Some thoughts on the Jacobian conjecture. I (Q2474468)

A pair \(f,g\in k[X,Y]\) is called a Jacobian pair if \(det(Jac(f,g))\in k^*\). The author reported on p. 181 of \textit{S. S. Abhyankar} [Proc. Indian Acad. Sci., Math. Sci. 104, No. 3, 515--542 (1994; Zbl 0812.13013)] the following: if \(f,g\) is a Jacobian pair and \noindent[1] either: \(H\leq 2\) (two characteristic pair case) \noindent[2] or: \(H=3\) with \(d_3\leq 4\) (plus epsilon case). \noindent Then \((f,g)\) is an automorphism. (The definition of \(H\) and \(d_3\) is quite involved, we refer to section 3 and 11 of part I.) The proof of the ``plus epsilon'' case was left out of the above quoted book, and the proof of this part is the main motivation for the two linked papers (and apparently a part III mentioned in the abstract of part II [J. Algebra 319, No. 3, 1154--1248 (2008; Zbl 1139.14047)]). A new proof of Jung's automorphism theorem is given. The reviewer wishes to point towards the very interesting problem 10.1 on page 1247 of Part II: Let \(f:=X+X^2Y\). Show that there do not exist \(g,h\in k[X,Y,Z]\) such that \(f,g,h\) is a Jacobian triple (i.e. \(\det(\text{Jac}(f,g,h))\in k^*\)). It is obvious that such a triple is not an automorphism, so anyone attempting to prove the Jacobian Problem should first prove that such a triple does not exist!