Continuation of analytic sets across real manifolds (Q1091519)

Let M be a k-codimensional \(C^ 2\) real submanifold of the domain \(\Omega \subset {\mathbb{C}}^ n\), \(k\geq 2\). To each \(\zeta\in M\) one can associate an integer s(\(\zeta)\) in the following way: - choose \(\rho_ 1,...,\rho_ k\) \(C^ 2\) defining functions for M in a small ball around \(\zeta\) ; - for each \(\alpha \in S^{k-1}\) define \(\rho_{\alpha}=\sum^{k}_{j=1}\alpha_ j\rho_ j\) and \(E_{\alpha}=\{z\in T^{{\mathbb{C}}}(M)| H(\rho_{\alpha},\zeta,z)=0\}\) where \(H(\rho_{\alpha},\zeta,z)\) is the Levi form of \(\zeta_{\alpha}\) in the point \(\zeta\). Then define \(s_{\alpha}=\max \{\dim_ 0L_{\alpha}|\) \(L_{\alpha}\subset E_{\alpha}\) is a complex linear subspace\(\}\) ; - put \(s(\zeta)= \max_{| \alpha | =1} s_{\alpha}\). The integer s(\(\zeta)\) does not depend on the choices made above. Let now \(A\subset \Omega \setminus M\) be a pure m-dimensional analytic set. The following two results are proved: Theorem 1. If \(s(\zeta)<m-1\) then A has an analytic continuation to a neighborhood of \(\zeta\). Theorem 2. If \(s(\zeta)<m-1\) for any \(\zeta\in \bar A\cap M\), then \(\bar A\) is an analytic subset of \(\Omega\). This last theorem improves (for \(C^ 2\) manifolds) a result of E. M. Chirca ``If \(m-1> \dim_{{\mathbb{C}}}T^{{\mathbb{C}}}_{\zeta}(M)\) for any \(\zeta\in \bar A\cap M\) then \(\bar A\) is analytic'' [see \textit{E. M. Chirca}, Complex analytic sets, Nauka, Moscow (1985; Zbl 0586.32013) p. 186].