The Jacobian conjecture and the extensions of polynomial embeddings (Q1195984)

Let \(X\subset\mathbb{C}^ n\) be an algebraic subset of \(\mathbb{C}^ n\). We say that the set \(X\) has the Abhyankar-Moh property (A.M.P.) if for any polynomial embedding \(\varphi:X\to\mathbb{C}^ n\) there exists a polynomial automorphism \(\Phi\) of \(\mathbb{C}^ n\) such that \(\varphi\) is a restriction of this automorphism to the set \(X\). Abhyankar and Moh proved that a line in \(\mathbb{C}^ 2\) has A.M.P. It can be easily deduced from this that the cross \(K_ 2=\{x\in\mathbb{C}^ 2:x_ 1\cdot x_ 2=0\}\) has A.M.P. The analogous questions in \(\mathbb{C}^ n\) \((n>2)\) for the hyperplane (the Abhyankar conjecture) and the \(n\)-cross \(K_ n=\{x\in\mathbb{C}^ n:x_ 1\cdot\dots\cdot x_ n=0\}\) (the McKay- Wang conjecture) are open, moreover in the case \(n\geq 3\) we don't know any example of a hypersurface which has A.M.P. --- In the present paper we consider the Abhyankar-Moh property for the \(n\)-cross \(K_ n=\{x\in\mathbb{C}^ n:x_ 1\cdot\dots\cdot x_ n=0\}\) in \(\mathbb{C}^ n\). Our main result is: Theorem. If the Jacobian conjecture holds then the \(n\)-cross \(K_ n=\{x\in\mathbb{C}^ n:x_ 1\cdot\dots\cdot x_ n=0\}\) has the Abhyankar- Moh property. If the \(n\)-cross has no property of the extension of polynomial embeddings in \(\mathbb{C}^ n\) then our theorem refutes the Jacobian conjecture and shows that the counterexample might be obtained by construction of a ``strange'' embedding of the \(n\)-cross into the \(n\)- space \((n\geq 3)\). This suggests that if the Abhyankar conjecture is false then the Jacobian conjecture is false, too. On the other hand we show at the end of the paper that the Abhyankar conjecture implies that the \(n\)- cross has the Abhyankar-Moh property.