A proof of A. Gabrielov's rank theorem (Q2050437)

Let \(\mathbb K\) be either the field of real or the field of complex numbers. In this review we let \(\mathbb K\) always be the field of complex numbers. Let \(\phi: {\mathbb K}\{\mathbf{x}\} \to {\mathbb K}\{\mathbf{u}\}\) -- where \(\mathbf{x}=(x_1,\ldots, x_n)\), \(\mathbf{u} =(u_1,\ldots, u_m)\) -- be a morphism of convergent power series rings. For such a morphism there are defined three kinds of rank for \(\phi\): the generic rank \(\mathbf r(\phi)\), the formal rank \(\mathbf r^{\mathcal F}(\phi)\) and the analytic rank \(r^{\mathcal A}(\phi)\). It is known that \(r(\phi)\leq r^{\mathcal F}(\phi)\leq r^{\mathcal A}(\phi)\) (cf. [\textit{S. Izumi}, Duke Math. J. 59, No. 1, 241--264 (1989; Zbl 0688.32008)]). Gabrielov's rank theorem in this case states: \(r(\phi)=r^{\mathcal F}(\phi)\) implies that \(r^{\mathcal F}(\phi)=r^{\mathcal A}(\phi)\). Now let \(\phi: A\to B\) be a morphism of \(\mathbf C\)-analytic algebras where \(B\) is an integral domain. The notion of formal rank, analytic rank and generic rank of \(\phi\) can be extended from the power series case to this case. This article provides a complete proof of Gabrielov's rank theorem which states: Let \(\phi: A\to B\) be a \(\mathbb C\)-analytic morphism, \(B\) being an integral domain. If \(r(\phi)=r^{\mathcal F}(\phi)\), then \(r^{\mathcal F}(\phi)=r^{\mathcal A}(\phi)\). [This is Theorem 1.4 of the paper.] In order to prove Theorem 1.4, it is enough to prove it for the case \(A=\mathbb C\{x_1,x_2,x_3\}\), \(B=\mathbb C\{u_1,u_2\}\) [cf. Section 3.2]. This is done in Sections 3--5. The authors make use of the Abhyankar-Moh reduction theorem, stated as Theorem 3.3, and a formal Bertini theorem, stated Theorem 3.4; they give a proof of that theorem. A factorization theorem [Theorem 5.8, named after Newton, Puiseux and Eisenstein] and its reformularion in Corollary 5.9, play an important role. Let us mention also the following interesting result. Let \(\phi: A\to B\) be a morphism of \(\mathbb C\)-analytic algebras; when does \(\widehat{\phi}(\widehat A)\cap B=\phi(A)\) hold true? If this equality is satisfied, then \(\phi\) is said to be strongly injective. In Theorem 1.6 it is shown; Let \(B\) be an integral domain. \(\phi\) is strongly injective iff \(r(\phi)=r^{\mathcal F}(\phi)=r^{\mathcal A}(\phi)\).