Counterexamples to the Jacobian conjecture in dimensions greater than two

arXiv:2608.00222 · math.AG · Submitted 2026-07-31 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: I'm Ada.

Emmy: And I'm Emmy, here with Kolya.

Ada: We're looking at the paper: "Counterexamples to the Jacobian conjecture in dimensions greater than two".

Emmy: Here's the gist of the paper. The paper presents counterexamples to the Jacobian conjecture in dimensions greater than two by developing a tangent-sweep mechanism generalized from plane curves to direction fields on hypersurfaces, producing explicit counterexamples in every dimension and of arbitrarily large geometric degree The Tangent-Sweep Mechanism The core idea is a classical mechanism where one sweeps the tangent lines of a plane curve, which is forced by projective duality to be many-to-one, and then conjugates by monomial maps so that the Jacobian factor picked up by the sweep is cancelled exactly The construction produces counterexamples in every dimension greater than two and, in each dimension, of arbitrarily large geometric degree The mechanism is isolated and named the tangent sweep by Speyer

14: Explicit Counterexamples The authors provide five new explicit maps: one three-dimensional of degree four, two four-dimensional of degrees five and ten, and two fivedimensional of degrees six and twelve These counterexamples provide explicit examples of ´etale coverings C n → C n that are not proper: they are everywhere unramified, and fail to be injective only through points escaping to infinity The construction propagates its own counterexamples along n 7→ n + 3 General Construction Framework A general framework is proposed for any dimension n > 2 by sweeping a tangent direction field on a parametrized hypersurface of dimension n − 2 in C n−1, padding by one multiplier variable, and twisting by monomial maps This framework is set up at once for arbitrary dimension n ≥ 3 The padded sweep is F0(x, γ, w) = γx X(w) + γ∆(w): C n → C n Fiber Structure and Results The fiber stratifications are classical, with contact orders of tangent objects through the target point For F this gives fiber sizes 3, 1, 0 (Theorem 3.4) For G, Theorem A.1 proves the sizes are exactly 4, 3, 2, 1, and 0 over every stratum The final maps F6 and F7 illustrate the two mechanisms available in dimension five: a mixed direction field on its flexible bottom branch and the curve-type field with transverse Keller-type data Open Problems Several questions suggest themselves, including whether the middle branch of the mixed direction field (1, w1, w2 1, w2)T is genuinely rigid The minimal geometric degree of a Keller counterexample in dimension n ≥ 4 remains an open question The paper also notes that for F6 and F7 the univariate fiber polynomials RY from the proofs of Theorems 4.5 and 4.13 reduce the stratifications to discriminants of a sextic and a degree-twelve polynomial in the target coordinates, which are left to be unwound Acknowledgements The computations reported here were carried out in exact rational arithmetic; scripts are available from the author The authors acknowledge Claude Fable 5 assisted in the proofs and in the writing up of the paper The main idea and framework are due to the author, and Claude Fable 5 assisted in the proofs and in the writing up of the paper Graph Ideals For a polynomial map Φ = (ϕ1,..., ϕn)T: C n → C n, the graph ideal is JΦ = ϕ1 − v1,..., ϕn − vn ⊆ Q

x1,.., xn, v1,.., vn: The element h1 generates the elimination ideal JF ∩ Q

x, v: : written by its (z, y, x)-grouping The leading coefficient is the curve equation: Resw W, X − p(w) = −1/64E and discw(W) = −1/64E for the tangency quartic The element h1 generates the elimination cubic referred to in Theorem 3.4(1); and this is the elimination cubic referred to in Theorem 3.4(1) References

1: H.

Ada: Bass, E.

Emmy: H.

Ada: Connell and D.

Paper summary: Ada: So, what’s the core idea here? It seems like they are tackling the Jacobian conjecture again.

Emmy: The paper claims they found counterexamples in every dimension larger than two, and these examples can have an arbitrarily large geometric degree.

Kolya: An arbitrarily large degree means that for any number you pick, there's a map that has more and more preimages of that point as you move around the space. That’s pretty significant if we think about the expected behavior of these maps.

Ada: Exactly, Kolya. The authors use a specific mechanism called the tangent sweep mechanism to show how this can happen in higher dimensions, generalizing it from plane curves to direction fields on hypersurfaces <ref:2608.00222#pg1>.

Emmy: They are developing this tangent sweep by taking the concept from plane curves—where it’s forced to be many-to-one due to duality—and applying it to direction fields on hypersurfaces, adding a multiplier variable and twisting by monomial maps <ref:2608.00222#pg3>.

Kolya: So, what does this mechanism actually do in terms of the output? I mean, does it produce counterexamples that are proper, or something else?

Ada: The results show they create explicit maps C n to C n that are everywhere unramified but not proper <ref:2608.00222#pg1>. This means they fail to be injective only through points escaping to infinity, which is a loophole the Jacobian hypothesis never excluded <ref:2608.00222#pg3>.

Emmy: They give five new explicit examples: one in three dimensions with degree four, two in four dimensions with degrees five and ten, and two in five dimensions with degrees six and twelve <ref:2608.00222#pg1>. These specific maps are a concrete way to see the mechanism working <ref:2608.00222#pg1>.

Kolya: And they mention that the construction propagates its own counterexamples along n seven to n plus three <ref:2608.00222#pg3>. That suggests a way to build off these examples in even higher dimensions, or perhaps just show that the construction is robust across different dimensions.

Ada: It seems like the real power of this paper is showing that you don't need dimension three to see this kind of behavior; it works in every dimension greater than two <ref:2608.00222#pg1>. The explicit maps they provide, like F6 and F7 in five dimensions, illustrate different ways this mechanism can be deployed <ref:2608.00222#pg3>.

Paper summary: Emmy: I’m curious about the fiber structure they find. They talk about classical stratifications based on contact orders of tangent objects through the target point <ref:2608.00222#pg3>. For one example, F, these gives fiber sizes of three, one, and zero <ref:2608.00222#pg3>.

Kolya: That fiber size information is interesting because it tells you about the local geometry around a specific point on the target space. It seems like they also have more detailed structure for another case, G, where they prove the sizes are exactly four, three, two, one, and zero over every stratum <ref:2608.00222#pg3>.

Ada: That level of detail in fiber counting is what allows them to verify these maps using exact rational arithmetic <ref:2608.00222#pg3>. They say the verification scripts are available from the author, which is a big deal for confirming these complex algebraic identities <ref:2608.00222#pg3>.

Emmy: I also noticed they discuss some open questions. They ask if the middle branch of this mixed direction field is genuinely rigid <ref:2608.00222#pg3>. Also, the minimal geometric degree of a Keller counterexample in dimension four remains an open question <ref:2608.00222#pg3>.

Kolya: So, even with these explicit examples and the general framework for any n greater than two, there are still lingering questions about rigidity and finding the absolute minimum degree for a specific dimension <ref:2608.00222#pg3>.

Ada: And they mention that for some of their maps, like F6 and F7, the univariate fiber polynomials reduce the stratifications to things like discriminants of a sextic or a degree twelve polynomial in the target coordinates, which they have left to be unwound <ref:2608.00222#pg3>. That’s where things get computationally heavy.

Emmy: It sounds like while they have built a powerful general framework for any dimension n greater than two, fully understanding the full structure of these counterexamples still requires solving some complex algebraic problems <ref:2608.00222#pg3>.

Kolya: For someone just listening to this, what does this all mean practically? It means we have a tool—this tangent sweep—that systematically generates examples that challenge the Jacobian conjecture in higher dimensions, and it shows we can keep finding new ones as the dimension increases <ref:2608.00222#pg1>.

Ada: The main point is that the construction produces counterexamples in every dimension greater than two and of arbitrarily large geometric degree <ref:2608.00222#pg1>. This suggests that if you're looking for a map where the Jacobian determinant is a non-zero constant but it doesn't have an inverse, this method gives you a systematic way to build increasingly complex ones.

Paper summary: Emmy: It’s about showing that the structure of these counterexamples isn't just limited to low dimensions; they can be scaled up indefinitely, which is what the paper describes by propagating its own examples along n seven to n plus three <ref:2608.00222#pg3>.

Kolya: So, for a listener, this means there's a reliable way to generate examples of the Jacobian conjecture failing in any dimension greater than two, and these examples can get geometrically very complicated <ref:2608.00222#pg1>. The caveat is that the full picture of the structure for every single example still has some open questions regarding rigidity and minimal degrees <ref:2608.00222#pg3>.

Ada: So, to sum up, this paper presents a general framework using a tangent sweep mechanism that produces explicit counterexamples in every dimension greater than two with arbitrarily large geometric degree <ref:2608.00222#pg1>. It’s about showing the scope of where the conjecture fails when you move beyond three dimensions <ref:2608.00222#pg1>.

Emmy: And they provide specific, verifiable maps—one in three dimensions of degree four, and others in four and five dimensions—to demonstrate this construction <ref:2608.00222#pg1>.

Kolya: We need to remember that the paper states its limitations regarding the full structure; it notes that for some maps, like F6 and F7, the univariate fiber polynomials reduce things to specific polynomial degrees, which they still need to work through <ref:2608.00222#pg3>.

Ada: That’s right. The paper is a detailed account of this tangent sweep mechanism and how it scales to higher dimensions <ref:2608.00222#pg1>. It’s about the systematic generation of non-proper maps that bypass the Jacobian hypothesis <ref:2608.00222#pg3>.

Emmy: So, it moves the study of this conjecture beyond dimension three by giving us a method to construct examples in every higher dimension <ref:2608.00222#pg1>.

Kolya: That’s the essence: a systematic construction for higher dimensions with increasing complexity <ref:2608.00222#pg1>.

Ada: We’ve covered the summary and the conclusion of this paper on "Counterexamples to the Jacobian conjecture in dimensions greater than two" <ref:2608.00222#pg1>.

Emmy: It shows that even when we move beyond three dimensions, there are still ways to construct maps that defy simple inversion, but it also points out where the structural understanding still needs more work <ref:2608.00222#pg3>.

Kolya: We’ve discussed the general mechanism and the explicit examples provided by Gao et al. <ref:2608.00222#pg1>.

Ada: That’s all for this discussion on the paper's summary and conclusions <ref:2608.00222#pg1>.

Conclusion: Ada: So we're wrapping up this look at this paper, "Counterexamples to the Jacobian conjecture in dimensions greater than two." Basically, they’ve shown that you don't need dimension three for these kinds of counterexamples; it works in every dimension larger than two.

Emmy: It’s about the tangent sweep mechanism—taking a concept from plane curves and generalizing it to direction fields on higher dimensional hypersurfaces—to systematically produce maps that fail to be proper. It’s the structure behind the failure that’s really interesting here.

Kolya: I see how they build these up, propagating their own examples along dimensions, showing you can keep making them more complex geometrically as you move from dimension seven to whatever comes next. What this means for us is that if we are looking for maps where the Jacobian determinant is a constant but the map isn't actually invertible in a nice way, this gives us a way to generate those examples systematically across any dimension above two.

Ada: Exactly. The authors provide explicit maps in three, four, and five dimensions, with increasing geometric degrees like four and ten for four dimensions <ref:2608.00222#pg1>. It’s about showing the scope of where the Jacobian conjecture fails when you go past three dimensions.

Emmy: That systematic generation is powerful because it connects this specific mechanism to a whole family of examples, which gives us a structural understanding <ref:2608.00222#pg1>. The fiber sizes they calculate, like three, one, and zero for one specific example <ref:2608.00222#pg3>, show how the local geometry behaves in these counterexamples.

Kolya: But you have to remember that while this framework is general for any dimension greater than two, there are still open questions about the rigidity of certain components, like that middle branch of a direction field <ref:2608.00222#pg3>. It means we’ve built the engine, but we haven't fully mapped out every single part of what it produces yet.

Ada: Right. So while they have this tool to generate these non-proper maps in higher dimensions, understanding the precise structure and minimizing the degree for any given dimension is still on the table <ref:2608.00222#pg3>. That leaves a lot of room for future investigation into what these constructions are actually capable of producing.

Clemson University

math.AG

Submitted: 2026-07-31

Updated: 2026-11-04

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 90/100

The gist: The paper presents counterexamples to the Jacobian conjecture in dimensions greater than two by developing a tangent-sweep mechanism generalized from plane curves to direction fields on…

Key concepts

Tangent-Sweep Mechanism
This is a core technique where one sweeps tangent lines of a curve. It's used to generate many-to-one maps and then uses monomial maps to precisely cancel out the Jacobian factor, leading to counterexamples.
Direction Fields on Hypersurfaces
Instead of working with simple plane curves, the method is extended to study direction fields defined on hypersurfaces in higher dimensions. This allows the construction of counterexamples in any dimension $n>2$ by sweeping tangent directions across these complex surfaces.
Fiber Stratifications
These are classical ways to categorize how the fibers of a map behave. The paper analyzes fiber sizes (like 3, 1, or 0) for different strata, helping to understand the structure of the counterexample maps.

Terminology

Summary

The paper presents counterexamples to the Jacobian conjecture in dimensions greater than two by developing a tangent-sweep mechanism generalized from plane curves to direction fields on hypersurfaces, producing explicit counterexamples in every dimension and of arbitrarily large geometric degree

The Tangent-Sweep Mechanism

The core idea is a classical mechanism where one sweeps the tangent lines of a plane curve, which is forced by projective duality to be many-to-one, and then conjugates by monomial maps so that the Jacobian factor picked up by the sweep is cancelled exactly The construction produces counterexamples in every dimension greater than two and, in each dimension, of arbitrarily large geometric degree The mechanism is isolated and named the tangent sweep by Speyer [14]

Explicit Counterexamples

The authors provide five new explicit maps: one three-dimensional of degree four, two four-dimensional of degrees five and ten, and two fivedimensional of degrees six and twelve These counterexamples provide explicit examples of ´etale coverings C n → C n that are not proper: they are everywhere unramified, and fail to be injective only through points escaping to infinity The construction propagates its own counterexamples along n 7→ n + 3

General Construction Framework

A general framework is proposed for any dimension n > 2 by sweeping a tangent direction field on a parametrized hypersurface of dimension n − 2 in C n−1, padding by one multiplier variable, and twisting by monomial maps This framework is set up at once for arbitrary dimension n ≥ 3 The padded sweep is F0(x, γ, w) = γx X(w) + γ∆(w): C n → C n

Fiber Structure and Results

The fiber stratifications are classical, with contact orders of tangent objects through the target point For F this gives fiber sizes 3, 1, 0 (Theorem 3.4) For G, Theorem A. p. 2

Where the paper says it

Appendix A defines graph ideals and analyzes exact fiber sizes: for F we give the full (six-element) reduced lexicographic Gr¨obner basis certifying the stratification of §3. 4; for G we prove a complete fiber-size theorem (Theorem A.

1 proves the sizes are exactly 4, 3, 2, 1, and 0 over every stratum The final maps F6 and F7 illustrate the two mechanisms available in dimension five: a mixed direction field on its flexible bottom branch and the curve-type field with transverse Keller-type data

Open Problems

Several questions suggest themselves, including whether the middle branch of the mixed direction field (1, w1, w2 1, w2)T is genuinely rigid The minimal geometric degree of a Keller counterexample in dimension n ≥ 4 remains an open question The paper also notes that for F6 and F7 the univariate fiber polynomials RY from the proofs of Theorems 4.5 and 4.13 reduce the stratifications to discriminants of a sextic and a degree-twelve polynomial in the target coordinates, which are left to be unwound

Acknowledgements

The computations reported here were carried out in exact rational arithmetic; scripts are available from the author The authors acknowledge Claude Fable 5 assisted in the proofs and in the writing up of the paper The main idea and framework are due to the author, and Claude Fable 5 assisted in the proofs and in the writing up of the paper

Graph Ideals

For a polynomial map Φ = (ϕ1,..., ϕn)T: C n → C n, the graph ideal is JΦ = ϕ1 − v1,. p. 6

Where the paper says it

The underlying curve K = (p, q) T is a rational cuspidal cubic with one cusp, at (p, q)( 2 3 ) = ( 4 3 , 4 27 ), and no node; the Pl¨ucker formula gives class m = 3 · 2 − 3 · 1 = 3, matching the fiber count. The fibers of F are completely described by the following stratification, which we established by exact computation (a lexicographic Gr¨obner basis of the graph ideal, together with symbolic lifting certificates over the function fields of the strata).

.., ϕn − vn ⊆ Q[x1,.., xn, v1,.., vn] The element h1 generates the elimination ideal JF ∩ Q[x, v]: written by its (z, y, x)-grouping The leading coefficient is the curve equation: Resw W, X − p(w) = −1/64E and discw(W) = −1/64E for the tangency quartic The element h1 generates the elimination cubic referred to in Theorem 3. p. 6
Where the paper says it

discx of the elimination cubic of the graph ideal equals −c3a 2 2 , and a 2 2−4(4−3v2v3) 3 = 108v 2 3 c3; 2. v 2 3 c3 = E ◦ σ where E(X, Y ) = X3 − X2 − 18XY + 16Y + 27Y 2 is the implicit equation of K, and a2 is (a constant multiple of) the equation of the cuspidal tangent line pulled back by σ; 3.

4(1); and this is the elimination cubic referred to in Theorem 3. p. 6
Where the paper says it

Let c3(v) = 27v 2 1 v 2 3 −18v1v2v3+16v1+v 3 2 v3−v 2 2 and a2(v) = 54v1v 2 3 −18v2v3+16 = ∂c3/∂v1, and let σ(v) = (v2v3, v1v 2 3 ) T. discx of the elimination cubic of the graph ideal equals −c3a 2 2 , and a 2 2−4(4−3v2v3) 3 = 108v 2 3 c3; 2.

4(1)

References

[1] H. Bass, E. H. Connell and D. p. 1

Where the paper says it

map) is a polynomial automorphism. The problem became widely known through the survey of Bass, Connell and Wright [1] and the book of van den Essen [5].

Wright, The Jacobian conjecture: reduction of degree and formal expansion of the inverse, Bull. p. 2
Where the paper says it

Wang [16] proved the conjecture for maps of degree at most 2 in any dimension. Bass–Connell–Wright [1] and Yagzhev [17] showed that it suffices to prove the conjecture for maps of the form x + H(x) with H cubic homogeneous (in all dimensions simultaneously), and Dru˙zkowski [4] reduced further to the case where each component of H is the cube of a linear form.

Amer. Math. Soc (N S) 7 (1982), 287–330 [2] A Bia lynicki-Birula and M Rosenlicht, Injective morphisms of real algebraic varieties, Proc. p. 2
Where the paper says it

Two structural facts frame the counterexample. A theorem of Bia lynicki-Birula and Rosenlicht [2] states that an injective polynomial map C n → C n is bijective with polynomial inverse; hence a Keller counterexample must fail injectivity.

Amer. Math Soc 13 (1962), 200–203 [4] L M Dru˙zkowski, An effective approach to Keller’s Jacobian conjecture, Math Ann 264 (1983), 303–313 [5] A van den Essen, Polynomial Automorphisms and the Jacobian Conjecture, Progress in Mathematics 190, Birkh¨auser, Basel, 2000 [6] Z Jelonek, The set of points at which a polynomial map is not proper, Ann. Polon. Math. 58 (1993), 259–266 [7] O.-H Keller, Ganze Cremona-Transformationen, Monatsh. p. 0
Where the paper says it

If F has a polynomial inverse, the chain rule forces det J(F) to be a nonzero constant. The Jacobian conjecture, raised by Keller [7], asserts the converse: a polynomial map with det J(F) ∈ C × (a Keller ∗AI disclosure: The main idea and framework are due to the author, and Claude Fable 5 assisted in the proofs and in the writing up of the paper.

Math Phys 47 (1939), 299–306 [8] T. p. 0
Where the paper says it

Counterexamples to the Jacobian conjecture in dimensions greater than two Shuhong Gao∗ School of Mathematical and Statistical Sciences Clemson University, Clemson, SC 29634-0975 Email: sgao@clemson. edu July 31, 2026 Abstract The Jacobian conjecture, open since 1939, asks whether every polynomial map of C n whose Jacobian determinant is a nonzero constant must have a polynomial inverse.

-T Moh, On the Jacobian conjecture and the configurations of roots, J Reine Angew. p. 2
Where the paper says it

Bass–Connell–Wright [1] and Yagzhev [17] showed that it suffices to prove the conjecture for maps of the form x + H(x) with H cubic homogeneous (in all dimensions simultaneously), and Dru˙zkowski [4] reduced further to the case where each component of H is the cube of a linear form. In dimension two, Moh [8] verified the conjecture for maps of degree at most 100; the two-dimensional case remains open and is untouched by the counterexamples discussed here, which exist only in dimension ≥ 3 (by Wang’s theorem, degree 2 examples are impossible, and the known constructions produce degree ≥ 3).

Math 340 (1983), 140–212 [10] P. p. 0
Where the paper says it

1 arXiv:2608. 00222v1 [math.

C Rosenbloom and D V Widder, Expansions in terms of heat polynomials and associated functions, Trans Amer Math Soc 92 (1959), 220–243 [11] L Alp¨oge, A counterexample to the Jacobian conjecture in dimension three, announcement, July 19, 2026 [12] A Gallagher, An infinite family of counterexamples to the Jacobian Conjecture in dimension three: every generic fiber degree n ≥ 3 occurs, preprint, July 20, 2026 [13] Secret Blogging Seminar, The new counterexample to the Jacobian conjecture, blog post, July 20, 2026 [14] D E Speyer, The geometry and structure of Gallagher’s counterexamples to the Jacobian conjecture, note, July 23, 2026 [15] T Tao, A digestion of the Jacobian conjecture counterexample, blog post, July 21, 2026 [16] S S-S Wang, A Jacobian criterion for separability, J Algebra 65 (1980), 453–494 [17] A V Yagzhev, On Keller’s problem, Sibirsk. p. 0
Where the paper says it

edu July 31, 2026 Abstract The Jacobian conjecture, open since 1939, asks whether every polynomial map of C n whose Jacobian determinant is a nonzero constant must have a polynomial inverse. It was refuted in dimension three by Alp¨oge on July 19, 2026, with an infinite family by Gallagher (July 20) and a geometric explanation by Speyer (July 23): the counterexample sweeps the tangent lines of a plane curve — a map that classical duality forces to hit most points several times.

Mat. Zh. 21 (1980), 141–150 [2608. p. 0
Where the paper says it

The Jacobian conjecture, raised by Keller [7], asserts the converse: a polynomial map with det J(F) ∈ C × (a Keller ∗AI disclosure: The main idea and framework are due to the author, and Claude Fable 5 assisted in the proofs and in the writing up of the paper. 1 arXiv:2608.

00222> The paper is arXiv:2608. p. 0
Where the paper says it

The Jacobian conjecture, raised by Keller [7], asserts the converse: a polynomial map with det J(F) ∈ C × (a Keller ∗AI disclosure: The main idea and framework are due to the author, and Claude Fable 5 assisted in the proofs and in the writing up of the paper. 1 arXiv:2608.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improve counterexample generation by implementing a general construction based on sweeping tangent direction fields on parametrized hypersurfaces for any dimension n greater than two, as described in section 4.

  2. Enhance symbolic manipulation capabilities to handle exact rational arithmetic and Gr¨obner bases for verifying exact fiber structures, allowing the system to determine exact fiber structures through Gr¨obner bases and verify identities in exact rational arithmetic.

  3. Develop a module capable of solving complex systems of polynomial equations arising from the stage equations (8), including finding solutions via finite-dimensional linear algebra at each order, which is crucial for constructing the maps F4 through F7.

  4. Implement specialized solvers to analyze the branch analysis in dimension five, specifically handling the mixed direction field (1, w1, w2 1, w2)T by identifying and analyzing rigidity phenomena where the middle branch turns out to be rigid on a natural locus.

  5. Integrate algorithms to efficiently calculate and verify the univariate fiber polynomial RY for maps F6 and F7, which are derived from the proofs of Theorems 4.5 and 4.13, enabling automated determination of generic fiber sizes based on the roots of this polynomial.

Abstract

The Jacobian conjecture, open since 1939, asks whether every polynomial map of n whose Jacobian determinant is a nonzero constant must have a polynomial inverse. It was refuted in dimension three by Alpöge on July 19, 2026, with an infinite family by Gallagher (July 20) and a geometric explanation by Speyer (July 23): the counterexample sweeps the tangent lines of a plane curve. We give a self-contained account of this tangent-sweep mechanism and generalize it from plane curves to direction fields on hypersurfaces. The resulting construction, including stabilization of the three-dimensional family, produces counterexamples in every dimension greater than two and, in each dimension, of arbitrarily large geometric degree (the number of preimages of a typical point). We work it out in five explicit maps: one three-dimensional of degree four, two four-dimensional of degrees five and ten, and two five-dimensional of degrees six and twelve. We show that our explicit examples in dimensions n at least4 are not polynomially left-right equivalent to the counterexamples of Alpöge and Gallagher when these are embedded in n by adjoining identity coordinates. The counterexamples provide explicit examples of étale morphisms n to n that are not proper: they are everywhere unramified, and fail to be injective only through points escaping to infinity. All identities were verified in exact rational arithmetic. Part proves the higher-dimensional inequivalence results, and Appendix A determines exact fiber structures for the three-dimensional examples. A complete classification of the fibers of F 4,F 5,F 6,F 7 remains open.

Related papers