. For PL metric weights, Proposition 19.18.8 additionally treats trivial valuation. The full transfinite source potential class is not identified here.

Remark A.5.2 (BMP26 Example 6.23). The negative-axis branch of the printed obstacle is discontinuous at −ϵ-\epsilon. Proposition 16.7.1 supplies a continuous correction, proves its envelope formula and computes all atoms of its measure. The example still disproves unrestricted orthogonality in that polyhedral category.

Remark A.5.3 (Cai and Mihatsch: a convention difference). Cai writes d′=dP′+∂′d'=d'_P+\partial', while Mihatsch writes d′=dP′−∂′d'=d'_P-\partial'. Section 15.9 gives the translation. This is not a false curvature formula in either source. The distinction is recorded separately from mathematical errors.

For the remaining CLD25 residual argument, Section 19.18 separates noncharging and the formal positivity inference from residual realization and the calibrated fibre formula. Theorem 19.18.12 completes this assembly in pure dimension, using explicitly cited lifting, integration and graded-reduction theorems; those inputs are not independently re-proved. Similarly, Lemma 11.6.1 sharpens the diagnosis of Proposition 8.1.11: upper semicontinuity suffices for uniform convergence on compact sets, but does not replace discontinuous predecessors by continuous ones.

Remark A.5.4 (The inherited negligibility exponent). CLD25 Proposition 18.3.7 repeats the exponent of Corollary 16.8.6. The metric form of the Gauss-point counterexample refutes this printed assertion too; see Section 19.18.3. The corrected exponent is Hn−p\mathcal H^{n-p}; Proposition 19.18.8 verifies the PL-metric scope over every complete base field.

Remark A.5.5 (Base-point normalization). In CLD25 paragraph 18.8.3, the residual bundles M~i\widetilde M_i belong to the base germ (Y,y)(Y,y). Formula (19.18.2) places the coefficient there, with its multiplicity and defect, and retains the tensor factor q−dq^{-d}. This is a correction of the displayed notation using Proposition 18.6.1, not a new fibre-integration theorem.

The prescribed graded comparison is proved in Proposition 19.18.20. Together with graded semiampleness and the fixed-tube criterion, it yields Theorem 19.18.22. This supplies the integral PL bridge while leaving the general continuous predecessor problem unchanged.

For CLD25 Proposition 8.1.11, Theorem 11.6.4 supplies a fixed-compact selection theorem but not the required sheaf globalization. The valid indexed inclusion is Proposition 11.6.6. The neighbouring Proposition 8.1.12 uses inherited local integrability and is proved in Proposition 11.6.7; it does not share the missing continuity step.

A.6 The local envelope comparison in Wu’s preprint

Remark A.6.1 (A negative barrier and the background metric). In [Wu26], Proposition 3.28, p. 22, the displayed Fubini–Study maximum includes the branch log⁡∣s0∣\log|s_0| with weight zero. Its local weight in the frame s0s_0 is therefore at least zero, whereas the next step requires v(x)<0v(x)<0 and uses negative sublevel sets. Formula (11.7.4) retains a negative constant branch and is consistent with those sublevels. The proof also changes from the initial background ω\omega to a constructed background θ\theta on the same line bundle. These backgrounds need not be equal. Proposition 11.7.3 records θ=ω+d′d′′b\theta=\omega+d'd''b and the compensating global term bb; Lemma 11.7.4 supplies a fixed collar for the decreasing tail. These corrections concern the displayed construction, not a counterexample to the proposition’s conclusion.

Remark A.6.2 (Version matching before importing the comparison). Wu’s Appendix A cites prepublication Propositions 5.20 and 5.22 of the work called APW26b there. The fixed October version [APW26] used in this book has a Lemma 5.20 instead; localization and local PL approximation are treated in Propositions 5.19 and 5.21. Thus the printed locators cannot simply be copied into a reference to the October version. Appendix A also uses Theorem 6.13 of a second, mixed-calculus preprint. The comparison in Corollary 11.7.6 is attributed to Wu’s Theorem A.3; it is not presented as a newly completed proof from those external ingredients. The content-level matching and a uniform bounded-family repair are now given in Section A.7; the mixed-calculus theorem and full common-domain passage remain unverified here.

A.7 APW localization and the common-domain audit

The following table matches the sources by mathematical content. The locators in Wu26 refer to a version preceding the October APW file; they are not the theorem numbers in the latter. No new numbering is assigned to the unavailable mixed-calculus theorem.

Locator in Wu26 Fixed source Content and status
APW26b, 5.20 APW26 v1, 5.19 Localization. Local part repaired in Proposition 11.8.6; extra majorization on all UU is false.
APW26b, 5.22 APW26 v1, 5.21 Local PL approximation. Its proof uses the local identity and global PL density, not the false whole-domain inequality.
APW26a, 6.13 Mixed-calculus preprint Analytic comparison invoked in Appendix A; primary full text not retrieved in this revision.

Remark A.7.1 (The scope of the counterexample). Example 11.8.5 satisfies the hypotheses of the fixed Proposition 5.19 in every positive dimension and contradicts its added whole-domain inequality. The failure is already visible from boundedness in Definition 4.3. Restricting the maximum construction to an inner compact domain repairs the local conclusion and gives uniform constants for a bounded family. We do not label strong semipositivity or Proposition 5.21 false on this basis. Its global PL density and bounded-growth correspondence remain source inputs.

Remark A.7.2 (What a generator comparison still needs). Wu26 Appendix A proves approximation assertions for individual generators before concluding the two inclusions of first-stage classes. Lemma 11.8.1 supplies the numerical diagonal step once a common domain is known; Proposition 11.8.2 states one sufficient domain hypothesis. Corollary 11.8.3 verifies it for sequences with the displayed finite logarithmic presentations. These are concrete cases, not a proof that arbitrary local presentations can be placed on one prescribed neighbourhood. Likewise, the corrected affine support in Lemma 11.8.4 fixes the centred differential and the tropical point, without establishing the analytic pullback comparison assigned to the other APW preprint. Theorem 11.9.5 now supplies one projective rational PL direction; Theorem 11.12.5 now supplies the relative rational PL identification, while the reverse comparison remains.

A.8 Convexity and model approximation in the local comparison

The comparison in Section 11.9 requires genuine convex PL generators and a nontrivially valued model category. Two source checks prevent those requirements from being lost in the notation.

Example A.8.1. (A missing contact condition). In [Wu26] (Definition 3.1(3), p. 14), the function is named ff, but the displayed supporting-affine condition uses uu and reads u(x)=ℓ(x)+C,u(x′)≥ℓ(x′)(x′∈V).u(x)=\ell(x)+C,\qquad u(x')\geq\ell(x')\quad(x'\in V). Even after correcting uu to ff, unrestricted CC makes this condition insufficient to define convexity. On an ordinary real open set every finite continuous function satisfies it locally. For example f(t)=−∣t1∣f(t)=-|t_1| on Rn\mathbf R^n, n≥1n\geq 1, satisfies it but is not convex.

Proof. Near any point xx, continuity supplies a constant cc with f≥cf\geq c. Take ℓ=c\ell=c and C=f(x)−cC=f(x)-c. These satisfy the displayed condition. For the example, the midpoint inequality along the first coordinate would require 0=f(0)≤(f(e1)+f(−e1))/2=−10=f(0)\leq(f(e_1)+f(-e_1))/2=-1, which is false. ◻

The intended touching-support condition is f(x)=ℓ(x)f(x)=\ell(x) and f≥ℓf\geq\ell near xx. Equivalently, retain CC in the first equation and put f(x′)≥ℓ(x′)+Cf(x')\geq\ell(x')+C in the second. The source’s supporting-function argument in Appendix A uses equality at the point, consistent with this repair. We treat the displayed definition as a typographical defect. It neither refutes the intended convex theory nor supplies the missing smooth-to-PL comparison. Whenever this book uses convex PL generators, it uses the actual convexity condition, and Section 11.9.1 gives an explicit local maximum presentation.

Remark A.8.2 (Separating current positivity from common-domain regularization). The proof of [Wu26] (Theorem 3.23) passes from its local PL approximation class to CLD current positivity and then to model nefness. For a finite continuous endpoint, Lemma 11.9.1 justifies the first passage directly by uniform convergence on each compact test support. It does not require Theorem A.3 or a resolution of the general predecessor-replacement issue in CLD25 Proposition 8.1.11. For the extensions in Proposition 11.9.3 even this endpoint step is unnecessary: their local weights already have convex finite-max presentations. The remaining model nefness theorem is an explicit geometric input.

The following table records the source boundaries of the new assembly. In particular, the trivial-valuation statements of Wu are not obtained by applying a theorem whose hypothesis says nontrivially valued.

Input Verified use Restriction retained
[GM] (Theorem 1.1) PL/formal/model identification Proper algebraic analytification, nontrivial valuation; rational version by a tensor power
[CLD25] (Corollary 18.8.8) Positive model curvature gives model nefness Original current tests and the cited residual geometry
[BE21] (Theorem 7.8) Uniform pure FS approximation of a nef model metric Ample line bundle, nontrivially valued field
Theorem 20.8.1 Original smooth approximation on the same line bundle Exact deletion at zeros and transition compatibility
Theorem 11.9.5 One ordinary neighbourhood for a decreasing sequence Explicit G\mathcal G generators; one inclusion only

For the generator class, Section 11.10 expands Wu’s Remark 3.2 and the Lipschitz step of Remark 3.9. The relative-image definitions 3.6–3.7 are kept separate; Example 11.10.9 explains the failure of a same-chart extension shortcut. For the numerical step, Proposition 11.10.7 uses [GK19] (Theorem 0.1) over an algebraically closed field and then the model descent results of [GM] (Lemmas 3.3, 3.7 and Proposition 3.5). The GM locators have been checked in the published 2019 text: an earlier author-hosted PDF carries the v2 arXiv stamp and must not be described as the v3 file.

Theorem 11.10.11 gives an alternative model-theoretic route. Its additional inputs are [GJR] (Proposition 7.1 and Definition 7.2), [GRPL] (Corollary 5.5), and the local-to-global semipositivity in [GM] (Theorem 1.1 and Proposition 3.11). The maximum result is used for rational PL functions on good strictly analytic spaces. It is not a theorem about arbitrary real-coefficient presentations or positive currents.

Section 11.11 gives a direct curvewise route from the corrected relative support condition to classical psh, and supplies the model-compatible PL identification on unit charts. Its source comparison fixes a locator mismatch: the companion-paper references Proposition 3.1 and Corollary 3.5 in the proof of GRPSH Theorem 4.3 (v2) correspond to Proposition 5.1 and Corollary 5.4 in the consulted GRPL v1. This is a reference-version issue, not a counterexample to either result. Section 11.12 subsequently establishes the extended-image PL identification under its common-branch and coefficient hypotheses, using Martin’s finite polyhedral image theorem.

Two further APW proof details are repaired in Section 11.11.5. Proposition 11.11.10 restores the interior lower-support hypothesis omitted from the auxiliary claim in APW Proposition 3.3. Lemma 11.11.11 replaces the erroneous zero set {f−ℓf≤0}\{f-\ell_f\leq 0\} in the proof of APW Proposition 5.10 by {g−ℓg≤0}\{g-\ell_g\leq 0\}; equation (11.11.4) gives the required modification bound. These repairs retain the intended local conclusions. Separately, Remark 11.11.12 explains why APW Proposition 5.21, stated for its real PL class, is not by itself a value-group-compatible Q-PL approximation theorem. Proposition 11.11.13 supplies a replacement for finite ambient convex maxima on bounded unit charts.

For coefficient-dependent closure statements, Proposition 11.12.11 reconciles GJR v3 Remark 7.11 with its own Example 5.3 and with GRPL v1 Corollary 5.5 and Theorem 5.6. Maximum and limit statements must retain PL membership; integral and Q-PL maxima satisfy that requirement automatically. This is a correction to the overbroad R-PL maximum wording, not a failure of classical psh maximum stability.

The additional model audit in Section 11.14 checks the published GM Lemma 4.12 on p. 353 and the cofinal blowup construction on p. 337. The counterexample concerns the fixed source model; Proposition 11.14.4 recovers the formal-model domination needed by BE21 Theorem 7.8. Corollary 11.14.5 supplies finite simultaneous polarizations; Proposition 11.14.8 supplies the exact Picard relation and bounded correction, with formal-domination scope. Section 11.13 instead uses GJR Corollary 5.9, Lemmas 6.11 and 6.14, and Proposition 7.10; its original-psh comparison additionally uses GRPSH Theorem 4.3.

Model morphisms and reduced fibres.

The source chain in Section 11.15 is FK arXiv v5, 0.9.2.7 and 0.8.5.25(2), then I.10.3.1 for algebraic morphisms; BE21 Theorem 4.20 for reduced fibres over algebraically closed fields; and the published BGM22 Lemma 1.1, p. 6223. BE21 Lemmas 6.1–6.3 and Theorem 6.4 distinguish exact norm equalities from bounded errors. None of these inputs uses a general positive-current BT theorem.

A separate convention audit is now given in Section 11.21.4. The last finite-selection step of APW Proposition 5.12 preserves ordinary real-polyhedral PL functions, but rational face directions and a rational simplicial presentation require additional control. Examples 11.21.9 and 11.21.11 separate these requirements. They identify a gap in applying that proof under the literal source conventions; they do not disprove the existence conclusion of Proposition 5.21. The germ and transport results in Section 11.21 provide specified forward interfaces, not an unconditional replacement for the source regularization theorem.

Bibliography

[AK]

F. Ardila and C. J. Klivans, The Bergman complex of a matroid and phylogenetic trees. Journal of Combinatorial Theory, Series B 96(1) (2006), 38–49. https://doi.org/10.1016/j.jctb.2005.06.004. https://arxiv.org/abs/math/0311370.

[BFJ15]

S. Boucksom, C. Favre, and M. Jonsson. Solution to a non-Archimedean Monge–Ampère equation. J. Amer. Math. Soc. 28(3) (2015), 617–667. https://arxiv.org/abs/1201.0188.

[BGJK]

J. I. Burgos Gil, W. Gubler, P. Jell, and K. Künnemann, A comparison of positivity in complex and tropical toric geometry, Mathematische Zeitschrift 299 (2021), 1199–1255. Version used: arXiv:2003.08644v3, 15 February 2021. https://arxiv.org/abs/2003.08644v3.

[BGJK25]

J. I. Burgos Gil, W. Gubler, P. Jell, and K. Künnemann. Pluripotential theory for tropical toric varieties and non-Archimedean Monge–Ampère equations. Kyoto J. Math. 65(1) (2025), 55–152. https://doi.org/10.1215/21562261-2024-0010. The section numbering used here is that of arXiv:2102.07392v1 (15 February 2021). https://arxiv.org/abs/2102.07392v1.

[BGK25]

J. I. Burgos Gil, W. Gubler, and K. Künnemann. A tropical formula for non-Archimedean local heights. Preprint, arXiv:2512.07431v1 (2025-12-08). https://arxiv.org/abs/2512.07431.

[BRg]

M. Baker and R. Rumely, Harmonic analysis on metrized graphs. Version used: arXiv:math/0407427v2, 27 July 2005. https://arxiv.org/abs/math/0407427v2.

[CL06]

A. Chambert-Loir. Mesures et équidistribution sur les espaces de Berkovich. J. Reine Angew. Math. 595 (2006), 215–235. https://doi.org/10.1515/CRELLE.2006.049. https://arxiv.org/abs/math/0304023.

[CLD12]

A. Chambert-Loir and A. Ducros. Formes différentielles réelles et courants sur les espaces de Berkovich. Original preprint (2012). https://arxiv.org/abs/1204.6277v1.

[CLD25]

A. Chambert-Loir and A. Ducros. Formes différentielles réelles et courants sur les espaces de Berkovich. Version dated 25 July 2025, 608 pages; CLD25 numbering refers to this version. https://webusers.imj-prg.fr/~antoine.chambert-loir/publications/papers/fdceb-20250725.pdf. https://arxiv.org/abs/1204.6277v2.

[Duc09]

A. Ducros, Les espaces de Berkovich sont excellents. Annales de l’Institut Fourier 59(4) (2009), 1443–1552. https://doi.org/10.5802/aif.2470.

[FGK]

Y. Fang, W. Gubler and K. Künnemann, On the non-archimedean Monge–Ampère equation in mixed characteristic, Nagoya Mathematical Journal 259 (2025), 548–561. Version used: arXiv:2203.12282v3 (final version, 8 December 2025). https://arxiv.org/abs/2203.12282v3.

[FS]

W. Fulton and B. Sturmfels. Intersection Theory on Toric Varieties. Topology 36(2) (1997), 335–353. Version used: arXiv:alg-geom/9403002v1 (1 March 1994). https://arxiv.org/abs/alg-geom/9403002v1.

[GH]

W. Gubler and J. Hertel, Local heights of toric varieties over non-archimedean fields, Publications mathématiques de Besançon, Algèbre et théorie des nombres (2017/1), 5–77. Version used: arXiv:1512.06574v2 (16 January 2017). https://arxiv.org/abs/1512.06574v2.

[GJKM]

W. Gubler, P. Jell, K. Künnemann and F. Martin, Continuity of plurisubharmonic envelopes in non-archimedean geometry and test ideals, with an appendix by J. I. Burgos Gil and M. Sombra, Annales de l’Institut Fourier 69 (2019), 2331–2376. Version used: arXiv:1712.00980v3. https://arxiv.org/abs/1712.00980v3.

[GJR]

W. Gubler, P. Jell, and J. Rabinoff. Forms on Berkovich spaces based on harmonic tropicalizations. Bulletin de la Société Mathématique de France 154(1) (2026), 135–238. https://doi.org/10.24033/bsmf.2919. Numbering used: arXiv:2111.05741v3, 7 March 2025; selected statements also checked in the published text. https://arxiv.org/abs/2111.05741v3.

[GJRc]

W. Gubler, P. Jell and J. Rabinoff, Dolbeault Cohomology of Graphs and Berkovich Curves. Version used: arXiv:2111.05747v1 (10 November 2021). https://arxiv.org/abs/2111.05747v1. The cohomology and duality results cited here are Theorems 6.4.3 and 6.4.5 in this version.

[GK17]

W. Gubler and K. Künnemann. A tropical approach to nonarchimedean Arakelov geometry. Algebra & Number Theory 11(1) (2017), 77–180. https://doi.org/10.2140/ant.2017.11.77. Published full text: https://msp.org/ant/2017/11-1/ant-v11-n1-p03-p.pdf.

[GK19]

W. Gubler and K. Künnemann. Positivity properties of metrics and delta-forms. J. Reine Angew. Math. 752 (2019), 141–177. https://doi.org/10.1515/crelle-2016-0060. https://arxiv.org/abs/1509.09079v2.

[GM]

W. Gubler and F. Martin, On Zhang’s semipositive metrics, Documenta Mathematica 24 (2019), 331–372. Version used: published 2019 text, pp. 331–372. https://doi.org/10.4171/DM/683. Published full text: https://ems.press/content/serial-article-files/26466. https://arxiv.org/abs/1608.08030v3.

[GRPL]

W. Gubler and J. Rabinoff. On semipositive piecewise linear functions in non-Archimedean analytic geometry. Preprint, arXiv:2503.05513v1 (7 March 2025). https://arxiv.org/abs/2503.05513v1.

[GRPSH]

W. Gubler and J. Rabinoff. Classically psh and pluriharmonic functions on Berkovich spaces. Preprint, arXiv:2506.13548v2 (2025-09-17). https://arxiv.org/abs/2506.13548v2.

[GRW16]

W. Gubler, J. Rabinoff, and A. Werner. Skeletons and tropicalizations. Adv. Math. 294 (2016), 150–215. https://arxiv.org/abs/1404.7044v3.

[GRW17]

W. Gubler, J. Rabinoff and A. Werner, Tropical Skeletons, Annales de l’Institut Fourier 67 (2017), 1905–1961. https://arxiv.org/abs/1508.01179.

[Gub07]

W. Gubler, The Bogomolov conjecture for totally degenerate abelian varieties, Inventiones Mathematicae 169 (2007), 377–400. https://gubler.app.uni-regensburg.de/Publikationen/arXiv2bc.pdf.

[Gub10]

W. Gubler, Non-archimedean canonical measures on abelian varieties, Compositio Mathematica 146 (2010), 683–730. Author’s version dated 14 December 2009, including the explicit correction following Theorem 1.1. https://gubler.app.uni-regensburg.de/Publikationen/archivmeasure_v3.pdf.

[Gub13]

W. Gubler. A guide to tropicalizations. In Algebraic and Combinatorial Aspects of Tropical Geometry, Contemp. Math. 589 (2013), 125–189. https://arxiv.org/abs/1108.6126v3.

[Gubler]

W. Gubler, Forms and currents on the analytification of an algebraic variety (after Chambert–Loir and Ducros), arXiv:1303.7364v3, 5 October 2021. https://arxiv.org/abs/1303.7364v3. This is the revised preprint of the chapter published in 2016; section and proposition locators here refer to v3.

[JSS]

P. Jell, K. Shaw and J. Smacka, Superforms, tropical cohomology, and Poincaré duality. Advances in Geometry 19(1) (2019), 101–130. Version used: arXiv:1512.07409v3. https://arxiv.org/abs/1512.07409v3.

[Lag12]

A. Lagerberg, Super currents and tropical geometry, Mathematische Zeitschrift 270 (2012), 1011–1050. Version used: arXiv:1008.2856v1 (17 August 2010). https://arxiv.org/abs/1008.2856v1. https://doi.org/10.1007/s00209-010-0837-8.

[Mih24]

A. Mihatsch. δ\delta-Forms on Lubin–Tate spaces. Duke Math. J. 173(14) (2024), 2809–2928. https://doi.org/10.1215/00127094-2023-0069. https://arxiv.org/abs/2112.10018v4.

[MS15]

D. Maclagan and B. Sturmfels. Introduction to Tropical Geometry. Graduate Studies in Mathematics 161, American Mathematical Society, 2015. https://doi.org/10.1090/gsm/161.

[Pay09]

S. Payne. Analytification is the limit of all tropicalizations. Math. Res. Lett. 16(3) (2009), 543–556. https://arxiv.org/abs/0805.1916.

[Yua08]

X. Yuan, Big line bundles over arithmetic varieties, Inventiones Mathematicae 173 (2008), 603–649. https://math.berkeley.edu/~yxy/preprints/big.pdf.

[Zh93]

S. Zhang, Admissible pairing on a curve, Inventiones Mathematicae 112 (1993), 171–193. https://web.math.princeton.edu/~shouwu/publications/apc.pdf.

[Ber90]

V. G. Berkovich, Spectral Theory and Analytic Geometry over Non-Archimedean Fields. Mathematical Surveys and Monographs 33, American Mathematical Society, 1990. https://doi.org/10.1090/surv/033.

[Ber93]

V. G. Berkovich, Étale cohomology for non-Archimedean analytic spaces. Publications Mathématiques de l’IHÉS 78 (1993), 5–161. https://doi.org/10.1007/BF02712916.

[BPR13]

M. Baker, S. Payne and J. Rabinoff, On the structure of non-Archimedean analytic curves. Contemporary Mathematics 605 (2013), 93–121. Numbering used: arXiv:1404.0279v1, 1 April 2014. https://arxiv.org/abs/1404.0279v1.

[Duc18]

A. Ducros, Families of Berkovich spaces. Astérisque 400 (2018). Numbering used: arXiv:1107.4259v6, 9 October 2017. https://arxiv.org/abs/1107.4259v6.

[Duc07]

A. Ducros, Variation de la dimension relative en géométrie analytique pp-adique. Compositio Mathematica 143 (2007), 1511–1532. https://doi.org/10.1112/S0010437X07003193.

[Tem00]

M. Temkin, On local properties of non-Archimedean analytic spaces. Mathematische Annalen 318 (2000), 585–607.

[Tem04]

M. Temkin, On local properties of non-Archimedean analytic spaces II. Israel Journal of Mathematics 140 (2004), 1–27. https://doi.org/10.1007/BF02786625. https://math.huji.ac.il/~temkin/papers/Local_Properties_II.pdf.

[TemIntro]

M. Temkin, Introduction to Berkovich analytic spaces. Version used: arXiv:1010.2235v2, 16 December 2011. All numbered locators in the foundation chapters refer to this version. https://arxiv.org/abs/1010.2235v2.

[APW26]

O. Amini, M. Piquerez, L. Wu. On solutions to tropical Monge–Ampère equations. arXiv:2610.25602v1, 19 October 2026. https://arxiv.org/abs/2610.25602v1.

[BMP26]

A. M. Botero, E. Mazzon, L. Pille-Schneider. Monge–Ampère measures on balanced polyhedral spaces. arXiv:2603.08664v1, 9 March 2026. https://arxiv.org/abs/2603.08664v1.

[Lin26]

C. Lin. Non-archimedean Monge–Ampère measures for toric metrics on subvarieties of toric varieties. arXiv:2609.12969v1, 11 September 2026. https://arxiv.org/abs/2609.12969v1.

[DT26]

A. Ducros, A. Thuillier. Les squelettes accessibles d’un espace de Berkovich. arXiv:2409.08755v5, 24 October 2026. https://arxiv.org/abs/2409.08755v5.

[Cai23]

Y. Cai. The push-forwards and pull-backs of δ\delta-forms and applications to non-archimedean Arakelov geometry. arXiv:2303.04978v1, 9 March 2023. https://arxiv.org/abs/2303.04978v1.

[Mih23]

A. Mihatsch. On tropical intersection theory. Selecta Math. 29 (2023), article 17. https://doi.org/10.1007/s00029-022-00818-7; version history: https://arxiv.org/abs/2107.12067v3.

[Liu20]

Y. Liu. Tropical cycle classes for non-Archimedean spaces and weight decomposition of de Rham cohomology sheaves. Ann. Sci. Éc. Norm. Supér. 53 (2020), 291–352. https://doi.org/10.24033/asens.2423. Published PDF: https://smf.emath.fr/system/files/filepdf/ens_ann-sc_53_291-352.pdf.

[AP-H]

O. Amini, M. Piquerez. Hodge theory for tropical varieties. arXiv:2007.07826v1 (2020). https://arxiv.org/abs/2007.07826v1.

[AP-F]

O. Amini, M. Piquerez. Hodge theory for tropical fans. arXiv:2310.15367v1 (2023). https://arxiv.org/abs/2310.15367v1.

[AP-FY]

O. Amini, M. Piquerez. Tropical Feichtner–Yuzvinsky and positivity criterion for fans. arXiv:2405.05014v1 (2024). https://arxiv.org/abs/2405.05014v1.

[AP-D]

O. Amini, M. Piquerez. Homological smoothness and Deligne resolution for tropical fans. arXiv:2405.05718v1 (2024). https://arxiv.org/abs/2405.05718v1.

[AP-CS]

O. Amini, M. Piquerez. Tropical Clemens–Schmid sequence and existence of tropical cycles with a given cohomology class. arXiv:2012.13142v1 (2020). https://arxiv.org/abs/2012.13142v1.

[GS23]

W. Gubler, S. Stadlöder. Monge–Ampère measures for toric metrics on abelian varieties. Publ. Math. Besançon (2023), 49–84. https://doi.org/10.5802/pmb.49; https://pmb.centre-mersenne.org/item/10.5802/pmb.49.pdf.

[CG26]

Y. Cai, W. Gubler. Abstract divisorial spaces and arithmetic intersection numbers. arXiv:2409.00611v2, 15 May 2026. https://arxiv.org/abs/2409.00611v2.

[BJ24]

S. Boucksom, M. Jonsson. Non-Archimedean Green’s functions and Zariski decompositions. C. R. Math. 362, special issue S1 (2024), 5–42. https://doi.org/10.5802/crmath.579.

[BJ-S]

S. Boucksom, M. Jonsson. Measures of finite energy in pluripotential theory: a synthetic approach. arXiv:2307.01697v1 (2023); Contemp. Math. 810 (2025). https://arxiv.org/abs/2307.01697; https://doi.org/10.1090/conm/810/16207.

[Dub26]

L. Dubocs. Non-archimedean Green currents for the zero-locus of a regular section. arXiv:2606.19160v1, 17 June 2026. https://arxiv.org/abs/2606.19160v1.

[BGJKM20]

J. I. Burgos Gil, W. Gubler, P. Jell, K. Künnemann, F. Martin. Differentiability of non-archimedean volumes and non-archimedean Monge–Ampère equations, with an appendix by R. Lazarsfeld. Algebraic Geometry 7 (2020), 113–152. https://arxiv.org/abs/1608.01919; author PDF: https://gubler.app.uni-regensburg.de/Publikationen/1608.01919.pdf.

[BGM22]

S. Boucksom, W. Gubler, F. Martin. Differentiability of relative volumes over an arbitrary non-Archimedean field. IMRN 2022, 6214–6242. https://doi.org/10.1093/imrn/rnaa314; https://arxiv.org/abs/2004.03847.

[Wu26]

L. Wu. On the local theory of non-archimedean psh functions. arXiv:2609.27610v1, 23 September 2026. https://arxiv.org/abs/2609.27610v1.

[BE21]

S. Boucksom and D. Eriksson, Spaces of norms, determinant of cohomology and Fekete points in non-Archimedean geometry, Advances in Mathematics 378 (2021), 107501. DOI: 10.1016/j.aim.2020.107501. Primary full text: https://research.chalmers.se/publication/522347/file/522347_Fulltext.pdf. Theorems 5.14, 7.8 and the proof of the latter are used for pure Fubini–Study approximation of nef model metrics over a nontrivially valued field.

[Mar13]

F. Martin, Constructibilité dans les espaces de Berkovich, doctoral thesis, Université Pierre et Marie Curie, defended 7 October 2013, Section 4.1.1 and Theorem 4.3.1. https://www.imj-prg.fr/theses/pdf/florent_martin.pdf.

[Mar15]

F. Martin, A note on tropicalization in the context of Berkovich spaces, Israel J. Math. 210 (2015), 323–334. https://doi.org/10.1007/s11856-015-1254-9.

[Addington21]

N. Addington. Worksheet 2, PRIMA Summer School on Brauer Classes, 3 August 2021, part 3. Explicit deformation of O(1)2\mathcal O(1)^2 to O⊕O(2)\mathcal O\oplus\mathcal O(2). https://www.pims.math.ca/files/N._Addington_Day_2_ws1.pdf.

[StacksPB]

The Stacks Project Authors. The Stacks Project, Tags 01OA (quotient convention for projective bundles) and 0B9S (cohomology and base change). Consulted 14 November 2026. https://stacks.math.columbia.edu/tag/01OA; https://stacks.math.columbia.edu/tag/0B9S.

[FK17]

K. Fujiwara and F. Kato. Foundations of Rigid Geometry I, arXiv:1308.4734v5, 28 February 2017. Locators in this manuscript refer to this fixed preprint: 0.8.5.25(2), 0.9.2.7, I.10.1.2, I.10.3.1 and II.9.5.2. https://arxiv.org/abs/1308.4734v5. Published book: EMS, 2018.

[BBS22]

A. M. Botero, J. I. Burgos Gil, M. Sombra. Convex analysis on polyhedral spaces. Mathematische Zeitschrift 301 (2022), 1631–1674. https://doi.org/10.1007/s00209-021-02869-6.

[FL00]

M. Finzel, W. Li. Piecewise affine selections for piecewise polyhedral multifunctions and metric projections. Journal of Convex Analysis 7 (2000), no.1, 73–94. https://doi.org/10.68381/jca07004.

[Bosch14]

S. Bosch, Lectures on Formal and Rigid Geometry, Lecture Notes in Mathematics 2105, Springer, 2014. https://doi.org/10.1007/978-3-319-04417-0.

[Yam26]

Y. Yamamoto, Non-archimedean SYZ fibrations via tropical contractions, Selecta Mathematica 32 (2026), article 56, published 12 June 2026. https://doi.org/10.1007/s00029-026-01167-5.

[BGR84]

S. Bosch, U. Güntzer and R. Remmert, Non-Archimedean Analysis: A Systematic Approach to Rigid Analytic Geometry. Grundlehren der mathematischen Wissenschaften 261, Springer, 1984. https://doi.org/10.1007/978-3-642-52229-1.

[GrP74]

L. Gruson and M. van der Put, Banach spaces. Mémoires de la Société Mathématique de France 39–40 (1974), 55–100, §5.

[Ser56]

J.-P. Serre, Géométrie algébrique et géométrie analytique. Annales de l’Institut Fourier 6 (1956), 1–42. Sections 13–17 consulted in the English translation. https://faculty.bicmr.pku.edu.cn/~qizheng/files/hodge/GAGA_en.pdf.

[Tem05]

M. Temkin, A new proof of the Gerritzen–Grauert theorem. Mathematische Annalen 333 (2005), 261–269. Proof page locators use the author’s nine-page preprint, checked 15 November 2026. https://math.huji.ac.il/~temkin/papers/Gerritzen_Grauert.pdf.

[StacksAlg]

The Stacks Project Authors, The Stacks Project. Tag 00E0, Lemma 10.17.2(2); Tag 00IQ, Lemma 10.51.5(1); Tag 00CM, Section 10.9 (module localization, immediately before Definition 10.9.6). Tag 00DF, Lemma 10.12.10; Tag 00D2, Lemma 10.12.7; Tag 010H, Lemma 12.5.17(2); Tag 00HD, Lemma 10.39.5(4); Tag 00HC, Lemma 10.39.4; Tag 00HB, Definition 10.39.1(1),(3). Official statements and the indicated proofs consulted 15 November 2026. These are declared external foundations, not internal proofs attributed to this project. https://stacks.math.columbia.edu/tag/00E0; https://stacks.math.columbia.edu/tag/00IQ; https://stacks.math.columbia.edu/tag/00CM; https://stacks.math.columbia.edu/tag/00DF; https://stacks.math.columbia.edu/tag/00D2; https://stacks.math.columbia.edu/tag/010H; https://stacks.math.columbia.edu/tag/00HD; https://stacks.math.columbia.edu/tag/00HC; https://stacks.math.columbia.edu/tag/00HB.

Subject index

  • Abhyankar equality 450, 451
  • Accessible skeleta 1056
  • Actual Green tests 1600, 1601
  • Actual-test scalar multiplication 1110, 1141
  • Adelic height variation 1587, 1604
  • Admissible blow-up 448, 458
  • Admissible graph measure 1524, 1526
  • Aeppli cohomology 1635, 1647
  • Affine functions, distributional kernel 1181
  • Affine potential cocycle 1364, 1364
  • Affinoid algebra 6, 6
  • Affinoid compactness 8
  • Affinoid presentation gluing 1322
  • Affinoid rational replacement 1323
  • Affinoid tropical purity 1053
  • Alexandrov measure 1539, 1575
  • Algebraic domination of models 1255
  • Ample extension, fixed-model obstruction 1248, 1249
  • Ample model, cofinal construction 1251
  • Analytic boundary 1167, 1450
  • Analytic domains 73, 75
  • Analytic fibres 79, 80
  • Analytic flatness 82, 82
  • Analytic realization of a fan 1261
  • Analytification and GAGA 84, 85
  • Annular deformation retraction 579
  • Annulus, Gauss point 1037, 451
  • Annulus, slope formula 1163, 1616
  • Approximation classes 1173, 1343
  • Arithmetic boundary completion 1617
  • Automorphy cocycle 1567
  • Balancing and closedness 1371, 1372
  • Balancing, geometric theorem 786
  • Banach algebra 3
  • Barrier, projective Fubini--Study 1203, 1204
  • Base change, scalar extension 1610, 1610
  • Bedford--Taylor operation 1333, 1343
  • Bergman fan 1055, 1055
  • Berkovich point 7, 8
  • Binomial initial degeneration 799, 799
  • Blowup and model measures 1455
  • Bochner--Martinelli form 1591, 1594
  • Bogomolov, tropical difference map 1552, 1553
  • Bott--Chern cohomology 1647, 1647
  • Boundary supports, uniform-limit failure 1312
  • Boundary factorization for relative finite selections 1229
  • Boundary mass, toric 1556, 1561
  • Boundary, relative interior 80, 446
  • Boundary-operator convention 1379
  • Bounded decreasing stability 1196
  • Bounded extension, fixed background 1194, 1195
  • Compact decreasing hierarchy 1326
  • Calibrated degree 797, 1611
  • Canonical height 1546
  • Canonical measure, mixed dimensions 1547, 1550
  • Canonical metric, contraction 1543, 1545
  • Canonical reduction 442, 448
  • Capacity, continuous admitted class 1353, 1353
  • Cellwise-convex approximation 1289
  • Chern action, gluing 1349
  • Chern mass 1515, 1516
  • Chow ring of a smooth fan 1373, 1374
  • Classical psh 1441, 1450
  • CLD12 and CLD25 comparison 1425, 1427
  • Closed currents, zero extension 1147, 1563
  • Coefficient detection, polyhedral 1368
  • Coefficient measures 1129, 1130
  • Cohomology, flat torus 1646, 1647
  • Cohomology, graph 1414, 1417
  • Collar gluing 1329, 1329
  • Common domain, diagonal approximation 1198, 1199
  • Compact support, radial homotopy 1641
  • Compact upper barrier 1186, 1187
  • Comparison principle, admitted metrics 1352
  • Compatible finite selections 1304
  • Complete intersection, length 1597, 1599
  • Completed residue field 8, 450
  • Completed tensor product 78, 79
  • Continuous multipliers 1132, 1110
  • Continuous obstacles 1192
  • Continuous stages, conditional collapse 1196
  • Controlled-slope extension 1291
  • Convex boundary extension 1288
  • Convex extension, fixed-chart obstruction 1262, 1261
  • Convex extension, positive tangential Hessian 1266
  • Convex representative 1179, 1180
  • Convexity, touching support 1682
  • Corner divisor 791, 1164
  • Cross, failure of Poincaré duality 1642
  • Cross, harmonic-kernel obstruction 1357, 1362
  • Current gluing 1146
  • Current restriction and extension 1433
  • Current topology 1127, 1140
  • Curve detection 1439, 1456
  • Cycle class, de Rham vanishing 1653
  • Decreasing row diagonal 1325
  • Delta-form quotient 1383, 1391
  • Delta-form, six operators 1395, 1395
  • Descent, representatives 1612, 1612
  • Difference maps 1553, 1554
  • Dirichlet energy, graph 1521, 1614
  • Discrete Monge--Ampère equation 1575, 1575
  • Distributional Hessian 1183, 1183
  • Duality and filtered limits 1637, 1637
  • Exact norm fractions 1321
  • Energy, concavity 1520
  • Energy, metric 1519, 1519
  • Energy topology, strong 1497
  • Envelope, local--global comparison 1193, 1197
  • Envelope, orthogonality 1465, 1466
  • Equidistribution, variational proof 1551, 1552
  • Exact support, Green equation 1601, 1592
  • Extended boundary, convex PL generators 1212, 1213
  • Extended tropicalization 1041, 1555
  • First-stage class comparison 1324
  • Faithful tropicalization 1038, 1051
  • Fibre coefficient, native normalization 1471
  • Finite flatness 1596, 1596
  • Finite support testing 1458
  • Finite-stage duality, failure 1637
  • Fixed reduction tube 1485
  • Fixed-fan rational density 1239, 1241
  • Fixed-refinement coefficient criterion 1284
  • Formal GAGA, model morphisms 1254
  • Formal metric, construction 456, 458
  • Formal models and tubes 446, 447
  • Foster identity 1525
  • Four types of points 575
  • Fourier contraction 1646
  • Fubini--Study metric 1505, 1513
  • Gauss norm 796, 443
  • Gauss section, negligibility 1477
  • Gaussian boundary limit and strict contact 1236, 1237
  • Generated products, commutation 1348
  • Generated psh classes 1174, 1348
  • Genus, graph and residual curves 1418, 1419
  • Germ approximation, moving centre 1297
  • Germ reduction 453, 455
  • Graded nef-plus-ample 1492
  • Graded reduction 452, 452
  • Graph Green kernel 1528, 1529
  • Graph harmonic morphism 1613, 1615
  • Graph harmonic weights 1361, 1362
  • Graph Laplacian 1162, 1527
  • Graph pullback 1380
  • Graph trace 1614, 1616
  • Green equation, actual tests 1601, 1601
  • Green objects, three categories 1605
  • G-topology and gluing 75, 76
  • Harmonic coordinates 1405, 1420
  • Harmonic functions, finite dimension 1451
  • Harmonic kernel, dependence on current 1181, 1362
  • Harmonic lower contact 1221
  • Hausdorff support estimate 1144, 1144
  • Height, toric roof 1574, 1574
  • Hessian, local potential 1070, 1183
  • Hodge index, graph 1523
  • Hodge index, real torus 1650
  • Homological smoothness 1654
  • Initial degeneration 784, 802
  • Integral PL, nefness and generated psh 1493, 1494
  • Integral tangent lattice 797, 796
  • Integration by parts, energy 1518, 1520
  • Intersection multiplicity, binomial 800
  • Inverse limit, differential obstruction 1636, 1048
  • Inverse limit, topology 1047, 1047
  • Irrational break directions 1302
  • Irrational constant norm 1475
  • Lattice norm and supnorm 1257, 1259
  • Lattice normalization 797, 1145
  • Laurent cover, graded gluing 1479
  • Laurent domain 75
  • Legendre transform 1571, 1572
  • Lelong-type radial mass 1143, 1143
  • Levine form 1602, 1603
  • Lipschitz approximation obstruction 1302
  • Local potential gluing 1363
  • Local residue fields 77
  • Locally integrable hierarchy 1190
  • Locally nef rational PL transfer 1219
  • Logarithmic branch 1330
  • Matroid rank and closure 1054, 1054
  • Maximum, contact mass 1346, 1347
  • Maximum principle, analytic 1167, 1450
  • Maximum principle, tropical 1166
  • Metric independence, model 1453, 1454
  • Mihatsch calculus, comparison 1392, 1393
  • Mixed products, continuity 1345, 1343
  • Model measure 1509, 1455
  • Model metric, numerical semipositivity 1438, 1453
  • Monomial maps 798, 798
  • Moving rectangle approximation 1285
  • Moving relative approximation 1294
  • Multitangent spaces 1639, 1640
  • Nefness, surjective pullback 1453
  • Negligibility exponent 1468
  • Newton polygon, edge weight 803
  • Nilpotents, local intersection 1597, 799
  • Noncharging, PL products over arbitrary fields 1478
  • Normalized residual maximum 1483
  • Norms, quotient and spectral 448, 9
  • One-scale transport 1298
  • Order zero 1128, 1139
  • Ordinary positivity 1062, 1064
  • Original smooth approximation, real-intercept germ 1301
  • Orthogonality, counterexample 1398
  • Payne inverse limit 1039, 1047
  • Piecewise smooth Hessian 1165, 1165
  • PL, absolute-value Lipschitz estimate 1215
  • PL, relative extended-image identification 1231
  • PL, relative support and ambient convexity 1217
  • PL, relative unit-chart identification 1222
  • PL convention, rational directions 1302
  • PL weights, explicit smooth approximation 1476
  • Plücker relation 1065, 1066
  • Pluriharmonic functions 1451, 1452
  • Poincaré lemma, real 1069, 1071
  • Poincaré--Lelong 1149, 1163
  • Pointwise closure, PL functions 1459, 1459
  • Polarization and mixed measures 1516, 1347
  • Polyhedral convexity 1398
  • Polyhedral Monge--Ampère atoms 1539
  • Positivity cones, distinctness 1066, 1131
  • Potential independence 1334, 1362
  • Premodels, prescribed realization 1489, 1490
  • Projection formula, BT 1350
  • Projection formula, toric 1565
  • Projective model nefness, functorial criterion 1216
  • Projective PL-to-smooth transfer 1208, 1209
  • Proper domination, numerical positivity 1480
  • Properness 80, 81
  • Psh, continuous approximation class 1343, 1343
  • R-PL germ, rational approximation 1245, 1247
  • Radial homotopy 1068, 1072
  • Radius extension field 10
  • Rational affine bumps 1306, 1307
  • Rational approximation, monotone convex composition 1271
  • Rational barrier 1330, 1331
  • Rational coefficient closure 1278
  • Rational domain 74, 74
  • Rational polytopal complex, coefficient density 1274
  • Rational positive closure 1281
  • Rational relative approximation, same chart 1270
  • Rational triangulation obstruction 1287
  • Rationalization, fixed-subdivision obstructions 1277, 1277
  • Reduced special fibre, cofinal models 1256
  • Reduction strata, dimension threshold 1473
  • Reduction, anti-continuity 443
  • Regular immersion 1597, 1598
  • Regularized maximum 1327, 1327
  • Relative-support regularization, continuous inputs 1310
  • Relative affine supports, classical psh 1221
  • Relative volume derivative 1495
  • Removal of singular sets 1147, 1563
  • Residual criterion, dependencies 1469
  • Residual criterion, pure dimension 1481
  • Residual curve and directions 581
  • Residual Picard obstruction 1421, 1422
  • Residual positivity 1462, 1463
  • Residual realization, simultaneous 1480
  • Residue tube, simultaneous finite presentation 1244
  • Resistance kernel 1524
  • Riemann--Zariski space 453, 453
  • Roof function 1542, 1574
  • Scalar extension, degree warning 1610, 1611
  • Scale propagation through edges 1287
  • Semiample residue, smooth approximation 1484
  • Semistable skeleton 1037, 450
  • Semistable vertex sets 585, 585
  • Shilov boundary 444, 1049
  • Shilov boundary 445, 1450
  • Shrinking stars 1300
  • Skeletal current 1355
  • Skoda--El Mir extension, toric 1563
  • Slope formula, sign 588, 588
  • Smooth approximation, periodic 1567
  • Smooth psh, chart-dependent convexity 1263
  • Soft maximum, boundary limitation 1539, 1542
  • Stable intersection 793, 1369
  • Star cohomology 1643
  • Stokes theorem 1082, 1371
  • Strata, closed-current decomposition 1564
  • Strict face supports 1282
  • Strictification obstruction 1475
  • Strictly integral PL, current positivity 1487
  • Strong positivity, span failure 1066, 1671
  • Superform signs 1061, 1062
  • Support, analytic versus tropical 1636, 1612
  • Symmetric Hessian potential 1070, 1070
  • Tate acyclicity 76
  • Tate algebra 4, 5
  • Tate contraction 1543
  • Test density and extension 1434
  • Test forms, quotient relations 1107, 1110
  • Toric boundary, lifting mass 1562, 1562
  • Toric Monge--Ampère transport 1559, 1541
  • Toric smoothness, counterexample 1557
  • Tropical Calabi--Yau, local conditions 1399
  • Tropical Dolbeault comparison 1640, 1642
  • Tropical measures, information loss 1432
  • Tropical multiplicity, local constancy 805
  • Truncation, Green forms 1590, 1594
  • Truncation, Monge--Ampère atoms 1560, 1560
  • Unit-presentation obstruction 1319
  • Uniform convergence, moving currents 1141
  • Uniform tropical localization 1201, 1201
  • Valuation extension, real weights 801
  • Variational solution, Monge--Ampère 1467, 1467
  • Vertical intersection matrix 1522, 1523
  • Weak positivity, order-zero failure 1131
  • Weakly smooth forms 1405, 1406
  • Weierstrass division 5, 6
  • Weighted tropical germ 1579
  • Wild scalar extension, norm discrepancy 1258
  • Wu's local approximation classes 1191, 1682
  • Zero coordinates, rational bumps 1313, 1314
  • Zero extension 1147, 1563

@mainmattertrue