Pith. sign in

REVIEW 1 major objections 5 minor 29 references

Stable Reduction via the Log Canonical Model

T0 review · 1 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Assuming two standard MMP conjectures, this paper proves stable reduction for log canonically polarized pairs in large characteristic, and recovers properness and projectivity of the moduli of stable surfaces.

desk verdict A genuinely attractive LCM-first proof of stable reduction with a clean tame base-change characterization; the mixed-characteristic claim in the surface corollary overreaches its citation, but the core is sound and deserves refereeing. read the letter →

arxiv 2411.17909 v1 pith:KDDRLXCL submitted 2024-11-26 math.AG

Tai-Hsuan Chung This is my paper · ORCID
classification math.AG MSC 14D0614E3014G1714J17
keywords stablereductionlogcanonicalmodelpositivecharacteristicmoduliofsurfacesminimalprogramboundednessvolumetamebasechangeDCCsets
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a higher-dimensional analogue of the classical stable reduction theorem for curves and proves it in large characteristic: when the residue characteristic $p$ exceeds the ratio $v/v(n,D(I))$ of the volume of the generic fibre to a universal minimal volume, every component of the central fibre of the log canonical model has multiplicity smaller than $p$, so the natural base change is tame and yields the stable model. The argument runs on two standard Minimal Model Program conjectures: existence of the log canonical model over the base, and a uniform lower bound on volumes of stable log varieties with coefficients in a DCC set. In dimension two both are theorems, so the paper recovers, unconditionally, the properness of the moduli stack and the projectivity of the coarse moduli space of stable surfaces of fixed volume over algebraically closed fields of large characteristic. A characteristic-free theorem characterizes stable reduction after a tame base change in terms of the multiplicities of the central fibre of the log canonical model, generalizing the classical curve criterion.

What carries the argument

The load-bearing object is the log canonical model $(X, \Delta+(X_k)_{\mathrm{red}})$ over the base DVR — the pair obtained by adding the reduced central fibre to the given boundary — together with the identity $v \ge \operatorname{vol}(K_{X_K}+\Delta_K) = \sum_i m_i \operatorname{vol}(K_{F_i}+B_i)$, where $F_i$ are the components of the central fibre with multiplicities $m_i$ and $B_i$ is the different on the normalization of $F_i$. This identity converts the total volume into a weighted sum of component volumes, each bounded below by $v(n,D(I))$ by Conjecture $V_n$, so each multiplicity is bounded by $v/v(n,D(I))$. The enabling criterion is Theorem 3.1: a log canonically polarized pair has stable reduction after a tamely ramified base change of degree $N$ if and only if there is an LCM whose central-fibre multiplicities have least common multiple dividing $N$. The forward direction uses the tame cover $\pi = t^N$ and a standard discrepancy lemma for ramified covers; the reverse direction builds the LCM as the quotient of the stable model by a cyclic Galois group.

What would settle it

Exhibit a DVR with algebraically closed residue field of characteristic $p > v/v(n,D(I))$ and a projective geometrically normal pair $(X_K,\Delta_K)$ over $K$ with $K_{X_K}+\Delta_K$ log canonical and ample, volume at most $v$, and coefficients in a DCC set $I$, for which after every finite extension $L/K$ there is no stable log model; this would refute Theorem 1.1, and in dimension two it would refute Corollary 1.2. A less global test is to find an LCM whose central fibre has a component of multiplicity at least $p$ while $(K_X+\Delta+(X_k)_{\mathrm{red}})^n\cdot F_i$ is still uniformly bounded below, which would violate the volume identity and show the characteristic threshold is sharp.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for fixed dimension $n$, volume bound $v$, and DCC coefficient set $I$, if Conjecture $\operatorname{LCM}_{n+1}$ (existence of a log canonical model of the pair together with the reduced central fibre over the base DVR) and Conjecture $V_n$ (a uniform positive lower bound on volumes of $n$-dimensional normal stable log varieties with coefficients in a DCC set) hold, then Conjecture $\operatorname{SR}_n$ holds for all pairs with volume at most $v$ and coefficients in $I$, provided $\operatorname{char} k > v/v(n,D(I))$. The proof starts with the log canonical model, writes the central fibre $X_k = \sum_i m_i F_i$, and uses the identity $\operatorname{vol}(K_{X_K}+\Delta_K) = \sum_i m_i \operatorname{vol}(K_{F_i}+B_i)$ to show each multiplicity $m_i$ is at most $v/v(n,D(I))$. When the characteristic exceeds this ratio, the least common multiple $N$ of the multiplicities is prime to $p$, so the base change $\pi = t^N$ is tame; the normalization of the pullback is then the stable log model, with reduced central fibre and log canonical total space. In the surface case both conjectures are known theorems, giving an unconditional stable reduction statement for log canonically polarized surfaces and recovering the properness of the moduli stack of stable surfaces and the projectivity of its coarse space over large characteristic fields.

Load-bearing premise

The argument collapses without Conjecture $\operatorname{LCM}_{n+1}$: there must exist a log canonical model of $(X, \Delta+(X_k)_{\mathrm{red}})$ over the DVR whose central fibre has components with bounded multiplicities, and in higher dimensions this existence is an open Minimal Model Program statement, while in the surface case it is supplied by a known existence theorem for threefolds in positive characteristic.

Editorial extensions

If this is right

  • Conjecture $\operatorname{SR}_2$ holds unconditionally for log canonically polarized surfaces over algebraically closed fields of characteristic greater than $\max\{5, v/v(2,D(I))\}$, for any fixed volume bound $v$ and DCC set $I$.
  • The moduli stack of stable surfaces of volume $v$ is proper, and its coarse moduli space is projective, over such fields — recovering the known properness theorem for stable surfaces in large characteristic.
  • The required base change is explicit: it is the tame cover $\pi = t^N$ with $N$ the least common multiple of the central-fibre multiplicities of the log canonical model, and one can always take $N = (p-1)!$.
  • The stable limit is explicitly described: if $F$ is a component of the central fibre of the LCM with multiplicity $m_F$, its preimage splits into $\gcd\{m_F, n_j\}_j$ reduced components, each a degree $m_F/\#F$ cover of $F$ with volume $(m_F/\#F)\cdot\operatorname{vol}(K_F+B)$.
  • In characteristic zero, the same circle of ideas proves Conjecture $\operatorname{SR}_n$ for klt pairs, because the log canonical model exists under mild assumptions.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If Conjectures $\operatorname{LCM}_{n+1}$ and $V_n$ are eventually proved in all dimensions, the same volume argument would yield stable reduction for log canonically polarized pairs in arbitrary dimension over large characteristic fields, with an explicit characteristic threshold.
  • The volume identity gives a practical way to detect when wild base change is unavoidable: by Corollary 3.3, an LCM component whose multiplicity is divisible by $p$ forces any stabilizing base change to have degree divisible by $p$, so the open problem of finding the right base change in that case is the main obstruction to removing the large-characteristic assumption.
  • Making the boundedness constant $v(n,D(I))$ effective would turn the theorem into an explicit arithmetic bound on the characteristic, which could be tested computationally on examples of log surfaces.
  • The description of stable limits in Corollary 3.10 suggests a concrete recipe for constructing higher-dimensional stable limits in practice: pass to the LCM, make the tame base change reducing all multiplicities to one, and read off the limit components as covers of the LCM's central components.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. This paper proposes a higher-dimensional stable reduction conjecture (SR_n) for log canonically polarized pairs over a DVR with algebraically closed residue field, and proves a conditional theorem. Assuming (i) existence of a log canonical model over Spec R for the pair with the reduced central fibre (Conjecture LCM_{n+1}) and (ii) a uniform lower volume bound for n-dimensional stable log varieties with coefficients in a DCC set (Conjecture V_n), Theorem 1.1 shows that if char k > v/v(n,D(I)) and vol(K_{X_K}+Δ_K) ≤ v, then SR_n holds: the author bounds multiplicities of the central fibre components by the volume, then makes a tame base change of degree the lcm of these multiplicities and invokes Proposition 3.2. For surfaces, V_2 is known via [HK19] and the proof invokes [HNT20] to supply LCM_3, yielding unconditional stable reduction (Corollary 1.2); with Posva's gluing and [ABP24] this recovers properness and projectivity of the moduli of stable surfaces over large characteristic fields (Theorem 1.3).

Significance. If correct, the conditional theorem is a clean and useful reduction: it replaces ad hoc stable reduction arguments with two standard MMP conjectures and gives an explicit characteristic bound and explicit base change. The proof of Theorem 1.1 is transparent, has no fitted parameters, and does not appear circular; the conjectures are honestly stated as hypotheses. The surface application is the main concrete evidence, while the higher-dimensional statement is conditional on substantial open problems. The paper is a well-written note whose value is conceptual and expository, and it recovers known properness and projectivity results for stable surfaces.

major comments (1)
  1. [§3, proof of Corollary 1.2; Remark 1.5] The proof of Corollary 1.2 discharges Conjecture LCM_3 by invoking [HNT20, Theorem 4.11] after passing to a log resolution, quoting it as a theorem 'over any perfect field k with char k > 5'. If that theorem is a statement about projective threefolds over a perfect field, it does not by itself produce a relative log canonical model over Spec R when R is a mixed-characteristic DVR, since Spec R is not a k-scheme and no k-algebra structure is available on R. The proof as written therefore establishes Corollary 1.2 only for equicharacteristic DVRs, and Remark 1.5 overclaims in asserting that the results hold in both mixed and equi-characteristic. This is not an internal inconsistency in Theorem 1.1, which explicitly assumes LCM_{n+1}, but it is load-bearing for the unconditional surface statement. Theorem 1.3 survives, because the valuative criterion over k only tests DVRs that are k-algebras; still, the author should either restrict Corollary 1.2 and Remark 1.5 to the equicharacteristic case or supply a valid mixed-characteristic reference for the lc model over Spec R (for instance [ABP24] or [BMP+23]).
minor comments (5)
  1. [Abstract] There is a typo: 'De ligne–Mumford' should be 'Deligne–Mumford'.
  2. [Theorem 3.1, equation (4)] The sign in the displayed equivalence and the subsequent application of [Kol13, 1.17] are terse; please state explicitly which variant of the negativity lemma is used and why E1−E2 satisfies its hypotheses.
  3. [Proof of Corollary 3.7] The displayed equality coeff(D_i^K+Δ_i^K)=coeff(Δ_i^K) is not literally correct, since the conductor D_i^K has coefficient 1; the needed inclusion coeff(D_i^K+Δ_i^K) ⊆ {1} ∪ coeff(Δ_i^K) ⊆ I nevertheless holds because I contains 1.
  4. [Corollary 3.10] The sentence 'Since R contains k = \bar{k}, g is Galois' assumes R is a k-algebra; if R is only a DVR with residue field k, this needs justification or a Henselization step.
  5. [Example 3.5] The example depends on an unpublished lecture [McQ]; please add a citable reference or explicitly mark the example as folklore.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main theorem is an explicitly conditional derivation from two stated conjectures, and the surface corollary discharges them via independent external results.

full rationale

The paper's central claim, Theorem 1.1, is explicitly conditional on Conjecture LCM_{n+1} (existence of a log canonical model over Spec R with generic fibre (X_K, Delta_K)) and Conjecture V_n (uniform positive lower bound on volumes of n-dimensional normal stable log varieties). The proof of Theorem 1.1 uses exactly these assumptions: it starts with the log canonical model supplied by LCM_{n+1}, expands the volume of the central fibre as a sum of multiplicities times volumes of components, applies V_n to bound each multiplicity by v/v(n,D(I)), and then invokes the tame base-change construction of Theorem 3.1 to produce the stable log model. None of these steps identifies the conclusion with an input by construction: stable reduction is not the same statement as existence of an LCM, and the tame base-change output is not the same statement as the volume lower bound. Corollary 1.2 discharges the two conjectures for surfaces by citing HNT20 Theorem 4.11 for the threefold relative log canonical model and HK19 Theorem 2 for boundedness of slc surfaces; these are external results distinct from the target theorem HK19 Theorem 4, and the properness conclusion additionally relies on Pos24 and ABP24. Thus the recovery of the Hacon-Kovacs theorem is not a self-citation chain. The possible mismatch between HNT20's equicharacteristic DVR setting and Remark 1.5's mixed-characteristic claim is a correctness or scope concern, not a circularity concern, and no equation in the paper reduces to its own input by construction. No circular step can be exhibited, so the score is 0.

Assumptions & free parameters 0 free parameters · 9 assumptions · 0 invented entities

The paper introduces no free parameters or new postulated mathematical entities. Its central claims are conditional on two explicit conjectures (LCM and V) and on a web of published theorems, all listed above. The only invented entity is the Stable Reduction Conjecture itself, which is a statement, not a new object.

assumptions (9)
  • domain assumption Conjecture LCM_{n+1}: for any (X_K, Δ_K) as in Conjecture SR_n, there exists a log canonical model (X, Δ+(X_k)_{red}) over Spec R with generic fibre isomorphic to (X_K, Δ_K).
    Assumed in Theorem 1.1 and used at the first line of its proof. In the surface case it is replaced by HNT20 Theorem 4.11, so Corollary 1.2 is unconditional; in higher dimensions it is a major open MMP conjecture.
  • domain assumption Conjecture V_n: the set of volumes of n-dimensional normal stable log varieties with coefficients in a DCC set C has a positive minimal element v(n,C).
    Used in Theorem 1.1 to bound the sum of central fibre multiplicities via equation (7). For n=2 it is known by Ale94 Theorem 8.2 and HK19 Theorem 2, which supplies v(2,D(I)) in Corollary 1.2.
  • standard math Properties of lc/klt under ramified covers (Kol13 2.42-2.43, quoted as Proposition 3.2).
    Used in the (2)=>(1) direction of Theorem 3.1 to deduce that the normalized base change is a stable log model, and in Corollary 3.8 to show klt-ness of the pulled-back pair.
  • standard math Negativity of contraction (Kol13 1.17).
    Used in the proof of Theorem 3.1, (1)=>(2), to show E1=E2 and that the Galois automorphism extends to the model.
  • standard math Existence of log resolutions for arithmetical threefolds (CP19 Theorem 1.1, CJS20 Corollary 1.5).
    Used in the proof of Corollary 1.2 to pass to a log resolution of (X, Δ+(X_k)_{red}) over Spec R.
  • standard math Existence of log canonical models for log canonical threefolds in characteristic greater than 5 (HNT20 Theorem 4.11).
    Replaces Conjecture LCM_3 in the proof of Corollary 1.2, giving the log canonical model over Spec R for surfaces.
  • standard math Boundedness of volumes for lc surfaces (Ale94 Theorem 8.2, HK19 Theorem 2).
    Supplies Conjecture V_2 and the constant v(2,D(I)) used in Corollaries 1.2, 3.7 and Theorem 1.3.
  • standard math Posva's gluing theorem for slc surfaces and threefolds in positive characteristic (Pos24) and the S2 result for stable surface families over Z[1/30] (ABP24 Theorem 4.12).
    Used in Corollary 3.7 and Theorem 1.3 to glue stable models of the normalization and to check condition (S2) for the central fibre.
  • standard math Finite generation of the relative log canonical ring for klt pairs over C (BCHM10 Theorem 1.2(3)).
    Used in Corollary 3.8 to obtain the log canonical model over Spec R in characteristic 0.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Stable Reduction via the Log Canonical Model." pith.science (2026). https://pith.science/paper/KDDRLXCL

@misc{pith2026241117909,
  author       = {Pith},
  title        = {Pith review of: Stable Reduction via the Log Canonical Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KDDRLXCL}},
  note         = {Machine review of arXiv:2411.17909}
}
abstract

We formulate a stable reduction conjecture that extends Deligne-Mumford's stable reduction to higher dimensions and provide a simple proof that it holds in large characteristic, assuming two standard conjectures of the Minimal Model Program. As a result, we recover the Hacon-Kov\'acs theorem on the properness of the moduli stack $\overline{\mathscr{M}}_{2,v,k}$ of stable surfaces of volume $v$ defined over $k=\overline{k}$, provided that $\operatorname{char}k>C(v)$, a constant depending only on $v$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 27 canonical work pages

  1. [1]

    Arvidsson, F

    E. Arvidsson, F. Bernasconi, and Zs. Patakfalvi, On the properness of the moduli space of stable surfaces over Z [1/30] , 2024, to appear in Moduli

    work page 2024

  2. [2]

    Alexeev, Boundedness and K ^2 for log surfaces , Internat

    V. Alexeev, Boundedness and K ^2 for log surfaces , Internat. J. Math. 5 (1994), no. 6, 779--810

    work page 1994

  3. [3]

    , Moduli spaces \( M _ g,n ( W )\) for surfaces , Higher-dimensional complex varieties (Trento, 1994) (1996), 1--22

    work page 1996

  4. [4]

    Artin, Wildly ramified Z /2 actions in dimension two , Proc

    M. Artin, Wildly ramified Z /2 actions in dimension two , Proc. Amer. Math. Soc. 52 (1975), 60--64

    work page 1975

  5. [5]

    Birkar, P

    C. Birkar, P. Cascini, C. Hacon, and J. M ^ c Kernan, Existence of minimal models for varieties of log general type, J. Amer. Math. Soc. 23 (2010), no. 2, 405--468

    work page 2010

  6. [6]

    Birkar, Existence of log canonical flips and a special LMMP , Publ

    C. Birkar, Existence of log canonical flips and a special LMMP , Publ. Math. IHÉS 115 (2012), 325--368

    work page 2012

  7. [7]

    Bhatt, L

    B. Bhatt, L. Ma, Zs. Patakfalvi, K. Schwede, K. Tucker, J. Waldron, and J. Witaszek, Globally + -regular varieties and the minimal model program for threefolds in mixed characteristic, Publ. Math. IHÉS 138 (2023), no. 1, 69--227

    work page 2023

  8. [8]

    Cossart, U

    V. Cossart, U. Jannsen, and S. Saito, Desingularization: Invariants and strategy, Lecture Notes in Mathematics, vol. 2270, Springer, Berlin, 2020

    work page 2020

Show all 29 references
  1. [9]

    Cossart and O

    V. Cossart and O. Piltant, Resolution of singularities of arithmetical threefolds, Journal of Algebra 529 (2019), 268--535. 3942183

    2019

  2. [10]

    Deligne and D

    P. Deligne and D. Mumford, The irreducibility of the space of curves of given genus, Publ. Math. IHÉS 36 (1969), 75--109

    1969

  3. [11]

    Lars Halvard Halle, Stable reduction of curves and tame ramification, Math. Z. 265 (2009), 529--550

    2009

  4. [12]

    Hartshorne, Algebraic geometry, Graduate Texts in Mathematics, vol

    R. Hartshorne, Algebraic geometry, Graduate Texts in Mathematics, vol. 52, Springer, New York, NY, 1977

    1977

  5. [13]

    Hacon and S

    C. Hacon and S. Kovács, On the boundedness of slc surfaces of general type, Ann. Sc. Norm. Super. Pisa Cl. Sci XIX (2019), no. 5, 191--215

    2019

  6. [14]

    Hashizume, Y Nakamura, and H

    K. Hashizume, Y Nakamura, and H. Tanaka, Minimal model program for log canonical threefolds in positive characteristic, Math. Res. Lett. 27 (2020), no. 4, 1003--1054

    2020

  7. [15]

    Hacon and C

    C. Hacon and C. Xu, Existence of log canonical closures, Invent. math. 192 (2013), 161--195

    2013

  8. [16]

    Kollár et al., Flips and abundance for algebraic threefolds, Astérisque, vol

    J. Kollár et al., Flips and abundance for algebraic threefolds, Astérisque, vol. 211, Soc. Math. de France, 1992

    1992

  9. [17]

    Kempf, F

    G. Kempf, F. Knudsen, D. Mumford, and B. Saint-Donat, Toroidal embeddings 1, Springer, Berlin, 1973

    1973

  10. [18]

    Kollár and S

    J. Kollár and S. Mori, Birational geometry of algebraic varieties, Cambridge University Press, Cambridge, 1998

    1998

  11. [19]

    Kollár, Singularities of the minimal model program, Cambridge Tracts in Mathematics, vol

    J. Kollár, Singularities of the minimal model program, Cambridge Tracts in Mathematics, vol. 200, Cambridge University Press, Cambridge, 2013, With the collaboration of S\' a ndor Kov\' a cs

    2013

  12. [20]

    , Families of stable 3-folds in positive characteristic, Epijournal Geom. Algebr. 7 (2023), no. 6, 1--11

    2023

  13. [21]

    , Families of varieties of general type, Cambridge University Press, Cambridge, 2023

    2023

  14. [22]

    Kollár and N.I

    J. Kollár and N.I. Shepherd-Barron, Threefolds and deformations of surface singularities, Invent. Math. 91 (1988), 299--338

    1988

  15. [23]

    Liu, Algebraic geometry and arithmetic curves, Oxford University Press, Oxford, 2002

    Q. Liu, Algebraic geometry and arithmetic curves, Oxford University Press, Oxford, 2002

    2002

  16. [24]

    McQuillan, Unpublished lecture on foliations

    M. McQuillan, Unpublished lecture on foliations

  17. [25]

    M ^ c Kernan and Y

    J. M ^ c Kernan and Y. Prokhorov, Threefold threshold, Manuscripta Math (2004), 281--304

    2004

  18. [26]

    Patakfalvi, On the projectivity of the moduli space of stable surfaces in characteristic p>5 , arXiv preprint (2017), 1--42, arXiv:1710.03818 https://arxiv.org/abs/1710.03818

    Zs. Patakfalvi, On the projectivity of the moduli space of stable surfaces in characteristic p>5 , arXiv preprint (2017), 1--42, arXiv:1710.03818 https://arxiv.org/abs/1710.03818

    2017 arXiv

  19. [27]

    Posva, Gluing theory for slc surfaces and threefolds in positive characteristic, Ann

    Q. Posva, Gluing theory for slc surfaces and threefolds in positive characteristic, Ann. Sc. Norm. Super. Pisa Cl. Sci. (5) XXV (2024), 811--886

    2024

  20. [28]

    Saito, Vanishing cycles and geometry of curves over a discrete valuation ring, Amer

    T. Saito, Vanishing cycles and geometry of curves over a discrete valuation ring, Amer. J. Math. 109 (1987), no. 6, pp. 1043--1085

    1987

  21. [29]

    The Stacks project authors , The stacks project, https://stacks.math.columbia.edu, 2024

    2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.