The Jacobian conjecture fails for pseudo-planes (Q1989702)

The famous and still open Jacobian Conjecture (JC) asserts that if \(F:\mathbb{C} ^{n}\rightarrow \mathbb{C}^{n}\), \(n\geq 2,\) is a polynomial (=regular) mapping with non-zero constant jacobian then \(F\) is a polynomial isomorphism (=biregular). It is well known that it suffices to prove \(F\) is a proper mapping. M. Miyanishi reformulated JC for a broader class of varieties in the following form: (Generalized Jacobian Conjecture (GJC)) If \(X\) is a smooth complex variety then all étale endomorphisms of \(X\) are proper. There are known classes of complex varieties for which GJC holds and the ones for which GJC does not hold. The authors consider a specific class of smooth complex surfaces: \(\mathbb{Q}\)-homology plane of negative logarithmic Kodaira dimension with an effective action of an infinite algebraic group \(G. \) They prove that \(G\)-invariant GJC holds for such surfaces unless \(G= \mathbb{C}^{\ast }\) and the surface is \(\mathbb{C}^{\ast }\)-equivariantly isomorphic to some \(S(k,r,a)\) for some \(k,r\geq 2\) and \(a\in \{1,\ldots ,k-1\}\) coprime to \(k\), where \(S(k,r,a)\) is a quotient of \(\widetilde{S}(k,r):= \{x^{r}y=z^{k}-1\}\subset \mathbb{C}^{3}\) by the free \(\mathbb{Z}_{k}\) -action \(\varepsilon \ast _{a}(x,y,z):=(\varepsilon x,\varepsilon ^{-r}y,\varepsilon ^{-a}z).\) As a corollary they obtain an important result: For every positive integer \(n\) there exists a smooth variety \(\mathbb{C}^{n}\)-fibered over \(\mathbb{C}\) (hence rational and of negative logarithmic Kodaira dimension) which is simply connected and for which the GJC fails.