hypersetup{ pdftitle={Log abundance in characteristic zero (v2)}, pdfauthor={OpenAI}, pdflang={en-US} } allowdisplaybreaks[1] setcounter{tocdepth}{2} # Log abundance in characteristic zero[2pt] large a rewritten, tightened edition maketitle Abstract: We prove the rational-boundary log abundance conjecture in every dimension over algebraically closed fields of characteristic zero: every nef Q-Cartier log canonical divisor on a projective log canonical pair with effective rational boundary is semiample. Over C, the same proof establishes smooth canonical nonvanishing in every dimension. The proof is a dimension induction; the hard input is nonvanishing, closed by a reduction to positive characteristic and a determinant argument. It is independent of logarithmic Iitaka subadditivity except for a single use in the Albanese reduction. tableofcontents newpage %====================================================================== # A guide for the reader addcontentsline{toc}{section}{A guide for the reader} %====================================================================== ## The statement we prove The log abundance conjecture says this: if K_X+B meets every curve of X nonnegatively, then some positive multiple of K_X+B is a genuine line bundle generated by its global sections. Nefness is a numerical condition on curve degrees. Generation by sections is an algebraic condition on actual functions. Log abundance says the first forces the second. [Log abundance, main theorem] Let (X,B) be a normal projective log canonical pair over an algebraically closed field of characteristic zero, with B an effective rational divisor and K_X+B Q-Cartier. If K_X+B is nef, then some positive Cartier multiple of K_X+B is globally generated. The body of the paper works over C. The passage to an arbitrary algebraically closed field of characteristic zero is bookkeeping and lives in Section~sec:completion. ## The shape of the argument The proof is a dimension induction. In dimension n it aims to produce two outputs, in this order. [label=(roman*)] Smooth canonical nonvanishing (NV_n). A smooth projective complex n-fold X with pseudo-effective K_X has H^0(X,mK_X)ne0 for some m>0. Good log minimal models in dimension n for all projective lc pairs with effective real boundary and pseudo-effective adjoint. On a good model the adjoint is semiample. This is the inductive assertion; it yields Theorem~thm:main in dimension n by comparing the pair with its good model. The two outputs are entangled. (ii) in dimension n-1 (the induction hypothesis) is used to prove (i) in dimension n. And (i) in dimension n, together with the signed-representative theorem (Sections~sg:section--sg:descent), is used to prove (ii) in dimension n. Together, (i)+(ii) in dimension n supply the induction hypothesis for the next dimension. The base case is a point. The engine that turns nonvanishing into abundance is the signed-representative theorem (Theorem~thm:signed). Its statement is subtle: it handles a nef L=K_X+D on a Q-factorial dlt pair, with D reduced and with L rational-linearly equivalent to a Q-divisor with either positive or negative coefficients on the components of D, provided (i) L-cD is pseudo-effective for some c>0, and (ii) L|_D is semiample. Under those hypotheses it produces abundance: kappa(L)=nu(L). The signed-representative theorem is what turns a numerically-nonnegative adjoint into a genuinely semiample one. Its proof (Sections~sg:section--sg:descent) uses roots, Hodge modules, and a formal deformation argument; it is the technical heart of the paper. ## Where the difficulty lives Smooth canonical nonvanishing is the bottleneck. Everything else in the induction is formal once NV_n is in hand: Section~sec:induction proves Proposition~prop:inductive-reduction, which makes good models of (X,Delta) for real boundary once NV_n holds. The signed- representative theorem fills the zero-Iitaka case; the case of positive Kodaira dimension is Fujino--Gongyo. Everything rests on NV_n. The proof of NV_n assumes a counterexample and derives a contradiction in four stages. This is the pivotal movement of the paper, and the earlier part of our reading guide has no more to say about it beyond a signpost to the relevant sections. medskipnoindent Stage 1: geometric exclusions (Section~sec:exclusions). The counterexample X must: (a) have irregularity zero, so numerical equivalence equals Q-linear equivalence on its rational divisors; (b) not be covered by proper subvarieties of non-general type; (c) admit no positive canonical current annihilating the fibers of a rational fibration (this rules out families of subvarieties pulled back from a smaller ambient); (d) admit no covering families of curves with uniformly bounded genus and bounded normalized degree (this rules out the ``multi-section'' cases that would otherwise survive the two-slot analysis); (e) and, on any terminal birational model, admit no family of birational correspondences between fixed finite covers whose graphs dominate the product. This last is Hanamura's birational-group theorem applied to a positive canonical current. medskipnoindent Stage 2: two opposing jet estimates (Sections~sec:common-tools and~sec:jets). At a very general point of X, every section of every Cartier multiple of a suitable normalized nef divisor P has bounded vanishing order (the scalar jet bound). But on the projective bundle over Xtimes X associated to two copies of the same line bundle, the very general Seshadri constant exceeds 2n+2, so jets vanish to very high order (the two-slot separation). The first bound consumes the curve exclusion (d); the second is what produces the determinant incompatibility of Stage 3. medskipnoindent Stage 3: the Frobenius determinant (Section~fr:section). After reduction modulo a large prime, the two-slot jets give an evaluation map of full generic rank on the product of two Frobenius-twisted copies of the model. A canonical filtration of its restriction to the diagonal of the Frobenius square cuts that rank down by a factor of four. The determinant of any maximal minor is a nonzero section of an external product of two line bundles; the scalar jet bound bounds its order at a diagonal point from above, while the rank deficit bounds it from below. Contradiction for large primes. medskipnoindent Stage 4: contradiction (Section~fr:section, end). The counterexample was impossible. NV_n holds. ## Reading paths If you want only the statement and the skeleton. Section~sec:introduction (main theorem, conventions), then this guide, then the statement of Theorem~thm:signed in Section~sg:section, then the statement of Theorem~thm:smooth-nonvanishing in Section~fr:section, then the proof of Theorem~thm:main in Section~sec:completion. Everything else is either a lemma filling a gap in those two theorems or a consequence downstream of the main theorem. If you want to understand the engine. Sections~sg:section--sg:descent: the root-and-adjunction construction, the Hodge-module lifting, the deformation off the boundary, the final descent. This is a self-contained subproof of Theorem~thm:signed and does not require other parts of the paper. If you want to understand the hard input. Sections~sec:exclusions--fr:section in that order. The section on jets depends on the section on exclusions; the section on Frobenius depends on both and closes the loop. The full proof, end to end. Read linearly. ## What is used from outside The proof uses three external inputs: [label=(alph*)] Logarithmic Iitaka subadditivity, stated below as Theorem~asm:iitaka. It is used once, in Lemma~ex:irregularity, and nowhere else. The zero-boundary case used there also follows from Hacon--Popa--Schnell and from Cao--Pu{a}un. Good minimal models of big klt adjoints, from Birkar--Cascini-- Hacon--McKernan [cite]. These are used in the signed- representative engine and in the reduction at positive Kodaira dimension. Ordinary negative-ample Kodaira--Saito vanishing and the Du~Bois machinery of Saito [cite], used in the Hodge step of the engine. No logarithmic Iitaka subadditivity is used in the engine, the exclusions, the jet construction, or the Frobenius comparison. The companion preprint [cite] enters only in the finite-generation corollary, after abundance has been proved. ## Notation (X,B) always denotes a log canonical pair with B effective, rational boundary unless stated otherwise. ``Nef'' means nonnegative on every curve. ``Semiample'' means a positive multiple is generated by global sections. ``Abundant'' means kappa(L)=nu(L), i.e.~the Iitaka dimension equals the numerical dimension. Real-boundary statements occur only in the induction (Section~sec:induction); the main theorem has rational boundary. All other conventions are collected in Section~sec:introduction, final subsection. newpage %====================================================================== # Introduction %====================================================================== Let (X,B) be a projective log canonical pair with rational boundary. Its log canonical divisor K_X+B packages the canonical class and the boundary. A rational Cartier divisor is nef if its degree on every curve is nonnegative; it is semiample if a positive Cartier multiple is generated by global sections. Such a multiple defines a morphism, so semiampleness turns numerical positivity into a fibration described by sections. Log abundance asserts that semiampleness holds whenever nefness does (Theorem~thm:main). The proof runs by induction on dim X. Its inductive hypothesis is stronger than the main theorem: every projective lc pair with effective real boundary and pseudo-effective real Cartier adjoint has a good log minimal model, on which the adjoint is semiample. Smooth canonical nonvanishing --- the statement that a pseudo-effective canonical class on a smooth projective variety has a nonzero pluricanonical section --- is proved at each dimension before the good-model output of that dimension is used. At the very end, descent of the evaluation map of an actual Cartier line bundle transfers the result over C to every algebraically closed field of characteristic zero (Section~sec:completion). Throughout, kappa(L) is the Iitaka dimension of a rational divisor L, with kappa(L)=-∞ if every positive integral system is empty; nu(L) is the numerical dimension. For nef L we always have kappa(L)lenu(L); abundance is the reverse ineuality. ## Consequences covered by this paper The consequences of the main theorem fall into four groups, each with its own section, each recording the exact hypotheses of the cited external theorems it consumes. Good models, linear triviality, slc, and dlt extension (Section~sec:consequences). Good models over C for real boundaries; the smooth-canonical good model; rational-linear triviality when kappa=0; abundance for semi-log-canonical pairs after normalization; supported extension of pluricanonical sections across the reduced dlt boundary. Finite generation of rounded log canonical rings (Section~sec:canonical-rings). Both absolute (projective over a field of characteristic zero) and relative (proper over an algebraic base over C). Classification-flavoured consequences (Section~sec:classification): rational connectedness via Mumford's criterion; base of the MRC fibration via cotangent tensors; finiteness of pi_1 when kappa=0 and chine0; quasi-projective universal covers in the sense of Claudon--H"oring--Koll'ar. Relative analytic abundance (Section~sec:analytic-abundance): projective surjective morphisms of normal complex analytic varieties with rational lc adjoint nef over the whole base, obtaining a relatively generated Cartier multiple in a neighborhood of each compact base subset. Each section states the theorem it applies precisely and separates the review literature from the contribution of this paper. ## Historical development The threefold case illustrates why nonvanishing and boundary geometry are central. Miyaoka [cite] proved the numerical-dimension one case by studying an effective pluricanonical divisor and its infinitesimal neighborhoods after suitable finite covers. Kawamata [cite] settled the remaining numerical-dimension two case for minimal threefolds; his Section~4 gives an alternative proof of Miyaoka's case via finite covers and Hodge theory. Keel--Matsuki-- McKernan [cite] proved the logarithmic theorem in dimension three. Fujino [cite] extended the statement to semi-log-canonical threefolds, where sections must agree along the conductor before they descend. This normalisation-and-gluing issue recurs in every dimension whenever the coefficient-one boundary is reducible. In higher dimensions, Birkar--Cascini--Hacon--McKernan [cite] established minimal models for klt pairs in the big-boundary or log-general-type range, and finite generation of the projective rational klt canonical ring without a bigness hypothesis cite[Corollary~1.1.2]{BCHM2010}. Two distinctions remain sharp on the boundary of the positive cone. [label=(alph*)] Nonvanishing produces sections but not semiampleness. Hashizume cite[Theorem~1.4]{Hashizume2018} reduces real-boundary lc nonvanishing and ordinary minimal-model existence in dimension le n to smooth canonical nonvanishing in dimension n. The induction must then make these models good. In positive Kodaira dimension the required step is Fujino--Gongyo cite[Lemma~3.5]{FujinoGongyoRings2014}. Kodaira dimension zero and numerical dimension zero are distinct hypotheses. Gongyo cite[Theorem~1.2]{Gongyo2013} proves that a numerically trivial rational lc adjoint is Q-linearly trivial. In the zero-Iitaka argument below, numerical triviality forces an already available effective representative to vanish. Liu and Xu [cite] handle numerical dimension at most one: good models for lc pairs of dimension at most five with nonnegative invariant Kodaira dimension and numerical dimension at most one cite[Theorem~5.1]{LiuXu2025}, and a reduction to smooth nonvanishing in that numerical range cite[Theorem~5.4]{LiuXu2025}. The induction here has no numerical-dimension restriction. The ordinary-pair specialisation of cite[Corollary~1.2]{OpenAIMinimalModels2026} concerns existence of a nef model and is not an input to the good-model induction; we use it, after abundance has been proved, in the finite-generation part of Section~sec:canonical-rings. ## The two tasks in the induction Assume good models exist in all dimensions below n. We must prove (i)~NV_n and (ii)~good models in dimension n. Both jobs use the same criterion for a signed representative on a reduced boundary, which we state here once and then prove in Sections~sg:section--% sg:descent. Let L be a nef divisor and nu(L) =max{j : L^jH^{n-j}>0} for H ample and n=dim X; this is the numerical dimension. The signed-representative criterion is [ L=K_X+D nef, L-cD pseudo-effective for some c>0, Lsim_{Q}textstylesum_i a_iD_i, L|_D semiample as a rational line bundle, ⇒ kappa(L)=nu(L). ] Here (X,D) is a projective Q-factorial dlt pair, and D=sum_iD_i is the decomposition of the reduced boundary into prime components. The rational coefficients a_i are allowed to be positive, negative, or zero. The hypothesis is semiampleness of L|_D on the entire reduced scheme D, so normalisation and conductor gluing are built in (they enter via cite[Theorem~1.5]{FujinoGongyo2014}). For the good-model task, positive Kodaira dimension is Fujino--Gongyo cite[Lemma~3.5]{FujinoGongyoRings2014}; the signed criterion handles the remaining zero-Iitaka case (Proposition~prop:inductive-reduction). The signed criterion is proved geometrically. A generated multiple of L|_D gives a morphism from D to projective space. Over a general point of the image of a positive-coefficient component, a root construction separates that component from the negative and coefficient-zero boundary. A filtered Hodge-module calculation lifts local parameters on this image, together with a parameter transverse to the lifted boundary, through all infinitesimal neighborhoods. Fixing the lifted base parameters and varying the transverse one deforms a boundary fiber off D to a projective subvariety. Its variations give a dominating family of subvarieties on which L is numerically trivial. The dimension of this family is n-nu(L), which bounds the base dimension of the nef reduction of L from below by nu(L); descent along the reduction then identifies L as the pullback of a big divisor on the base, so kappa(L)genu(L), hence abundance. The formal-deformation shape has a direct precedent in Miyaoka's construction, who credits Reid with the suggestion to analyse a pluricanonical divisor cite[p.~220]{Miyaoka1988}. Boundary extension provides an alternative approach: Demailly--Hacon--Pu{a}un cite[Theorem~1.7,Corollary~1.8]{DemaillyHaconPaun2013} use a singular- metric refinement of Ohsawa--Takegoshi extension under the hypothesis that the boundary is purely log terminal and the effective representative contains the coefficient-one divisor; Chan--Choi cite[Theorem~1.6.1]{ChanChoi2021} drop the containment condition. In the present paper the representative may have both signs, and compatibility is required across the whole reducible boundary. We therefore construct compatible formal lifts on the whole reduced boundary. Semiampleness on the boundary is Fujino--Gongyo cite[Theorem~1.5]{FujinoGongyo2014}. The ingredients are familiar. Normalised cyclic covers and their eigenspace decomposition follow Esnault--Viehweg cite[Section~3]{EsnaultViehweg1992}. Deligne's logarithmic Hodge theory [cite] is the ambient framework. On the singular boundary we use Saito's projective strictness and graded de Rham vanishing with his comparison to the Du~Bois complex cite[Theorem~2.14,Proposition~2.33]{Saito1990} cite[Theorem~0.2,Corollary~0.3]{Saito2000}. The paper's contribution here is to identify the obstruction to lifting and to show that these ordinary-variety statements apply to the covering construction; Section~sec:hodge carries out exactly this identification. ## Smooth nonvanishing The proof of NV_n assumes a counterexample X (smooth projective n-fold, K_X pseudo-effective, kappa(X,K_X) =-∞) and proceeds through four stages. Exclusions: (Section~sec:exclusions). The zero-boundary case of logarithmic Iitaka subadditivity (Theorem~asm:iitaka) rules out positive irregularity via the Albanese map. This is its only use. Geometric preparations then exclude three ways the counterexample could be covered by lower-dimensional subvarieties: families carrying a positive current pulled back from a smaller variety; families of curves of uniformly bounded genus and uniformly bounded degree, after normalising the nef divisors to a fixed small volume; families of birational correspondences between fixed finite covers whose graphs dominate the product. The first two use positive currents, algebraic webs, and an ascending chain condition for Lelong numbers at valuations of bounded discrepancy. The third uses Hanamura's birational-group theorem. The signed- representative theorem plays a second role here: it forces the full reduced support of any signed canonical representative to have a big logarithmic adjoint (Proposition~prop:signed-alternative). This controls the multisections of the two-slot construction below. Jets: (Sections~sec:common-tools and~sec:jets). The exclusions produce two incompatible estimates for vanishing at a moving point. At a very general point, every nonzero section of every Cartier multiple kP of a suitable normalised nef divisor P has vanishing order bounded by k times a small normalised constant. On the other hand, on the projective bundle over Xtimes X associated to two copies of a line bundle, the Seshadri constant at a very general point exceeds 2n+2, so jets of very high order exist. The two summands let the argument move in either factor. The first estimate fails only if the corresponding jet kernel has a proper moving base component. Slicing and the curve exclusions leave only components that project generically finitely in both factors, hence give correspondences; their branch divisors cannot sweep either factor, and a fixed branch complement with bounded degree leaves only finitely many covers. After fixing the covers, the correspondences form a family of birational maps whose graphs dominate the product. Hanamura then makes a fixed cover birational to an abelian variety; the ramification-and-current argument of Section~sec:exclusions returns a pluricanonical section on the counterexample, a contradiction. This closes the two-slot estimate and therefore the jet separation. Frobenius: (Section~fr:section). Once the projective bundle, its divisors, and the required finite collection of jet sections are fixed, reduction modulo a large prime gives an evaluation map on the product of two copies of the Frobenius-twisted model. That map has full generic rank. Its restriction to the ordinary diagonal is controlled by sections on the Frobenius thickening of the diagonal, where the two relative Frobenius maps agree; a canonical filtration of those sections cuts the rank by a factor of four. The determinant of any nonzero maximal minor is a section of an external product of two line bundles. The first estimate --- expressed by a fixed movable curve on a blowup --- bounds the order of the determinant at the selected diagonal point. The rank deficit bounds the same order from below by 3/4 of the rank. Contradiction for large primes. Theorem~common:finite-data-frobenius states this incompatibility for a fixed smooth variety, fixed jet sections, and a fixed movable curve bound. Its proof opens Section~fr:section and uses neither abundance nor subadditivity. The geometric application follows in the same section, using the fixed inputs from the exclusions and from the jet constructions. The numerical sub-adjunction and moving-kernel results needed for those constructions are proved in Section~sec:common-tools. The numerical part combines two established methods. The argument following a moving base component uses parameter differentiation as in Ein--K"uchle--Lazarsfeld cite[Proposition~2.3]{EinKuchleLazarsfeld1995}; the paper supplies the tracking and adjunction estimates needed here. After reduction to positive characteristic, ordinary jets are compared to Frobenius jets using the local containment recorded by Mustac{t}u{a}--Schwede cite[Equation~(2.8)]{MustataSchwede2014}. The resulting evaluation map is studied through Sun's canonical Frobenius filtration cite[Theorem~3.7]{Sun2008} (its curve case is treated in Joshi--Ramanan--Xia--Yu cite[Section~5.3]{JoshiRamananXiaYu2006}), using the description by powers of the diagonal ideal due to Kitadai--Sumihiro cite[Definition~3.1,Remark~3.2]{KitadaiSumihiro2008}. Langer's instability estimate cite[Corollary~2.5]{Langer2004} controls the slopes of the graded bundles. ## The subadditivity input The all-dimensional argument uses the following companion result. Its application is confined to Lemma~ex:irregularity. [Logarithmic Iitaka subadditivity, cite[Corollary~6.2]{OpenAIIitaka2026}] Let f:Xto Y be a surjective morphism with connected fibres between smooth connected projective complex varieties. Let D_X,D_Y be reduced effective SNC divisors, allowing zero boundaries, with Supp(f^*D_Y)⊆Supp(D_X). For a very general smooth fibre F, put D_F=D_X|_F. Then [ kappa(X,K_X+D_X)gekappa(F,K_F+D_F)+kappa(Y,K_Y+D_Y). ] Here D_F is reduced with simple normal crossings and K_F+D_Fsim(K_X+D_X)|_F. The convention (-∞)+b=-∞ also holds when b=-∞, and the zero divisor on a point has Iitaka dimension zero. The support inclusion allows extra components in D_X; the morphism need not be smooth away from D_X. The theorem imposes no nefness, abundance, bigness, or good-model condition. Its zero-boundary case is used in the Albanese reduction of Section~sec:exclusions; this is the only subadditivity input in the proof. The companion's characteristic-zero specialisation cite[Corollary~6.3]{OpenAIIitaka2026} uses a geometric generic fibre; the present paper keeps the complex, very-general-fibre statement distinct. No fractional-boundary, orbifold, non-projective, Hodge-conjectural, or arithmetic extension of Theorem~asm:iitaka is assumed here. The zero-boundary application in Lemma~ex:irregularity also follows from the external maximal-Albanese-dimension theorem of Hacon--Popa--Schnell cite[Theorem~1.1]{HaconPopaSchnell2017}: the smooth base there maps generically finitely onto a subvariety of an abelian variety, and Cao--Pu{a}un cite[Theorem~1.1]{CaoPaunAbelian2016} prove the underlying abelian-base case. Those cover the application, not the full logarithmic statement. ## Reading order Section~sec:induction sets out the dimension induction and proves the passage from nonvanishing to good models. Sections~sg:section--% sg:descent prove the signed-boundary criterion: first the root and adjunction construction, then the infinitesimal lifting, then the final descent. Section~sec:exclusions derives the geometric exclusions. After the local tools of Section~sec:common-tools, Section~sec:jets constructs the scalar and two-slot jets. Section~fr:section proves the independent Frobenius comparison and then uses it to establish smooth nonvanishing. Sections~sec:completion and~sec:consequences finish the induction, descend to characteristic zero, and record the boundary consequences. Section~sec:canonical-rings proves finite generation. Section~sec:classification records the cotangent, fundamental-group, and quasi-projective-cover consequences. Section~sec:analytic-abundance records the relative analytic statement. A reader using only the finite-data comparison can start with Theorem~common:finite-data-frobenius and its proof in Section~fr:section; neither assumes the counterexample geometry. ## Conventions A variety is integral. Canonical divisors are chosen compatibly on birational models. For a divisor E on a smooth model p:Wto X, the log discrepancy is [ a(E;X,B)=1+ord_E(K_W-p^*(K_X+B)). ] Log canonical means a(E;X,B)ge0 for every divisor over X; klt means a(E;X,B)>0. We use the standard dlt, plt, and slc conventions. For non-nef pseudo-effective divisors, statements about numerical dimension use Nakayama's kappa_sigma as specified in the relevant reduction; these are not silently interchanged with the intersection-theoretic numerical dimension nu. For a divisor with nonempty positive system, the Iitaka dimension is the maximum dimension of the images of its positive systems under the associated rational maps. We write equiv for numerical equivalence and sim_{Q} for rational-linear equivalence. Unless a different base field is specified, constructions are over C. A very general point of a variety avoids a countable union of proper closed subvarieties. The transfer to arbitrary algebraically closed characteristic-zero fields is made in Section~sec:completion only after the complex argument has been completed. %====================================================================== # The inductive framework %====================================================================== The dimension induction has two outputs in each dimension: smooth canonical nonvanishing (NV_n) and good log minimal models for projective lc pairs with effective real boundary and pseudo-effective adjoint. The present section proves the passage from the first output to the second, given good models strictly below the current dimension. All varieties here are projective over C. ## Definitions and the induction hypothesis For an effective real boundary Delta, a good log minimal model of (X,Delta) is a log minimal model on which the adjoint divisor is semiample. For a real boundary, semiampleness means real-linear equivalence to the pullback of an ample real divisor by a contraction. We use the standard definition of log minimal model in which extracted prime divisors are included with coefficient one. For a pseudo-effective divisor A we write kappa_sigma(X,A) for Nakayama's numerical dimension. When A is nef it agrees with the intersection-theoretic numerical dimension nu(X,A). Fix an integer nge1. [Lower-dimensional good models] Every projective log canonical pair of dimension 0 with mL Cartier. For every rational [ 0<ε<(1)/(4nm+2), ] every step of an LMMP for L-ε P is L-trivial. On every model of the program, the transform of L is nef, and its multiple by the same integer m is Cartier. Proof. Both Delta-ε P and Delta-1/2P are effective klt boundaries. If R is a ray negative for L-ε P, nefness of L makes R negative for L-1/2P. Choose a rational curve C spanning R and satisfying the length bound [ -(L-1/2P)· Cle2n. ] Writing a=L· Cge0 and p=P· C, we have [ ple2a+4n, a<ε p, (1-2ε)a<4nε. ] The choice of ε gives a<1/m. Since ma is a nonnegative integer, a=0. For the driving klt contraction c:Zto B, the line bundle OO_Z(mL) is numerically trivial over B. The line-bundle clause of the contraction theorem gives its descent to a line bundle cA_B on B with no change of m cite[Theorem~3.74(3)]{Fujino2017}. (One may also see the unchanged index directly from relative basepoint freeness: two consecutive sufficiently large powers descend, so their quotient descends.) The line bundle cA_B is nef because its pullback OO_Z(mL) is nef. For a flip c^+:Z^+to B, its pullback (c^+)^*cA_B is the line bundle of the transformed mL. This proves, separately, that the transformed L is nef and that mL remains Cartier with the same m. The driving boundary remains klt throughout its LMMP. The transformed Delta-1/2P remains effective and is smaller than the driving boundary, so it too is klt. The same estimate and the same integer m apply inductively at every subsequent step. ## The good-model implication We now assume smooth nonvanishing in the current dimension and prove the good-model implication. The argument first reduces real nef boundaries to rational ones. Positive Kodaira dimension is then handled by lower-dimensional good models; what remains is the zero-Iitaka case and its klt perturbations. [Inductive reduction to smooth nonvanishing] Assume Assumption~mmp:lower-models. Assume also that every smooth projective n-fold with pseudo-effective canonical divisor has nonnegative Kodaira dimension. Then every projective lc n-fold with real boundary and pseudo-effective adjoint has a good log minimal model. In particular, every nef rational lc adjoint in dimension n is semiample. Proof. Hashizume's reduction cite[Theorem~1.4]{Hashizume2018} gives nonvanishing and log minimal models for projective lc pairs with real boundary in dimensions at most n. Pass to a Q-factorial dlt log minimal model. Fix the support of its boundary, impose the rational affine constraints that its coefficient-one components remain equal to one, and take a sufficiently small rational polytope around the given boundary within those constraints. On a fixed log resolution witnessing dlt, the strict discrepancy ineualities remain strict on this neighbourhood, so its boundaries remain dlt. The rational-polytope theorem for nef adjoints cite[Remark~3.1, Proposition~3.2(3)]{Birkar2011II}, applied to all extremal rays and intersected with this neighbourhood, expresses the given real boundary as a finite rational simplex of dlt boundaries with nef adjoints. It is enough to prove semiampleness for these rational adjoints: their positive real combinations are semiample by multiplication of the associated contractions. (One may also rationalise semiampleness of a rational adjoint directly: the finite system of linear equations expressing its divisor and principal divisors has rational data, so a real solution with positive coefficients has a rational one.) We therefore work with a rational dlt pair (Z,Delta) and L=K_Z+Delta nef. Real nonvanishing implies rational nonvanishing here: writing a real effective representative in terms of finitely many divisors and principal divisors and solving the resulting rational linear system with nonnegative rational coefficients produces a rational one. If kappa(Z,L)ge1, lower-dimensional good models give a good minimal model by cite[Lemma~3.5]{FujinoGongyoRings2014}, and Lemma~mmp:comparison transfers semiampleness back to Z. It remains to treat kappa(Z,L)=0. The non-pseudo-effective perturbation.: Put P=Delta in the klt case and P=floor(Delta) otherwise. Suppose L-ε P is not pseudo-effective for every sufficiently small ε>0. Choose rational ε as in Lemma~mmp:trivial-perturbation, and run the klt LMMP with scaling to a Mori fibre space [ (Z,Delta-ε P)dashrightarrow(Z',Delta'-ε P') xrightarrow{f}T. ] Existence and termination in the non-pseudo-effective case are the established klt results of [cite]. The program is crepant for L. The line-bundle descent in Lemma~mmp:trivial-perturbation, applied also to the final contraction, gives a nef rational divisor A on T with [ K_{Z'}+Delta'sim_{Q}f^*A. ] The base T has dimension less than n. If (Z,Delta) is klt, the transformed original pair (Z',Delta') is klt as well. Ambro's descent theorem cite[Theorem~0.2]{Ambro2005} gives an effective rational boundary Delta_T such that (T,Delta_T) is klt and [ Asim_{Q}K_T+Delta_T. ] All the hypotheses hold: the total pair is globally klt and effective, f is a projective contraction, and the adjoint is rationally linearly pulled back. In particular this use requires no conjecture on semiampleness of a moduli divisor. Assumption~mmp:lower-models and Lemma~mmp:comparison make A semiample. Otherwise P' is effective and relatively ample, since K_{Z'}+Delta'-ε P' is relatively antiample. Some component of the floor therefore dominates T. Take a crepant dlt blowup b:(widetilde Z,widetildeDelta)to(Z',Delta') and choose the strict transform S of such a component. The restriction f_S=(fcirc b)|_S:Sto T is surjective. Crepancy and adjunction give an lc pair on S whose nef adjoint is rationally linearly equivalent to f_S^*A; it is semiample by the lower-dimensional hypothesis. Semiampleness of a rational line bundle descends along a proper surjection: use the equality of sections after Stein factorisation for the connected fibre, and norms for the remaining finite morphism. For the latter: choose a section of a basepoint-free multiple nonzero simultaneously at all points of the chosen finite fibre; such a section exists because each vanishing condition is a proper linear subspace, and finitely many proper linear subspaces do not cover a complex vector space. Its norm is nonzero at the downstairs point. Hence A is semiample in this case as well, and crepancy transfers the conclusion back to L. The klt pseudo-effective perturbation.: We next settle the klt case where L-εDelta is pseudo-effective for some small rational ε>0. Nonvanishing gives [ Lsim_{Q}G_0:=H_0+εDelta, H_0ge0. ] Take a resolution p:Wto Z with reduced SNC divisor D containing all exceptional divisors and the strict support of H_0+Delta. Then [ K_W+D=p^*L+R, Rge0. ] Every component of D has positive coefficient in G_W:=p^*G_0+Rsim_{Q}K_W+D. Indeed, at strict boundary components the difference between coefficient one and a klt boundary coefficient is positive, and at exceptional components the coefficient contributed by the adjoint comparison is the positive log discrepancy. The pushforward p_*G_W is bounded coefficientwise by a fixed multiple of G_0, and sections inject under birational pushforward, so [ 0lekappa(W,K_W+D)lekappa(Z,G_0)=0. ] Take a Q-factorial dlt log minimal model (V,D_V) of (W,D), and fix a common resolution [ Wxleftarrow{q}Uxrightarrow{v}V. ] Its boundary is reduced. On U the comparison divisor of Lemma~mmp:comparison is exceptional over V; pushing q^*G_W forward by v gives [ K_V+D_Vsim_{Q}G_V=sum_ib_iD_i, b_i>0, Supp G_V=D_V. ] (For completeness: at a component extracted on V, positivity also holds since its log discrepancy over (W,D) is zero, its centre lies in D, and the pullback of the full-support effective G_W has positive order there. At a prime retained on V, there is no comparison-divisor coefficient.) Effectivity and exceptionality in the comparison preserve the adjoint section ring, so the nef adjoint on V still has Iitaka dimension zero. The small-perturbation hypothesis of Theorem~thm:signed is now explicit. If D_Vne0, choose rational c with 00 with 2εledelta, small enough to keep the perturbed boundaries effective and klt. Convexity of the pseudo-effective cone, applied between L and L-delta P, makes both [ L-ε P, L-2ε P, P=floor(Delta), ] pseudo-effective klt adjoints. Their Iitaka dimensions are zero: nonvanishing gives the lower bound, and adding the effective perturbation bounds each by kappa(Z,L)=0. The preceding klt case gives good models for both of these perturbed adjoints, so their Nakayama numerical dimensions are zero. Since [ 2(L-ε P)=L+(L-2ε P), ] and the last summand has an effective rational representative, monotonicity and homogeneity of kappa_sigma give [ kappa_sigma(Z,L)lekappa_sigma(Z,2(L-ε P))=0. ] The nef divisor L is therefore numerically trivial. Nonvanishing is now already available in this finishing case: write Lsim_{Q}Ege0. For an ample divisor H, E· H^{n-1}=0 forces E=0. Hence Lsim_{Q}0, without any further nonvanishing or abundance premise. This settles the remaining zero-Iitaka case, so every rational nef dlt adjoint considered above is semiample. Returning to the finite rational simplex proves the real-boundary good-model assertion, and Lemma~mmp:comparison descends semiampleness to every original nef rational lc pair. %====================================================================== # The signed-representative engine %====================================================================== The induction, at its key step, has a nef log canonical divisor L=K_X+D on a Q-factorial dlt pair (X,D) with D reduced. The hypothesis of the signed criterion is that some rational-line representative of L --- not necessarily the boundary itself --- is supported on D with either signs. The difficulty is that L|_D is semiample but L itself is only nef. The engine of this section converts that hypothesis into abundance. The outline, already given on page~pageref{:engineering}, is repeated here in numerical form so the reader knows what the individual lemmas are for. The semiample L|_D gives a morphism [ phi:D→ P:=PP^s, N_0|_D~=phi^*OO_P(1) ] for a sufficiently divisible positive integer m, where N_0:=OO_X(mL). Write s_0 for the rational section of N_0 having divisor mG, where G=G_+-G_- comes from the signed representative and G_pm,Ege0. Let r+1 be the dimension of the image of a positive-coefficient component of D in P; a lemma below shows that every other component has image of dimension at most r, and that G_+ restricted to Supp G_-cup D_0 has image of dimension less than r. The root construction of Proposition~prop:signed-construction produces, from this data, a projective variety X with a reduced Cartier divisor SsubsetX and a finite representable morphism Sto Dtimes_PP whose image covers the positive boundary, such that S is (generically) its own root line and carries a Stein factorisation [ Sxrightarrow{g}Y{p}PP^s ] with g finite representable and p a projective bundle. The factorisation is the geometric output that the Hodge step of Section~sec:hodge consumes. That step lifts functions on the positive boundary to every infinitesimal thickening, deformes them off D, and produces a fixed moveable family of subvarieties on which L is trivial. Descent along the resulting nef reduction (Section~sg:descent) then proves abundance. The full theorem is: [Signed-boundary abundance criterion] Let X be a projective Q-factorial dlt variety of dimension n, let D=sum_iD_i be a reduced divisor whose support is the coefficient-one part of the boundary, and let L=K_X+D. Suppose: [label=(roman*)] L is nef; L-cD is pseudo-effective for some rational c>0; L|_D is semiample as a rational line bundle on the reduced scheme D; the support D is nonempty and 00, semiampleness of L|_D identifies the numerical dimension of each D_i's image in P with the numerical dimension of L|_{D_i}, which equals the intersection D_iL^{nu(L)-1}H^{n-nu(L)}=dimphi(D_i). This intersection is zero by the vanishing just proved, so the image dimension of every D_i is at most r=nu(L)-1. Part (ii). By the same vanishing, [ 00 and at least one term is strictly positive. The corresponding component of G_+ attains image dimension exactly r. Part (iii). Consider the symmetric matrix Q_{ij}=D_iD_jL^rH^{n-r-2}. Distinct effective Q-Cartier divisors have effective intersection cycles, so the off-diagonal entries are nonnegative. The mixed Hodge index theorem gives the ambient intersection form at most one positive direction: approximate by ample classes and apply the ordinary Hodge index theorem on complete-intersection surfaces. The class L is isotropic and orthogonal to every D_i (this is the vanishing from part (i)), while its pairing with H is strictly positive: [ Q(L,L)=L^{r+2}H^{n-r-2}=0, Q(L,D_i)=D_iL^{r+1}H^{n-r-2}=0, Q(L,H)=L^{r+1}H^{n-r-1}>0. ] The intersection form on the orthogonal complement of L is therefore negative semidefinite, and this holds on the span of the D_i. Pulling back to a resolution preserves the conclusion, so smoothness of X is not needed. But any strictly positive off-diagonal entry would contribute a positive direction to the span; the entries therefore have image dimension strictly less than r wherever their images have r-dimensional support, forcing the locus Supp G_+ cap(Supp G_-cup D_0) to have phi-image of dimension less than r. We will also place the Stein image in a smooth target: this allows differential operators and projective Hodge-module direct images to be used on ordinary scheme charts. The infinitesimal neighborhoods will later be cut down to a projective fiber before they are algebraized. Choose ell divisible by m and every nonzero integer |ma_i|, and set a=ell/m. There is a normal tame Deligne--Mumford stack X, considered on a neighborhood of a reduced Cartier divisor S of pure dimension N=n-1, together with a morphism to X and a reduced boundary T having no component in common with S, with the following properties. [label=(roman*)] The pair (X,S+T)X,S+T) is log canonical, and a fixed isomorphism of reflexive sheaves is given by [ omega_{X}+T)~=OO_{X} tag{(ast_{adj})} ] The support of T is locally set-theoretically principal. The ambient space and S are Cohen--Macaulay on scheme charts. Off T, the ambient space is Gorenstein and the displayed isomorphism is ordinary Cartier adjunction data. Put J=(Scap T)_{red}. Both S and J are Du Bois on scheme charts, the ideal I_{J/S} is maximal Cohen--Macaulay, and [ B:=B:= Hom_S( I_calI_{J/S},omega_S) ~=OO_S(aS). ] The identification agrees off J with the residue convention specified by (sg:ambient-adjunction). There is a finite representable morphism [ S→ Dtimes_PP ] whose image covers Supp G_+. The line OO_S(S) is the pullback of C. The image of J in P has dimension less than r, whereas the image of S has dimension r. There is a smooth projective bundle p:Y PcalP and a representable projective morphism g:StoY. This is a morphism from the reduced divisor S; no extension to a neighborhood of S in X is asserted here. Writing h=pcirc g and T={P}*OO_S, the morphism g factors through a closed embedding T↪T↪Y, and [ g_*OO_S=OO_{T}, OO_S(S)=g^*C, B=g^*C^a.B=g^*C^a. ] For a separated quasi-projective scheme chart UtoY that is etale and of finite type, put S_U=Stimes_{Y} The etale morphism S_Uto S extends compatibly to every nilpotent thickening qS in X. Only this etale chart is extended; an extension of g to qS is not needed to define the chart. The resulting qS_U are quasi-projective schemes, separated and quasi-finite over X. All the adjunction and duality identifications are compatible with changes of etale chart. Proof. We construct the divisor and its adjunction data first, then place its Stein image in a smooth projective bundle. The final step verifies that the infinitesimal neighborhoods have the scheme charts required for formal lifting. smallskipnoindentRoots and a fixed adjunction isomorphism. Let rho:R X be the normalized simultaneous root stack of the Cartier divisors mG_+, mG_-, and mD_0, taking an ell-th root of each divisor together with its section; the construction allows empty divisors. Denote the tautological root divisors by S_+,S_-,S_0 on R. Their reducedness can be checked at a generic prime of downstairs multiplicity b. The chosen divisibility conditions give bmidell. A normalized chart of y^ell=x^b has ramification index ell/b, and y has order one. Since m divides ell, this also applies to D_0. The divisors are Cartier, so normality and generic reducedness give their reducedness everywhere. Put B=S_++S_-+S_0 on R. With this full reduced boundary, log ramification gives a log canonical pair and the relation [ K_{R}+Bsim_{Q}a(S_+-S_-). ] The ambient charts are klt. They are klt off the boundary, and, on a log resolution, decreasing every boundary coefficient slightly removes all zero-discrepancy places. The difference between the two sides is a torsion integral Weil class, represented by [ F= (omega_{R} ⊗ OO_{R}-a(S_+-S_-)))^{**}. ] Choose a periodicity isomorphism F^{[q]}~=OO_{R}, where square brackets denote reflexive powers. It defines multiplication in the finite algebra [ A=⊕_calA=⊕_{i=0}^{q-1}F^{[i]}. ] Let tau:R^ RcalR be the normalization of Spec_{R} A.A. In codimension one this is the ordinary cover of a torsion line bundle and is etale. The cover preserves log canonicity, klt ambient singularities, and the reduced Cartier boundary. From now on S_+,S_-,S_0 denote their pullbacks to R^ Tautological evaluation on this cover trivializes the pullback of F in codimension one and hence fixes the adjoint isomorphism. Because this evaluation is a sheaf morphism on the cover stack itself, the adjoint isomorphism is equivariant on every atlas. This equivariance will be needed for the quotient below. smallskipnoindentSeparating the positive and negative parts. Let p_1:Rtilde R^R^dagger be the normalized blowup of the ideal of S_+cap S_-. The spaces constructed so far are related by [ Rtilde{p_1}R^ xrightarrow{tau}R{rho}X. ] The blown-up ideal is locally generated by two elements, so the ordinary blowup embeds in a relative projective line. Its fibers have dimension at most one, as do the fibers after finite normalization. Hence each exceptional divisor lies over a codimension-two component of the intersection. Downstairs such a component is a dlt stratum, generically SNC. Separate normalized Kummer charts and purity for the index cover at the smooth generic locus give the same description upstairs. The tautological exceptional divisor E is consequently reduced Cartier, and the divisors [ S'_+=p_1^*S_+-E, S'_-=p_1^*S_--E ] are reduced Cartier and disjoint on Rtilde. Since no codimension-two two-branch stratum is contained in S_0, its pullback p_1^*S_0 is reduced Cartier and has no exceptional component. The total boundary [ S'_++S'_-+E+p_1^*S_0 ] is crepant and log canonical. To check the ambient singularities as well, denote this boundary on a chart V' by B'. The crepant equality gives log canonicity for every valuation. Outside B', the blowup is an isomorphism to the old klt boundary complement. A divisor F with a(F;V',B')=0 therefore has center in Supp B'. Since B' is effective Cartier, ord_F(B')>0, whence [ a(F;V',0)=a(F;V',B')+ord_F(B')>0. ] Positive pair discrepancies also remain positive after the boundary is dropped. Thus V' is klt even at its higher-codimension singular loci. On Rtilde, put S=S'_+ and T=S'_-+E+p_1^*S_0. Near S, the section of the disjoint divisor S'_- is a unit, and the fixed linear equivalence reads [ omega(S+T)~=OO(aS). ] Klt singularities make the ambient charts Cohen--Macaulay. All the boundary divisors in this display are Cartier, so the isomorphism also makes these charts Gorenstein. We also need finiteness of Sto S_+ on R^ Locally write S_+=(u=0) and S_-=(v=0). Before normalization, the strict transform of S_+ lies in the blowup chart with ratio u/v, where its equation is u/v=0. Thus each fiber of this strict transform over S_+ has at most one point. Properness and finite normalization prove finiteness of Sto S_+. A codimension-one point of Scap T therefore lies over a codimension-two intersection downstairs. The generic SNC description shows that T|_S is generically reduced. Being Cartier on the Cohen--Macaulay scheme S, it is reduced everywhere. smallskipnoindentRetaining the required root line. The construction so far has supplied reduced Cartier divisors and fixed adjunction data. We now remove the root characters that are unnecessary, while preserving the line attached to S. On Rtilde the line OO(S-S'_-) has ell-th power equal to the pullback of N_0 and hence defines a morphism to the gerbe of ell-th roots of N_0 on X. Denote this gerbe by G, and let kappa:Rtilde XcalX be the relative coarse-space morphism over G. On a trivializing chart, we quotient by the kernel of the finite-group character on the indicated root line. The quotients are ordinary quasi-projective schemes: the covers used above are finite, the blowup is projective, and finite quotients preserve quasi-projectivity. These chartwise quotients patch. Before taking the quotient, the section of the disjoint divisor S'_- is a unit near S. There the retained line OO(S-S'_-) identifies with OO(S), so the line of S and its section both descend through the kernel quotient. Continue to write S,T for their reduced images under kappa; they now lie on X. In particular S remains Cartier. The finite-group norm of a local equation for T upstairs makes its image set-theoretically principal; Cartierness of the reduced image T is not needed. Divisorial ramification occurs only on the boundary. Log ramification thus preserves log canonicity and identifies the descended adjoint isomorphism with (sg:ambient-adjunction). The ambient chart and S are Cohen--Macaulay because their structure sheaves are invariant direct summands of the finite CM covers. More explicitly, the kernel acts trivially on the retained root line and hence on OO(aS). Equivariance of the fixed isomorphism gives the same kernel character on omega(S+T). Taking invariants descends the isomorphism, and log ramification identifies the invariant log dualizing sheaf. Off T, this isomorphism makes the canonical sheaf invertible, giving the asserted Gorenstein property. smallskipnoindentDuality on the boundary. The reduced supports S and J are unions of log canonical centers, since intersections of lc centers are again unions of lc centers. Both are therefore Du Bois by cite[Theorems~1.4 and~1.7]{KollarKovacs2010}. To identify the ideal and its dual, work on a quotient chart and denote the finite cover above by f:S^{up}to S. Reducedness of T^{up}|_{S^{up}} gives [ I_{J/S} = (f_*OO_{S^{up}} (-T^{up}|_{S^{up}}))^{inv}. ] An invariant function vanishes on the reduced quotient image exactly when its pullback vanishes on this reduced preimage. The displayed invariant sheaf is maximal Cohen--Macaulay: the upstairs sheaf is invertible on a CM scheme, finite pushforward preserves its depth over the quotient, and in characteristic zero invariants form a direct summand. Apply finite-map duality with trace and then take invariants: [ Hom_S( I_calI_{J/S},omega_S) = (f_*omega_{S^{up}} (T^{up}|_{S^{up}}))^{inv} ~=OO_S(aS). ] The last isomorphism comes from Cartier adjunction upstairs and descent of the line of S. Finite-map duality applies in the presence of ramification; it requires no invertibility of omega_S itself. Off J, the identification is the ordinary residue fixed by the ambient isomorphism. Normalized trace preserves this residue convention, so the identifications agree on overlaps. smallskipnoindentThe target of the lifting argument. On S, the root gerbe of N_0 is Dtimes_PP. Finiteness of the strict divisor, proved above, and removal of the relative inertia kernel give a finite representable map Sto Dtimes_PP. Its image covers the positive boundary, and the root line is OO_S(S)=C. Lemma~sg:numerical-boundary therefore gives image dimension r for S in P. To check the smaller image of J, recall that before the quotient S'_- is disjoint from S, the exceptional divisor E maps into Supp G_+capSupp G_-, and p_1^*S_0 maps into D_0. Taking the quotient changes none of these images in X. The image of J in D is consequently contained in [ Supp G_+cap(Supp G_-cup D_0). ] The same lemma gives image dimension less than r for this locus in P. We have thus obtained a representable projective morphism h:StoP; it remains to embed its Stein image in the required smooth bundle. The Stein algebra defines a finite stack T={P}*OO_S over P. This coherent algebra is generated as a module by vector bundles on P. Indeed, decompose it by the finitely many inertia characters, twist each summand by a power of C so that it descends to P, and use Serre generation there. The resulting vector-bundle surjection embeds T in a vector bundle over P. Since T is proper over the base, it remains closed in the projective completion. This completion is Y and gives the map g:StoY from the reduced divisor. The Stein construction yields g_*OO_S=OO_{T}, and the line identities follow from those already established. Finally, start with the etale chart S_Uto S obtained from g. Invariance of the etale site under nilpotent thickening extends this chart uniquely to qS_Uto qS, compatibly as q varies. This construction does not use a map qStoY; lifting the map from S is the task of the next section. The reduction S_U is a scheme; its extension qS_U is an algebraic space with trivial inertia. The chosen chart is separated, the map to the root gerbe is finite, and that gerbe has finite diagonal, so qS_U is separated over X. The finite representable map Sto Dtimes_PP and the etale map S_Uto S show that the finite-type geometric fibers of S_Uto X are zero-dimensional. Nilpotent thickening changes neither geometric points nor fiber dimensions. The morphism qS_Uto X is therefore quasi-finite as well. Zariski's main theorem embeds qS_U in a scheme finite over X, making it a quasi-projective scheme as required. # Hodge theory and formal lifting Our goal is to lift functions and a transverse parameter through every infinitesimal neighborhood of the divisor constructed in Proposition~prop:signed-construction. We recall its data. (P=[ell]{OO_{PP^s}(1)}) is a root gerbe, (C) is its tautological line bundle, and [ Sxrightarrow{g}Y{p}P, h=pcirc g. ] Here (g) is representable and projective, and (p) is a smooth projective bundle. The Stein factor of (h) is finite over (P; it has been embedded as (T YcalY) so that (g_*OO_S=OO_{T} The reduced Cartier divisor (SsubsetX has dimension (N=n-1). If (J=(Scap T)_{red}), then (S,J) are Du Bois, (I_{J/S}) is maximal Cohen--Macaulay, and B:=B:= Hom_S( I_calI_{J/S},omega_S) ~= OO_S(aS)=g^*C^a, OO_S(S)=g^*C, for an integer (a>0). Off (T), the fixed ambient isomorphism (omega_{X}~=OO_{X} supplies these identifications by adjunction. Put (I=OO_{X}-S)). For a separated quasi-projective scheme etale chart (UtoY of finite type, write (S_U=Stimes_{Y} and again (g:S_Uto U) for the induced projective morphism. Denote by (qS_U) the unique extension of (S_Uto S) to the nilpotent thickening (qS). These extensions are compatible in (q). Proposition~prop:signed-construction makes them quasi-projective schemes: they are separated and quasi-finite over (X), and hence open in schemes finite over the projective variety (X). Powers and quotients of (I) below are pulled back to the appropriate thickenings. On these quotients, (g_*) denotes direct image for the common underlying topological map; a ringed-space map from a thickening to (U) has not yet been constructed. For every such chart (U), every (jge0), and every (kge1), the transition [ g_*(I^j/I^{j+k+1}) → g_*(I^j/I^{j+k}) ] is surjective as a map of sheaves of abelian groups. Its kernel is (OO_{T_U} ⊗ C^{-j-k}), where (T_U= TcalTtimes_{Y} The statements are compatible with further etale changes of charts. On an affine chart, these transitions are also surjective on global sections. In their isolated-component case with numerical dimension one, Liu--Xu use finite covers and Du Bois cohomology to obtain infinitesimal motion cite[Theorem~3.1]{LiuXu2025}. Here the semiample boundary image may have positive dimension, so its local functions and the transverse parameter must be lifted together. For (jge0) and (kge1), the kernel of the transition from length (k+1) to length (k) in (I^j/I^{j+k+1}) is (OO_S ⊗ C^{-j-k}). The corresponding connecting homomorphism therefore takes values in [ R^1g_*OO_S ⊗ C^{-j-k}. ] We will identify this sheaf with a negative twist of the lowest filtered piece (F_0M) of a Hodge-module direct image. Lifting local functions gives Cech cocycles in this coherent sheaf. A calculation with their graphs shows that the coordinate cocycle lies in the kernel of the differential-operator symbol map. Negative-twist vanishing will make that kernel have no global sections. The same graph calculation then kills the error in lifting the transverse parameter. ## The lowest Hodge piece We work with graded-polarizable mixed Hodge modules on ordinary complex algebraic varieties. The inputs are projective strictness and the usual functorial operations cite[Theorem~2.14 and Section~4]{Saito1990}, the comparison with the Du Bois complex and its coherent dual cite[Theorem~0.2, Corollary~0.3 and Proposition~5.3]{Saito2000}, and Kodaira--Saito vanishing cite[Proposition~2.33]{Saito1990}. On a smooth stack, we use compatible systems of these objects on scheme etale charts and descend their filtered differential modules. Thus the argument requires no Hodge-module theory on stacks. All differential modules in this section are right modules with an increasing filtration. In the Spencer de Rham complex the module itself occupies degree zero. Accordingly, on a smooth chart (V), the term in degree (-i) of (Gr^F_kDR_V(M)) is [ Gr^F_{k-i}M ⊗ _{OO_V}∧^iT_V. ] When (F_{<0}M=0), the lowest graded de Rham complex therefore consists of the sheaf (F_0M) in degree zero. We now identify a Hodge module whose lowest piece contains the coherent lifting obstructions. On each chart (U), write (J_U=Jtimes_{Y} For (j:S_U∖ J_U↪ S_U), put [ A^bullet_U=D!( j_!Q^H_{S_U∖ J_U}[N]), A_{0,U}={}^{p}!H^0A^bullet_U, M_U={}^{p}!H^1g_*A_{0,U}. ] Here (D) is Hodge-module duality, ({}^pH^i) denotes perverse cohomology, and (g_*) in the definition of (M_U) is the derived Hodge-module direct image. Later, (g_+) will denote the corresponding direct image of differential modules. These constructions commute with etale restriction. Thus the (M_U) form a system (M) supported on (T The following identifications hold compatibly on the charts: [ F_{<0}A_{0,U}=0, F_0A_{0,U}=B|_calB|_{S_U}, F_{<0}M_U=0, F_0M_U=R^1g_*(B|_calB|_{S_U}). ] In a smooth ambient embedding (S_U↪ V) of codimension (c), the underlying module of (A_{0,U}) is (H^c_{S_U}(omega_V)), localized off (J_U). The lowest-piece inclusion is the adjunction, or Ext-to-support, inclusion off (J_U), extended by meromorphic localization. Proof. The proof must identify not just a coherent sheaf but its actual inclusion into the differential module: the graph calculation will differentiate classes in that inclusion. We first compute the sheaf by duality. Du Bois comparison applied to the difference triangle for the constant objects of (S_U) and (J_U) identifies its degree-zero graded de Rham complex with (I_{J_U/S_U}). Coherent duality and the maximal-Cohen--Macaulay property give Gr^F_kDR(A^bullet_U)=0quad(k<0), Gr^F_0DR(A^bullet_U)=B|_calB|_{S_U}[0]. Here ([N]) cancels the dimension shift of the dualizing complex. In the smooth case, our convention gives (D(D(Q^H[N])=Q^H[N](N)); its right differential module has (omega) as its lowest piece, at index zero. To pass to perverse cohomology, work locally on a fixed chart and let (q) be the smallest filtration index occurring in any perverse cohomology object of (A^bullet_U). Boundedness of the complex and goodness of the filtrations provide such a lower bound. At index (q), each of these objects has a graded Spencer complex consisting only of its degree-zero term. The spectral sequence for perverse truncation has a single nonzero row at that index, and therefore gives [ H^i =F_q,{}^p!H^iA^bullet_U. ] If (q<0), this contradicts Equation~(hg:derived-lowest). It follows that all these objects have (F_{<0}=0). The same argument at index zero yields (F_0A_{0,U}=B|_calB|_{S_U}), while the lowest pieces of the other perverse cohomology objects vanish. This passage does not require the shifted constant complex to be perverse. We next identify the underlying unfiltered module. Ambient duality in (V) identifies the dual of the reduced constant complex with [ RGamma_{[S_U]}(omega_V)[c]. ] The degree-zero module is the first nonzero support cohomology, (H^c_{S_U}(omega_V)). The support of (J_U) is locally set-theoretically principal in (S_U) by the geometric construction. Lift a local defining function to (V) and invert it. This exact meromorphic localization realizes (j_*) on the underlying module and gives the description of (A_{0,U}) in the statement. The two descriptions must agree as inclusions, with the fixed residue normalization. On the smooth locus of (S_U∖ J_U), filtered duality and smooth graph pushforward give the ordinary dualizing sheaf and its residue inclusion. This is precisely Ext-to-support with the normalization specified by ambient adjunction. Naturality in smooth etale coordinates gives the same scalar on each chart. The agreement extends throughout (S_U∖ J_U), because the first support-cohomology module has no sections supported on a smaller-dimensional subset. One can see this from the Cousin resolution of the smooth dualizing sheaf: its first term supported on (S_U) is a sum of injective modules at the codimension-(c) generic points. Localization extends the agreement across (J_U). We have therefore identified the actual lowest-piece inclusion, including its residue convention. We have now identified (F_0A_{0,U}) and its inclusion before direct image. To identify (F_0M_U), compute (g_+) by a graph embedding followed by a smooth projection. At filtration zero, only the top term of the relative Spencer complex survives: it is the direct image of (B|_calB|_{S_U}) on the graph. Projective strictness yields [ F_{<0}M_U=0, F_0M_U=R^1g_*(B|_calB|_{S_U}), ] and identifies this sheaf as a coherent subsheaf of the unfiltered degree-one direct image. All constructions in this identification are canonical under etale changes of charts. The obstruction sheaf for the transition indexed by (j,k) is thus (F_0M ⊗ C^{-(a+j+k)}), by Equation~(hg:geometric-lines) and projection formula. We do not need all its global sections to vanish. The lifting proof will show that the coordinate obstruction lies in the kernel of a symbol map, and will apply the following vanishing to that kernel. ## Negative-twist vanishing The support of the system (M) just constructed is finite over (P We prove the required vanishing for any system with this support property. Let (M) be a compatible system of mixed Hodge modules in perverse degree zero on the scheme etale charts of (Y Suppose its support is finite over (P Then, for every integer (k), every positive integer (b), and every (j<0), [ H^j!( Y,Gr^F_kDR_{Y} ⊗ p^*C^{-b} )=0. ] Here the graded de Rham complex is the descended coherent complex. Proof. The line (C) lives on the root gerbe, so we first push the complex to (P We will then pull back to projective space, where (C) becomes (OO(1)) and ordinary Kodaira--Saito vanishing applies. Since (p) is finite on the support of (M), its direct image is concentrated in perverse degree zero; call the resulting system on (P (M'). On ordinary scheme charts, projective strictness gives Rp_*Gr^F_kDR_{Y} ~= Gr^F_kDR_{P}'). For the desired global hypercohomology vanishing, this identity must hold for the descended complexes. We verify that compatibility before passing to a finite cover. For right differential modules, use the transfer bimodule [ D_{Y P}P = OO_{Y} ⊗ _{p^{-1}OO_{P}} p^{-1}D_{P}, ] with its order filtration. Derived tensor over (D_{Y}, followed by derived direct image, forms the filtered direct image; equivalently, these operations can be carried out on Rees modules. De Rham on the base is a further derived tensor with (OO_{P}, filtered from degree zero. The filtered Spencer resolution, associativity of derived tensor, and projection formula identify this with direct image of the de Rham complex upstairs. Indeed, tensoring the transfer module over the base differential operators with the base structure sheaf gives (OO_{Y} These sheaf constructions are canonical on the etale site, and the Spencer resolutions are locally free over differential operators. They commute with chart changes already before taking associated gradeds. The ordinary projective direct-image theorem on the base charts supplies strictness and concentration in perverse degree zero. Passing to the etale site does not change coherent cohomology on those charts. This proves Equation~(hg:graded-direct-image) globally, as an identity computing coherent hypercohomology. The same verification will apply to the finite representable map below. To pass from the gerbe to a scheme, use the representable finite surjective morphism [ v:PP^s_z→P ] defined by the power map ([z_0:...:z_s]|->[z_0^ell:...:z_s^ell]) and the root (OO_{PP^s_z}(1)); thus (v^*C=OO_{PP^s_z}(1)). Trivial roots on the standard affine opens of (PP^s) give scheme etale charts (check U_itoP The restriction of (v) to (z_ine0) factors through (check U_i). On each such chart take the derived Hodge-module inverse image of (M') and then its degree-zero perverse cohomology. Thus the object being constructed chartwise is [ H={}^pH^0(v^*M'). ] The chosen root and functoriality identify these objects on overlaps, producing a system on the Zariski opens of (PP^s_z). We check that (H) is a global graded-polarizable mixed Hodge module, so that ordinary vanishing applies to it. The perverse, filtered, and weight data glue, as do the strict-support decompositions of the pure weight gradeds, by uniqueness. On a smooth connected dense stratum of an irreducible support, a polarization on a nonempty Zariski open extends to a flat pairing: the fundamental group of that open surjects onto the fundamental group of the stratum. Hodge compatibility extends by continuity, and the extended flat pairing remains nondegenerate and positive. The local Hodge modules already give boundary quasi-unipotence. Saito's extension theorem for polarizable variations cite[Section~3.b]{Saito1990} and uniqueness of strict-support extension identify the global pure object with the glued object. Projectivity of (PP^s_z) leaves no additional boundary at infinity. The weight extensions give the required mixed object. Vanishing for (H) will imply vanishing for (M') once we realize (M') as a retract of (v_+H). On a chart one must use the whole base-changed finite cover. Over (check U_i), this is a disjoint union of copies of the corresponding affine coordinate chart of (PP^s_z). The unit on unshifted constants and its dual trace, formed using smooth duality in equal dimensions, compose to multiplication by the degree. This can first be checked on the finite etale locus, where the trace sums over the sheets. On each connected smooth base chart, the shifted constant Hodge module is the rank-one intersection complex, with endomorphism ring (Q), and its endomorphisms are determined on a dense open. The composition therefore equals the degree globally, with no contribution supported on the branch locus. This argument permits ramification. Tensor the unit and trace with (M'), apply proper projection formula, and take perverse cohomology in degree zero. Finite direct image is perverse exact, so it commutes with this degree-zero cohomology. We obtain maps [ M'→ v_+H→ M' ] whose composite is multiplication by the degree of the whole base-changed cover. Dividing the second map by that degree gives the retraction. Each map is canonical under etale base change; the retraction consequently descends on filtered differential modules as well. Applying the finite version of Equation~(hg:graded-direct-image) and projection formula now exhibits the hypercohomology in the statement as a direct summand of [ H^j!( PP^s_z,Gr^F_kDR(H) ⊗ OO_{PP^s_z}(-b) ). ] Ordinary negative-ample Kodaira--Saito vanishing makes this group zero for (j<0). For a mixed object, apply the vanishing to the pure weight gradeds and then to the weight-filtration exact sequences, whose Hodge filtrations are strict. ## Lifting through the infinitesimal neighborhoods We now combine the lowest-piece identification with negative-twist vanishing. The remaining step is to identify the relation satisfied by the coordinate and conormal obstructions. Proof. [Proof of Proposition~prop:formal-lifting] We induct on (k), simultaneously for all powers (I^j). The connecting homomorphisms will factor through the first graded layers and together form a derivation. We must show that this derivation vanishes both on functions from (U) and on a local generator of (I/I^2). The graph calculation relates these two errors in the differential module (M_U). Its order-one symbol, together with negative-twist vanishing, first kills the function error. The original relation then kills the conormal error. Define [ E_{j,k}=g_*(I^j/I^{j+k}), jge0, kge1. ] The first graded layer is [ E_{j,1}=OO_{T_U} ⊗ C^{-j}. ] For adjacent lengths, the exact sequence has kernel (g_*OO_{S_U} ⊗ C^{-j-k}), and its connecting homomorphism takes values in [ R^1g_*OO_{S_U} ⊗ C^{-j-k}. ] smallskipnoindentThe obstruction derivation. Assume inductively that all shorter transitions are surjective. A section of (E_{j,k}) with zero length-one term comes from (E_{j+1,k-1}) when (k>1). Induction lifts it locally from (E_{j+1,k}), so its connecting class vanishes. Induction also makes (E_{j,k}to E_{j,1}) surjective. The obstruction thus factors through the length-one layer; for (k=1) this factorization is immediate. More precisely, these factorizations give additive maps [ D_{k,j}:OO_{T_U} ⊗ C^{-j} → R^1g_*OO_{S_U} ⊗ C^{-j-k}. ] Together they define (D_k) on the graded algebra (⊕_{jge0}OO_{T_U} ⊗ C^{-j}), with graded shift (k); no (OO_U)-linearity is asserted. Its values are computed by choosing local lifts and taking Cech differences. For a product, this difference is the sum of the two first-order differences: the product of two errors vanishes in the layer under consideration. Thus (D_k) is a (C)-derivation, with values in the graded module whose degree-(j) term is (R^1g_*OO_{S_U} ⊗ C^{-j}). Restrict its degree-zero part along (OO_UtoOO_{T_U} This derivation factors through (Ω_U^1), defining a section of (T_U ⊗ R^1g_*OO_{S_U} ⊗ C^{-k}). As the chart varies, these sections descend to ein H^0!( Y,T_calY,T_{Y} ⊗ R^1g_*OO_S ⊗ C^{-k} ) = H^0!( Y,T_calY,T_{Y} ⊗ F_0M ⊗ C^{-(a+k)} ). The equality follows from Equation~(hg:geometric-lines), Lemma~hg:lowest-piece, and projection formula. For the descent assertion, function lifts pull back along the unique etale thickenings. Ordinary flat base change on the reductions pulls back their Cech classes in the coherent obstruction sheaves. Differentials also pull back and generate under an etale map, giving the required compatibility. smallskipnoindentCoordinate errors and adjunction. We now compute this derivation in local coordinates. After shrinking (U), choose etale coordinates (t_1,...,t_d), with (d=dim U), and a frame of (C^{-1}). Write (y) for its pullback, a conormal frame. Induction lifts the coordinates along (g) modulo (I^k) and lifts (y) to (I/I^{k+1}). Lift them one step further on an open cover of (S_U), obtaining functions (h_{i,lambda}) modulo (I^{k+1}) and generators (y_i) modulo (I^{k+2}). The conormal generator must be lifted one order farther than the coordinates: its next error lies in (I^{k+1}/I^{k+2}), whereas a coordinate error lies in (I^k/I^{k+1}). By formal etaleness of the coordinate map, each (h_i) defines a local map from ((k+1)S_U) to (U). Write the overlap errors as h_{j,lambda}-h_{i,lambda} =e_{ij,lambda}y_i^k, y_j-y_i=eta_{ij}y_i^{k+1}. Thus (D_k(t_lambda)) is represented by ((e_{ij,lambda}y^k)), and (D_k(y)) by ((eta_{ij}y^{k+1})). The powers of (y) record the target line bundles; (e_{ij,lambda}) and (eta_{ij}) are their coefficients on the reduction. Denote the induced frame of (B|_calB|_{S_U}) by (sigma=y^{-a}). Off (J_U), the fixed adjunction isomorphism gives [ omega_{(k+1)S_U} ~=OO_{X}+k)S)|_{(k+1)S_U}. ] Via this isomorphism, the local equation (y_i) defines a dualizing generator, denoted (b_i=y_i^{-(a+k)}). This notation denotes a frame of the line bundle (OO_{X}+k)S)) restricted to the thickening; it does not invert the nilpotent function (y_i) in its structure sheaf. Realize the dualizing sheaves in support cohomology. Under the trace inclusion (omega_{S_U}↪omega_{(k+1)S_U}), we have b_j-b_i=-(a+k)eta_{ij}sigma, y_i^kb_i=sigma, y_i^{k+1}b_i=0. The first identity is the binomial expansion of ((1+eta_{ij}y_i^k)^{-(a+k)}): all quadratic terms vanish because (2kge k+1). In the last two identities, (sigma) is viewed in the dualizing sheaf of the thickening through the trace inclusion. They express Cartier trace and the annihilator of the thickening. smallskipnoindentThe graph calculation. We next put both errors in the same differential module. Embed ((k+1)S_U) as a closed subscheme of a smooth open (Zsubsetoverlineoverline Z=Z=PP^m). The graph of the reduced map (g) is closed in (overlineoverline ZZtimes U), since it is proper over (U). Let (iota:S_U↪overlineoverline ZZtimes U) be this reduced graph embedding, and let (N_ be the underlying unfiltered differential module of (iota_+A_{0,U}). This module is distinct from the direct-image module (M_U) on (U). For each local graph lift (h_i), Ext-to-support cohomology sends the dualizing section (b_i) to a section (delta_i(b_i)) of (N_ All the lifted graphs have the same reduced support, namely the graph of (g), so these sections lie in one support-cohomology module rather than in different modules for the different lifts. These sections need not lie in its lowest filtered piece. We will prove the identity delta_j(b_j)-delta_i(b_i) = -(a+k)eta_{ij}sigma +sum_{lambda=1}^d (e_{ij,lambda}sigma)·∂_{t_lambda}. On the right, each reduced dualizing section (sigma) is included in (N_ by the lowest-piece map of Lemma~hg:lowest-piece. The right differential-operator action is taken in this graph module. The calculation also applies when (S_U) is singular. Put (c_0=dim Z-N). Ext-to-support identifies the dualizing sheaf of the thickening with its annihilator submodule in (H^{c_0}_{S_U}(omega_Z)). To see this, use the support-cohomology spectral sequence. Support cohomology vanishes below (c_0), so in total degree (c_0) the Ext group consists of homomorphisms from the structure sheaf of the thickening to the first support-cohomology module. In particular, no local complete-intersection hypothesis on (S_Usubset Z) is needed. Lift the (h_{i,lambda}) locally to functions on (Z). For a dualizing section (b) of the thickening, its graph inclusion in (Ztimes U) is the generalized fraction [ frac{b,dt_1wedge...wedge dt_d} {prod_{lambda=1}^d(t_lambda-h_{i,lambda})}. ] It represents successive support cohomology in the additional coordinate equations, equivalently the Koszul residue for the graph, and can be computed etale-locally near the graph branch. After taking support cohomology from (Z), the translated new coordinates form a regular sequence; their support cohomology is consequently concentrated in the degree equal to their number. Either set of translated coordinates gives the same localization of this support-torsion module, because the two translations differ nilpotently on every section. The change from (h_i) to (h_j) thus has a finite Taylor expansion on (b_j). Again using (2kge k+1), products of two differences from Equation~(hg:coordinate-errors) annihilate (b_j). The linear terms are determined by [ (h_{j,lambda}-h_{i,lambda})b_j =e_{ij,lambda}sigma. ] The right action on a top differential form satisfies [ ((dt_lambda)/(t_lambda-h)) ·∂_{t_lambda} =(dt_lambda)/((t_lambda-h)^2). ] The doubled pole therefore contributes the positive sign in Equation~(hg:graph-identity). Combining these terms with Equation~(hg:dualizing-errors) proves the identity. So far the calculation is off (J_U). Both its sections and its identities extend meromorphically across (J_U) in the unfiltered localized module, where arbitrary finite pole orders are allowed. smallskipnoindentVanishing of the obstruction. Push the graph identity along the projection (pi:overlineoverline ZZtimes Uto U), using the relative Spencer complex. Closed direct image is perverse exact, and (pi_+iota_+A_{0,U}=g_+A_{0,U}). Thus degree-one holonomic differential-module cohomology of this Spencer direct image is the underlying module of (M_U={}^{p}!H^1g_*A_{0,U}). The cochain ((delta_i(b_i))) belongs to the top, degree-zero term, from which there is no outgoing relative Spencer differential. Its Cech difference is a boundary in total degree one. A sufficiently refined ambient cover around the closed graph computes this boundary; all the sheaves in question are supported there. By Lemma~hg:lowest-piece, the reduced cocycles in Equation~(hg:graph-identity) represent their (R^1g_*B classes in (F_0M). Right differentiation in the (U)-coordinates acts on the pushforward complex. Consequently the Cech classes satisfy [ -(a+k)[eta_{ij}sigma] +sum_{lambda=1}^d [e_{ij,lambda}sigma]·∂_{t_lambda}=0 in M_U. ] Here each bracketed class belongs to (R^1g_*B=F_0M_U; after differentiation, its image belongs to (F_1M_U). Thus every term of the relation lies in (F_1M), and the (eta)-term lies in (F_0M). The equality itself was proved in the unfiltered differential module. The inclusion (F_1Msubset M) is an inclusion of actual sheaves, so vanishing in (M) implies vanishing in (F_1M). Taking the order-one symbol shows that the global section (e) from Equation~(hg:obstruction-section) is killed by T_{Y} ⊗ F_0M → Gr^F_1M, after twisting by (C^{-(a+k)}). Only the relation is used in this filtered step; no bound on the Hodge filtration of the auxiliary cochain ((delta_i(b_i))) has been imposed. Because (F_{<0}M=0), the entire complex at filtration index one is [ Gr^F_1DR(M)= [T_{Y} ⊗ F_0M →Gr^F_1M], ] in degrees (-1) and (0), with differential the symbol map in Equation~(hg:symbol-map). After the indicated twist, its degree-(-1) hypercohomology is therefore exactly the space of global sections of the kernel of that map. The twist is negative, since (a+k>0). Lemma~hg:negative-vanishing annihilates this space, and hence (e=0). It remains to kill the conormal part of the derivation. With coordinates and a line trivialization fixed locally, (e=0) says that every class ([e_{ij,lambda}sigma]) vanishes in (F_0M), and thus in (M). Its actual differential-operator image vanishes as well, before passing to symbols. Returning to the unfiltered relation gives [ (a+k)[eta_{ij}sigma]=0. ] Injectivity of the lowest-piece inclusion and (a+k>0) then give (D_k(y)=0). The degree-zero derivation also vanishes: its restriction to (OO_U) is the coordinate error (e), and (OO_UtoOO_{T_U} is surjective because (T_U is a closed subscheme of (U). The coefficient algebra and the conormal frame (y) generate the full graded length-one algebra locally. Thus (D_k=0) in every degree, completing the induction and proving surjectivity of all the sheaf transitions. The kernels are the stated coherent sheaves (OO_{T_U} ⊗ C^{-j-k}). If (U) is affine, their first cohomology vanishes, so the exact sequences of sheaves of abelian groups give surjectivity on global sections as well. Successive choices of lifts then give compatible formal lifts of any chosen functions on (T_U and, after trivializing (C), of the transverse conormal frame. # Compact null families and descent We now complete Theorem~thm:signed. Retain 00. It follows that G_Y=pi^*G is vertical. It is also relatively nef because G_Ysim_{Q}pi^*L. Lemma~sg:vertical-descent now gives [ pi^*Lsim_{Q}f^*G_B ] for a Q-divisor G_B on the smooth base. Every curve in B is dominated by a curve in Y, where the pullback has nonnegative degree; hence G_B is nef. Numerical dimension is unchanged by these pullbacks; the intersection formula gives [ v=nu(X,L)=nu(Y,pi^*L)=nu(B,G_B)le b. ] Together with ble v, this yields b=v=nu(B,G_B). Since G_B is nef on the b-dimensional base, it follows that G_B^b>0, so G_B is big. Pullback of sections then gives kappa(X,L)ge b=v. The ineuality kappa(X,L)lenu(X,L) for nef divisors provides the reverse bound and proves abundance. # Geometric exclusions for a nonvanishing counterexample We now prepare the smooth nonvanishing step. Assume only the lower-dimensional good-model hypothesis of Assumption~mmp:lower-models, and suppose that a smooth projective (n)-fold (X), with (n>0), satisfies K_X is pseudo-effective, kappa(X,K_X)=-∞. We first restrict its moving proper subvarieties and positive canonical currents, and then exclude covering curves of bounded genus and normalized degree. Two further conclusions will handle the remaining jet configurations: the reduced-boundary adjoint of a signed canonical representative is big, and fixed finite covers admit no family of birational maps whose graphs dominate their product. The argument takes place over (C), and the two displayed properties persist on smooth projective birational models. Numerical dimension for a pseudo-effective divisor means Nakayama's numerical dimension (kappa_sigma); for a nef divisor it agrees with the intersection definition used in Theorem~thm:signed. ## The Albanese reduction and algebraic webs Subadditivity first shows that the Albanese map is constant. This will let us pass from an effective numerical representative of a rational divisor to an effective rational-linear representative. We then construct the rational quotient generated by a moving family of subvarieties. Every smooth projective birational model of (X) has irregularity zero. Consequently, on such a model, or on a projective (Q)-factorial terminal birational model, numerical equivalence of rational divisors implies their (Q)-linear equivalence. Proof. If (q(X)>0), resolve the Stein factorization of the Albanese map to obtain a morphism (f:Wto B) with connected fibers, with (W,B) smooth projective and (dim B>0). The map from (B) to a subvariety of (Alb(X)) is generically finite. Wedges of invariant one-forms on the abelian variety therefore give a nonzero section of (K_B): at the generic point, choose (dim B) independent pulled-back one-forms. Thus (kappa(B,K_B)ge0). For a very general smooth fiber (F), the restriction of the pseudo-effective class (K_W) is pseudo-effective. For example, restrict a positive current representing it to almost every fiber, or use effective approximations with arbitrarily small ample error. Adjunction identifies this restriction with (K_F). If (F) has positive dimension, the induction hypothesis gives (kappa(F,K_F)ge0); for a point the same assertion holds by convention. Applying Theorem~asm:iitaka with both boundaries empty gives (kappa(W,K_W)ge0), contradicting (ex:counterexample). This is the only use of Theorem~asm:iitaka in the proof. Irregularity is a smooth birational invariant. Finally, when (Pic^0=0), a numerically trivial line bundle is torsion, since the group of numerically trivial line bundles modulo (Pic^0) is finite. Clear denominators for rational divisors. On a terminal (Q)-factorial model, pull back to a smooth resolution and then descend the rational linear equivalence. To handle subvarieties through very general points, parameterize them in countably many Hilbert families, then resolve and stratify. On a suitable open parameter space the domains form a smooth projective family with integral fibers, and evaluation is dominant. When each fiber maps generically finitely onto its image, general ample cuts of the parameter space make total evaluation generically finite while preserving dominance. We resolve and compactify this total evaluation before using its ramification divisor. Invariance of plurigenera for smooth projective families cite[Theorem~1]{Paun2007} permits the same construction for the Iitaka fibers of general members: choose a divisible pluricanonical system on the general member, spread it by base change, and resolve its relative rational map. When the member has nonnegative Kodaira dimension but is not of general type, the smooth general fibers of this map have positive dimension. They form the new family used below. [Algebraic web quotient] Let a smooth projective variety be dominated generically finitely by the total space of a family with smooth integral projective general fibers. On a dense regular open, form the distribution generated by the tangent spaces of the fiber images, taking spans, saturation, and Lie brackets. Its general leaves are dense opens of algebraic subvarieties. Their closures are the general fibers of a rational map, which has connected general fiber after Stein factorization. Proof. Shrink to an open (U) where the evaluation has finitely many etale sheets, the parameter map is smooth, and the generated involutive distribution (F has constant rank. Its construction is algebraic: spans and brackets stabilize at the generic point after finitely many operations, and descend from the etale sheets. The nonempty open parts of the integral parameter fibers are connected. A chain step moves between two points in the image of one such open fiber. We use only steps lying in (U), so each step remains in one analytic leaf of (F For (xin U), endpoints of chains of at most (k) steps form a constructible set, by the algebraic fiber-product construction and Chevalley's theorem. Every endpoint lies in the analytic leaf through (x). A locally closed smooth variety contained in an analytic leaf has dimension at most (rkF Indeed, in a Frobenius chart the leaf meets at most countably many plaques: use finite plaque chains in a countable foliation atlas. The transverse coordinate map on any connected smooth piece then has countable image and is constant. There is therefore a maximal dimension among the closures of all such endpoint loci; it is attained for some finite (k). The loci are nested because we allow at most (k) steps, including the identity step. Choose an irreducible component (V) of maximal dimension and a dense open (Osubset V) of reachable points. On each etale sheet, the parameter map defines the algebraic relation consisting of pairs in the same fiber. Restrict its first coordinate to the inverse image of (O), take the component through the diagonal, and map the second coordinate to (U). The closure of this image contains (O), hence (V), while its points are reachable in at most (k+1) steps. Maximality forces that closure to equal (V). At a general diagonal point, smoothness of the parameter map identifies its vertical tangent with the corresponding web tangent. Every web tangent is consequently tangent to (V), and so is every bracket. Hence (dim VgerkF, while the reverse ineuality was proved above. A plaque is thus open in (V). The equations of (V), pulled back to the connected immersed leaf, vanish on a nonempty open and hence on the entire leaf. Its closure is exactly (V). These invariant subvarieties of dimension (rkF have countably many Hilbert or Chow parameter spaces. Tangency on the regular locus is an algebraic condition after stratification, so one such family dominates. An invariant irreducible variety meeting (U) contains the leaf through any of its points in (U): the tangent vector fields preserve its reduced ideal, first on its smooth locus and then everywhere by closure. A leaf closure of dimension (rkF is therefore unique through a general point. Assigning that closure gives the desired rational map. Resolve its graph and take the Stein factorization. For positive closed ((1,1))-currents we shall use this quotient in the following way. If a current vanishes on almost every parameter fiber away from a fixed removed divisor, it annihilates the corresponding fiber tangents on a dense open. In submersion coordinates, choose locally integrable plurisubharmonic potentials. Fubini's theorem makes the pure fiber Hessian zero as a distribution; positivity then makes the mixed coefficients in fiber directions zero as well. A generically etale evaluation transfers this assertion to the target. Annihilation passes to brackets by the identities [ L_v T=d(iota_vT)+iota_v(dT), iota_{[v,w]}T=L_v(-iota_w(L_vT). ] Pullbacks by dominant morphisms and restrictions to almost every smooth parameter fiber are defined by the same local potentials. ## The general-type exclusion The next proposition turns the lower-dimensional good-model hypothesis into a restriction on every moving proper subvariety. Its proof passes from Iitaka fibers to the algebraic quotient of their tangent directions and then back to nonvanishing. Every positive-dimensional proper subvariety through a very general point of (X) is of general type on resolution. The assertion holds also on every projective birational model. Proof. Otherwise take a covering family of nongeneral-type subvarieties, with smooth domains (V), and slice its parameters as above. Restricting the ramification formula of its generically finite total evaluation shows that (K_V) dominates the restriction of the pulled-back (K_X) by an effective divisor. Thus (K_V) is pseudo-effective. The induction hypothesis gives a good minimal model of (V), so (kappa(V)ge0). Since (V) is not of general type, its general Iitaka fibers have positive dimension and Kodaira dimension zero. Spreading these fibers gives another covering family with smooth domains (G), of dimension less than (n), and (kappa(G)=0). Each (G) has a good minimal model and (kappa_sigma(K_G)=0). To exploit this zero-Iitaka-dimension family, fix a positive current (T) representing (K_X). After slicing and resolving the total evaluation of the (G)'s, its restriction to almost every parameter fiber, plus the effective restricted ramification divisor, is a positive current in (K_G). Every positive current in that class is the current of the fixed canonical divisor of (G). To see this, let (R) be such a current and take a common resolution (u:Hto G), (v:Hto G_{min}) with the good minimal model. The current [ u^*R+[K_H-u^*K_G] ] is positive and represents (K_H). Since (K_{G_{min}}) is torsion, this class is represented by an effective (v)-exceptional divisor. Its intersection with the ((dim G-1))-st power of an ample pullback from (G_{min}) is zero. That form is strictly positive off the exceptional locus, so the current is supported there. The support theorem makes it divisorial, and independence of exceptional divisor classes, by negativity, determines its coefficients. Pushing forward by (u) determines (R) itself. After shrinking the parameter space, these fixed canonical divisors are the fixed divisor of a sufficiently divisible relative pluricanonical system. Remove this divisor and the excluded loci of the construction from the total family. Their images under generically finite evaluation are proper algebraic subsets of (X). It follows that (T) annihilates the web distribution of the (G)'s on a dense open. If that distribution has full rank, (T) is supported on a proper algebraic set. The support theorem for positive closed ((1,1))-currents expresses it as a finite nonnegative real combination of prime divisors [cite]. Its rational cohomology class then has a nonnegative rational representative: solve the finite rational linear system for the coefficients on the same face of the nonnegative orthant. Lemma~ex:irregularity converts numerical effectivity to (Q)-linear effectivity, a contradiction. Otherwise Lemma~ex:web gives a rational quotient with general smooth fiber (F) on a smooth resolved graph, where (00). Consequently, if (alphane0) is a supporting functional of the pseudo-effective cone with (alpha(K_Y)=0), then (alpha(P)>0). Proof. Resolve a moving subsystem of a multiple of (P). If its map is generically finite, (P) is big and the assertion follows from pseudoeffectivity of (K_Y). Otherwise its general fiber is a proper positive-dimensional subvariety through very general points and is of general type by Proposition~prop:general-type. The canonical divisor of the smooth resolution is relatively big, so adding a sufficiently large ample pullback from the image makes it big. Since that pullback is bounded by a multiple of the resolved moving subsystem, pushforward gives (K_Y+aP) big for some (a>0). Convexity with the pseudo-effective (K_Y), and then addition of the pseudo-effective (P), gives the assertion for every (c>0). A nonzero nonnegative functional on a closed convex cone is strictly positive on its interior. Applying it to (K_Y+cP) proves the last assertion. ## Valuations and positive currents The next two exclusions involve currents whose numerical intersection gaps tend to zero. A limiting ineuality alone is insufficient: we need one family on which the gap is exactly zero. The ACC statement below supplies that step for a fixed current and bounded discrepancies. For a positive closed ((1,1))-current (T) on a smooth projective variety (V), let (nu_E(T)) denote the generic Lelong number of its pullback at a divisorial valuation (E). The normalization is such that a reduced smooth divisor has generic number one. We write (A_V(E)=a(E;V,0)) for log discrepancy. For a fixed (T) and a fixed (M), the set [ {nu_E(T): A_V(E)<= M} ] satisfies the ascending chain condition. Proof. We induct on the integer part of (M); discrepancies over a smooth variety are positive integers, so the assertion is empty when (M<1). Suppose that a strictly increasing sequence exists, and discard initial terms so its members exceed a fixed positive number (c). Skoda integrability and change of variables give nu_E(T)<= A_V(E)nu_x(T) at a general point (x) of the center. Indeed every exponent smaller than (1/nu_x(T)) is locally integrable, whereas integrability after pullback imposes the corresponding discrepancy bound. Hence all centers lie in the fixed proper Siu locus ({nu_x(T)>= c/M}), which is algebraic by Siu analyticity and projectivity [cite]. Passing to a subsequence, all centers lie in one irreducible component of this locus. If, after passage to a subsequence, the centers lie in a fixed codimension-at-least-two subvariety, principalize its ideal. All lifted centers lie in the exceptional locus, where the relative canonical divisor has positive integral order. The discrepancy bound for the new smooth ambient space has therefore decreased by at least one. Pull back (T) and apply induction. Otherwise the centers lie in a fixed prime divisor (D) of the Siu locus. Write (T=c_D[D]+T') with (T') positive and with zero generic Lelong number along (D). The integers (ord_E(D)) are bounded: the fixed effective divisor (D) has positive log canonical threshold, and (lct(D)ord_E(D)<= A_V(E)). Pass to a constant order. The residual numbers (nu_E(T')) are still strictly increasing, and after discarding a term have a positive lower bound. Equation~(ex:skoda-valuation) places their centers in a fixed Siu locus of (T'), which does not contain (D). Their intersection with (D) has codimension at least two, so the preceding case applies. To compare those Lelong numbers with movable algebraic systems, use the following multiplier-ideal approximation cite[Theorems~5.11, 6.27, and 14.2]{Demailly2012}. On a smooth projective variety, if ({T}) is a real algebraic class, one can choose integral line bundles (L_k) such that (c_1(L_k)-k{T}) stays in a bounded, uniformly sufficiently ample set and (L_k ⊗ J(kT) is globally generated. Round the coefficients of (k{T}) in a fixed integral basis and add a fixed sufficiently positive integral class. Nadel vanishing and Castelnuovo--Mumford regularity give generation. For every divisorial valuation, ord_EJ(kT)<= knu_E(T). The local Bergman weight is bounded below by the original weight up to a constant; this ineuality persists after pullback and gives (ex:multiplier-order). These are statements for a fixed current, not uniform assertions over all currents. Let (Y) be a fixed projective (Q)-factorial terminal birational model of (X), with (K_Y) nonbig. A positive current in the pullback class of (K_Y) on a smooth resolution cannot, on a dense regular Zariski open, annihilate the tangent spaces of the general fibers of a rational fibration of positive relative dimension. Proof. Fix a current (T) as in the statement and suppose such a fibration exists. Consider all dominant rational maps from (Y) with connected general fiber of positive dimension for which the pulled-back current annihilates the fiber tangents on a dense regular open of a resolved graph. Choose one with smallest base dimension (b). We will descend the current to this base and find a covering family of curves on which the descended current vanishes. Their quotient will give a member of the same collection with smaller base dimension. If (b=0), the current vanishes on a dense open; the divisorial-current argument in Proposition~prop:general-type contradicts (ex:counterexample). Thus (00), constant on (B^0), by closedness. The projection formula therefore gives [ c^{-1}h_*(T'wedgeomega^d)|_{B^0}=S^0. ] The left side defines a global positive closed extension. Its class is real algebraic, since ({T'}) is the algebraic class (p^*K_Y) minus real divisor classes and algebraic intersection and pushforward preserve such classes. Remove its divisorial parts along (B∖ B^0), and call the result (S). On a dense open (T=h^*S). Their difference is again closed of order zero, so the same support theorem leaves only a signed divisor sum. Over (B^0), this difference consists only of the horizontal components removed above. Its coefficient at a prime dominating a base prime that meets (B^0) is therefore zero. Along a base prime contained in (B∖ B^0), the current (S) has zero generic Lelong number by construction. At a prime upstairs dominating it, the generic local map is a transverse power map times a submersion. Local target balls of a radius equal to a fixed power of the source radius give the Lelong comparison, so (h^*S) also has zero generic number there. Positivity of (T) then makes the coefficient of (T-h^*S) nonnegative. Thus all coefficients at primes dominating base primes are nonnegative. The only possible negative coefficients lie over base codimension at least two and are exceptional over (Y). We have therefore obtained a positive closed current (S) in a real algebraic class on (B), with (T-h^*S) a divisor current whose only possible negative coefficients are (p)-exceptional. Push this difference forward by (p). The exceptional terms disappear and the remaining terms are effective, so K_Y-T_B{S}quadis pseudo-effective. The horizontal components removed during descent contribute effective divisors to the same difference. noindentA negative canonical direction on the base. Choose a nonzero supporting functional (alpha) of the pseudo-effective cone at (K_Y), and put (eta=T_B^*alpha). The corresponding functionals (eta') on higher smooth base models are movable classes by pseudo-effective/movable duality cite[Theorem~0.2]{BDPP2013}. They annihilate base-exceptional divisors. Since (S) is positive, (eta·{S}ge0). Applying (alpha) to (ex:descent-domination) gives the reverse ineuality, because (alpha(K_Y)=0). Hence (eta·{S}=0). If (c_D[D]) is a positive Siu component of (S), both (S-c_D[D]) and ([D]) are positive. It follows that (eta· D=0), or equivalently (alpha(T_BD)=0). Before applying multiplier approximation, we need to control the divisorial part of this descended current. Only finitely many prime divisors on (B) have positive divisorial Lelong coefficient for (S). Otherwise their nonzero strict pullbacks are effective divisors on (Y) killed by (alpha), with no common prime components. Indeed, a nonexceptional prime of (Y) cannot map into the intersection of two base primes. Only finitely many base primes have zero strict pullback, since their total pullbacks are supported in the finite exceptional locus of (p). In the finite-dimensional rational space of divisor classes, infinitely many remaining strict pullbacks have a rational relation. This relation must have both positive and negative coefficients: a nonzero effective sum has positive intersection with an ample ((n-1))-fold product on (Y), so cannot be numerically zero. The two sides are thus nonzero effective rational divisors with no common prime component. Lemma~ex:irregularity makes them (Q)-linearly equivalent. Clearing denominators gives two distinct linearly equivalent effective integral divisors and hence a pencil, whose class is killed by (alpha). This contradicts Corollary~ex:supporting. We claim that K_B·eta<0. The smooth general fiber (F) of (h) is of general type by Proposition~prop:general-type. The relative-positivity theorem of Kov'acs--Patakfalvi cite[Theorem~9.9]{KovacsPatakfalvi2017} applies to a log canonical fiber space with smooth base, SNC total pair, and log-general-type geometric generic fiber; for a rational base divisor (M) with (kappa(M)ge0), it gives relative subadditivity with the term (h^*M). Here both spaces are smooth projective, the boundary is zero, and we take (M=0). Invariance of plurigenera on the smooth locus identifies the generic-fiber Kodaira dimension with that of (F). Thus kappa(W,K_W-h^*K_B)>= kappa(F,K_F)=n-b>0. This is an established general-type-fiber theorem, not another use or strengthening of Theorem~asm:iitaka. Choose compatible canonical divisors and set (P=K_Y-T_B(K_B)=p_*(K_W-h^*K_B)). For every sufficiently divisible (m), pushing forward an effective divisor (div_W(phi)+m(K_W-h^*K_B)) gives (div_Y(phi)+mPge0). The resulting injection of sections preserves their ratios, so (kappa(Y,P)ge1). Corollary~ex:supporting now gives (00). On (B_k), choose an ample divisor (H_k) and a sufficiently small rational (ε_k>0) so that the ample rational class [ Q_k=P_k+k^{-1}r_k^*H+ε_k H_k ] satisfies (Q_k·eta_k=O(k^{-1})). This is possible because (eta_k· r_k^*H=eta· H) is fixed. Approximate (eta_k) by a positive sum (gamma_k=sum_{i=1}^{N_k} a_i[C_i]), with (a_i>0), of covering-curve classes, closely enough that [ Q_k·gamma_kle C/k, -K_{B_k}·gamma_kgedelta/2 ] for a constant (C) independent of (k). Keep only terms with (K_{B_k}· C_i<0). Their total anticanonical degree is at least (delta/2), while their total (Q_k)-degree is at most (C/k). Taking the weighted average of the ratios therefore gives one covering curve (Gamma_k) with [ frac{Q_k·Gamma_k}{-K_{B_k}·Gamma_k} <=(2C)/(delta k). ] The quantitative bend-and-break estimate cite[Theorem~5]{MiyaokaMori1986} gives a rational curve through a general point of (Gamma_k) with (Q_k)-degree at most (2dim B) times this ratio. Parameterization and uncountability give a covering family of such rational curves, denoted (R_k). Since (Q_k-P_k-k^{-1}r_k^*H) is ample and both (P_k) and (r_k^*H) are nef, we obtain P_k· R_k=O(k^{-1}), H· r_{k*}R_k=O(1). Shrink each parameter space so that the curve class and its contacts with the exceptional divisors and the strict transforms of the Siu divisors of (S) are constant. The general parametrized member is free in characteristic zero. Consequently (K_{B_k}· R_kle-2). Put (barbar R_k=r_R_k=r_{k*}R_k). Its bounded (H)-degree bounds (K_B·barbar R_kR_k), and effectivity of the relative canonical divisor gives 0le K_{B_k/B}· R_k =K_{B_k}· R_k-K_B· r_{k*}R_k=O(1). In particular, the upper bound is independent of (k). If an exceptional prime (E) has coefficient (a_E=A_B(E)-1) in (K_{B_k/B}), a positive contact with it contributes (a_E(E· R_k)) to this bounded sum. Both factors are positive integers. Thus the number of exceptional primes met by (R_k), their log discrepancies (A_B(E)=a_E+1), and their contact degrees are uniformly bounded. Contacts with the finitely many strict transforms of divisorial Lelong components of (S) are bounded by their fixed degrees against (barbar R_kR_k). Subtracting all these divisorial parts from (r_k^*S) leaves a positive current. Its restriction to almost every member of the family is defined and positive, by the local-potential and Fubini argument following Lemma~ex:web. Its degree is the first difference below; this proves that difference is nonnegative. Using (ex:multiplier-order) and (ex:small-curves) gives 0&<= {S}·barbar R_kR_k -sum_Enu_E(S),E· R_k &<= {S}·barbar R_k-(F_k/k)R_k-(F_k/k)· R_k <= O(k^{-1}). The sum includes exceptional primes and those strict divisorial components; terms with zero contact are irrelevant. For the last bound, the expression involving (F_k) equals (P_k· R_k+({S}-L_k/k)·barbar R_kR_k). The first term is (O(k^{-1})) by (ex:small-curves); the second has the same bound because (L_k-k{S}) stays bounded and the (H)-degrees of (barbar R_kR_k) are bounded. Bounded ample degree gives only finitely many numerical classes of the integral cycles (barbar R_kR_k). Pass to one class, so ({S}·barbar R_kR_k) is now one fixed real number. The current (S) on (B) also remains fixed. By Lemma~ex:valuation-acc, the sums in (ex:zero-gap) belong to an ACC set: there are a bounded number of terms, their nonnegative integral multiplicities are bounded, and all relevant discrepancies are bounded. Finite sums of this kind preserve ACC, by successively taking nonincreasing subsequences of their entries. We can now replace the limiting gap by an exact equality: the nonnegative gaps in (ex:zero-gap) must equal zero for some covering families. Otherwise a sequence tending to zero has a strictly decreasing positive subsequence, giving a strictly increasing sequence of the complementary sums. For a family with zero gap, restrict the residual current to almost every parameter fiber as above. Its integral on that complete curve is zero. A positive current on a curve is a positive measure, so zero integral forces it to vanish. Off the removed divisors the residual current agrees with (r_k^*S). The discussion following Lemma~ex:web and that lemma itself yield a positive-relative-dimensional rational quotient of (B) whose vertical directions are annihilated by (S) on an open. Composing with (h) and taking the Stein factorization gives a quotient of smaller base dimension whose vertical directions are annihilated by (T). It belongs to the collection minimized at the start, contradicting the choice of (b). ## Excluding normalized bounded curves We apply the same exact-zero mechanism to curves on the small modifications of a fixed late model. The normalization by volume is chosen so that its scale tends to zero, while the current and the discrepancy lattice stay fixed. Run a canonical LMMP on (X) with scaling of a fixed very ample rational divisor (A). For rational (t>0), its positive-threshold stages give terminal (Q)-factorial models (Y_t) on which (K_t+tA_t) is nef, big, and semiample [cite]. There are only finitely many divisorial contractions, since each lowers Picard number. Fix a late model (Y) after the last such contraction. All subsequent maps (Ydashrightarrow Y_t) are small, and are canonical nonpositive birational maps. If the program terminates, use the last model repeatedly. Define mu_t=vol(K_X+tA)^{1/n}. Since (K_X) is not big, (mu_tto0). In particular (K_Y) is nonbig. The transform (A_t) is effective up to rational linear equivalence; all intersections below use general covering curves, which are not contained in a chosen such representative. [Exceptional current comparison] Fix a positive current (T) on a smooth resolution in the pullback class of (K_Y). On a common smooth resolution of (Ydashrightarrow Y_t), write [ p^*K_Y=p_t^*K_t+J_t, J_tge0, ] where (J_t) is exceptional over (Y_t). Then (nu_E(T)gecoeff_E J_t) for every prime on that resolution. Proof. Pass to a common smooth model (V) also dominating the fixed resolution carrying (T). Choose on the fixed resolution a sufficiently positive line bundle (H). For sufficiently divisible (k), multiplier approximation gives a globally generated (OO(kp^*K_Y+H) ⊗ J(kT) Pull this system to (V). For the finite collection of primes on the common resolution whose coefficients we are comparing, choose a general member (D_k) of the underlying effective divisor system. At every such prime (E), it has multiplicity (ord_EJ(kT); global generation after removal of the ideal ensures no additional generic vanishing. Here the notation (H) also denotes its pullback to (V). Put (D'_k=p_{t*}D_k) and (H_t=p_{t*}H). The first is effective, and both are (Q)-Cartier because (Y_t) is (Q)-factorial. Pushing the rational linear relation gives [ D'_ksim_Q kK_t+H_t. ] Thus the exceptional divisor (D_k-p_t^*D'_k) is rationally equivalent to [ kJ_t+H-p_t^*H_t. ] The two exceptional divisors are equal: their difference is numerically trivial and exceptional, so the negativity lemma applied with both signs makes it zero. Since (p_t^*D'_k) is effective, [ coeff_E D_k >= kcoeff_E J_t +coeff_E(H-p_t^*H_t). ] The last coefficient is fixed for this model and (t). Combining with (ex:multiplier-order), dividing by (k), and letting (kto∞) proves the assertion. The same proof can be made on each common model, so the comparison applies to all divisorial valuations needed later. No uniformity in (t) of the fixed error is required. There are no sequences (t_j↓0) and covering families of curves on (Y_{t_j}), with smooth projective domains (C_j), for which both [ g(C_j), frac{(K_{t_j}+t_jA_{t_j})· C_j}{mu_{t_j}} ] are bounded by a fixed constant. Degrees mean degrees on the domains, or equivalently intersection with their pushforward cycles. Proof. Suppose such families exist. Fix once and for all a smooth resolution (p:Wto Y) and a positive current (T) in the class (p^*K_Y); pseudoeffectivity supplies this current. Neither (W) nor (T) will depend on (t). Let (g_0) be a common upper bound for the genera, and choose an integer (m>0) such that (mK_Y) is Cartier. We first compare the curves on these fixed models without changing their domains. Slice the parameter space so that the smooth total family has dimension (n), the parameter space has dimension (n-1), and evaluation is generically finite. It has rational maps to (Y), to (W), and to the common models comparing (Y) and (Y_t). Each target is proper, so each map extends at codimension-one points by the valuative criterion. Its indeterminacy locus has dimension at most (n-2), and its image cannot dominate the parameter space. Remove these finitely many images from that space, then resolve and compactify away from the remaining complete fibers. Denote the resulting maps by (rho_t:U_tto W) and (q_t=pcircrho_t:U_tto Y). The smooth general parameter fiber is still the original (C_t), with the same genus. The current we use on this varying space is always (rho_t^*T). For a prime (Z) of (U_t) exceptional over (Y), the restricted valuation of its function field is (r_Zord_E) for a divisorial valuation (E) over (Y). The coefficient of (Z) in the ramification difference is coeff_Z(K_{U_t}-q_t^*K_Y) =r_Za(E;Y,0)-1. This follows by factoring through a model extracting (E) and calculating the tame ramification of DVRs. It is positive and bounded away from zero: (Y) is terminal, so (a(E;Y,0)>1), and the index (m) puts these log discrepancies in (m^{-1}Z). In particular (a(E;Y,0)ge1+1/m). The remaining ramification coefficients are nonnegative. On the general parameter fiber, [ K_{U_t}· C_t=2g(C_t)-2, q_t^*K_Y· C_tge0. ] The latter ineuality follows from pseudoeffectivity and the covering property. Thus [ 0le (K_{U_t}-q_t^*K_Y)· C_tle 2g_0-2. ] Each exceptional contact contributes ((r_Za(E;Y,0)-1)(Z· C_t)) to this bounded sum. The first factor is at least (1/m), and the second is a positive integer. Formula~(ex:ramification-cost) therefore bounds the number of contacted exceptional primes, their log discrepancies, their ramification indices, and their contact degrees. In particular the integers (r_Z(Z· C_t)) are uniformly bounded. Also [ q_t^*K_Y· C_tin m^{-1}Zcap[0,2g_0-2], ] so these degrees take only finitely many values. The ramification estimate has bounded the possible contacts. We next compare their Lelong contributions with the canonical differences. The generic Lelong number of (rho_t^*T) at (Z) is (r_Znu_E(T)). Extract (E), remove its generic divisorial part, and use the transverse-power-map comparison for the zero-generic-number remainder. Lemma~ex:current-comparison shows that these coefficients dominate the coefficients of the pulled-back canonical difference (J_t). Its support is exceptional over (Y) as well as (Y_t), since the map between these models is small. Subtracting all exceptional divisorial parts of (rho_t^*T) leaves a positive current. Shrink the parameter space so that the relevant divisor contacts are constant, and restrict this residual current to almost every fiber. As in (ex:zero-gap), its degree is the first difference below and is nonnegative. Consequently 0&<= q_t^*K_Y· C_t -sum_{Z {exceptional}} r_Znu_E(T),Z· C_t &<= K_t· C_t <= (K_t+tA_t)· C_t =O(mu_t). All divisor contacts used here are nonnegative, since the curves cover the total space. On the fixed smooth resolution (p:Wto Y), one has [ a(E;W,0)=a(E;Y,0) -ord_E(K_W-p^*K_Y)<= a(E;Y,0), ] because the relative canonical divisor is effective. Thus Lemma~ex:valuation-acc applies to the one fixed current (T) on (W), with a fixed discrepancy bound. The sums in (ex:bounded-gap) form an ACC set: their Lelong numbers belong to this ACC set, and both the number of terms and their integral multiplicities (r_Z(Z· C_t)) are bounded. Pass to a fixed value of (q_t^*K_Y· C_t). Since (mu_tto0), the exact-zero argument of (ex:zero-gap) gives a family for which the left gap is zero. For that family, the restriction of the residual positive current to almost every parameter fiber is a positive measure of total mass zero, so it vanishes. On the dense open away from the removed exceptional divisors it agrees with (rho_t^*T). Apply Lemma~ex:web to the family evaluated by (rho_t:U_tto W). Its curve tangents generate a rational fibration of positive relative dimension on the smooth projective variety (W), and (T) annihilates the fiber tangents on a dense open. Transport this rational fibration through (p:Wto Y) on their common isomorphism open. This contradicts Proposition~prop:no-pullback. ## Special termination for signed representatives To turn a signed canonical representative into a big logarithmic adjoint, we will run a program whose negative rays meet the reduced boundary. The following special-termination statement uses only the lower-dimensional hypothesis; its proof keeps the programs on the boundary separate from the ambient one. [Special termination over a point] Assume Assumption~mmp:lower-models. Let ((V,Delta)) be a projective (Q)-factorial dlt (n)-fold with rational boundary and pseudo-effective adjoint. Choose an effective ample rational divisor (A) such that ((V,Delta+A)) is dlt and (K_V+Delta+A) is nef. The LMMP for (K_V+Delta) with scaling of (A) has special termination: if the program is infinite, its flipping loci are eventually disjoint from the round-down of the boundary. In particular, if at every stage the adjoint is rationally linearly equivalent to a signed divisor supported on that round-down, the program terminates at a log minimal model. Proof. In an infinite ample-scaling program the scaling limit is zero: a positive limit is covered by the termination theorem for a dlt pair with an ample summand in the scaling divisor cite[Theorem~1.9(ii), equivalently Theorem~4.1(ii)] {Birkar2012Special}. Discard the finitely many divisorial contractions and write the flips as (V_idashrightarrow V_{i+1}/Z_i), with scaling numbers (lambda_i>0) tending to zero. Fix a component (S_1) of (floor(Delta_1)), write (S_i) for its normal strict transform, and let (T_i) be the normalization of its image in (Z_i). Since the ambient contraction is small, (S_ito T_i) is projective birational. The standard discrepancy argument makes (S_idashrightarrow S_{i+1}) an isomorphism in codimension one after finitely many steps cite[proof of Lemma~3.6]{Birkar2010}. To identify the lower-dimensional input, consider the relative program induced on this one component of the floor. Take a small projective (Q)-factorialization (p_i:S'_ito S_i), and put [ D_i=K_{S'_i}+B_i=p_i^*((K_{V_i}+Delta_i)|_{S_i}), C_i=p_i^*(A_i|_{S_i}). ] Adjunction gives a dlt pair ((S'_i,B_i)), with (B_i,C_ige0), and ((S'_i,B_i+lambda_i C_i)) is lc. The scaling divisor has no floor component, so its restriction is effective. The number (lambda_i) is rational, being the ratio of intersections of rational divisors on the contracted rational curve. The globally nef divisor (L_i=K_{V_i}+Delta_i+lambda_i A_i) descends rationally linearly across (V_ito Z_i). Indeed, for sufficiently small rational (delta>0), the pair ((V_i,(1-delta)Delta_i)) is klt and its negative adjoint is ample over (Z_i). Relative basepoint freeness applies to a Cartier multiple of (L_i); its numerical triviality and the connected fibers give descent. Restricting and pulling back yields [ D_i+lambda_i C_isim_{Q}0/T_i. ] The descended divisor on T_i is nef: its pullback to S'_i is the pullback of the restriction of the globally nef L_i. By cite[Theorem~1.1(3)]{Birkar2012Special}, a (D_i)-LMMP over (T_i) with scaling of a fresh ample divisor terminates. The theorem applies to the effective rational lc boundary (B_i+lambda_i C_i), with (lambda_i C_i) rational Cartier and the driving pair (Q)-factorial dlt. Since the base map is birational, the endpoint is a log minimal model, not a Mori fiber space. Comparison with the relatively ample model (S_{i+1}) identifies this endpoint with a small (Q)-factorialization (S'_{i+1}), as in cite[Remarks~2.9--2.10]{Birkar2012Special}. The next point is that these finite relative programs concatenate into an absolute program with scaling of the transforms of (C_1). Throughout the (i)-th piece, (D_i+lambda_i C_i) remains the pullback of a nef divisor on (T_i), hence is globally nef. Here D_i and C_i also denote their transforms along this relative program. A contracted negative ray R satisfies [ (D_i+lambda_i C_i)· R=0, D_i· R<0. ] It follows that C_i· R>0. Every smaller coefficient of C_i leaves this ray negative, while D_i+lambda_i C_i is globally nef. Thus the absolute scaling threshold at that step is exactly lambda_i. The relative cone is a face of the absolute cone, cut out by the pullback of an ample divisor on the projective base (T_i); its extremal rays are therefore absolute extremal rays. Thus the pieces concatenate as asserted in the cited remarks. Rescaling the initial scaling divisor by (lambda_1), if necessary, makes its sum with the initial boundary lc and nef. If the concatenation were infinite, its positive scaling numbers would tend to zero. Fix a late scaling value lambda_j. On every earlier ray R contracted at threshold lambda_i, the calculation above gives [ (D_i+lambda_j C_i)· R =(lambda_j-lambda_i)C_i· Rle0. ] Thus all preceding steps are nonpositive for the adjoint with this fixed coefficient lambda_j. On a common resolution of the initial and the j-th models, the pullback of D_1+lambda_j C_1 equals the pullback of the late nef divisor D_j+lambda_j C_j plus an effective comparison divisor. Push this equality down to the initial model. The pushforward of the nef pullback is pseudo-effective, as is the effective comparison term, so D_1+lambda_j C_1 is pseudo-effective. Closedness of the pseudo-effective cone and lambda_jto0 show that D_1 is pseudo-effective. Assumption~mmp:lower-models therefore gives an absolute log minimal model for that initial pair. The concatenation terminates by cite[Theorem~1.9(iii)]{Birkar2012Special}, since its zero limit is never attained. The floor-component argument of cite[Lemma~3.6]{Birkar2010} now gives special termination. The only good-model input in this argument concerns the absolute pair on the lower-dimensional floor component. In particular it does not require an effective representative in dimension (n). For the final assertion, a curve disjoint from the round-down has degree zero against any signed divisor supported there. It therefore cannot generate an adjoint-negative flipping ray. Special termination rules out all sufficiently late flips. There are only finitely many divisorial contractions, since each lowers the Picard number, and pseudo-effectivity excludes a Mori fiber space as the endpoint. Thus the endpoint is nef. ## The signed alternative The jet construction in Section~sec:jets will produce signed rational representatives of (K_t). The following result turns such a representative into a big logarithmic adjoint on a resolution. On any model (Y_t), let a signed rational divisor represent (K_t) up to (Q)-linear equivalence. On a log resolution (p:Wto Y_t), let (D) be the reduced SNC divisor consisting of its strict support and all exceptional divisors. Then (K_W+D) is big. Proof. Put (P=K_W+D), and suppose it is not big. It cannot have Iitaka dimension at least one: by Corollary~ex:supporting, (K_W+cP) would be big, and ((1+c)P=(K_W+cP)+D) would then be big as well. Hence (kappa(W,P)le0). Run the dlt LMMP for (P) with ample scaling. The divisor (P) has a signed rational representative supported on the floor (D), and its transforms retain that property. Every negative flip must meet the floor: on its complement the representing rational section is nowhere vanishing, so the adjoint has zero degree on every complete curve there. By Proposition~prop:special-termination, an infinite sequence would eventually have all flips disjoint from the floor, a contradiction. Divisorial contractions are finite in number. Thus the program terminates at a (Q)-factorial dlt log minimal model ((Z,D_Z)). The boundary is reduced, (K_Z) is pseudo-effective as the birational pushforward of (K_W), and the nef adjoint has a signed rational representative supported on (D_Z). Its restriction to (D_Z) is semiample by Proposition~prop:boundary-restriction. Apply Theorem~thm:signed with (L=K_Z+D_Z) and (c=1), so that the required pseudo-effective class (L-cD_Z) is (K_Z). It gives (kappa(Z,L)=nu(Z,L)ge0). The Iitaka dimension is at most zero, as proved above, so both dimensions are zero. Intersecting (K_Z+D_Zequiv0) with an ample ((n-1))-fold product, and using pseudoeffectivity of (K_Z), forces (D_Z=0). The signed relation then gives (K_Zsim_Q0). The signed representative of (K_W) is supported on (D): pull back the given representative of (K_t) and add the relative canonical divisor. Since (D_Z=0), every component of this support is contracted over (Z). On a common resolution (a:Vto W), (b:Vto Z), we therefore have (a^*K_Wsim_Q E) for a signed (b)-exceptional divisor (E). The class (a^*K_W) is pseudo-effective. A positive current in this class has zero intersection with the ((n-1))-st power of an ample pullback from (Z), so is supported on the exceptional locus. The support theorem makes it effective divisorial. Independence of exceptional divisor classes, by negativity, identifies its coefficients with those of (E); hence (Ege0). Pushing forward by (a) shows that (K_W) is (Q)-linearly effective, contradicting (ex:counterexample). ## Birational families between fixed covers The last obstruction concerns fixed finite covers of a birational model of the counterexample. If birational maps between two such covers form an algebraic family whose graphs dominate their product, then one fixed cover is birational to an abelian variety. The positive canonical current rules this out. [Fixed-cover obstruction] Let (Y) be a projective (Q)-factorial terminal birational model of the counterexample (X) in Equation~(ex:counterexample), and let (T_ito Y), (i=1,2), be fixed finite surjective morphisms from normal integral projective varieties. There is no integral parameter variety (A_0) and closed subvariety [ Gammasubset A_0times T_1times T_2 ] flat over (A_0), with each fiber the integral graph of a birational map (T_1dashrightarrow T_2), for which the projection (Gammato T_1times T_2) is dominant. Proof. Suppose such a family exists, and denote its birational maps by (phi_a:T_1dashrightarrow T_2). Choose one map (phi_{a_0}) in this family and use it to identify (T_2) birationally with (T_1). The maps (g_a=phi_{a_0}^{-1}circphi_a) are then birational selfmaps of (T_1). The evaluation ((a,u)|->(u,g_a(u))) is dominant onto (T_1times T_1); in particular, the images of a general point under these selfmaps are dense. We use the birational automorphism group to identify this fixed cover. The cover is non-uniruled, being generically finite over the non-uniruled (Y). Hanamura's non-uniruled birational-group theorem gives a smooth projective birational model for which the reduced birational group is a group scheme, locally of finite type, and its identity component is an abelian variety of regular automorphisms cite[Theorems~2.1--2.2]{Hanamura1988}; see also cite[Proposition~3.7 and Theorems~3.8--3.9]{Blanc2017}. After conjugating the maps to that model, shrink the irreducible parameter variety anew so that their graph closures are flat with integral fibers and both graph projections remain birational on every fiber. This is a family in the flat graph functor of cite[Definition~3.4]{Blanc2017}, which is represented by the birational scheme cite[Proposition~3.7]{Blanc2017}. It therefore defines a morphism to that scheme; because the parameter variety is reduced, this morphism factors through its reduction. Its connected image lies in a single component, hence in a translate of the identity component. Translate the family by one member so that it lies in that identity component; this preserves dominance of evaluation. Thus the identity component, an abelian variety, has a dense orbit. That orbit is the quotient by the stabilizer of a general point. A quotient of an abelian variety is again an abelian variety, so the fixed cover is birational to an abelian variety. It remains to see why this abelian cover contradicts the original counterexample. Resolve the generically finite rational map from that abelian variety to (X): on a smooth model (U), ramification gives [ K_U=f^*K_X+R, Rge0. ] The canonical class of (U) also has the effective representative (E_A) exceptional over the abelian variety: choosing a nowhere-zero top form on that variety gives (K_Usim E_A). This second choice of representative need not be the canonical divisor compatible with the displayed ramification formula. Let (T) be a positive current in (K_X). The current (f^*T+[R]) is positive and represents (K_U). Let (alpha) be an ample class on the abelian variety. The intersection of (K_U) with the ((n-1))-st power of its pullback is zero, since (K_Usim E_A) and (E_A) is exceptional. The positive current (f^*T+[R]) therefore has zero mass against that power. On the open where the birational map to the abelian variety is an isomorphism, the pulled-back ample form is strictly positive, so the current vanishes there. Its support is consequently exceptional over the abelian variety. The positive summand (f^*T) is supported there as well, and the support theorem makes it divisorial. Push forward this summand alone: (f_*(f^*T)) is a positive divisorial current whose class is (deg(f)c_1(K_X)) by the projection formula. Rationality of the numerical class supplies a rational effective representative, and irregularity zero turns numerical effectivity into nonvanishing, as in Lemma~ex:irregularity. This contradicts (kappa(X,K_X)=-∞). # Tools for moving base components This section follows the base components of a complete kernel of jets as the marked point moves. Their multiplicities give degree and canonical-intersection bounds. A separate finiteness lemma treats the covers that arise when a component projects generically finitely onto a fixed base. These arguments use neither nonvanishing, abundance, minimal models, nor logarithmic subadditivity. Section~sec:jets applies them to moving components produced by failure of the scalar or two-slot jet estimates. All varieties in the moving-center arguments are over C. A rational Cartier divisor is a rational divisor some positive integral multiple of which is Cartier. A requirement that kP be Cartier means that kP is an actual integral Cartier divisor. For a nef class on an integral subvariety, intersections are computed after pullback to its normalization and a resolution. For a linear subseries of a line bundle, its base ideal is the image of its evaluation map after tensoring by the inverse line bundle. ``Isolated at the generic point of V'' means that this ideal is primary for the maximal ideal of O_{Z,eta_V}. All powers of ideals below are ordinary powers. ## Numerical subadjunction on the center The first result converts a generic log canonical center into a canonical-intersection ineuality. Its testing class is allowed to live on the center itself. The proof uses dlt adjunction and the klt-trivial canonical bundle formula; it does not require the auxiliary pair to be log canonical away from the generic center. [Numerical generic subadjunction] Let Z be a projective Q-factorial klt variety over C, let Θge0 be a rational divisor, and let Vsubset Z be an integral positive-dimensional subvariety. Suppose that (Z,Θ) is lc at the generic point of V and that a divisor of log discrepancy zero has center V. Let f=dim V, let P be a nef rational Cartier divisor on V, and let rho:V^*to V be a smooth projective resolution, factoring through the normalization. Then K_{V^*}·(rho^*P)^{f-1} <= (K_Z+Θ)|_V· P^{f-1}. The right side is the intersection on V of the restricted rational Cartier class with P^{f-1}. No Q-Gorenstein hypothesis on the normalization of V is required. Proof. Apply the dlt blowup theorem for an effective rational divisor cite[Theorem~2.10]{FujinoHashizume2023}. It gives a projective birational morphism p:Qto Z with Q Q-factorial and p^*(K_Z+Θ)=K_Q+B_Q+E_Q, where (Q,B_Q) is dlt and E_Qge0 is the surplus over the non-lc locus. In this construction the coefficients of the strict boundary are truncated at one and the relevant extracted divisors are reduced; the coefficients left over form E_Q. The surplus is absent over the open where (Z,Θ) is lc. In particular it is absent over a neighborhood of the generic point of V. There are lc strata of (Q,B_Q) mapping onto V. Indeed over that lc neighborhood the construction is crepant, and the center of a log-discrepancy-zero divisor on the dlt model is an lc stratum. Choose a stratum S minimal among those mapping onto V. Iterated dlt adjunction makes S normal and gives an effective rational dlt different on S; its lc centers are precisely the smaller ambient strata in S [cite]. The divisor E_Q does not contain S. Since Q is Q-factorial, a Cartier multiple of E_Q restricts to an effective Cartier divisor on S. Adding that restriction to the different gives an effective rational divisor B_S such that K_S+B_Ssim_{Q}(p|_S)^*((K_Z+Θ)|_V). Over the generic point of V this pair is klt: the surplus is absent there, and a non-klt center of the different dominating V would be a smaller stratum, contradicting the choice of S. Factor the map through the normalization V^nu and its Stein factorization: [ Sxrightarrow{q}V'xrightarrow{eta}V^nuxrightarrow{nu}V. ] Here q has connected fibers and eta is finite. Put D_V=nu^*((K_Z+Θ)|_V) and D'=eta^*D_V. We apply the canonical bundle formula on the connected-fiber base V', then push the resulting intersection ineuality through the finite map eta. Its ramification will contribute an effective term. The final comparison is between the canonical cycle on V^nu and the canonical divisor on V^*. We verify the klt-trivial-fibration hypotheses for q:(S,B_S)to V'. The generic pair is sub-klt, and Equation~(common:subadjunction-stratum) is the required rational-linear pullback expression. For the rank condition, take a log resolution over the generic point of V'. Since the boundary on S is effective and the generic pair is klt, the round-up of its discrepancy divisor has coefficient zero along every strict boundary component and has nonnegative coefficients only on exceptional divisors. Over this dense klt open, its pushforward to the normal S is O_S: a rational function allowed poles only over exceptional centers extends across codimension at least two. Connectedness in the Stein factorization therefore gives generic rank one for the required discrepancy round-up pushforward. This verifies the rank condition, rather than deducing it from connected fibers alone. The canonical bundle formula, with Ambro's moduli positivity in the form recalled by Fujino--Gongyo, now gives D'sim_{Q}K_{V'}+B_{V'}+M_{V'}, where M is a b-nef moduli divisor cite[Definition~3.1 and Theorem~3.5]{FujinoGongyoModuli2014}. The definition requires sub-klt over the generic point; it does not require the pair to be globally sub-lc. Thus the possible non-lc fibers caused by E_Q|_S are permitted. The discriminant B_{V'} on this initial base is effective. To check this at a prime divisor Fsubset V', work over its generic point, where the base is regular. A component of its inverse image on S has multiplicity m_Fge1 and nonnegative boundary coefficient b_F. Its discrepancy bounds the generic lc threshold t_F from above by [ t_Fle(1-b_F)/(m_F)le1. ] Thus the discriminant coefficient 1-t_F is nonnegative. If the fiber is already non-lc, t_F can be negative; this only increases that coefficient. Let P'=eta^*nu^*P. On a higher projective base mu:Vhat'V'to V' where the moduli divisor is nef, its trace satisfies M_{V'}=mu_*M_{Vhat'}V'. The projection formula gives [ M_{V'}·(P')^{f-1} =M_{Vhat'}V'·(mu^*P')^{f-1}ge0. ] Only this intersection is asserted to be nonnegative; the trace M_{V'} need not itself be nef. Likewise B_{V'}·(P')^{f-1}ge0 by effectivity and nefness. All intersections with Weil divisors here mean their cycle intersections with rational Cartier powers, so they are defined even when the separate canonical or discriminant divisor is not Q-Cartier. Choose compatible canonical Weil divisors. Since eta is finite and separable, the codimension-one ramification formula, followed by pushforward, is [ eta_*K_{V'}=(degeta)K_{V^nu}+eta_*R_eta, R_etage0. ] This identity of Weil cycles uses the discrete valuation rings at generic prime divisors and does not require K_{V^nu} to be Q-Cartier. Intersect Equation~(common:subadjunction-cbf) with (P')^{f-1} and discard the nonnegative discriminant and moduli intersections. Pushforward through eta and the nonnegative ramification intersection then give [ (degeta)D_V·(nu^*P)^{f-1} >=(degeta)K_{V^nu}·(nu^*P)^{f-1}. ] Finally the pushforward of K_{V^*} to V^nu is its compatible canonical cycle. All additional divisors on the resolution have image of codimension at least two, so their intersection with f-1 pulled-back Cartier powers is zero. Dividing by degeta proves Equation~(common:subadjunction-ineuality). When f=1 there are no birational exceptional divisors on the normal curve, and the same calculation applies directly. ## Degree and canonical bounds for a base component We next apply numerical subadjunction to a boundary built from an actual linear system. Its local multiplicity controls both the degree of the center and the coefficient of that boundary. [A base component of high multiplicity] Let Z be a projective Q-factorial klt variety of dimension d, let P be a nef rational Cartier divisor, and let k>0 be an integer for which kP is Cartier. Let 0neH H^0(Z,O_Z(kP)) be a linear subspace with proper base locus and base ideal b.b. Suppose that V is an integral base-locus component of dimension f0}), ] then P^f· Vle (k/h)^aP^d. There are a rational number c with 00, on a smooth resolution V^* one consequently has K_{V^*}· P^{f-1} <= (K_Z+cP)|_V· P^{f-1}. The degree estimate includes f=0; the canonical estimate is asserted only for positive-dimensional V. Proof. The local ring R=O_{Z,eta_V} is regular of dimension a. Isolation means that bR is m-primary.m-primary. Choose a general sections of H. Their local equations form a system of parameters in R, since an m-primaryfrakm-primary ideal is contained in no smaller prime. Let q be the parameter ideal they generate. It is contained in m^h.m^h. Regularity and the multiplicity comparison give [ length(R/q)q) =e(q,R)q,R)>= e(m^h,R)=h^a.m^h,R)=h^a. ] This is the generic intersection multiplicity along V. The global intersection can be performed without assuming that the series has no other base components. Before each cut, discard the components of the current effective cycle that lie in the base locus. Before the final cut their dimensions are greater than f, so none can contain V: otherwise V would not be a base-locus component. Discarding them therefore leaves the calculation at eta_V unchanged. A general next section meets the remaining components properly. Nefness of P shows that discarding effective components can only decrease their total P-degree. After a cuts the surviving effective cycle has P-degree at most k^aP^d, and contains V with multiplicity at least h^a. This proves Equation~(common:base-component-degree). Let lambda=lct_{eta_V}(b)b) be the generic ideal threshold. It is positive and rational. The valuation from blowing up the smooth generic center has log discrepancy a and ideal order at least h, so [ 00, Lemma~common:numerical-subadjunction gives Equation~(common:base-component-canonical), and c=klambdale ak/h. ## Differentiating the full moving jet kernel Fixing a high order at a varying point produces a family of linear subspaces of one fixed section space. A component common to two successive base loci acquires high multiplicity, because parameter derivatives of the higher kernel still belong to the lower kernel. We first state the parameter-family conclusion. Local restriction to a fiber is treated separately below. The differentiation argument is related to the method of Ein--K"uchle--Lazarsfeld cite[Proposition~2.3]{EinKuchleLazarsfeld1995}; the ordinary-power and primary-ideal details needed here are included. [Moving multiplicity] Let Z be an integral projective variety of dimension d, and let P be a nef, big, semiample rational Cartier divisor. Fix an integer k with kP Cartier and positive real numbers [ tau_00. ] For a smooth point x, let H_i(x) be the entire subspace of sections of kP vanishing at x to order at least lceil ktau_irceil. Suppose that for very general x all these subspaces are nonzero and the base locus of H_0(x) has a positive-dimensional component through x. If kdeltage4, then, after an algebraic parameter extension and restriction to nonempty opens, there are an irreducible parameter space T, a fixed adjacent pair of levels, and a family of integral subvarieties V_t common to the two base loci such that the incidence [ Gamma={(t,z):zin V_t}subset Ttimes Z ] dominates Z. Each general V_t has dimension in [1,d-1], is an isolated component of the higher base locus at its generic point, and its entire higher-series base ideal is contained there in I_{V_t}^{h}, h=lceil kdelta/2rceil, k/hle2/delta. These are ordinary local ideal powers in the smooth ambient open. The incidence base ideal lies in I_ on a dense open of Gamma. If, in addition, Z is Q-factorial and klt, then for a general member, with f=dim V_t, one has P^f· V_t&le(2/delta)^{d-f}P^d, K_{V_t^*}P^{f-1}&le(K_Z+cP)|_{V_t}P^{f-1}, 00) with (mK) Cartier. ## Normalization and two local tools We first normalize the small adjoint divisors so that their volumes stay bounded away from zero. Fix a small rational number (ε>0). An integer (q>0) will be chosen after (ε) and before (t). As rational (t↓0), choose positive integers (r=r(t)) with rmu_t→ε, L=r(K+tA), L_b=L+bmKquad(0le ble q). Only rational weights (b) are used for model constructions; volume functions in integrals are extended continuously to real weights. The positive part of (L_b) is the nef semiample class [ N_b=(r+bm)(K_s+sA_s) quadhbox{on }Y_s, s=(rt)/(r+bm). ] When emphasizing the weight, denote this axis model by (Y(b)=Y_s) and write (A_b) for its transform of (A). All references to positive parts below mean these classes on their scaling models, or their pullbacks to common resolutions. Monotonicity of volume in pseudo-effective order gives vol(L)le vol(L_b)le (1+(qm)/(r))^nvol(L). Indeed (L_b-L=bmK) is pseudo-effective, and ((1+bm/r)L-L_b=bmtA) is effective up to equivalence. Consequently (N_b^n) is bounded above and bounded away from zero, uniformly for (bin[0,q]), and tends uniformly to (ε^n). More explicitly, for every fixed (q), once (t) is sufficiently small depending on (q), (ε^n)/(2)le N_b^nle2ε^n (0le ble q). The constants in this display are independent of (q); only the threshold for (t) depends on the fixed choice of (q). All unspecified constants in this section may depend on (n,ε,q,Y,A,m), but not on sufficiently small (t) or on (b). This normalization gives a convenient form of the curve exclusion. A family of curves with uniformly bounded geometric genus and (N_b)-degree cannot cover the corresponding (Y_s) for arbitrarily small (t). In fact (s/tto1) uniformly in (b), and [ (r+bm)mu_s=vol(L_b)^{1/n} ] is bounded above and away from zero. Thus such curves would have uniformly bounded ((K_s+sA_s))-degree divided by (mu_s), contrary to Proposition~prop:bounded-curves. We call this consequence the normalized curve exclusion. We will apply the local estimates of Section~sec:common-tools on several different models. In each application, let (Z) denote the ambient model of dimension (d), let (P) be its nef, big, semiample class, and let (V) be a moving base component of dimension (f). Lemma~common:moving-kernel applies to the entire kernels of jets in one Cartier degree (kP), at levels separated by a gap (delta). It gives vanishing along (V) in the ordinary ideal power of order [ h=lceil kdelta/2rceil, kdeltage4, ] and, when (Z) is (Q)-factorial and klt, the estimates P^f· Vle (2/delta)^{d-f}P^d, K_{V^*}P^{f-1}le(K_Z+cP)|_V P^{f-1}, 00. Exceptional components not meeting that open are not assigned this positivity. The separate fiber-restriction Lemma~common:fiber-restriction will be used only at a general incidence point where the scheme fiber is reduced and the selected base component is isolated. To apply normalized curve exclusion, we need curves with bounded genus as well as bounded degree. The preceding intersection estimates provide them on components of general type. Suppose a moving (f)-fold (V) is of general type, (P|_V) is nef and big, and 01), cut by (f-1) general members of (|H|). Their (P)-degree is (H^{f-1}Ple C_2^{f-1}C_0), and their images under the birational morphism defined by (H) have degree (H^fle C_2^fC_0). A general plane projection is birational on such an image curve; the arithmetic genus of the plane image therefore bounds its geometric genus in terms of this degree. For (f=1), the canonical-degree bound already bounds the genus. We will also use this construction for a log smooth pair of log general type with reduced boundary, using coefficients in the fixed set ({1}) in effective birationality. If such components sweep an ambient space mapping to (Y), and their curves project nonconstantly with (N_b)-degree bounded by their (P)-degree, we obtain forbidden covering curves on (Y_s). When the component has positive-dimensional image in the base, general complete intersections from the big birational system project nonconstantly. The components sweep, and their chosen incidence points can lie outside any prescribed countable union of proper closed subsets. Hilbert schemes and spaces of maps have only countably many components, so these curves can be parameterized in covering algebraic families. Thus, whenever Equation~(jet:curve-input) is established below, it remains to check general type and the stated covering and projection conditions to invoke normalized curve exclusion. We first apply the two local tools to scalar sections. Excessive vanishing produces a proper moving component, whose general type turns the numerical estimates into the forbidden curves. [Scalar jet bound] For sufficiently small (t), let (P) be the positive part of (2r(K+2tA)), and put (rho=(P^n)^{1/n}). At a very general point, every nonzero section of every Cartier multiple (kP) has order at most (4krho). The same statement holds on a common resolution after adding an effective exceptional divisor that does not change the section spaces. Moreover [ 2rmu_tlerhole4rmu_t, ] so in particular (rhole8ε) for small (t). Proof. First, pseudo-effectivity of (K) gives the volume bounds: [ vol(K+tA)levol(K+2tA)le2^nvol(K+tA). ] Suppose the asserted order bound fails for arbitrarily small (t). There are countably many Cartier degrees, and nonzero jet kernels in each degree are detected by algebraic rank loci. Thus failure at a very general point gives a degree in which a violating section exists on a common constant-rank marked-point open. Taking powers makes this degree (k) arbitrarily large and divisible. Choose (n+1) levels strictly between (rho) and (4rho), with equal gaps comparable to (rho). The full jet kernels at these levels are nonzero. Write (tau_0) for the lowest level. If the marked point were isolated in its base locus, local Bezout for (n) general kernel sections would contribute at least ((ktau_0)^n>k^nP^n), exceeding the total intersection. Lemma~common:moving-kernel therefore supplies a moving proper positive-dimensional component (V). We can now apply the curve construction. The component is of general type by Proposition~prop:general-type. Here let (Z) be the scaling model for (K+2tA), so that (P=2r(K_Z+2tA_Z)). The effective transform (A_Z) does not contain a general such component, and (K_Z|_Vle(2r)^{-1}P|_V) in effective order. Equations~(jet:tracking-estimates) consequently imply Equation~(jet:curve-input) with constants uniform in (t); the gap is bounded above and away from zero because (rho) is comparable to the fixed (ε). The resulting bounded-genus, bounded-normalized-degree covering curves contradict Proposition~prop:bounded-curves. The argument applies to every Cartier multiple, not only the divisible degrees used for tracking: a section violating the bound in any Cartier degree could be raised to a power in an arbitrarily divisible degree. Its order and its degree increase by the same factor. Finally, exceptional additions preserve the section spaces and the orders at general points. This proves the assertion on resolutions. ## The two-slot bundle We next seek large jet separation by placing two copies of the normalized adjoint divisor on a projective bundle. The weight decomposition of its sections will let us compare moving components with either copy of the base. On (Ytimes Y), let Z_0=PP(mK_1 ⊕ mK_2), xi=OO_{Z_0}(1), M=L_1+L_2+qxi. We use the convention that sections of (kxi) are symmetric powers of the indicated direct sum. Fixing one base slot gives the one-slot bundle (PP(OO_Y ⊕ mK)), with class (L+qxi), up to a constant line from the fixed slot. We call it a slice. The torus open in either bundle is the complement of its two axes. [Models, volume, and weights] Both the total class (M) and the slice class have big semiample positive parts (P) on (Q)-factorial klt models (Z), with [ K_Z+Deltasim_{Q}c_tP, Deltage0, ] where (c_t) is bounded independently of small (t). Their volumes satisfy, respectively, vol(M)&=((2n+1)!)/((n!)^2) int_0^qvol(L_b)vol(L_{q-b}),db, vol(L+qxi)&=(n+1)int_0^qvol(L_b),db. In particular both volumes are bounded above and below by positive constants times (q). For each rational weight, on common resolutions, Psim_{Q}N_{b,1}+N_{q-b,2}+E_b, E_bge0, with the analogous one-term formula on slices. At a generic point of a base divisor of a slice, and on a general torus fiber, the full system has no fixed subtraction. Proof. We construct the semiample models as models of big klt adjoints. The weight decomposition then gives both the volume formulas and the comparison with the axis models. The product (Ytimes Y) is (Q)-factorial. To see this, take a smooth resolution (widetilde Yto Y). Its irregularity is zero by Lemma~ex:irregularity, so every line bundle on (widetilde Ytimeswidetilde Y) is a tensor product of line bundles from the two factors. The strict transform of a Weil divisor on (Ytimes Y) is therefore linearly equivalent to a sum of two factorwise pullbacks. Push this relation down to (Ytimes Y). The resulting factor divisors are (Q)-Cartier because (Y) is (Q)-factorial, and hence so is the original divisor. Products of canonical singularities are canonical, and the projective bundles are klt. The two disjoint Cartier axes form a plt boundary (D), and [ K_{Z_0}+D=K_{sum}, Dsim2xi-mK_{sum}. ] The notation (K_{sum}) means (K_1+K_2), or (K) on a slice. For (00) for (Xi_t) on the original space. The same reasoning on each axis gives, after decreasing (eta), a klt threshold bound for the restricted divisor there. The first bound will be used away from the axes; the restricted bounds allow inversion of adjunction along the coefficient-one axes. Now choose rational (00), there is an integer (q>0) such that, for all sufficiently small rational (t>0) and (r) as in Equation~(jet:normalization), the positive part (P) of Equation~(jet:two-slot) has [ ε(P;x)>2n+2 ] at a very general smooth point of its torus open. In particular, for some sufficiently divisible integer (k>0), the complete system (|kP|) separates jets of order strictly greater than ((2n+2)k) at such a point. The section spaces and these general-point jets agree on the original bundle and on common birational resolutions. Proof. The tracking lemma reduces failure of the assertion to a moving base component. We will first analyze its fibers over the two base factors, and then treat components that are generically finite over both. Write (d=2n+1). Choose a gap (delta>0) large enough that [ 2(n+1)ε^n(2/delta)^n<1. ] This ineuality will exclude a component that is a whole projective-line fiber of a slice. Now choose (q) sufficiently large that [ (((2n+1)!)/(4(n!)^2)qε^{2n})^{1/d} >2n+3+(d+1)delta. ] For all sufficiently small (t), the (d+1) levels (tau_i=2n+3+(i+1)delta), (0le ile d), then lie below the volume root of the total positive part. Both choices precede the choice of (t). Assume that the Seshadri assertion fails for arbitrarily small (t). Because the highest level is below the volume root, jet counts show that the full kernels at all these levels are nonzero in sufficiently large divisible degree. A curve through the marked point with degree-to-multiplicity ratio below the lowest level is contained in that system's base locus. Such a curve exists from the assumed Seshadri bound and the choice (tau_0>2n+3), whether or not the infimum defining the constant is attained. Thus the lowest kernel has a positive-dimensional base component, and Lemma~common:moving-kernel supplies a moving proper component (V) and both estimates (jet:tracking-estimates). The numerical estimates already rule out components of general type. For all moving components under consideration, the effective boundary (Delta) of Lemma~jet:bundle-models does not contain the component. Thus (K_Z|_Vle c_tP|_V) in effective order. Whenever (V) is of general type, the curve construction following Equation~(jet:curve-input) and weight domination give a contradiction. To treat the remaining components, we begin with their fibers over a base factor. Restriction to a slice.: Resolve the rational projection from the total model to the first copy of (Y), and work on the open where this resolution and the original bundle agree. Fixing the first coordinate there gives the opposite slice. Write (Z_{sl}) for its model from Lemma~jet:bundle-models and (P_{sl}) for its positive part; their dimension and volume are (n+1) and Equation~(jet:slice-volume), respectively. We use these slice objects until returning to the total model below. Choose one sufficiently divisible Cartier degree for the total model, the slice model, and their comparison on a common resolution. Choose a general incidence point on the common isomorphic torus open, away from the other base components and where the ordinary-power containment of Lemma~common:moving-kernel holds. Generic smoothness allows us also to require that (V) be smooth over its actual image in the first slot there. Let (F) be the reduced component of the scheme fiber through that point, and suppose (00). The ruled component (F) itself need not be of general type, so we seek the adjunction estimate on its base (W). We obtain it from the coefficients of the slice sections. Use the section-space identification to view the subseries on the original slice, where the toric weight decomposition is defined. The full fixed factors are units at the chosen generic torus point, so this identification preserves the order (h). In degree (k), expand each of these sections into its toric weights. At the generic point of (W), let (I_l) be the ideal generated by the coefficients of the nonzero weight (l) over all these sections. Put (h=lceil kdelta/2rceil). All (I_l) lie in (m_W^hfrakm_W^h): vanishing along the generic torus fiber says that the coefficient of every Laurent monomial vanishes to this order. Their sum is (m_W)-primary. Otherwise its larger common zero germ, times the generic torus fiber, would contradict isolation of (F). Our aim is to construct an effective boundary (Θsim_{Q}c'L_b), for some rational weight (b), such that (W) is an lc center near its generic point and (c'le2(n-w)/delta). Numerical subadjunction will then give a canonical-intersection bound on (W), even though (F) is ruled. Work in the regular local ring at the generic point of (W), of dimension (c=n-w). Resolve the finitely many coefficient ideals simultaneously. The exponents for which their product defines an lc ideal pair form the rational polytope [ u_lge0, sum_lu_lord_E(I_l)le a(E;Y,0). ] Here (E) runs through the divisors needed to test log canonicity on the simultaneous resolution, and (a(E;Y,0)) is the log discrepancy. Include the exceptional divisor of the blowup of the closed point. Since every (I_lsubsetm_W^hfrakm_W^h), it gives (sum_lu_lle c/h). The polytope is compact, and (sum_lu_l) has a positive rational maximum on it. Fix a rational maximizing point. For each coordinate (l), some equality at this point must involve a valuation with (ord_E(I_l)>0); otherwise (u_l) could be increased while staying in the polytope. Such a valuation has log discrepancy zero for the ideal pair. Its center (C_l) is contained in (V(I_l)) and contains the generic point of (W). Their intersection is contained in (V(sum_l I_l)), so it is exactly (W) locally. To apply the intersection property for lc centers, we first realize these ideal lc places by an actual divisor. Choose an integer (M_0>max_l u_l) and (M_0) general coefficient divisors (D_{l,j}) from each system. Set (Θ=sum_l(u_l/M_0)sum_{j=1}^{M_0}D_{l,j}). On the simultaneous resolution its fixed crepant coefficients are at most one, and its general moving divisors meet transversely with coefficients below one. Thus the pair is lc near the generic point of (W), and every active valuation remains an lc place. Each (C_l) is consequently an lc center of this pair. The intersection property cite[Theorem~1.7]{KollarKovacs2010}, applied to the identity morphism on its lc open, shows that (W) is an lc center generically. This also applies when some maximizing coordinate (u_l) is zero: the active valuation blocking that coordinate still has center in (V(I_l)). Set [ b=(sum_lu_l(l/k))/(sum_lu_l), c'=ksum_lu_lle(kc)/(h). ] Since a weight-(l) coefficient divisor has class (kL_{l/k}), the constructed boundary has class (c'L_b). Also (c'le kc/hle2(n-w)/delta). Transform it and (W) to the small scaling model (Y(b)); our general incidence choice places the generic point of (W) in their common isomorphism open. On this model the boundary has class (c'N_b). Numerical subadjunction and (N_b=(r+bm)K_{Y(b)}+rtA_b) give [ K_{W^*}N_b^{w-1} le(frac1{r+bm}+c')N_b^w.W. ] Indeed a general (W) is not contained in a fixed effective representative of (A_b), so its contribution to this intersection is nonnegative. The slice degree estimate and Equation~(jet:saturated-degree) give, in turn, [ 00 and real numbers r>1, ε>0. Suppose H^nle(2ε)^n, K_WH^{n-1}le2H^n/r, (2L+qS)H^{n-1}le4H^n, (14ε)^n<1/4, 80ε<3/4. Assume that the following fixed data exist. [label=(roman*)] A smooth integral complete-intersection flag from W to a curve C, whose successive divisor classes are h_i=l_iH for fixed positive integers l_i, with mu_{min}(Ω_W^1|_C)ge0. On the projective bundle [ Z=P(P(O_W(S)_1 ⊕ O_W(S)_2)→O_W(S)_2)→ Wtimes W, xi=c_1(O_Z(1)), M=L_1+L_2+qxi, ] use the convention that the direct image of O_Z(a is the a-th symmetric power of the displayed sum. A smooth point z_* in the complement of the two axes and finitely many sections of k_0M, for some integer k_0>0 with k_0L Cartier, generate jets through order (2n+2)k_0 at z_*. A point xin W, its blowup pi_x:WWtohat W with exceptional divisor J, and a curve class [ gamma=f_*(A_1... A_{n-1}), ] where f:VtoWWhat is one fixed birational morphism, V is smooth integral projective, and the A_i are ample integral Cartier divisors, such that J·gamma>0 and pi_x^*(L+K_W/2+lambda S)·gamma le40ε J·gamma (0lelambdale q). It suffices to check the two endpoints in this affine ineuality. These data cannot coexist. There is no relation required between x and the two base coordinates of z_*. The divisors L,S,K_W are not assumed nef, and K_W is not assumed pseudo-effective. The theorem starts with the actual flag and movable witness, rather than with geometric conditions from which one might construct them. All data in its statement are fixed before the residual characteristic tends to infinity. ### Spreading the fixed witnesses We prove the theorem in the next five steps. Put Λ=prod_{i=1}^{n-1}l_i. The flag satisfies [C]=Λ H^{n-1}, h_i· C=l_iΛ H^n. Choose one sufficiently ample integral Cartier divisor T_0 so that every T_0+aS, 0le ale(k_0-1)q, has a section nonvanishing at each of the two base coordinates of z_*. Fix these finitely many sections. All models, morphisms, divisors, points, the flag, the complete-intersection witness for gamma, and the jet and filler sections spread over an integral finitely generated Z-algebrabfZ-algebra of characteristic zero. Shrink its spectrum so that the fibers and flag are smooth projective and geometrically integral, the birational morphisms remain birational, the fixed ample bundles remain ample, and the finite jet surjection and nonvanishings persist. Birationality is preserved by spreading an isomorphism between dense opens. All displayed intersection numbers are constant. We also preserve mu_{min}(Ω_W^1|_C)ge0. One way is to spread its characteristic-zero Harder--Narasimhan filtration, make all its factors locally free, and use openness of their semistability on curves; their degrees and ranks are fixed. Equivalently, a single relative ample twist globally generates the curve bundle. Every quotient of rank a then has degree bounded below by -ad for a fixed integer d. A negative-degree quotient has one of finitely many ranks and degrees. The proper relative Quot schemes for those Hilbert polynomials have images missing the generic point: after removal of torsion, a negative-degree coherent quotient there would give a negative-degree vector-bundle quotient. Removing those finitely many images proves the same openness assertion directly. This base has closed points in arbitrarily large characteristics. Take algebraic closures of their residue fields and exclude denominator primes. Retain the notation for the reductions. The curve gamma continues to pair nonnegatively with every effective divisor: pull that divisor back under the fixed birational morphism and intersect with its fixed ample divisors. This tests effective divisors on the reduction itself. No specialization of a characteristic-zero effective cone is used. For a sufficiently large residual characteristic p, set [ N_p=floor( p/k_0), B_p=k_0N_pL+T_0. ] The sections of pqxi+B_{p,1}+B_{p,2} generate jets through order (2n+2)k_0N_p at z_*. In particular, they surject onto the quotient of its local ring by the p-th powers of all regular parameters. Proof. Multiply N_p copies of the fixed jet system. Every monomial of total degree at most (2n+2)k_0N_p is a product of N_p monomials of degree at most (2n+2)k_0. Taking sections with those leading monomials gives a triangular system, ordered by total degree. Starting at degree zero and removing the error in each successive degree proves the required jet surjection. This argument takes place in the local ring and never divides by factorials. Put d_p=(p-k_0N_p)q, so 0le d_ple(k_0-1)q. Among the sections fixed before reduction, choose [ u_0in H^0(W,O_W(T_0)), v_{d_p}in H^0(W,O_W(T_0+d_pS)) ] nonzero at the first and second base coordinates of z_*, respectively. Their tensor product belongs to the summand of first-slot weight zero in the projective-bundle formula; its fiber monomial is the d_p-th power of the second coordinate. It defines a section of d_pxi+T_{0,1}+T_{0,2} nonzero at z_*, since z_* lies off both axes. Multiplying the product jet system by this section changes its divisor from N_pk_0M to pqxi+B_{p,1}+B_{p,2}. Multiplication is invertible on the local jet algebra. Only the finitely many previously fixed values of d_p occur. Since dim Z=2n+1, the quotient by the p-th powers of regular parameters has top total degree (2n+1)(p-1). For large p this is at most (2n+2)k_0N_p. The underlying ordinary/Frobenius ideal comparison is also recorded in cite[proof of Proposition~2.12, Equation~(2.8)]{MustataSchwede2014}. ### Full rank off the diagonal Let F:Wto W' be relative Frobenius to the base-field twist. It is finite flat because W is smooth. Write S' for the twist of S, so F^*S'=pS, and define [ E=F_*E=F_*O_W(B_p), s=rkE=p^n, R=s^2, U_a=H^0(W,O_W(B_p+paS)). ] Projection formula gives an evaluation map on W'times W': ⊕_{substack{a+b=qa,bge0}} U_a ⊗ U_b ⊗ O(-aS'_1-bS'_2) → E E.E. The map in Equation~(common:two-slot-evaluation) has generic rank R. Proof. Trivialize the two summands defining Z near the selected base points. Let z be the torus coordinate, whose value at z_* is z_0ne0. By the projective-bundle formula, the sections in Lemma~common:positive-characteristic-jets are weight polynomials in z with weight-j coefficients in [ H^0(W,B_p+jS) ⊗ H^0(W,B_p+(pq-j)S), 0le jle pq. ] Let A_* be the tensor product of the local algebras of the two base Frobenius fibers. The local quotient in Lemma~common:positive-characteristic-jets is [ A_*[z]/(z^p-z_0^p), ] free over A_* with basis 1,z,...,z^{p-1}. Project to the coefficient of 1 in this basis. Exactly the weights j=pa survive, with multipliers z_0^{pa}. Their coefficients are the values of the corresponding summands of Equation~(common:two-slot-evaluation), in the chosen frames and the induced twisted frames. As the full jet map is surjective, these coefficients span A_*, of dimension p^{2n}=R. Thus the evaluation has full rank at this pair of base points, and hence generically. The calculation uses a basis over the possibly nonreduced algebra A_*, so reducedness of the Frobenius fibers is unnecessary. The jet calculation at z_* proves generic full rank; the diagonal estimate below will be tested at (x',x'), obtained from the separately fixed point x. ### The diagonal filtration The Frobenius fiber product has the cartesian square Delta_F=Wtimes_{W'}W arrow[r,"pr_2"] arrow[d,"pr_1"'] & W arrow[d,"F"] W arrow[r,"F"'] & W'. Its reduced subscheme is the diagonal W. Put [ L_p= O_{Delta_F}(B_{p,1}+B_{p,2}+pqS_1). ] Write h=Fcircpr_1=Fcircpr_2: Delta_Fto W'. On this fiber product, [ O(pS_1)~= h^*O_{W'}(S')~=O(pS_2). ] Consequently every pair of weights a+b=q gives the same line bundle: [ O_{Delta_F}(B_{p,1}+B_{p,2}+paS_1+pbS_2) ~=L_p. ] On the diagonal of W'times W', every source twist of Equation~(common:two-slot-evaluation) becomes -qS'. Finite base change and projection formula identify the twisted target as [ (E ⊗ E)(qS')E)(qS')~= h_*L_p. ] The product sections defining the columns restrict to global sections of this one bundle L_p. Their values at a diagonal point therefore lie in the image of H^0(Delta_F,L_p) in the corresponding fiber of h_*L_p. This proves that the rank at every diagonal point is at most h^0(Delta_F,L_p). For a rank-n bundle V in characteristic p, denote by A_j(V) the degree-j part of its symmetric algebra modulo the p-th powers of local generators. This construction is independent of frame: the p-th power of a linear combination is the sum of the p-th powers in characteristic p. Each A_j(V) is locally free and a quotient of V^{ ⊗ j}. Set [ u=n(p-1), e_j=rk A_j(V), sum_{j=0}^u e_j=p^n. ] Filtering L_p by powers of the ideal of the reduced diagonal in Delta_F gives the graded bundles G_j=+pqS) ⊗ A_j(Ω_W^1), 0le jle u. Indeed, in smooth local coordinates the algebra is generated by parameter differences with their p-th powers zero, and its degree-one conormal is Ω_W^1. This coordinate calculation can be made étale locally or in completions and identifies the global associated graded. This is the diagonal-ideal form of the canonical Frobenius filtration cite[Corollary~3.5]{KitadaiSumihiro2008}. The same calculation, using only the bundle from the second factor, filters F^*E with pieces cite[Theorem~3.7]{Sun2008} O_W(B_p) ⊗ A_j(Ω_W^1). Complementary multiplication is a perfect pairing, giving A_{u-j}(V)~= A_j(V)^vee ⊗ (det V)^{p-1}. On monomials, each exponent vector pairs with its complement to (p-1,...,p-1). The top line transforms by the character (det V)^{p-1}: this is immediate on diagonal matrices and hence identifies the character of GL_n on that line. ### One slope bound and one flag polynomial We bound the sections of every graded bundle in Equation~(common:diagonal-graded) by its rank times the same scalar polynomial in p. This uniformity lets us sum the ranks of the graded pieces without an additional factor for their number. We begin with a slope bound on the fixed curve; here slope means ordinary degree divided by rank. There is a constant cge0, independent of the sufficiently large residual characteristic p and of j, such that for V=Ω_W^1|_C one has [ mu_{max}(A_j(V))le(p-1)K_W· C+nc. ] Proof. Write F_C for absolute Frobenius of C. Langer's instability estimate cite[Corollary~2.5]{Langer2004} gives L_{min}(V):= lim_{eto∞}p^{-e}mu_{min}(F_C^{e*}V) ge-(c)/(p-1). To check the uniform constant, the discrepancy from mu_{min}(V)ge0 is at most (rk V-1)deg A_C/(p-1), where A_C is a nef line bundle such that T_C ⊗ A_C is globally generated. Its degree can be bounded in terms of the fixed genus, using a sufficiently positive line bundle. Here the rank is n, irrespective of the fact that the base is a curve. We also need, for every integer ige0, mu_{min}(V^{ ⊗ i})ge iL_{min}(V). Fix p and i. Choose a line bundle A_e of degree -mu_{min}(F_C^{e*}V)+O(1), rounded up with a fixed genus-dependent additive constant, so that F_C^{e*}V ⊗ A_e is globally generated. This follows from Serre duality: after subtracting any point, its minimum slope can be made greater than 2g(C)-2, which annihilates H^1 and makes evaluation at that point surjective. For every vector-bundle quotient Q of V^{ ⊗ i}, the bundle F_C^{e*}Q ⊗ A_e^{ ⊗ i} is then globally generated. Consequently [ p^emu(Q)+ideg A_ege0. ] Divide by p^e and let eto∞ to prove Equation~(common:tensor-min). This limit is taken for each fixed p,i; there is no interchange of Frobenius-iteration and characteristic limits. Since A_{u-j}(V) is a quotient of V^{ ⊗ (u-j)}, the perfect pairing Equation~(common:socle-pairing) yields mu_{max}(A_j(V)) &= (p-1)degdet V-mu_{min}(A_{u-j}(V)) &le(p-1)K_W· C+((u-j)c)/(p-1) &le(p-1)K_W· C+nc. The divisors B_p-pL=T_0-(p-k_0N_p)L range over a fixed finite list. Combining Equation~(common:intersection-bounds), Equation~(common:flag-degrees), Lemma~common:truncated-slope therefore gives, uniformly for 0le jle u, mu_{max}(G_j|_C)G_j|_C) &le p(4+2/r)Λ H^n+O(1)notag &le M_p:=7pΛ H^n+c_0, where c_0ge0 is fixed. The coefficient 7 provides harmless room in this common bound. For all sufficiently large p, [ h^0(Delta_F,L_p) le R(7^nH^n+O(1/p))M_p, because then its maximum slope is negative. Each nonzero summand has at most e_j(M_p+1) sections. Indeed, evaluation at floor( M_p)+1 distinct points is injective; the kernel has negative maximum slope after subtracting those points. By Equation~(common:flag-degrees) it follows that h^0(W,G_j) &le e_j(M_p+1) prod_{i=1}^{n-1} (1+(M_p)/(l_iΛ H^n))notag &=e_jp^n(7^nH^n+O(1/p)). The leading coefficient follows from Λ=prod_i l_i. More importantly, the whole scalar polynomial on the first line is the same for every j. Sum over the filtration in Equation~(common:diagonal-graded). The ranks satisfy sum_j e_j=p^n, so Equation~(common:common-flag-polynomial) gives the asserted bound with R=p^{2n}. There is no extra factor for the number of graded pieces. Finally [ 7^nH^nle(14ε)^n<1/4 ] by Equation~(common:intersection-bounds), proving the strict ineuality for all large p. ### The determinant has incompatible orders Choose R generically independent columns of Equation~(common:two-slot-evaluation) from bases of its source vector spaces. Their determinant is a nonzero section [ 0nesigmain H^0(W'times W',Q_1 Q_2)Q_2) ] of an actual external-product line bundle. To specify its factors, let m_i be the sum of the weights of the R chosen columns in slot i. The target determinant is (detE)^s, and the chosen source determinant is O(-m_1S')boxtimesO(-m_2S'). Thus [ Q_i=( E)^E)^{ ⊗ s} ⊗ O_{W'}(m_iS'), lambda_i=m_i/R. ] Every column weight lies between 0 and q, so 0lelambda_ile q independently of the columns chosen. Under the numerical identification with the base-field twist, frac{c_1(Q_i)} =frac{c_1(E)}+lambda_iS =frac{B_p}p+(p-1)/(2p)K_W+lambda_iS, 0lelambda_ile q. The Frobenius first-Chern-class identity is also cite[Lemma~4.2]{Sun2008}; it is an identity of rational divisor classes. For the second equality, use Equation~(common:Frob-pullback-graded). Complementary degrees in Equation~(common:socle-pairing) have total first Chern class e_j(p-1)K_W. Summing gives [ sum_jc_1(A_j(Ω_W^1))=frac{s(p-1)}2K_W. ] Thus c_1(F^*E)=sB_p+s(p-1)K_W/2,E)=sB_p+s(p-1)K_W/2, and Frobenius pullback of twisted divisor classes multiplies by p, as required. For 0lelambdale q, the fixed movable ineuality gives [ (L+K_W/2+lambda S)·gammale40ε J·gamma, ] where pullback by pi_x is understood. This concerns only intersections of fixed divisors and the fixed complete-intersection witness, so it remains true on the reduction. The class in Equation~(common:slot-determinant-class) differs from its left-hand divisor at lambda=lambda_i by [ frac1p(B_p-pL-K_W/2). ] Its numerator belongs to the fixed finite list T_0-aL-K_W/2, 0le a0, it follows that frac{ord_{x'}(tau_i)}Rle40ε+O(1/p). The constant is uniform in the columns and in tau_i. This argument tests sections on the reduction itself; it does not require them to lift to characteristic zero. K"unneth identifies the space containing sigma with H^0(W',Q_1) ⊗ H^0(W',Q_2). Choose bases adapted to the orders of vanishing at x' in each factor. In each degree their leading terms are linearly independent. Tensor products of these leading terms remain independent in each bidegree of the two disjoint sets of parameters; terms of different bidegrees cannot cancel. Every nonzero tensor consequently has order at most the sum of the two maximum basis orders. Applying Equation~(common:slot-order), we obtain ord_{(x',x')}sigmale R(80ε+O(1/p)). On the other hand, Equation~(common:diagonal-rank-upper), Lemma~common:diagonal-sections bound the matrix rank at (x',x') by a number strictly smaller than R/4. Trivialize its line bundles near that point. Invertible constant row operations make more than 3R/4 rows vanish at the point. Each entry of those rows belongs to the maximal ideal, so every term of the determinant has order at least 3R/4. Therefore ord_{(x',x')}sigmage3R/4. Equations (common:determinant-order-upper) and (common:determinant-order-lower) contradict 80ε<3/4 for sufficiently large p. This contradiction proves Theorem~common:finite-data-frobenius. The full-rank calculation used z_*, while the diagonal estimate holds at every point and was tested at the separately chosen x. The errors were controlled by one fixed finite list of divisor classes and one fixed flag polynomial. Thus their constants do not depend on p, the filtration index, or the chosen columns. ## The geometric witnesses and smooth nonvanishing We return to the notation and the hypothetical counterexample of Section~sec:jets. Theorem~common:finite-data-frobenius asks for fixed divisors, a small polarization and flag, finite two-slot jets, and one actual strongly movable curve. We construct all these inputs in characteristic zero. The positive parts on the scaling models, rather than the original non-nef classes, provide the small polarization. No minimal-model statement or pseudo-effective cone is specialized to positive characteristic. [The smooth nonvanishing step] Assume the lower-dimensional real-boundary good-model hypothesis of Assumption~mmp:lower-models. If (X) is a smooth projective complex variety of dimension (n) and (K_X) is pseudo-effective, then (kappa(X,K_X)ge0). In dimension zero the assertion is immediate. On a smooth curve, pseudo-effectivity of (K_X) gives (g(X)ge1), hence (h^0(X,K_X)=g(X)>0). We therefore suppose (nge2) and argue by contradiction. We use the late terminal model (Y), the divisors (K=K_Y) and (A=A_Y), and the scaling models of the preceding section. Fix (m>0) with (mK) Cartier, and write [ mu_t=vol(K_X+tA_X)^{1/n}, L=r(K+tA). ] As before, (r) is a positive integer chosen so that (rmu_t→ε) as (t↓0). Choose the rational number (ε>0) sufficiently small for Propositions~prop:scalar-jets and~prop:large-seshadri, and also so that (14ε)^nqm+1). Once those estimates hold, (t) and (r) remain fixed through all the subsequent characteristic-zero constructions and reductions. ### A small polarization and a cotangent flag A small ample polarization and a complete-intersection flag will control the number of sections on the Frobenius thickening of the diagonal. We construct them from the positive part on a scaling model. Choose a smooth common resolution (W) of (Y) and the scaling model for (K+tA). In this application (K,A,L) also denote their pullbacks to (W), and [ K_W=K+E, Ege0. ] Let (H_0) be the pullback of the nef semiample positive part (N_0) of (L) on its scaling model. The comparison of canonical pullbacks and moving parts gives H_0^n→ε^n, LH_0^{n-1}=H_0^n, K_WH_0^{n-1}=KH_0^{n-1}<=(H_0^n)/(r). To justify these intersections, recall that the maps (Ydashrightarrow Y_t) are small by the choice of the late model. Every divisor exceptional over (Y) is therefore also exceptional over (Y_t). The differences discarded when passing to the moving parts are exceptional over the scaling model, and the transform of (A) is effective up to rational linear equivalence. Intersecting with (H_0^{n-1}) gives the displayed equalities and ineuality. In particular, since (qm/r<1), [ (2L+qmK)H_0^{n-1} le(2+qm/r)H_0^n<3H_0^n. ] Choose (t) small enough that also (H_0^n<(2ε)^n), and now fix (t,r) and the resolution (W). For a fixed ample divisor (H_{amp}), put (H=H_0+eta H_{amp}), where (eta>0) is rational. The strict margins just obtained, together with (K_WH_0^{n-1}le H_0^n/r), allow us to choose (eta) small enough that continuity gives H^nle(2ε)^n, K_WH^{n-1}<=(2H^n)/(r), (2L+qmK)H^{n-1}le4H^n. We next choose the flag for the section estimates. The canonical divisor of (W) is pseudo-effective. Generic semipositivity, applied with empty boundary cite[Theorem~2.1]{CampanaPaun2015}, and restriction to a sufficiently general complete-intersection curve give the following fixed data. Apply restriction to the Harder--Narasimhan filtration and its factors cite[Theorem~6.1 and Remark~6.2]{MehtaRamanathan1982}. Choose successively large divisible integers l_i such that h_i=l_iH are integral very ample classes, and choose a smooth integral complete-intersection flag [ W=W_nsupset W_{n-1}supset...supset W_1=C, W_{i-1}in |h_{n-i+1}|_{W_i}, ] for which mu_{min}(Ω_W^1|_C)ge0. Here and below slopes on (C) mean degree divided by rank. One may obtain the flag by applying restriction to the finitely many Harder--Narasimhan quotients of (Ω_W^1). ### A movable witness for all determinant weights For the vanishing-order estimate, choose a very general point (xin W), away from the exceptional loci, and let (pi_x:WWtohat W) be its blowup, with exceptional divisor (J). Set [ D^+=2r(K+2tA)+2E. ] This big divisor has the scalar sections controlled by Proposition~prop:scalar-jets: pushing to (Y) identifies the section spaces, since the added exceptional part does not change them. If (rho=vol(2r(K+2tA))^{1/n}), then [ rhole4rmu_tle8ε ] for our sufficiently small (t). Thus every section of a sufficiently divisible multiple of (D^+) has normalized order at (x) at most (4rhole32ε). We convert this order bound into a curve ineuality. It follows that (pi_x^*D^+-40ε J) is not pseudo-effective. Otherwise its convex combination with the big class (pi_x^*D^+) would make (pi_x^*D^+-cJ) big for some rational (32ε0. The strictness lets us fix an algebraic family that will survive reduction. More precisely, the negative pairing can first be detected by a strongly movable curve: take the pushforward of a general complete intersection of very ample divisors on one fixed smooth birational model of (WWhat). Fix this model, its divisors, and the resulting covering family. Thus the choice consists of finite algebraic data, rather than a limiting real movable class. Positivity of (J·gamma) follows from the strict ineuality and pseudo-effectivity of (pi_x^*D^+). To turn this curve into the witness required by Theorem~common:finite-data-frobenius, set S=mK on W; it is an integral Cartier divisor. We verify all its weight ineualities before proceeding to the jet construction. Put Q_lambda=L+K_W/2+lambda S for 0lelambdale q. The equality K_W=K+E gives D^+-Q_lambda =(r-1/2-lambda m)K+3rtA+3/2E. This class is pseudo-effective in characteristic zero: r>qm+1, K is pseudo-effective, and A,E are effective up to the stated equivalences. Pairing with the fixed strongly movable class therefore gives [ pi_x^*Q_lambda·gamma le pi_x^*D^+·gamma <40ε J·gamma, J·gamma>0. ] These are the required movable ineualities. Only the two endpoint ineualities need to be retained as finite data, since the pairing is affine in lambda. The common theorem spreads these fixed numerical ineualities and the actual birational complete-intersection witness; it does not spread the pseudo-effective assertion used to obtain them. ### A finite two-slot jet system We must similarly realize the two-slot jet estimate by finitely many sections before reducing. Their products will then supply the jet orders needed for arbitrarily large characteristics. On (Wtimes W), let [ Z=PP(mK_1 ⊕ mK_2), xi=OO_Z(1), M=L_1+L_2+qxi, ] with the symmetric-power convention for the projective bundle. Proposition~prop:large-seshadri gives a semiample big model whose Seshadri constant exceeds (2n+2) at a very general torus point. On the open set where its birational contraction is an isomorphism, the jet interpretation of the Seshadri constant consequently gives an integer (k_0>0) such that (|k_0M|) generates jets of order at least ((2n+2)k_0) at a fixed smooth torus point (z_*in Z). Indeed, on the ample model choose a rational number (c) strictly between (2n+2) and its Seshadri constant. On the blowup of the selected smooth point, the pullback of the ample class minus (c) times the exceptional divisor is ample. Serre vanishing and the exceptional-divisor sequence then give jets of order (kc-1) for all sufficiently divisible large (k). Pulling sections back on the isomorphic open gives the asserted bound. Enlarge (k_0) to clear every denominator, in particular that of (L), and fix finitely many sections realizing this jet surjection. ### Applying the comparison All the preceding choices are now fixed. With S=mK, the dimension, intersection and flag hypotheses of Theorem~common:finite-data-frobenius are Equations~(fr:epsilon), (fr:intersection-bounds), and~(fr:generic-nef). The curve construction gives its movable witness and both endpoint ineualities, and the finite jet system gives the required surjection at z_*. This point need not lie over the separately chosen testing point x. The proof of the common theorem also fixes its auxiliary divisor T_0 and finitely many filler sections before choosing any residual characteristic. The common theorem now rules out these fixed data. Its contradiction compares two orders of the same nonzero determinant. After reduction, the Frobenius direct image (E used there has rank (p^n), so (E EcalE) has rank (R=p^{2n}). The fixed jets give full evaluation rank (R), while the small polarization bounds the rank on the diagonal by (R/4). Thus a nonzero maximal minor vanishes to order at least (3R/4) at ((x',x')), where (x') is the Frobenius image of the testing point (x). Its line bundle is an actual external product. Applying the fixed curve ineuality to each factor bounds the same order above by (R(80ε+O(p^{-1}))), a contradiction for large (p). The theorem proves these determinant estimates for the fixed witnesses we have now supplied; it requires no additional geometric input here. All constants are fixed before the residual prime varies. The contradiction excludes the assumed counterexample and proves Theorem~thm:smooth-nonvanishing. # Completion and change of ground field Smooth nonvanishing is now available in the order required by the induction. We first finish the complex proof, including descent from the good model to the original normal variety. We then descend the globally generated Cartier multiple to an arbitrary algebraically closed field of characteristic zero. Proof. [Proof of Theorem~thm:main over (C)] Use dimension induction in the real-boundary category of Proposition~prop:inductive-reduction. In dimension zero the variety is a point. Suppose the good-model assertion holds in all smaller dimensions. The signed-representative argument (Theorem~thm:signed), the exclusions, and the jet estimates are then available with exactly that lower-dimensional hypothesis. Theorem~thm:smooth-nonvanishing establishes smooth nonvanishing in the present dimension. Proposition~prop:inductive-reduction therefore proves the real-boundary good-model assertion in the present dimension. This is precisely the hypothesis needed to continue to the next dimension. Apply this to the rational lc pair ((X,B)) of Theorem~thm:main. Let (f:(X_d,B_d)to(X,B)) be a crepant dlt model and let ((V,B_V)) be its good log minimal model. Take a common resolution (a:Wto X_d), (q:Wto V), and put (p=fcirc a). Then [ p^*(K_X+B)=a^*(K_{X_d}+B_d)=q^*(K_V+B_V). ] The first equality is crepancy. The second is the nef comparison in Lemma~mmp:comparison, since (K_{X_d}+B_d=f^*(K_X+B)) is nef. The right side is semiample because ((V,B_V)) is a good model, so the pullback of (K_X+B) is semiample. Normality gives (p_*OO_W=OO_X). Choose a sufficiently divisible Cartier multiple (m(K_X+B)) whose pullback is globally generated. The projection formula identifies its sections with those of its pullback. Surjectivity of evaluation upstairs implies surjectivity downstairs: otherwise a base point downstairs would make every pulled-back section vanish on its nonempty fiber. Hence (K_X+B) is semiample. [Change of algebraically closed field] The conclusion of Theorem~thm:main over (C) implies its conclusion over every algebraically closed field of characteristic zero. Proof. Let ((X,B)) be defined over such a field (k). Choose a finitely generated subfield (k_1subset k) over which the projective variety, the rational boundary, a Cartier multiple of its adjoint, and a log resolution are all defined. After enlarging (k_1) if necessary, these data base change to the given data. Let (k_0) be its algebraic closure inside (k), and denote the resulting pair by ((X_0,B_0)). Choose an embedding (k_0↪C), and write ((X_{C},B_{C})) for the complex base change. The chosen log resolution tests log canonicity after either algebraically closed field extension: its boundary remains simple normal crossing and all discrepancy coefficients remain unchanged. The descended pair is consequently lc over both (k_0) and (C). Nefness is also invariant under such extensions. Indeed, a curve over an extension is represented by a point of a relative Hilbert scheme after its finite defining data have been spread over a finite-type parameter scheme. The degree of a fixed line bundle is constant in the resulting flat family. Specializing to a closed point over the algebraically closed smaller field preserves a negative degree, if one existed; some irreducible component of the specialized curve would then have negative degree. This contradicts nefness over that field. Conversely, curves over the smaller field remain available after extension. Apply the converse to (k_0subset k): nefness of (K_X+B) gives nefness of (K_{X_0}+B_0). Apply the forward direction to (k_0↪C): its complex base change is nef as well. The complex case supplies an integer (m>0), enlarged to a multiple of the Cartier index fixed over (k_0), for which the Cartier line bundle (M=OO_{X_0}(m(K_{X_0}+B_0))) becomes globally generated over (C). Proper flat base change for sections identifies [ H^0(X_0,M) ⊗ _{k_0}C = H^0(X_{C},M_{C}). ] Tensoring the evaluation map for (M) with (C) therefore gives the surjective complex evaluation map. Its coherent cokernel is zero by faithful flatness. Extending from (k_0) to (k) preserves this surjectivity, so the same integer (m) gives a globally generated line bundle on the original (X). Together with Proposition~prop:field-descent, the complex proof establishes Theorem~thm:main in its stated generality. # Good models, linear triviality, and boundary consequences The real-boundary induction and the final line-bundle descent have different natural scopes. We first record good-model existence over (C), including the precise smooth canonical comparison used by fiber-space applications. We then deduce actual rational-linear triviality in Iitaka dimension zero over every field covered by Theorem~thm:main. Finally, over (C), normalization gluing and a section-extension theorem give two consequences for reducible boundaries and nonnormal pairs. [Real-boundary good models over (C)] Every projective log canonical pair ((X,B)) over (C), with effective real boundary and pseudo-effective real Cartier adjoint (K_X+B), has a good log minimal model. Real semiampleness has the meaning fixed in Section~sec:induction. Proof. The simultaneous induction in Section~sec:completion starts with dimension zero and proves the full real-boundary conclusion of Proposition~prop:inductive-reduction at each dimension. In particular, its conclusion applies to the given pair before any rational-boundary or nefness specialization. [Smooth canonical good models] Let (X) be a smooth projective complex variety with pseudo-effective (K_X). There is a normal projective (Q)-factorial klt variety (V) and a (K_X)-negative birational contraction (Xdashrightarrow V) such that (K_V) is (Q)-Cartier and semiample. On a common smooth resolution (p:Wto X), (q:Wto V), compatible canonical divisors satisfy [ p^*K_X=q^*K_V+E, Ege0quadand E is q-exceptional. ] In particular (kappa(X,K_X)=kappa(V,K_V)). Proof. Apply Theorem~thm:good-models-complex to the klt pair ((X,0)). Choose its (Q)-factorial log minimal model, as in the construction of Proposition~prop:inductive-reduction. The klt clause of Lemma~mmp:comparison shows that it extracts no divisor, remains klt, and has the displayed effective exceptional comparison. Its boundary is zero, being the strict transform of the zero boundary with no extracted divisors. Discrepancies strictly improve at contracted divisors, which is the (K_X)-negative condition; semiampleness is the good-model conclusion. For sufficiently divisible (m>0), the effective (q)-exceptional divisor (mE) satisfies (q_*OO_W(mE)=OO_V), since (V) is normal. Projection formula and the displayed comparison identify the pluricanonical section spaces. Their ratios agree in the common function field, proving the assertion about Iitaka dimension. Let (Z) be a nonempty integral projective variety over an algebraically closed field, and let (A) be a semiample (Q)-Cartier divisor. If (kappa(Z,A)=0), then (Asim_{Q}0). Proof. Choose a positive integer (m) such that the line bundle (L=OO_Z(mA)) is globally generated. Its complete system gives [ phi:Z→PP(H^0(Z,L)^*), L~=phi^*OO(1). ] The image has dimension at most (kappa(Z,A)=0), so it is a point. A linear form nonzero at that point pulls back to a nowhere-vanishing section of (L). Thus (L~=OO_Z), an actual line-bundle isomorphism, and (Asim_{Q}0). Let ((X,B)) satisfy the hypotheses of Theorem~thm:main. If (kappa(X,K_X+B)=0), then (K_X+Bsim_{Q}0). In particular, every nef rational adjoint of Iitaka dimension zero on a projective klt fivefold over an algebraically closed field of characteristic zero is (Q)-linearly trivial. Proof. Theorem~thm:main gives semiampleness on the original normal variety. Lemma~lem:semiample-zero gives a positive Cartier multiple with trivial associated line bundle. Klt pairs are lc, so the fivefold assertion is a special case. Thus the nef adjoint also has numerical dimension zero. This is a consequence of the trivializing Cartier multiple, not an additional hypothesis. The multiple may depend on the pair; no uniform index is asserted. ## Semi-log-canonical abundance and dlt extension We use the standard semi-log-canonical convention, in which the underlying projective scheme may be reducible or nonnormal. It is reduced, satisfies (S_2), is pure-dimensional, and has ordinary double crossings in codimension one. No component of the effective boundary is contained in the codimension-one singular locus. If (nu:X^nuto X) is the normalization, the boundary (Θ) defined by [ K_{X^nu}+Θ=nu^*(K_X+Delta) ] includes the conductor boundary, and ((X^nu,Θ)) is componentwise log canonical. These are the conventions of cite[Definition~2.5]{FujinoGongyo2014}. [Semi-log-canonical abundance over (C)] Let ((X,Delta)) be a projective semi-log-canonical pair over (C), where (Delta) is an effective rational divisor and (K_X+Delta) is (Q)-Cartier. If (K_X+Delta) is nef, then it is semiample. Proof. Write (X^nu=coprod_i X_i^nu) for the normalization and put (Θ_i=Θ|_{X_i^nu}). Each ((X_i^nu,Θ_i)) is a normal projective log canonical pair with effective rational boundary. The adjoint (K_{X_i^nu}+Θ_i) is the pullback of (K_X+Delta), so it is nef. Theorem~thm:main makes each of these adjoints semiample. There are only finitely many components, hence one common positive Cartier multiple is globally generated on their disjoint union. Thus (nu^*(K_X+Delta)) is semiample. The normalization gluing theorem cite[Theorem~1.5]{FujinoGongyo2014} now gives semiampleness on (X). [Supported extension for dlt pairs over (C)] Let ((X,Delta)) be a projective dlt pair over (C), where (Delta) is an effective rational divisor, and put (S=floor(Delta)). Assume that (K_X+Delta) is nef and that, for an effective rational divisor (D), [ K_X+Deltasim_{Q}Dge0, S⊆Supp D. ] Then, for every integer (mge2) for which (m(K_X+Delta)) is Cartier, the restriction map [ H^0(X,OO_X(m(K_X+Delta))) → H^0(S,OO_X(m(K_X+Delta))|_S) ] is surjective. Proof. Set (B=Delta-S). The pair ((X,S+B)) is dlt, (B) is an effective rational divisor, and (floor( S+B)=S). Since a dlt pair is log canonical, Theorem~thm:main makes its nef adjoint semiample. The pair and the displayed effective representative therefore satisfy the hypotheses of cite[Proposition~5.12]{FujinoGongyo2014}, which gives the stated restriction surjectivity for every Cartier multiple with (mge2). This is the supported extension formulation in cite[Conjecture~5.8]{FujinoGongyo2014}. The first corollary settles the rational-boundary form of the semi-log-canonical abundance conjecture cite[Conjecture~1.3]{FujinoGongyo2014} in every dimension over (C). The second settles the supported dlt extension conjecture cite[Conjectures~1.10 and~5.8]{FujinoGongyo2014} with the precise Cartier-multiple range supplied by Proposition~5.12. Both are applications of the cited implications after normal abundance has been established here. The support condition (S⊆Supp D) is retained; no arbitrary restriction-surjectivity statement is asserted. # Finite generation of rational log canonical rings The preceding results also settle finite generation for rational lc pairs. The finite-generation conjecture asks whether finitely many rounded pluricanonical sections generate the log canonical ring cite[Conjecture~A]{FujinoGongyoRings2014}. The assertion below concerns the full rounded ring, not merely a sufficiently divisible subring, and requires no nefness assumption. For the projective absolute statement, we combine the ordinary minimal-model theorem with log abundance and keep track of every rounded degree. For the proper relative statement over (C), we apply Fujino--Gongyo's published reduction from good minimal models. In that relative statement the morphism is only required to be proper; the source need not be projective over (C). For an integral Weil divisor (G) on a normal variety (X), (OO_X(G)) denotes its divisorial sheaf. For a rational divisor (D) on (X) and nonnegative integers (m,n), the inclusions (floor( mD)+floor( nD)lefloor((m+n)D)) give the usual multiplication on the section algebras below. [Finite generation for rational lc pairs] leavevmode [label=(roman*)] Let (k) be an algebraically closed field of characteristic zero, let ((X,Delta)) be a normal projective log canonical pair over (k), and assume that (Deltage0) is rational and (D=K_X+Delta) is (Q)-Cartier. Then [ R(X,D)=⊕_{mge0} H^0(X,OO_X(floor( mD))) ] is a finitely generated (k)-algebra. Let ((X,Delta)) be a log canonical pair over (C), with (Deltage0) rational and (K_X+Delta) (Q)-Cartier. For every proper morphism (f:Xto S) onto an algebraic variety (S), the relative log canonical algebra [ R(X/S,K_X+R(X/S,K_X+Delta) =⊕_{mge0}f_*OO_X(floor( m(K_X+Delta))) ] is a finitely generated (OO_S)-algebra. Proof. In dimension zero, (X) is a point, and in (ii) so is (S). The boundary and canonical divisor are zero, and the two algebras are (k[t]) and (OO_S[t]), respectively, with (deg t=1). We therefore assume (dim Xge1). We first prove (i). If (D) is not pseudo-effective, a nonzero section (s) in degree (m>0) would give [ div(s)+mD =(div(s)+floor( mD)) +(mD-floor( mD))ge0. ] This would make (D) pseudo-effective. Thus all positive graded pieces vanish in this case. Since (H^0(X,OO_X)=k), we have (R(X,D)=k). Suppose now that (D) is pseudo-effective. The ordinary-pair conclusion of the companion minimal-model theorem cite[Corollary~1.2]{OpenAIMinimalModels2026} gives a normal projective lc pair ((Y,Delta_Y)) and a birational map (phi:Xdashrightarrow Y) extracting no divisors, with (Delta_Y=phi_*Delta) and (D_Y=K_Y+Delta_Y) (Q)-Cartier and nef. Log discrepancies do not decrease. Theorem~thm:main makes the rational adjoint (D_Y) semiample. On a common smooth resolution (p:Wto X), (q:Wto Y), choose compatible canonical divisors. Then p^*D=q^*D_Y+E, Ege0, E is q-exceptional. Indeed, the coefficient of (E) at a prime divisor (F) is (a(F;Y,Delta_Y)-a(F;X,Delta)), which is nonnegative by the discrepancy comparison. Suppose (F) is not (q)-exceptional. Since (phi) extracts no divisors, (F) is the strict transform of a divisor on (X). The pushed-forward boundary gives equality of these discrepancies. This proves the asserted support. For any normal variety (Z), rational Cartier divisor (G), and (mge0), the nonzero global sections of (OO_Z(floor( mG))) are exactly the rational functions (s) satisfying [ div_Z(s)+mGge0. ] This is equivalent to the usual ineuality with (floor( mG)) because the coefficients of (div_Z(s)) are integers. If the ineuality holds on (X), pull back its effective (Q)-Cartier left side to (W) and push forward by (q). Equation~(eq:ring-comparison) and (q_*E=0) give the ineuality on (Y). Conversely, pull back the ineuality on (Y), add (mE), and push forward by (p). Thus, inside the common function field, [ H^0(X,OO_X(floor( mD))) =H^0(Y,OO_Y(floor( mD_Y))) (mge0). ] These identifications preserve multiplication, so they identify the full graded rings. It remains to use semiampleness on (Y). Choose (r>0) such that (rD_Y) is Cartier and globally generated. Its complete linear system defines (psi:YtoPP^N_k) with (OO_Y(rD_Y)~=psi^*OO_{PP^N}(1)). Let (A=k[T_0,...,T_N]) act on (R(Y,D_Y)) by pulling back the coordinate sections, each in degree (r). For (0le i0. ] Proof. For positive-dimensional (X), the theorem of Boucksom--Demailly--P{u a}un--Peternell cite[Corollary~0.3 and Theorem~2.6]{BDPP2013} says that (X) is non-uniruled exactly when (K_X) is pseudo-effective. Corollary~cor:smooth-good-model then gives (kappa(X)ge0): the semiample canonical divisor on the good model has a nonzero section in a positive multiple, and the pluricanonical section spaces agree. Conversely, a nonzero pluricanonical section makes (K_X) pseudo-effective. This proves (i). For (ii), rational connectedness forces all positive covariant tensor differentials to vanish, as recalled by Lazi{'c} and Peternell. For the converse, their Proposition~2.4 reduces Mumford's criterion in dimension (n) to tensor nonvanishing for smooth projective varieties of positive dimension at most (n) whose canonical class is pseudo-effective cite[Proposition~2.4]{LazicPeternell2017}. If (Y) is such a variety of dimension (d), the preceding good-model argument supplies a nonzero section of (mK_Y) for some integer (m>0). Antisymmetrization embeds (Ω_Y^d) into ((Ω_Y^1)^{ ⊗ d}) over (C); tensoring this injection (m) times gives the required nonzero tensor differential. This is the observation in cite[Remark~2.5]{LazicPeternell2017}, so their reduction applies in every dimension. Both assertions for a point follow from our conventions. For a smooth connected projective complex variety (X), let (R(X)) be a smooth projective model of the base of its maximal rationally connected (MRC) fibration. This base is determined up to birational equivalence and is non-uniruled unless it is a point. We use (Ω_X^0=OO_X). [Pluriforms and the rational quotient] [label=(roman*)] For every smooth connected projective complex variety (X), [ dim R(X)= max_{substack{pinZ_{ge0},,kinZ_{>0} H^0(X,Sym^k(Ω_X^p))ne0}} p. ] A compact connected K"ahler manifold (X) is rationally connected if and only if [ H^0(X,Sym^k(Ω_X^p))=0 quadfor every pair of integers p,k>0. ] Proof. Write (r=dim R(X)) in (i). Resolve the MRC map to a morphism (f:X'to R(X)), with (X') smooth and projective. For a smooth general fiber (F), the cotangent sequence is [ 0→OO_F^{ ⊕ r} →Ω_{X'}^1|_F →Ω_F^1→0. ] If (p>r), every graded piece of the induced exterior-power filtration on (Ω_{X'}^p|_F) contains a positive exterior power of (Ω_F^1): at most (r) factors can come from (OO_F^{ ⊕ r}). Consequently, for (k>0), the graded pieces of the induced filtration on (Sym^k(Ω_{X'}^p)|_F) are direct summands of sums of positive tensor powers of (Ω_F^1). Here we use antisymmetrization and symmetrization over (C). Rational connectedness of (F) makes all their sections vanish, by Corollary~cor:uniruled-mumford. A global section therefore vanishes on every sufficiently general fiber, hence on (X'). Birational invariance of global covariant tensor differentials on smooth projective varieties gives the same vanishing on (X). This is the upper-bound argument of cite[Section~3]{BrunebarbeCampana2015}. If (r>0), the smooth non-uniruled base has (kappa(R(X))ge0) by Corollary~cor:uniruled-mumford. A nonzero section of (kK_{R(X)}) for some (k>0) pulls back to a nonzero section of (Sym^k(Ω_X^r)). The pullback is taken on (X') and descends to (X) by the same birational invariance. If (r=0), the term (p=0) gives a nonzero section and the same equality follows. This is the equality (r=r^-) predicted in the remark ending cite[Section~3]{BrunebarbeCampana2015}. For (ii), the forward vanishing is the rational-curve argument recalled by Brunebarbe and Campana in the same section. Conversely, the stated vanishing gives (H^{2,0}(X)=H^0(X,Ω_X^2)=0). Kodaira's projectivity criterion then makes (X) projective: the real degree-two cohomology is of type ((1,1)), so a rational class sufficiently close to a K"ahler class is still K"ahler, and clearing denominators gives an integral K"ahler class. This is the projectivity reduction in cite[Section~3]{BrunebarbeCampana2015}. Part (i) now applies. Its maximum is zero because every term with (p>0) vanishes, so the MRC base is a point and (X) is rationally connected. We next use ordinary cotangent invariants, with no boundary. If (L is a coherent rank-one subsheaf of a vector bundle on a smooth projective variety, its double dual is a line bundle; put (kappa(L)= L^L^{**})). Define [ omega(X)= max_{substack{m>0 LΩ_X^1)^{ ⊗ m} rankL=1}} kappa(L), ] where (m) is an integer and the value is (-∞) if no such sheaf has nonnegative Iitaka dimension. The inclusion of (L extends to its double dual: away from codimension two these sheaves agree, and a map into a locally free sheaf extends across that subset. Taking tensor powers then realizes the systems (H^0(X,(L^{**})^{ ⊗ a})), for integers (a>0), inside higher cotangent tensor powers. Thus allowing all (m) makes this equivalent to Campana's definition using the dimensions of the rational maps defined by (H^0(X,L) For positive-dimensional (X), also put [ kappa^+(X)= max_{substack{p>00neFΩ_X^p}} kappa(detF). ] Here (p) is an integer, (F is coherent, and (detF=(∧^F=(∧^{rkF} F)^F)^{**}). These are the invariants used in Campana's Appendix~A to Taji [cite]. [Cotangent invariants and the MRC base] Let (X) be a smooth connected projective complex variety. Then [ omega(X)=omega(R(X))= -∞,&if R(X) is a point, kappa(R(X)),&if dim R(X)>0. ] In particular, if (X) is positive-dimensional and non-uniruled, then (omega(X)=kappa(X)). If (X) is positive-dimensional and (kappa(X)ge0), then [ kappa^+(X)=kappa(X). ] Proof. Campana's birationally invariant fibration properties cite[Appendix~A, property~4]{Taji2014} give (omega(X)=omega(R(X))): the general MRC fiber is rationally connected and has (omega=-∞). Indeed any rank-one subsheaf with nonnegative Iitaka dimension would give a section of a positive tensor power, contrary to Corollary~cor:uniruled-mumford; a point has (omega=-∞) by the empty maximum. Birational invariance permits resolving the MRC map. If (R(X)) is a point, the displayed formula therefore follows without changing our convention (kappa(pt)=0). Suppose that (R=R(X)) has positive dimension. It is non-uniruled, so (kappa(R)ge0) by Corollary~cor:uniruled-mumford. To apply Campana's abundance deduction cite[Conjecture~A.2 and Remark~A.3]{Taji2014}, take a resolved Iitaka fibration (R'to Z), with (R') and (Z) smooth and projective, (R') birational to (R), and (dim Z=kappa(R)). Its smooth general fiber (F) has (kappa(F)=0). If (F) is positive-dimensional, Corollary~cor:smooth-good-model and Lemma~lem:semiample-zero give a normal birational good model (V_F) with (K_{V_F}sim_{Q}0). In particular its canonical divisor is numerically trivial on its smooth locus. Campana's criterion cite[Appendix~A, Examples~A.1(2)]{Taji2014} says that the existence of such a normal birational model implies (omega(F)=0); it does not require (K_F) itself to be trivial. Property~4 of the same appendix and birational invariance now give (omega(R)=omega(R')ledim Z=kappa(R)). If (F) is a point, this upper bound is simply (omega(R)ledim R=kappa(R)). The reverse ineuality (omega(R)gekappa(R)) is one of the defining cotangent ineualities recorded in the same appendix. This proves the piecewise formula. A non-uniruled (X) has (R(X)) birational to (X), giving the first stated specialization. Campana also records (omega(X)gekappa^+(X)gekappa(X)) cite[Appendix~A, property~2]{Taji2014}. When (kappa(X)ge0), Corollary~cor:uniruled-mumford makes (X) non-uniruled, and the two outer terms are equal by the formula just proved. Thus (kappa^+(X)=kappa(X)). This is the abundance consequence recalled by Taji cite[Theorem~1.3]{Taji2014}, who attributes it to Campana cite[Proposition~3.10]{Campana1995}. [Finite fundamental group in Kodaira dimension zero] Let (X) be a positive-dimensional smooth connected projective complex variety. If (kappa(X)=0) and (chi(X,OO_X)ne0), then its ordinary topological fundamental group (pi_1(X)) is finite. Proof. Corollary~cor:cotangent-mrc gives (kappa^+(X)=0). Campana's criterion cite[Corollary~5.3]{Campana1995}, in the form recalled by Taji cite[Theorem~1.6]{Taji2014}, gives the conclusion. ## Quasi-projective covers Here a quasi-projective cover means a complex analytic covering space biholomorphic to a quasi-projective variety; its deck transformations need not be algebraic. The following application describes such covers in terms of bundles over abelian varieties. [Quasi-projective covering spaces] Let (X) be a connected normal projective complex variety. [label=(roman*)] Its universal cover is biholomorphic to a quasi-projective variety if and only if there is a finite 'etale Galois cover (X'to X) and a locally trivial holomorphic fiber bundle (X'to A), where (A) is an abelian variety and the fiber is simply connected and projective. If (widetilde Xto X) is an infinite 'etale Galois cover and (widetilde X) is biholomorphic to a quasi-projective variety, there are a finite 'etale Galois cover (X'to X), a locally trivial holomorphic fiber bundle (alpha:X'to A) over an abelian variety, and an 'etale cover (widetilde Ato A) with no positive-dimensional compact analytic subvarieties, such that [ widetilde X~= X'times_Awidetilde A ] as covering spaces over (X). Proof. Theorem~thm:main with zero boundary on a smooth projective complex variety gives precisely Conjecture~1.2 of Claudon--H"oring--Koll'ar [cite]. Their Theorem~1.1 and Corollary~1.5 therefore give (i) and (ii), respectively. The simply connected fiber condition belongs to (i), not to the general-cover statement (ii). Local triviality is holomorphic; no global product after a finite cover or compact K"ahler extension is asserted. # Relative abundance for projective analytic morphisms The complex rational case of Theorem~thm:main supplies the algebraic fiber input in Fujino's relative analytic reduction cite[Theorem~1.10 and its proof]{FujinoRelativeAnalytic2025}. We record only its normal rational case. Throughout this section, complex analytic spaces are Hausdorff and second-countable, and a complex variety means a reduced and irreducible complex analytic space, as in cite[p.~5]{FujinoRelativeAnalytic2025}. Here a projective analytic morphism is proper and possesses a relatively ample line bundle. Rational Cartierness is local on the source; a single global Cartier index is not part of that definition. A divisor is (pi)-nef over (Y) if it has nonnegative intersection with every projective integral curve contracted by (pi), over every point of (Y). These are the conventions of cite[Definitions~2.28, 2.32 and~2.48]{FujinoAnalyticMMP2022}. [Relative analytic abundance] Let (pi:Xto Y) be a projective surjective morphism of normal complex analytic varieties. Let ((X,Delta)) be a log canonical pair with (Delta) an effective rational divisor and (D=K_X+Delta) (Q)-Cartier. Assume that (D) is (pi)-nef over all of (Y). For every compact subset (Wsubset Y), there exist an analytic open neighborhood (Usupset W) and a positive integer (m) such that (mD|_{X_U}) is Cartier, where (X_U=pi^{-1}(U)), and the line bundle [ L=OO_{X_U}(mD|_{X_U}) ] is generated relative to (pi_U:X_Uto U). Equivalently, the evaluation map [ pi_U^*(pi_U)_*L→ L ] is surjective. Proof. Since (pi) is proper, (pi^{-1}(W)) is compact. Local rational Cartier indices therefore have a common multiple (m_0) on an open neighborhood (O) of (pi^{-1}(W)). The image of the closed set (X∖ O) is closed and disjoint from (W), so there is an open neighborhood (Vsupset W) with (pi^{-1}(V)subset O). Thus (m_0D) is Cartier over (V). Fix (Pin W). Following cite[the proof of Theorem~1.10, p.~30]{FujinoRelativeAnalytic2025}, work over a sufficiently small neighborhood of (P) contained in (V) and take a crepant dlt blow-up (p:(X',Delta')to(X,Delta)). Put (pi'=picirc p) and (D'=K_{X'}+Delta'=p^*D). For every log canonical stratum (S) of ((X',Delta')), including (X') itself, adjunction gives an effective rational boundary (Delta_S) with [ K_S+Delta_S=D'|_S. ] The strata are normal, and this adjoint is nef over (pi'(S)). On each normal projective component (F) of an analytically sufficiently general fiber of (Stopi'(S)), restriction gives a log canonical pair ((F,Delta_F)) with effective rational boundary and nef adjoint (K_F+Delta_F=(K_S+Delta_S)|_F). Theorem~thm:main makes this adjoint semiample, so its Iitaka and numerical dimensions agree. Thus (D'|_S) is abundant over (pi'(S)) for every stratum: this is precisely (pi')-log abundance in cite[Definitions~2.10 and~2.13]{FujinoRelativeAnalytic2025}. The nef-and-log-abundant theorem cite[Theorem~1.4]{FujinoRelativeAnalytic2025} now gives a relatively generated multiple of (D') over a neighborhood (U_Psubset V) of (P). Enlarge its integer to a multiple (m_P) of (m_0); tensor powers preserve relative generation. Normality gives (p_*OO_{X'}=OO_X), and projection formula identifies the direct images over (U_P) of (OO_X(m_PD)) and (OO_{X'}(m_PD')). Their evaluation maps are related by pullback. Generation upstairs therefore implies generation downstairs: a base point downstairs would make every pulled-back section vanish on its nonempty fiber. Choose finitely many of these neighborhoods (U_{P_1},...,U_{P_s}) covering (W), and let (m) be a common multiple of (m_0,m_{P_1},...,m_{P_s}). Put (U=∪_i U_{P_i}). The divisor (mD) is Cartier over (U), and its line bundle is relatively generated on each (U_{P_i}). Surjectivity of evaluation is local on the base, so it holds over (U), as asserted. The base (Y) need not be algebraic, compact, or Stein, and (W) need not be a Stein compact subset. The integer (m) may depend on the pair, the morphism, and (W); no single Cartier index or relatively generated multiple over all of a noncompact base is asserted. The nefness premise is over all of (Y), not only over (W). Projectivity cannot be replaced here by mere properness or by a K"ahler hypothesis: when (Y) is a point it requires (X) to be projective. The statement does not include real boundaries or semi-log-canonical analytic pairs. %====================================================================== % references.bib (moved to the end of this file so a reader can start % at the guide and the paper. For a LaTeX build, restore the block % below to the preamble, immediately before the document environment.) %====================================================================== % [overwrite]{references.bib} % @article{EinKuchleLazarsfeld1995, % author = {Ein, Lawrence and K{"u}chle, Oliver and Lazarsfeld, Robert}, % title = {Local positivity of ample line bundles}, % journal = {Journal of Differential Geometry}, % volume = {42}, number = {2}, year = {1995}, pages = {193--219}, % doi = {10.4310/jdg/1214457231}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Proposition 2.3} % } % @misc{OpenAIIitaka2026, % author = {{OpenAI}}, % title = {{Orbifold and logarithmic Iitaka subadditivity}}, % howpublished = {OpenAI Math Release preprint % OAI:Orbifold-and-logarithmic-Iitaka-subadditivity-September-26-2026}, % year = {2026}, % note = {Corollaries 6.2 and 6.3} % } % @article{ClaudonHoeringKollar2013, % author = {Claudon, Beno{^i}t and H{"o}ring, Andreas and Koll{'a}r, J{'a}nos}, % title = {Algebraic varieties with quasi-projective universal cover}, % journal = {Journal f{"u}r die reine und angewandte Mathematik}, % volume = {679}, % year = {2013}, % pages = {207--221}, % doi = {10.1515/crelle.2012.017}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{Hashizume2018, % author = {Hashizume, Kenta}, % title = {On the non-vanishing conjecture and existence of log minimal models}, % journal = {Publications of the Research Institute for Mathematical Sciences}, % volume = {54}, % number = {1}, % year = {2018}, % pages = {89--104}, % doi = {10.4171/PRIMS/54-1-3}, % eprint = {1609.00121}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % url = {./pdf/unknown-nd-1609.00121v4-url.pdf}, % note = {Theorem 1.4 in arXiv:1609.00121v4} % } % @incollection{FujinoGongyoRings2014, % author = {Fujino, Osamu and Gongyo, Yoshinori}, % title = {On log canonical rings}, % booktitle = {Higher Dimensional Algebraic Geometry in honour of Professor Yujiro Kawamata's sixtieth birthday}, % series = {Advanced Studies in Pure Mathematics}, % volume = {74}, % publisher = {Mathematical Society of Japan}, % address = {Tokyo}, % year = {2017}, % pages = {159--169}, % eprint = {1302.5194}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Conjectures A--C, Theorem 1.1, Corollary 1.4, and Lemma 3.5} % } % @article{Birkar2011II, % author = {Birkar, Caucher}, % title = {On existence of log minimal models {II}}, % journal = {Journal f{"u}r die reine und angewandte Mathematik}, % volume = {658}, % year = {2011}, % pages = {99--113}, % doi = {10.1515/crelle.2011.062}, % eprint = {0907.4170}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % url = {./pdf/unknown-nd-0907.4170v1-url.pdf}, % note = {Remark 3.1 and Proposition 3.2(3) in arXiv:0907.4170v1} % } % @article{Birkar2010, % author = {Birkar, Caucher}, % title = {On existence of log minimal models}, % journal = {Compositio Mathematica}, % volume = {146}, % number = {4}, % year = {2010}, % pages = {919--928}, % doi = {10.1112/S0010437X09004564}, % eprint = {0706.1792}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % url = {./pdf/unknown-nd-0706.1792v3-url.pdf}, % note = {Lemma 3.6 in arXiv:0706.1792v3} % } % @article{Birkar2012Special, % author = {Birkar, Caucher}, % title = {Existence of log canonical flips and a special {LMMP}}, % journal = {Publications Math{'e}matiques de l'IH{'E}S}, % volume = {115}, % year = {2012}, % pages = {325--368}, % doi = {10.1007/s10240-012-0039-5}, % eprint = {1104.4981}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % note = {Theorem 1.9, restated as Theorem 4.1} % } % @article{FujinoGongyo2014, % author = {Fujino, Osamu and Gongyo, Yoshinori}, % title = {Log pluricanonical representations and the abundance conjecture}, % journal = {Compositio Mathematica}, % volume = {150}, % number = {4}, % year = {2014}, % pages = {593--620}, % doi = {10.1112/S0010437X13007495}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Theorem 1.5, restated as Theorem 4.3} % } % @article{Ambro2005, % author = {Ambro, Florin}, % title = {The moduli b-divisor of an lc-trivial fibration}, % journal = {Compositio Mathematica}, % volume = {141}, % number = {2}, % year = {2005}, % pages = {385--403}, % doi = {10.1112/S0010437X04001071}, % note = {Theorem 0.2} % } % @article{BCHM2010, % author = {Birkar, Caucher and Cascini, Paolo and Hacon, Christopher D. and McKernan, James}, % title = {Existence of minimal models for varieties of log general type}, % journal = {Journal of the American Mathematical Society}, % volume = {23}, % number = {2}, % year = {2010}, % pages = {405--468}, % doi = {10.1090/S0894-0347-09-00649-3}, % eprint = {math/0610203}, % archivePrefix = {arXiv}, % primaryClass = {math.AG} % } % @misc{Fujino2017, % author = {Fujino, Osamu}, % title = {Introduction to the log minimal model program for log canonical pairs}, % year = {2009}, % howpublished = {Unpublished manuscript}, % eprint = {0907.1506}, % archivePrefix = {arXiv}, % primaryClass = {math.AG}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Version 6.01, January 8, 2009; Theorem 3.74(3), p. 109} % } % @article{GongyoLehmann2013, % author = {Gongyo, Yoshinori and Lehmann, Brian}, % title = {Reduction maps and minimal model theory}, % journal = {Compositio Mathematica}, % year = {2013}, volume = {149}, number = {2}, pages = {295--308}, % doi = {10.1112/S0010437X12000553}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Read with the numerical-dimension erratum on Lehmann's research page} % } % @article{Fujino2019Numerical, % author = {Fujino, Osamu}, % title = {Corrigendum to ``On subadditivity of the logarithmic {Kodaira} dimension''}, % journal = {Journal of the Mathematical Society of Japan}, % year = {2020}, volume = {72}, number = {4}, pages = {1181--1187}, % doi = {10.2969/jmsj/82568256}, % eprint = {1904.11639}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {https://www.jstage.jst.go.jp/article/jmath/72/4/72_1181/_article/-char/ja} % } % @article{KovacsPatakfalvi2017, % author = {Kov{'a}cs, S{'a}ndor J. and Patakfalvi, Zsolt}, % title = {Projectivity of the moduli space of stable log-varieties and subadditivity of log-{Kodaira} dimension}, % journal = {Journal of the American Mathematical Society}, % year = {2017}, volume = {30}, number = {4}, pages = {959--1021}, % doi = {10.1090/jams/871}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Theorem 9.9 in the author version} % } % @article{BDPP2013, % author = {Boucksom, S{'e}bastien and Demailly, Jean-Pierre and P{u a}un, Mihai and Peternell, Thomas}, % title = {The pseudo-effective cone of a compact {K{"a}hler} manifold and varieties of negative {Kodaira} dimension}, % journal = {Journal of Algebraic Geometry}, % year = {2013}, volume = {22}, number = {2}, pages = {201--248}, % doi = {10.1090/S1056-3911-2012-00574-8}, % eprint = {math/0405285}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-math0405285-url.pdf}, % note = {Theorem locators refer to arXiv version 1} % } % @article{LazicPeternell2017, % author = {Lazi{'c}, Vladimir and Peternell, Thomas}, % title = {Rationally connected varieties---on a conjecture of {Mumford}}, % journal = {Science China Mathematics}, % year = {2017}, volume = {60}, number = {6}, pages = {1019--1028}, % doi = {10.1007/s11425-016-9039-2}, % eprint = {1608.04706}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-1608.04706-url.pdf}, % note = {Theorem locators refer to arXiv version 1} % } % @misc{BrunebarbeCampana2015, % author = {Brunebarbe, Yohan and Campana, Fr{'e}d{'e}ric}, % title = {Fundamental group and pluridifferentials on compact {K{"a}hler} manifolds}, % year = {2015}, % eprint = {1510.07922}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-1510.07922-url.pdf}, % note = {Version 2} % } % @article{Campana1995, % author = {Campana, Fr{'e}d{'e}ric}, % title = {Fundamental group and positivity of cotangent bundles of compact {K{"a}hler} manifolds}, % journal = {Journal of Algebraic Geometry}, % year = {1995}, volume = {4}, number = {3}, pages = {487--502} % } % @article{CampanaPaun2015, % author = {Campana, Fr{'e}d{'e}ric and P{u a}un, Mihai}, % title = {Orbifold generic semi-positivity: an application to families of canonically polarized manifolds}, % journal = {Annales de l'Institut Fourier}, % year = {2015}, volume = {65}, number = {2}, pages = {835--861}, % doi = {10.5802/aif.2945}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{Langer2004, % author = {Langer, Adrian}, % title = {Semistable sheaves in positive characteristic}, % journal = {Annals of Mathematics}, % year = {2004}, volume = {159}, number = {1}, pages = {251--276}, % doi = {10.4007/annals.2004.159.251}, % url = {https://annals.math.princeton.edu/2004/159-1/p04} % } % @article{Sun2008, % author = {Sun, Xiaotao}, % title = {Direct images of bundles under {Frobenius} morphism}, % journal = {Inventiones Mathematicae}, % year = {2008}, volume = {173}, number = {2}, pages = {427--447}, % doi = {10.1007/s00222-008-0125-y}, % eprint = {math/0611360v2}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-math0611360-url.pdf}, % note = {Theorem 3.7 and Lemma 4.2 in arXiv:math/0611360v2} % } % @article{FujinoHashizume2023, % author = {Fujino, Osamu and Hashizume, Kenta}, % title = {Existence of log canonical modifications and its applications}, % journal = {European Journal of Mathematics}, % year = {2023}, volume = {9}, number = {1}, % note = {Paper No. 13, 32 pp.; Theorem 2.10 in the accepted author version}, % doi = {10.1007/s40879-023-00598-0}, % url = {./pdf/fujino-hashizume-2023-existence-of-log-canonical-modifications-and-its-applications.pdf} % } % @article{FujinoGongyoModuli2014, % author = {Fujino, Osamu and Gongyo, Yoshinori}, % title = {On the moduli b-divisors of lc-trivial fibrations}, % journal = {Annales de l'Institut Fourier}, % year = {2014}, volume = {64}, number = {4}, pages = {1721--1735}, % doi = {10.5802/aif.2894}, % url = {./pdf/unknown-nd-url.pdf} % } % @incollection{HMX2018, % author = {Hacon, Christopher D. and McKernan, James and Xu, Chenyang}, % title = {Boundedness of varieties of log general type}, % booktitle = {Algebraic Geometry: Salt Lake City 2015, Part 1}, % series = {Proceedings of Symposia in Pure Mathematics}, % volume = {97.1}, pages = {309--348}, % publisher = {American Mathematical Society}, year = {2018}, % doi = {10.1090/pspum/097.1/01677}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Published Theorem 4.0.1; Theorem 3.0.1 in arXiv version 2} % } % @article{Saito1990, % author = {Saito, Morihiko}, % title = {Mixed {Hodge} modules}, % journal = {Publications of the Research Institute for Mathematical Sciences}, % year = {1990}, volume = {26}, number = {2}, pages = {221--333}, % doi = {10.2977/PRIMS/1195171082}, % url = {https://ems.press/journals/prims/articles/3552} % } % @article{Saito2000, % author = {Saito, Morihiko}, % title = {Mixed {Hodge} complexes on algebraic varieties}, % journal = {Mathematische Annalen}, % year = {2000}, volume = {316}, number = {2}, pages = {283--331}, % doi = {10.1007/s002080050014}, % eprint = {math/9906088}, archivePrefix = {arXiv}, % url = {./pdf/unknown-nd-math9906088-url.pdf}, % note = {Theorem 0.2, Corollary 0.3 and Proposition 5.3 in arXiv:math/9906088v1} % } % @incollection{Coste2007, % author = {Coste, Michel}, % title = {Real Algebraic Sets}, % booktitle = {Arc spaces and additive invariants in real algebraic and analytic geometry}, % series = {Panoramas et Synth{`e}ses}, % volume = {24}, % publisher = {Soci{'e}t{'e} Math{'e}matique de France}, % year = {2007}, % pages = {1--32}, % note = {Theorem 1.10 in the author's March 23, 2005 notes}, % url = {./pdf/unknown-nd-url.pdf} % } % @book{SGA1, % author = {Grothendieck, Alexander and others}, % title = {Rev{^e}tements {'e}tales et groupe fondamental ({SGA} 1)}, % series = {Lecture Notes in Mathematics}, % volume = {224}, % publisher = {Springer-Verlag}, % year = {1971}, % eprint = {math/0206203v2}, % archivePrefix = {arXiv}, % url = {./pdf/unknown-nd-math0206203-url.pdf}, % note = {Expos{'e} X, Theorem 3.1, and Expos{'e} XII, Theorem 5.1} % } % @article{MustataPopa2022, % author = {Musta{c t}{u a}, Mircea and Popa, Mihnea}, % title = {{Hodge} filtration on local cohomology, {Du Bois} complex and local cohomological dimension}, % journal = {Forum of Mathematics, Pi}, % year = {2022}, volume = {10}, % doi = {10.1017/fmp.2022.15}, % note = {Paper No. e22; Theorem C and Proposition 13.1}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{KollarKovacs2010, % author = {Koll{'a}r, J{'a}nos and Kov{'a}cs, S{'a}ndor J.}, % title = {Log canonical singularities are {Du Bois}}, % journal = {Journal of the American Mathematical Society}, % year = {2010}, volume = {23}, number = {3}, pages = {791--813}, % doi = {10.1090/S0894-0347-10-00663-6}, % url = {./pdf/unknown-nd-url.pdf} % } % @incollection{BauerEtAl2002, % author = {Bauer, Thomas and Campana, Fr{'e}d{'e}ric and Eckl, Thomas and Kebekus, Stefan and Peternell, Thomas and Rams, S{l}awomir and Szemberg, Tomasz and Wotzlaw, Lorenz}, % title = {A reduction map for nef line bundles}, % booktitle = {Complex Geometry: Collection of Papers Dedicated to {Hans Grauert}}, % publisher = {Springer}, address = {Berlin}, % year = {2002}, pages = {27--36}, % doi = {10.1007/978-3-642-56202-0_2}, % url = {./pdf/unknown-nd-math0106147-url.pdf}, % note = {Theorem 2.1 in arXiv:math/0106147v1} % } % @misc{StacksExistence, % author = {{The Stacks Project Authors}}, % title = {The {Stacks} Project: {Grothendieck}'s existence theorem, {I}}, % year = {2026}, % howpublished = {https://stacks.math.columbia.edu/tag/087V}, % note = {Tag 087V, Lemmas 30.24.1 and 30.24.3; accessed September 26, 2026} % } % @article{Taji2014, % author = {Taji, Behrouz}, % title = {Birational positivity in dimension 4}, % journal = {Annales de l'Institut Fourier}, % year = {2014}, volume = {64}, number = {1}, pages = {203--216}, % doi = {10.5802/aif.2845}, % url = {https://aif.centre-mersenne.org/articles/10.5802/aif.2845/}, % note = {With an appendix by Fr{'e}d{'e}ric Campana} % } % @article{Hanamura1988, % author = {Hanamura, Masaki}, % title = {Structure of birational automorphism groups, {I}: non-uniruled varieties}, % journal = {Inventiones Mathematicae}, % year = {1988}, volume = {93}, pages = {383--403}, % doi = {10.1007/BF01394338}, % note = {Theorems 2.1--2.2; statements also reproduced in Blanc, Theorems 3.8--3.9} % } % @incollection{Blanc2017, % author = {Blanc, J{'e}r{'e}my}, % title = {Algebraic structures of groups of birational transformations}, % booktitle = {Algebraic Groups: Structure and Actions}, % series = {Proceedings of Symposia in Pure Mathematics}, % volume = {94}, year = {2017}, pages = {17--30}, % publisher = {American Mathematical Society}, % doi = {10.1090/pspum/094/01620}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Theorems 3.8--3.9} % } % @article{Siu1974, % author = {Siu, Yum-Tong}, % title = {Analyticity of sets associated to {Lelong} numbers and the extension of closed positive currents}, % journal = {Inventiones Mathematicae}, % year = {1974}, volume = {27}, pages = {53--156}, % doi = {10.1007/BF01389965} % } % @book{Demailly2012, % author = {Demailly, Jean-Pierre}, % title = {Analytic Methods in Algebraic Geometry}, % series = {Surveys of Modern Mathematics}, % volume = {1}, year = {2012}, % publisher = {International Press and Higher Education Press}, % address = {Somerville, MA, and Beijing}, % isbn = {978-1-57146-234-3}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Theorem numbering refers to the author manuscript dated July 2011} % } % @article{MiyaokaMori1986, % author = {Miyaoka, Yoichi and Mori, Shigefumi}, % title = {A numerical criterion for uniruledness}, % journal = {Annals of Mathematics}, % year = {1986}, volume = {124}, number = {1}, pages = {65--69}, % doi = {10.2307/1971387}, % url = {https://annals.math.princeton.edu/1986/124/p03} % } % @article{Paun2007, % author = {P{u a}un, Mihai}, % title = {{Siu}'s invariance of plurigenera: a one-tower proof}, % journal = {Journal of Differential Geometry}, % year = {2007}, volume = {76}, number = {3}, pages = {485--493}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Theorem 1} % } % @article{Miyaoka1988, % author = {Miyaoka, Yoichi}, % title = {Abundance conjecture for 3-folds: case {nu=1}}, % journal = {Compositio Mathematica}, % volume = {68}, number = {2}, year = {1988}, pages = {203--220}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{Kawamata1992, % author = {Kawamata, Yujiro}, % title = {Abundance theorem for minimal threefolds}, % journal = {Inventiones Mathematicae}, % volume = {108}, year = {1992}, pages = {229--246}, % doi = {10.1007/BF02100604}, % url = {https://link.springer.com/article/10.1007/BF02100604} % } % @article{KeelMatsukiMcKernan1994, % author = {Keel, Sean and Matsuki, Kenji and McKernan, James}, % title = {Log abundance theorem for threefolds}, % journal = {Duke Mathematical Journal}, % volume = {75}, number = {1}, year = {1994}, pages = {99--119}, % doi = {10.1215/S0012-7094-94-07504-2}, % note = {Read together with the authors' 2004 corrections} % } % @article{KeelMatsukiMcKernan2004Correction, % author = {Keel, Sean and Matsuki, Kenji and McKernan, James}, % title = {Corrections to ``Log abundance theorem for threefolds''}, % journal = {Duke Mathematical Journal}, % volume = {122}, number = {3}, year = {2004}, pages = {625--630}, % doi = {10.1215/S0012-7094-04-12236-5}, % url = {https://doi.org/10.1215/S0012-7094-04-12236-5} % } % @misc{Matsuki2003Correction, % author = {Matsuki, Kenji}, % title = {A correction to the paper ``Log abundance theorem for threefolds''}, % year = {2003}, % eprint = {math/0302360v1}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-math0302360-url.pdf}, % note = {arXiv:math/0302360v1} % } % @article{Fujino2000, % author = {Fujino, Osamu}, % title = {Abundance theorem for semi log canonical threefolds}, % journal = {Duke Mathematical Journal}, % volume = {102}, number = {3}, year = {2000}, pages = {513--532}, % doi = {10.1215/S0012-7094-00-10237-2}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{Gongyo2013, % author = {Gongyo, Yoshinori}, % title = {Abundance theorem for numerically trivial log canonical divisors of semi-log canonical pairs}, % journal = {Journal of Algebraic Geometry}, % volume = {22}, number = {3}, year = {2013}, pages = {549--564}, % doi = {10.1090/S1056-3911-2012-00593-1}, % eprint = {1005.2796v2}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-1005.2796v2-url.pdf} % } % @article{DemaillyHaconPaun2013, % author = {Demailly, Jean-Pierre and Hacon, Christopher D. and P{u a}un, Mihai}, % title = {Extension theorems, non-vanishing and the existence of good minimal models}, % journal = {Acta Mathematica}, % volume = {210}, number = {2}, year = {2013}, pages = {203--259}, % doi = {10.1007/s11511-013-0094-x}, % url = {./pdf/unknown-nd-url.pdf} % } % @book{EsnaultViehweg1992, % author = {Esnault, H{'e}l{`e}ne and Viehweg, Eckart}, % title = {Lectures on Vanishing Theorems}, % series = {DMV Seminar}, volume = {20}, year = {1992}, % publisher = {Birkh{"a}user}, address = {Basel}, % doi = {10.1007/978-3-0348-8600-0}, % url = {./pdf/unknown-nd-url.pdf} % } % @article{Deligne1971, % author = {Deligne, Pierre}, % title = {Th{'e}orie de {Hodge}. {II}}, % journal = {Publications Math{'e}matiques de l'IH{'E}S}, % volume = {40}, year = {1971}, pages = {5--57}, % doi = {10.1007/BF02684692}, % url = {./pdf/unknown-nd-url.pdf} % } % @misc{OpenAIMinimalModels2026, % author = {{OpenAI}}, % title = {{Minimal models and Mori fibre spaces for generalized log canonical ℚ-pairs}}, % howpublished = {OpenAI Math Release preprint % OAI:Minimal-models-and-Mori-fibre-spaces-for-generalized-log-canonical-Q-pairs-September-24-2026}, % year = {2026}, % note = {Corollary 1.2} % } % @misc{LiuXu2025, % author = {Liu, Jihao and Xu, Zheng}, % title = {Non-vanishing implies numerical dimension one abundance}, % year = {2025}, % eprint = {2505.05250v2}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-2505.05250v2-url.pdf}, % note = {arXiv:2505.05250v2, July 30, 2025; Theorems 3.1, 5.1 and 5.4} % } % @article{KitadaiSumihiro2008, % author = {Kitadai, Yukinori and Sumihiro, Hideyasu}, % title = {Canonical filtrations and stability of direct images by {Frobenius} morphisms}, % journal = {Tohoku Mathematical Journal}, % series = {2}, volume = {60}, number = {2}, year = {2008}, % pages = {287--301}, doi = {10.2748/tmj/1215442876}, % url = {https://doi.org/10.2748/tmj/1215442876}, % note = {Definition 3.1, Remark 3.2, and Corollary 3.5} % } % @article{MustataSchwede2014, % author = {Musta{c t}{u a}, Mircea and Schwede, Karl}, % title = {A {Frobenius} variant of {Seshadri} constants}, % journal = {Mathematische Annalen}, % volume = {358}, number = {3-4}, year = {2014}, pages = {861--878}, % doi = {10.1007/s00208-013-0976-4}, % eprint = {1203.1081v1}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-1203.1081v1-url.pdf}, % note = {Section 2 and Proposition 2.12 in arXiv:1203.1081v1} % } % @article{JoshiRamananXiaYu2006, % author = {Joshi, Kirti and Ramanan, S. and Xia, Eugene Z. and Yu, Jiu-Kang}, % title = {On vector bundles destabilized by {Frobenius} pull-back}, % journal = {Compositio Mathematica}, % volume = {142}, number = {3}, year = {2006}, pages = {616--630}, % doi = {10.1112/S0010437X05001788}, % url = {./pdf/joshi-ramanan-2006-on-vector-bundles-destabilized-by-frobenius-pull-back.pdf}, % note = {Section 5.3} % } % @misc{ChanChoi2021, % author = {Chan, Tsz On Mario and Choi, Young-Jun}, % title = {Extension with log-canonical measures and an improvement to the plt extension of {Demailly--Hacon--Pu{a}un}}, % year = {2021}, eprint = {1912.08076v5}, archivePrefix = {arXiv}, % url = {./pdf/unknown-nd-1912.08076v5-url.pdf}, % doi = {10.1007/s00208-021-02152-3}, % note = {arXiv:1912.08076v5; Theorem 1.6.1} % } % @book{Kollar2013, % author = {Koll{'a}r, J{'a}nos}, % title = {Singularities of the Minimal Model Program}, % series = {Cambridge Tracts in Mathematics}, volume = {200}, % publisher = {Cambridge University Press}, address = {Cambridge}, year = {2013}, % note = {With the collaboration of S{'a}ndor Kov{'a}cs} % } % @article{MehtaRamanathan1982, % author = {Mehta, V. B. and Ramanathan, A.}, % title = {Semistable sheaves on projective varieties and their restriction to curves}, % journal = {Mathematische Annalen}, volume = {258}, year = {1982}, % pages = {213--224}, note = {Theorem 6.1 and Remark 6.2} % } % @misc{StacksSectionModules, % author = {{The Stacks Project Authors}}, % title = {The {Stacks} Project: Coherent sheaves on {Proj}, {I}}, % year = {2026}, % howpublished = {https://stacks.math.columbia.edu/tag/01YS}, % note = {Tag 01YS, Lemma 30.14.1(5); accessed September 27, 2026} % } % @misc{FujinoRelativeAnalytic2025, % author = {Fujino, Osamu}, % title = {On finiteness of relative log pluricanonical representations}, % year = {2025}, % howpublished = {Author manuscript}, % url = {./pdf/unknown-nd-url.pdf}, % note = {Version 0.14, May 29, 2025; Theorem 1.10 and its proof on p. 30} % } % @misc{FujinoAnalyticMMP2022, % author = {Fujino, Osamu}, % title = {Minimal model program for projective morphisms between complex analytic spaces}, % year = {2022}, % eprint = {2201.11315v1}, archivePrefix = {arXiv}, primaryClass = {math.AG}, % url = {./pdf/unknown-nd-2201.11315v1-url.pdf}, % note = {Definitions 2.28, 2.32 and 2.48} % } % @misc{CaoPaunAbelian2016, % author = {Cao, Junyan and P{u a}un, Mihai}, % title = {{Kodaira} dimension of algebraic fiber spaces over abelian varieties}, % year = {2016}, % eprint = {1504.01095v3}, % archivePrefix = {arXiv}, % url = {./pdf/unknown-nd-1504.01095v3-url.pdf}, % note = {Version 3; Theorem 1.1 gives the abelian-base case} % } % @misc{HaconPopaSchnell2017, % author = {Hacon, Christopher D. and Popa, Mihnea and Schnell, Christian}, % title = {Algebraic fiber spaces over abelian varieties: around a recent theorem by {Cao} and {P{u a}un}}, % year = {2017}, % eprint = {1611.08768v2}, % archivePrefix = {arXiv}, % url = {./pdf/unknown-nd-1611.08768v2-url.pdf}, % note = {Version 2; Theorem 1.1} % } %