Papers
Topics
Authors
Recent
Search
2000 character limit reached

On endomorphisms of affine spaces and the Jacobian problem

Published 4 Sep 2026 in math.AG | (2609.05746v1)

Abstract: Let pp be a prime. We provide examples which show that étale endomorphisms of affine planes over an algebraically closed field kk of characteristic pp can have fibers of arbitrary finite cardinal. Let (l,m)N×N<sup>(l,m)\in\mathbb N\times\mathbb N<sup>{\ast}. We provide examples of such étale endomorphisms whose images have complements of cardinality ll and whose geometric degrees are pmpm. Several conjectures are disproved, and in particular we provide an analog over kk of Kulikov's counterexample to the `Generalized Jacobian Conjecture' of Bass for surfaces. Surjective (resp.\ surjective and non-surjective) counterexamples to Adjamagbo's Jacobian Conjecture over each kk are included in dimension $2$ (resp.\ in all dimensions at least $3$); for instance, if p=2p=2, then we show for each mm there exist surjective étale endomorphisms of the affine spaces over kk of dimension at least $3$ of geometric degree mm. If e:XXe:X\rightarrow X is an endomorphism of a variety over an algebraically closed field KK, then we show that there exists nNn\in\mathbb N such that $\Imm(e<sup>n)=\Imm(e<sup>{n+1})$ provided either (i) ee is quasi-finite or (ii) dim(X)2\dim(X)\le 2 and $X\setminus\Imm(e)$ is finite. We provide examples of endomorphisms of affine planes of dimensions at least $3$ over KK whose images have complements of cardinality ll and for all nNn\in\mathbb N we have $\Imm(e<sup>n)\neq\Imm(e<sup>{n+1})$. We prove that all affine moduli schemes of étale endomorphisms of affine spaces with Jacobian matrices of determinant 1 over KK are connected and classify all those that are smooth. We prove that the Jacobian Conjecture over C\mathbb C and Adjamagbo's analog of it over kk hold for étale endomorphisms of affine spaces that are composites gfg\circ f, where ff is quasi-finite and a locally closed embedding in codimension $1$ outside a specific finite subset and gg is a projection that omits one coordinate.

Summary

  • The paper introduces a comprehensive framework of invariants, including geometric and algebraic degrees, fiber cardinalities and geometric invariants to analyze the structure of étale endomorphisms.
  • Disproves that there are fundamental limitations in positive characteristic settings and that Había counterexamples contradict assumptions.
  • Provides comprehensive modularity on maps.

Scope and mathematical setting

The paper studies polynomial self-maps of affine space whose Jacobian determinant is a nonzero constant, with particular emphasis on étale endomorphisms over algebraically closed fields of positive characteristic. Such maps are called Kraus–Keller maps. The central issue is that the Jacobian condition controls infinitesimal behavior but, unlike in characteristic zero under the conjectural Jacobian principle, does not force global invertibility or even surjectivity in positive characteristic.

The authors introduce a systematic collection of invariants for an étale endomorphism

e:AKnAKn.e:\mathbb A^n_K\longrightarrow \mathbb A^n_K.

These include the geometric degree, algebraic degrees up to left, right, or two-sided composition by automorphisms, the possible cardinalities of non-generic fibers, the cardinality of the complement of the image, stabilization behavior under iteration, and geometric invariants obtained from the normalization of the target in the source function field. The latter construction associates to ee a normal variety XeX_e containing the source affine space as an open subvariety and a finite self-map

ψe:XeXe,\psi_e:X_e\longrightarrow X_e,

called the associated Jacobian variety. The irreducible components of XeAnX_e\setminus\mathbb A^n define the class-rank invariants ρ(e)\rho(e) and ρeˊt(e)\rho_{\mathrm{\acute et}}(e).

This framework permits the paper to treat global image defects, fiber degenerations, iteration, normalization, singularities, and moduli-theoretic properties within a single geometric formalism. The main results are constructive, although some constructions are presented existentially after passing through normalizations, finite covers, or affine models.

Frobenius systems and the failure of positive-characteristic surjectivity

A basic endomorphism in characteristic pp is, up to polynomial automorphisms, a map of the form

(x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).

Its Jacobian matrix is the identity, so every basic endomorphism is étale. The associated preimage problem for the origin is the Frobenius system

xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.

A motivating conjecture proposed that Frobenius systems over algebraically closed fields should always be consistent. The paper disproves this in dimension two by constructing explicit inconsistent systems. In particular, the authors exhibit a basic étale endomorphism of ee0 whose image omits a point. Thus even the positive-characteristic analogue of the elementary surjectivity statement for one-variable maps fails in dimension two.

The construction is not merely a pathological example. It produces Frobenius systems with controlled numbers of solutions. Theorem 1(1) shows that for every ee1 there is an ee2-system with exactly ee3 solutions, including the case ee4. Consequently, for every prime ee5, there are étale endomorphisms of the affine plane having fibers of arbitrary prescribed finite cardinalities below their generic degree. This is a strong obstruction to any positive-characteristic surjectivity principle based only on the Jacobian condition.

The construction uses affine surfaces and open embeddings rather than only direct manipulation of polynomial equations. A key geometric ingredient is the surface

ee6

together with explicit open embeddings of ee7 into this surface. The complement of the image can be arranged to be a prescribed disjoint union of affine lines and points. Finite étale projections from these surfaces then yield endomorphisms of ee8 with controlled degree, image complement, and normalization geometry.

Arbitrary gaps, degrees, and fiber structures in dimension two

Theorem 1(2) gives the most precise numerical family in the paper. If ee9 and XeX_e0 satisfy XeX_e1, there exists an étale endomorphism XeX_e2 of XeX_e3 defined over XeX_e4 with

XeX_e5

where XeX_e6 is the cardinality of the complement of the image. The associated Jacobian variety is étale, and

XeX_e7

The left and right algebraic degrees satisfy

XeX_e8

with XeX_e9 bounded above by the same quantity.

The identity

ψe:XeXe,\psi_e:X_e\longrightarrow X_e,0

follows. Thus, even while fixing the geometric degree to a multiple of ψe:XeXe,\psi_e:X_e\longrightarrow X_e,1, one can prescribe an arbitrary finite image defect. The implication is that geometric degree and surjectivity are largely independent invariants in positive characteristic.

Theorem 1(3) supplies a complementary family. If ψe:XeXe,\psi_e:X_e\longrightarrow X_e,2 divides ψe:XeXe,\psi_e:X_e\longrightarrow X_e,3, there are étale endomorphisms of degree ψe:XeXe,\psi_e:X_e\longrightarrow X_e,4 with

ψe:XeXe,\psi_e:X_e\longrightarrow X_e,5

When ψe:XeXe,\psi_e:X_e\longrightarrow X_e,6, the iteration index is ψe:XeXe,\psi_e:X_e\longrightarrow X_e,7; when ψe:XeXe,\psi_e:X_e\longrightarrow X_e,8, necessarily ψe:XeXe,\psi_e:X_e\longrightarrow X_e,9, and XeAnX_e\setminus\mathbb A^n0. The associated Jacobian variety is regular, but it is non-étale for XeAnX_e\setminus\mathbb A^n1. This separates regularity of the normalization from étaleness of the associated self-map.

The paper also obtains explicit examples with arbitrary finite fiber cardinalities. In particular, the non-generic fiber set XeAnX_e\setminus\mathbb A^n2 can contain several distinct values, including zero-dimensional fibers, one-point fibers, and fibers of cardinality close to the generic degree. The examples show that fiber stratification in the positive-characteristic étale category is substantially more flexible than suggested by characteristic-zero intuition.

Counterexamples to positive-characteristic Jacobian conjectures

The paper disproves Adjamagbo’s characteristic-XeAnX_e\setminus\mathbb A^n3 Jacobian Conjecture, which asserted that an étale endomorphism of affine XeAnX_e\setminus\mathbb A^n4-space with geometric degree not divisible by XeAnX_e\setminus\mathbb A^n5 should be an automorphism. The conjecture fails in every dimension XeAnX_e\setminus\mathbb A^n6.

In dimension two, the authors construct families with degrees not divisible by XeAnX_e\setminus\mathbb A^n7, using substitution–étale tame perturbations. In dimension at least three, the linear–tame–perturbation method produces broader families, including both surjective and non-surjective examples. These families are obtained by starting with an étale map in two variables, adjoining a trivial coordinate, composing with a tame automorphism, and perturbing one coordinate by a XeAnX_e\setminus\mathbb A^n8-th power. Because the perturbation does not change the Jacobian matrix, étaleness is preserved, while the induced function-field extension can acquire degrees outside XeAnX_e\setminus\mathbb A^n9.

The resulting contradiction to Adjamagbo’s conjecture is particularly explicit in characteristic two. For every ρ(e)\rho(e)0 and every dimension ρ(e)\rho(e)1, the paper constructs a surjective étale endomorphism of geometric degree ρ(e)\rho(e)2. It also constructs non-surjective examples for broad classes of degrees. Therefore, in characteristic two, surjectivity does not constrain the geometric degree in the way proposed by the conjecture.

The dimension-two situation remains more constrained. The paper proves that for ρ(e)\rho(e)3, every positive integer occurs as the geometric degree of an étale endomorphism of ρ(e)\rho(e)4. For general ρ(e)\rho(e)5, however, the authors do not classify all possible degrees prime to ρ(e)\rho(e)6. They formulate this classification explicitly as an open problem. The weak Jacobian conjecture proposed in the paper remains unresolved: for fixed ρ(e)\rho(e)7, it asks whether sufficiently large characteristic excludes degrees in the interval from ρ(e)\rho(e)8 through ρ(e)\rho(e)9 for étale endomorphisms of the affine plane.

Iteration and stabilization of images

A second theme concerns the sequence

ρeˊt(e)\rho_{\mathrm{\acute et}}(e)0

for an endomorphism of a variety. The stabilization index ρeˊt(e)\rho_{\mathrm{\acute et}}(e)1 is the least ρeˊt(e)\rho_{\mathrm{\acute et}}(e)2 such that ρeˊt(e)\rho_{\mathrm{\acute et}}(e)3, if such an index exists.

The paper proves that if ρeˊt(e)\rho_{\mathrm{\acute et}}(e)4 is quasi-finite, then ρeˊt(e)\rho_{\mathrm{\acute et}}(e)5 is finite. This result applies over arbitrary algebraically closed fields and is obtained by reduction to positive characteristic. It generalizes earlier results that required stronger hypotheses, such as affine varieties over ρeˊt(e)\rho_{\mathrm{\acute et}}(e)6 and finite image complements.

For an arbitrary variety ρeˊt(e)\rho_{\mathrm{\acute et}}(e)7 of dimension at most two, if ρeˊt(e)\rho_{\mathrm{\acute et}}(e)8 is finite, then image stabilization also occurs. The proof uses the monotonicity of the successive increments

ρeˊt(e)\rho_{\mathrm{\acute et}}(e)9

which form a non-increasing sequence. In particular, if the image complement of pp0 is finite, then the complement of every iterate remains finite and grows in a controlled manner.

The behavior changes sharply in dimension at least three. The authors construct endomorphisms whose image complements have prescribed finite cardinality but for which

pp1

for every pp2. Thus quasi-finiteness or finite image complement is essential in the stabilization theorems. The examples show that finite initial defect does not by itself imply eventual stabilization once the dimension is sufficiently large.

The set-theoretic estimates are also structurally informative. If the smallest nonempty fiber has cardinality pp3, then the image-defect sequence is bounded by a geometric-series estimate, yielding an explicit bound on pp4. Infinite iteration index forces pp5 for every positive iterate. Hence persistent non-stabilization can occur only through the repeated appearance of one-point fibers.

Geometry of the associated Jacobian varieties

For an étale endomorphism pp6, the normalization pp7 of the target in the source function field compactifies the source relative to the finite morphism induced by pp8. The open immersion

pp9

has complement of pure codimension one, and the induced finite self-map (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).0 encodes the failure of (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).1 to be finite.

In dimension two, every irreducible component of the complement is shown to be isomorphic to (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).2. The associated affine surface is therefore a sphere-like surface: a normal affine surface containing an open copy of (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).3 whose complement is irreducible. Such surfaces admit an (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).4-fibration over (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).5, with the embedded affine plane equal to the inverse image of (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).6.

Theorem 10 establishes several strong properties of sphere-like surfaces. The boundary curve is rational and has one point at infinity; in fact, it is isomorphic to (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).7. There is at most one singular point. Every singularity is rational, and in characteristic zero a singularity is log-terminal. If the unique non-reduced fiber has multiplicity prime to the characteristic, the singularity is a quotient singularity by a subgroup of a cyclic group. In the wild case, it is still a quotient singularity by a finite group, although the group need not be cyclic.

These results provide geometric control over the normalization spaces attached to étale endomorphisms of the affine plane. The canonical affine cover by sphere-like surfaces decomposes according to the behavior of (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).8: components in the étale locus give regular sphere-like surfaces, while components outside it give quasi-spheres or pseudo-spheres. The numbers of these components are measured by the refined class-rank invariants.

The tangent bundle criterion is especially significant. For an affine open subvariety (x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\longmapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).9 containing the embedded affine space, the following are equivalent: xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.0 lies in the étale locus of xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.1; xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.2 is regular with trivial tangent bundle; and xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.3 is a complete intersection. In characteristic xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.4, these are also equivalent to the coordinate ring possessing a xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.5-basis or a differential basis. If xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.6 is regular but non-étale, then xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.7 cannot be a complete intersection. This gives a geometric obstruction to complete-intersection presentations that is detected by the ramification of the associated Jacobian variety.

Finite étale covers and torsorial constructions

The paper constructs finite étale covers of xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.8 in dimensions xi+fi(x1p,,xnp)=0.x_i+f_i(x_1^p,\ldots,x_n^p)=0.9 and ee00 using pullbacks of Lang torsors associated with ee01. The resulting covers are defined over ee02 and factor through finite étale morphisms of degrees

ee03

For ee04 and ee05, the construction gives an intermediate rational cover of degree ee06 with a further Galois cover of degree ee07. For ee08 and ee09, there is an intermediate rational cover of degree ee10 with a further quaternionic Galois cover of order ee11. Since the intermediate varieties have only constant units, these covers imply non-factoriality through Kummer theory.

The dimension-two examples arise as quotients of the dimension-three constructions by a left ee12-action. This provides explicit positive-characteristic counterexamples to conjectures asserting that certain rational affine varieties or affine spaces have trivial prime-to-ee13 étale fundamental group. It also demonstrates that finite étale covers of affine space in characteristic ee14 can arise from highly nontrivial torsorial geometry while retaining rationality of the covering varieties.

Moduli of Jacobian maps

The authors study affine moduli schemes parametrizing étale endomorphisms of ee15 defined by polynomial tuples of bounded degree and with Jacobian determinant equal to ee16. For every degree bound ee17, the corresponding moduli affine scheme is connected.

The smoothness classification is unusually rigid. The moduli scheme is smooth precisely when

ee18

or in the exceptional case

ee19

Thus bounded-degree spaces of Jacobian maps are generally connected but singular. The implication is that connectedness does not reflect a simple geometric interpolation between automorphisms and non-invertible maps: the parameter spaces retain singularities except in low-degree or characteristic-two exceptional cases.

The paper also disproves additional conjectures concerning the geometry of these moduli spaces. Because the parameter spaces are affine schemes rather than merely sets of polynomial tuples, the result provides deformation-theoretic information about families of Jacobian maps, although it does not yield a classification of their irreducible components beyond connectedness.

Structural factorization results

One positive result concerns a restricted factorization class. Suppose an étale endomorphism factors as

ee20

where the first map is quasi-finite and a locally closed embedding in codimension one away from a finite set, and the second map is a coordinate projection. The paper proves that the classical Jacobian Conjecture holds for such maps in characteristic zero. The corresponding characteristic-ee21 degree restriction also holds in positive characteristic.

This result isolates a geometric mechanism under which the conjectural conclusion survives: the first factor has sufficiently controlled codimension-one behavior, while the second factor is a projection. It does not address arbitrary étale endomorphisms, and the finite exceptional subset in the embedding hypothesis is essential to the stated argument.

Limitations and open questions

The paper leaves several central classification problems unresolved. Most importantly, it does not settle the weak positive-characteristic Jacobian conjecture in dimension two. The constructions produce many degrees and arbitrary finite image gaps, but they do not determine the complete set of degrees prime to ee22 that can occur for étale endomorphisms of ee23.

The paper also does not resolve the surjectivity conjecture over ee24 in dimension two. Its positive-characteristic examples cannot be transferred directly to characteristic zero, and the authors explicitly identify this as an open problem. Similarly, the classification of all possible sequences of image-complement cardinalities under iteration remains open, despite the general monotonicity and boundedness results.

The geometric classification of the normalizations ee25 is also incomplete. When the associated Jacobian variety is regular but non-étale, the paper proves that ee26 is not a complete intersection, but no general classification of these varieties is available. The hypersurface and sphere-like constructions describe substantial families rather than all possibilities.

Finally, the moduli results establish connectedness and characterize smoothness, but they do not describe the singularities, normalization, irreducible components, or deformation theory of the moduli schemes in general. These questions are especially relevant because singularity of the moduli space is the generic behavior outside the stated exceptional cases.

Conclusion

The paper develops a broad geometric and algebraic theory of étale endomorphisms of affine spaces, with emphasis on invariants that detect image defects, fiber cardinalities, iteration, normalization geometry, and singularities. Its principal conclusions are negative for several positive-characteristic Jacobian conjectures: étale maps can have arbitrary finite image gaps in dimension two, degrees not divisible by the characteristic, and highly nontrivial associated covers. At the same time, the paper proves positive structural results for quasi-finite iteration, sphere-like normalizations, complete-intersection criteria, moduli connectedness, and restricted factorizations. The unresolved dimension-two degree and surjectivity problems delimit the scope of the constructions and identify the main questions left by the theory (2609.05746).

Whiteboard

Explain it Like I'm 14

1. What is this paper about?

This paper studies polynomial functions that send an affine space to itself. An affine space is the algebraic version of ordinary space: for example,

  • A1\mathbb A^1 is like a line,
  • A2\mathbb A^2 is like a plane,
  • A3\mathbb A^3 is like ordinary three-dimensional space.

The authors focus especially on maps whose Jacobian determinant is a nonzero constant. Informally, this means the map is locally well-behaved: near every point, it does not suddenly squash space in a bad way.

These maps are connected to the famous Jacobian Conjecture, which asks:

If a polynomial map of complex space looks locally reversible everywhere, must it also have a polynomial formula for its global inverse?

The paper investigates related questions, especially over fields with positive characteristic pp. These are mathematical systems where adding a number to itself pp times gives zero. For example, in characteristic $2$, we have $1+1=0$.

2. What questions do the authors ask?

The paper has several main goals.

Understanding “locally reversible” maps

The authors study polynomial maps that are locally like reversible changes of coordinates. In the paper, these are called étale endomorphisms or Kraus–Keller maps.

They ask whether such maps must be globally reversible or at least cover the whole space.

Testing versions of the Jacobian Conjecture

The authors examine several conjectures that are supposed to extend the Jacobian Conjecture to positive characteristic. Their questions include:

  • Must an étale polynomial map be surjective, meaning that every point has at least one point mapping to it?
  • If the map has a small number of preimages for a typical point, must it be an automorphism?
  • Can maps have constant Jacobian but still fail to have a polynomial inverse?
  • How complicated can the missing points and fibers of these maps be?

A fiber is the set of points that map to one chosen point. It is similar to asking, “Which starting locations end at this destination?”

Studying special maps involving the Frobenius operation

In characteristic pp, raising something to the ppth power behaves unusually. For example,

(a+b)p=ap+bp.(a+b)^p=a^p+b^p.

This operation is called the Frobenius map. The authors study special maps of the form

(x1,,xn)(x1+f1(x1p,,xnp),,xn+fn(x1p,,xnp)).(x_1,\ldots,x_n)\mapsto \bigl(x_1+f_1(x_1^p,\ldots,x_n^p),\ldots, x_n+f_n(x_1^p,\ldots,x_n^p)\bigr).

They ask whether equations built from these maps always have solutions.

3. How did the authors investigate these questions?

The research is mainly theoretical. Instead of collecting data from experiments, the authors construct and analyze explicit mathematical examples.

Constructing polynomial maps

The authors build many polynomial maps in two or more variables. They carefully calculate properties such as:

  • the number of preimages of a typical point,
  • the number of points that are never reached,
  • whether the map is surjective,
  • whether the map is étale,
  • how the map behaves when applied repeatedly.

This is similar to testing a machine by examining exactly how many inputs produce each output and whether some outputs can never occur.

Using characteristic pp

The authors take advantage of special rules in characteristic pp, especially the Frobenius operation xxpx\mapsto x^p. These rules allow them to create maps that look locally reversible but behave unexpectedly on a global scale.

For instance, they give a system of equations that has no solution at all, even though it has the special Frobenius form. This disproves the earlier hope that every such system would always have a solution.

Studying geometric objects

The paper also uses ideas from algebraic geometry. Polynomial equations can describe shapes called varieties. For example, an equation such as

x2x+yz=0x^2-x+yz=0

describes a surface in three-dimensional space.

The authors study surfaces, coverings, and missing parts of these geometric objects. One important technique is to enlarge the space associated with a polynomial map so that the map becomes finite and easier to understand. This is somewhat like replacing a complicated map with a larger diagram in which the hidden structure becomes visible.

Defining numerical and geometric invariants

An invariant is a property that stays unchanged when the coordinates are altered in a harmless way. For example, changing from meters to centimeters changes the numbers used to describe a length, but not the actual length.

The authors introduce several invariants, including:

  • Algebraic degree: how complicated the polynomial formulas are.
  • Geometric degree: how many preimages a typical point has.
  • Fiber information: which unusual numbers of preimages can occur.
  • Gap or missed-points number: how many points are not in the image.
  • Iterate invariant: whether repeated applications eventually have the same image.
  • Class rank: information about the boundary or missing pieces of a related geometric space.

These measurements help compare polynomial maps that may look very different.

4. What are the main findings?

Frobenius systems can have no solutions

A major result is that some Frobenius systems in two variables have no solution over any field of characteristic pp.

This is important because it disproves the expectation that these systems are always solvable. It also shows that maps with a constant Jacobian can fail to reach certain points.

Étale maps can miss many points

The authors construct étale endomorphisms of the affine plane whose images leave out any prescribed finite number of points.

In simple terms, for every number ll, they can build a locally well-behaved polynomial map that misses exactly ll points of the plane.

This is surprising because “locally well-behaved” does not guarantee that the map covers the entire space.

The number of preimages can be highly flexible

They construct maps whose typical fibers have many different possible sizes. In particular, in characteristic pp, they produce maps with geometric degree pmpm for many choices of mm.

Thus, a map can have a constant Jacobian and still have several preimages for a typical point instead of being one-to-one.

Several proposed conjectures are false

The paper disproves important proposed versions of the Jacobian Conjecture in positive characteristic.

In particular, the authors show that certain étale maps are not automorphisms even when their degrees satisfy conditions that had been suggested as sufficient. They also disprove related conjectures connected with the work of Adjamagbo and with a generalized conjecture of Bass.

Some of the counterexamples already occur in dimension $2$, while others require dimension $3$ or higher.

Repeated images can behave in complicated ways

For many quasi-finite maps, the image eventually stops changing when the map is repeatedly applied. The paper proves this in several situations.

However, the authors also construct examples in dimensions at least $3$ where

Im(en)Im(en+1)\operatorname{Im}(e^n)\neq \operatorname{Im}(e^{n+1})

for every nn.

In everyday language, repeatedly using the same map can keep changing the set of reachable points forever.

The authors create covers with interesting symmetry

The paper constructs finite coverings of affine spaces with symmetry groups such as SL2(Fpq)\mathrm{SL}_2(\mathbb F_{p^q}), cyclic groups, and the quaternion group Q8Q_8.

A covering is a map that resembles several copies of one space laid over another. The symmetry group describes how those copies can be rearranged without changing the covering.

These examples show that affine spaces can have complicated hidden algebraic structures, especially in positive characteristic.

Some positive results remain

The paper is not only a collection of counterexamples. It also proves positive statements. For example, it shows that certain specially structured polynomial maps do satisfy the classical Jacobian Conjecture over C\mathbb C, and related positive-characteristic versions also hold for maps built in a particular way.

The authors also prove that certain moduli spaces—spaces that classify families of maps—are connected and determine which of them are smooth.

5. Why does this research matter?

The Jacobian Conjecture has been studied for more than a century because it asks a basic question about polynomial maps:

If a map is locally reversible everywhere, is it globally reversible?

This paper shows that the answer becomes much more complicated in positive characteristic. A map can have a constant Jacobian and still:

  • fail to be one-to-one,
  • fail to be onto,
  • miss any chosen finite number of points,
  • have many preimages,
  • and behave unpredictably under repetition.

These examples are valuable because they tell mathematicians which possible shortcuts cannot work. A false conjecture is still useful: it reveals the limits of a theory and helps researchers search for a more accurate statement.

The paper also develops tools for measuring how polynomial maps fail to be invertible. These tools may help with the classical Jacobian Conjecture over the complex numbers, which remains unresolved.

Simple takeaway

The paper investigates polynomial maps that look reversible when viewed locally. The authors show that, especially in systems involving a prime number pp, local reversibility does not guarantee global reversibility. They build precise examples with missing points, multiple preimages, and complicated geometric behavior. At the same time, they prove some positive results for specially constructed maps.

Overall, the research helps mathematicians understand both the power and the limits of the Jacobian Conjecture and its possible generalizations.

Knowledge Gaps

Knowledge gaps, limitations, and open questions

The paper establishes numerous counterexamples and existence results, but leaves the following issues unresolved:

  • Weak Jacobian Conjecture in characteristic pp: It remains unknown whether, for each m>1m>1, there is a bound N(m)N(m) such that every two-dimensional Kraus–Keller map over an algebraically closed field of characteristic pN(m)p\ge N(m) has geometric degree $1$ or greater than mm.
  • Effective bounds for N(m)N(m): Even if the Weak Jacobian Conjecture is true, no explicit or asymptotically meaningful formula is known for N(m)N(m); the proposed possibility N(m)=m+1N(m)=m+1 is unresolved.
  • Classification of attainable degrees in dimension $2$: For a fixed prime pp, it is unknown which integers mp+1m\ge p+1 relatively prime to pp occur as the geometric degree of an étale endomorphism of the affine plane over an algebraically closed field of characteristic pp.
  • Surjectivity over C\mathbb C: The paper does not determine whether every complex Kraus–Keller map in dimension $2$ is surjective. It also remains unclear whether the constructions developed here can be adapted either to prove or to disprove this Surjectivity Conjecture.
  • Iterated image-complement sequences: For characteristic pp, there is no classification of the non-decreasing eventually constant sequences (an)(a_n) that can arise as the cardinalities of the complements of the images of successive iterates ene^n of an étale endomorphism of the affine plane.
  • Higher-dimensional iterate behavior: The paper gives examples in dimensions at least $3$ for which Im(en)Im(en+1)\operatorname{Im}(e^n)\neq\operatorname{Im}(e^{n+1}) for every nn, but does not characterize which varieties, dimensions, or complement geometries permit such non-stabilizing behavior.
  • Classification of associated Jacobian varieties: A general classification of the Jacobian varieties ψe\psi_e associated with étale endomorphisms of affine spaces remains open, particularly in dimensions n3n\ge 3.
  • Classification of finite covers XeX_e: Equivalently, finite covers of affine space arising as normalizations XeX_e are not classified. The difficulty is especially pronounced when the associated Jacobian variety is regular but non-étale.
  • Structure of non-complete-intersection examples: When ψe\psi_e is regular and non-étale, the associated XeX_e is shown not to be a complete intersection, but the possible structures, embeddings, and defining equations of such varieties remain unexplored.
  • Minimal embeddings of XeX_e: The paper does not provide a general method for determining the smallest embedding dimension or constructing explicit embeddings of XeX_e into affine or projective space, particularly when the auxiliary parameter ded_e is at least $2$.
  • Intrinsic classification via hypersurfaces: Hypersurface models are introduced as an approach to studying XeX_e, but the paper shows that hypersurfaces alone do not suffice to classify the invariant Jn(K)\mathbb J_n(K). A broader intrinsic description is still lacking.
  • Classification of affine open embeddings: The pairs (f,g)(f,g) for which the hypersurface Hf,g,Kn\mathbb H^n_{f,g,K} is normal and contains an open subvariety isomorphic to AKn\mathbb A^n_K have not been classified.
  • Classification of smooth finite étale hypersurface covers: In positive characteristic, the finite étale covers of AKn\mathbb A^n_K arising from smooth hypersurfaces Hf,g,Kn\mathbb H^n_{f,g,K} are not classified, even within the specified polynomial families.
  • Relation between the six invariants: The paper studies algebraic degree, geometric degree, non-generic fibers, iterate stabilization, image gaps, and Jacobian varieties, but does not establish a general set of relations determining one invariant from the others or characterize which combinations of values are realizable.
  • Realizability beyond the constructed examples: The existence theorems produce broad families of examples, but they do not determine the full range of possible tuples

(π(e),deg(e),F(e),ι(e),φ(e),ρ(e)).\bigl(\pi(e),\deg(e),\mathcal F(e),\iota(e),\varphi(e),\rho(e)\bigr).

In particular, compatibility constraints among these invariants remain largely unknown.

  • Explicitness and uniformity of constructions: Although the proofs are described as constructive, several results are existential and do not yield uniform formulas for the corresponding polynomial maps. More explicit constructions and degree bounds are needed, especially in arbitrary dimension.
  • Generalization beyond algebraically closed fields: Many principal results are formulated over algebraically closed fields, and the behavior of the constructions over non-algebraically closed fields, finite fields, or more general base schemes is not systematically determined.
  • Characteristic-zero analogues: The methods produce strong phenomena in characteristic pp, but it remains unclear which constructions or invariant behaviors have genuine characteristic-zero analogues and which depend essentially on Frobenius.
  • Frobenius systems: The paper disproves consistency of Frobenius systems in general, but does not identify broad necessary or sufficient conditions—such as degree, sparsity, or geometric conditions on the fif_i—under which an FnF_n-system must be consistent.
  • Basic endomorphisms versus general étale endomorphisms: The extent to which the behavior of basic endomorphisms represents all étale endomorphisms is unresolved. In particular, it is unknown whether the observed fiber, degree, and image-complement phenomena can occur outside the basic-endomorphism framework in fundamentally different ways.
  • Fundamental groups of the constructed varieties: The examples demonstrate nontrivial prime-to-pp covers in selected cases, but there is no general determination of the étale fundamental groups of the intermediate varieties or of the varieties XeX_e associated with arbitrary étale endomorphisms.
  • Cancellation and rigidity questions: The paper does not determine when two varieties or covers arising from different endomorphisms are isomorphic, stably isomorphic, or distinguishable by the proposed invariants; corresponding cancellation and rigidity problems remain open.
  • Scope of the available results: The supplied text ends during the development of the surface constructions, so any conclusions about later sections, additional examples, or further open problems cannot be assessed from the provided material.

Practical Applications

Immediate Applications

  • Benchmarking and stress-testing the Jacobian Conjecture in positive characteristic (mathematics and symbolic computation).
    • Jacobian determinants and étaleness;
    • polynomial-map invertibility;
    • injectivity, surjectivity, and fiber structure;
    • normalization and finite-field extensions;
    • behavior under composition and iteration.
    • Dependencies: Implementations must work over algebraically closed fields or suitable finite-field extensions and must correctly account for positive-characteristic phenomena, especially inseparability and Frobenius powers.
  • Construction of counterexample libraries for automated conjecture testing.
    • the Surjectivity Conjecture in characteristic pp;
    • Adjamagbo’s Jacobian Conjecture analog;
    • generalized and unirational Jacobian conjectures.
    • Such a library could support symbolic regression, theorem-prover evaluation, and automated detection of invalid assumptions such as “constant Jacobian implies surjectivity” in characteristic pp.
    • Dependencies: The examples must be translated into machine-readable polynomial tuples, together with metadata for the characteristic, degree, non-generic fibers, and image complement.
  • Testing and validation of computer algebra routines for polynomial maps. The invariants introduced in the paper—algebraic degree, geometric degree, non-generic fiber sets, iterate index, gap size, and Jacobian-variety class rank—suggest a practical test suite for systems such as SageMath, Magma, Singular, and custom algebraic-geometry software. A workflow could:

    1. input an endomorphism e:AnAne:\mathbb A^n\to\mathbb A^n;
    2. compute its Jacobian matrix and geometric degree;
    3. determine exceptional fibers;
    4. normalize the associated coordinate ring;
    5. estimate or compute the complement of the image;
    6. compare the result with the paper’s predicted invariants. Dependencies: Exact computation of image complements and normalizations can be expensive; many procedures are practical only in low dimensions or for the explicitly constructed families.
  • Positive-characteristic training examples for algebraic geometry education.

    • Frobenius and pp-basis methods;
    • étale morphisms and quasi-finite maps;
    • normalization and finite covers;
    • failures of characteristic-zero intuition;
    • the distinction between generic and global behavior.
    • Dependencies: Examples should be accompanied by simplified computations, since the full formulas can be highly complicated.
  • Design of algebraic-geometry research workflows based on invariant extraction. The six invariant families proposed in the paper can be used now as a standardized descriptive vocabulary for polynomial endomorphisms. Researchers can record an endomorphism using a profile such as
    1
    2
    
    (algebraic degree, geometric degree, exceptional fibers,
     iterate index, image-gap size, Jacobian-variety class)
    This facilitates comparison of examples across fields, dimensions, and characteristics. Dependencies: Some invariants are existential or require normalization, so practical use may initially be limited to explicitly tractable examples.
  • Finite-cover and torsor construction in arithmetic geometry.
    • the study of étale fundamental groups;
    • nontrivial finite covers of rational affine varieties;
    • torsors and descent;
    • examples of non-factorial affine varieties;
    • computational investigations of fundamental groups in characteristic pp.
    • Dependencies: The constructions are field- and characteristic-sensitive, and their geometric interpretation may change after base extension.
  • Testing assumptions in symbolic and numerical algebraic geometry. The examples show that a polynomial map may have constant Jacobian determinant and still possess nontrivial fibers or fail to be surjective in positive characteristic. They can therefore be used to prevent software or modeling pipelines from inferring global invertibility solely from local Jacobian information. This is relevant to symbolic elimination, algebraic system solving, and formal verification of polynomial transformations. Dependencies: The result concerns exact algebraic settings; numerical analogies over R\mathbb R or C\mathbb C require separate validation.
  • Construction of finite-field and algebraic-geometry datasets for machine learning.
    • degree;
    • fiber cardinalities;
    • image deficits;
    • étale or non-étale associated varieties;
    • iteration behavior.
    • These datasets could train or evaluate models that classify polynomial maps, predict normalization properties, or identify exceptional loci.
    • Dependencies: Dataset generation must preserve exact field arithmetic and avoid treating finite-field behavior as representative of characteristic zero.

Long-Term Applications

  • A computational classification platform for étale endomorphisms of affine space.
    • normalize polynomial tuples;
    • compute canonical or semi-canonical representatives;
    • compare geometric degrees and exceptional-fiber profiles;
    • classify associated Jacobian varieties;
    • search for maps with prescribed image complements or iteration sequences.
    • Dependencies: This requires major advances in effective normalization, isomorphism testing, invariant computation, and classification of affine varieties.
  • A systematic theory of image defects and iterated polynomial dynamics. The examples with prescribed finite complements and non-stabilizing sequences

Im(en)Im(en+1)\operatorname{Im}(e^n)\neq \operatorname{Im}(e^{n+1})

suggest a broader theory of image dynamics for algebraic endomorphisms. Potential applications include: - classification of eventual image stabilization; - algebraic analogs of absorbing states and reachable-state analysis; - study of polynomial transition systems; - analysis of how exceptional loci evolve under iteration. Dependencies: A useful theory would require new structural results beyond the low-dimensional constructions and would need to distinguish set-theoretic, scheme-theoretic, and geometric image behavior.

  • Algorithmic detection of positive-characteristic obstructions to invertibility.
    • solving Frobenius systems;
    • computing pp-bases;
    • detecting Artin–Schreier structures;
    • examining induced function-field extensions.
    • Dependencies: General decision procedures are likely difficult or undecidable in broad settings. Practical algorithms would probably require bounded degree, fixed dimension, or restricted polynomial families.
  • Transfer of techniques to arithmetic and geometric cryptography.
    • algebraic pseudorandom mappings;
    • structured finite-field coverings;
    • hard-to-invert polynomial transformations;
    • cryptographic use of nontrivial torsors and non-factorial varieties.
    • Dependencies: The paper does not establish cryptographic security. Any application would require resistance to inversion, collision, and algebraic attacks, as well as careful treatment of field-size and implementation constraints.
  • Improved models for algebraic system solving over finite and imperfect fields.
    • coding theory;
    • finite-field constraint systems;
    • arithmetic geometry;
    • polynomial dynamical systems;
    • algebraic statistics over finite fields.
    • Dependencies: The systems in the paper can be deliberately inconsistent even over algebraically closed fields, so algorithms must not assume consistency merely from étaleness or a constant Jacobian.
  • Classification of affine hypersurfaces and generalized Danielewski-type varieties.
    • databases of explicit affine varieties and their covers;
    • algorithms for detecting normality, smoothness, and affine-space embeddings;
    • new examples relevant to cancellation problems;
    • structural results on automorphism groups and group actions.
    • Dependencies: The classification of the associated normalizations and embeddings is explicitly identified as an open and technically difficult problem.
  • New approaches to the classical Jacobian Conjecture over characteristic zero.
    • testing the Surjectivity Conjecture over C\mathbb C;
    • identifying which positive-characteristic phenomena have characteristic-zero analogs;
    • constructing new obstructions or reduction strategies for the classical Jacobian Conjecture.
    • Dependencies: Positive-characteristic counterexamples cannot be transferred directly to characteristic zero. Any such application requires a valid lifting, specialization, or deformation argument.
  • Policy and research-infrastructure implications for mathematical software.
    • separability versus inseparability;
    • étaleness versus finiteness;
    • generic versus global degree;
    • detected exceptional fibers;
    • assumptions used in invertibility claims.
    • Dependencies: This requires agreement on interoperable metadata standards and the development of certified algorithms for the relevant invariants.

Glossary

  • Affine space: A geometric space modeled on a finite-dimensional vector space, defined algebraically by a polynomial coordinate ring. “the affine spaces over kk
  • Affine variety: A variety that can be realized as a closed subset of affine space. “Let XX be a non-empty affine variety over kk
  • Algebraically closed field: A field in which every nonconstant polynomial has a root, equivalently every polynomial factors completely into linear factors. “over an algebraically closed field kk of characteristic pp
  • Artin–Schreier cover: A covering in characteristic pp defined by an additive equation of the form upu=au^p-u=a. “is of Artin--Schreier type”
  • Automorphism: An invertible morphism from a mathematical object to itself. “of automorphisms of ARn\mathbb A^n_R over SpecRSpec R
  • Basic endomorphism: An endomorphism that, up to composition with automorphisms, is defined by a Frobenius system. “By a basic endomorphism of ARn\mathbb A^n_R we mean an endomorphism eEndn(R)e\in End_n(R) such that an element of GAn(R)eGAn(R)GA_n(R)eGA_n(R) is defined by an FnF_n-system over RR.”
  • Cohen–Macaulay: A property of a commutative ring or scheme whose depth equals its Krull dimension at every local ring. “the local rings of XeX_e of dimension $2$ are Cohen--Macaulay”
  • Complete intersection: A variety or scheme defined locally by the minimum possible number of equations. “with the help of other smooth affine hypersurfaces that contain affine spaces as open dense subvarieties”
  • Conjugacy class: The collection of objects obtained from one object by conjugation by invertible elements. “in the conjugacy class {aea1aGAn(K)}\{aea^{-1}|a\in GA_n(K)\}
  • Dominant morphism: A morphism whose image is dense in the target. “a dominant map ((x_1,\ldots,x_n)\mapsto \bigl(f_1(x_1,\ldots,x_n),\ldots,f_n(x_1,\ldots,x_n)\bigr)\”
  • Endomorphism: A morphism from a mathematical object to itself. “Let Endn(R)End_n(R) be the monoid of endomorphisms”
  • Étale morphism: A morphism that is algebraically analogous to a local isomorphism and is both smooth of relative dimension zero and unramified. “étale endomorphisms of affine planes”
  • Finite morphism: A morphism corresponding algebraically to a finite module extension of coordinate rings. “the composite of the finite morphism XeAK,tnX_e\rightarrow \mathbb A^n_{K,t}
  • Finite field extension: An extension of fields in which the larger field has finite dimension as a vector space over the smaller field. “the degree of the finite field extension Kt(x1,,xn)Ks(x1,,xn)K_t(x_1,\ldots,x_n)\rightarrow K_s(x_1,\ldots,x_n)
  • Flat morphism: A morphism whose associated ring map makes the target ring a flat module over the source ring. “the finite morphism XeAK,tnX_e\rightarrow\mathbb A^n_{K,t} is flat”
  • Frobenius system: A system of polynomial equations in characteristic pp involving pp-th powers of the variables. “we call a Frobenius system”
  • Fundamental group: An invariant encoding the connected finite étale covers of a space. “its prime-to-$2$ fundamental group is non-trivial”
  • Galois cover: A finite covering associated with a Galois extension of function fields and equipped with a transitive group of deck transformations. “There exists a finite Galois cover”
  • Geometric degree: The separable degree of the function-field extension induced by a morphism, equivalently the number of points in a generic fiber in the stated setting. “whose geometric degree is not a multiple of pp
  • Hypersurface: A subvariety defined by a single equation in an ambient variety. “the hypersurface of Spec(C[t,x,y,z])Spec(\mathbb C[t,x,y,z]) defined by the equation x+x2y+z2+t2=0x+x^2y+z^2+t^2=0
  • Ind-group scheme: A group object represented by a direct system of schemes, often used for infinite-dimensional transformation groups. “the ind-group scheme Aut(SK2)Aut(\mathbb S^2_K)
  • Invariant: A quantity or structure unchanged under a specified class of transformations. “We use the following formalism of invariants for such endomorphisms.”
  • Jacobian Conjecture: The conjecture that a polynomial map with constant nonzero Jacobian determinant has a polynomial inverse. “The classical Jacobian Conjecture”
  • Jacobian determinant: The determinant of the matrix of first partial derivatives of a polynomial map. “has a non-zero constant Jacobian determinant”
  • Jacobian variety: In this paper, an endomorphism of a normal connected variety whose image is open, whose induced map to its image is finite, and whose restriction to the image is étale. “By a Jacobian variety over KK of dimension n2n\ge 2 we mean an endomorphism”
  • Kraus–Keller map: A polynomial self-map with nonzero constant Jacobian determinant; in the paper, an étale endomorphism of affine space. “the corresponding polynomial map, to be called a (complex) Kraus--Keller map”
  • Lang torsor: A torsor arising from the Lang map on an algebraic group over a finite field, typically gg1F(g)g\mapsto g^{-1}F(g). “the pullback of Lang 2(Fpq)_2(\mathbb F_{p^q})-torsor”
  • Locally closed embedding: An embedding whose image is the intersection of an open and a closed subvariety. “a locally closed embedding in codimension $1$”
  • Monoid: An algebraic structure with an associative multiplication and an identity, but not necessarily inverses. “Let Endn(R)End_n(R) be the monoid of endomorphisms”
  • Normalization: The process of replacing an integral scheme by the normal scheme obtained by taking the integral closure of its coordinate ring. “Let XeX_e be the normalization of AK,tn\mathbb A^n_{K,t}
  • Normal variety: A variety whose local rings are integrally closed domains. “an endomorphism ψ:XX\psi:X\rightarrow X of a normal connected variety”
  • Quasi-finite morphism: A morphism whose fibers are finite. “all quasi-finite endomorphisms of ARn\mathbb A^n_R
  • Quasi-isomorphism: In the paper, a pair of isomorphisms relating the source and target of two Jacobian varieties while preserving their images and maps. “a pair (ı~s,ı~t)(\tilde\imath_s,\tilde\imath_t) of isomorphisms”
  • Reduction to positive characteristic: A method that studies characteristic-zero objects by reducing their defining data modulo primes. “whose original proof is by reduction to positive characteristic”
  • Reduced scheme: A scheme whose coordinate rings contain no nonzero nilpotent elements. “For a scheme WW, let WredW_{red} be its associated reduced scheme.”
  • Regular locus: The subset of a variety consisting of its regular, or nonsingular, points. “let Reg(X):=XSing(X)Reg(X):=X\setminus Sing(X) be its regular locus”
  • Separable polynomial: A polynomial with no repeated roots, equivalently one relatively prime to its derivative over a field. “a monic separable polynomial”
  • Separable degree: The degree of the separable part of a finite field extension. “deg(e)\deg(e) is the separable degree of the finite field extension”
  • Smooth variety: A variety with no singularities in the scheme-theoretic sense. “the affine smooth surface”
  • Spec: The spectrum construction assigning a scheme to a commutative ring through its prime ideals. “An=Spec(Z[x1,,xn])\mathbb A^n=Spec(\mathbb Z[x_1,\ldots,x_n])
  • Tangent bundle: The vector bundle whose fiber at each point is the tangent space at that point. “If XX is regular, let TXT_X be the tangent bundle over XX.”
  • Torsor: A space with a group action that is locally, but not canonically, isomorphic to the group itself. “The finite Galois cover VknAkn\mathbb V_k^n\rightarrow\mathbb A^n_k
  • Unimodular tuple: A tuple of elements generating the unit ideal. “the nn-tuple (g1,,gn)Rn(g_1,\ldots,g_n)\in\mathcal R^n is unimodular”
  • Variety: A reduced scheme of finite type over a field, as specified by the paper. “by a variety XX over KK we mean a reduced scheme of finite type”
  • Zariski dense: A subset whose closure in the Zariski topology is the entire ambient space. “the union indexed by qNq\in\mathbb N^{\ast} of the set of solutions”
  • Zariski’s Main Theorem: A theorem describing quasi-finite separated morphisms as factoring through an open immersion followed by a finite morphism. “we have an open embedding ıe:AK,snXe\imath_e:\mathbb A^n_{K,s}\rightarrow X_e by Zariski's Main Theorem”

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Tweets

Sign up for free to view the 3 tweets with 357 likes about this paper.