On the Fontaine-Mazur conjecture for number fields and an analogue for function fields (Q1970616): Difference between revisions

The conjecture of Fontaine-Mazur says that if \(k\) is a number field, \(l\) a prime, and \(M\) an unramified \(l\)-adic analytic \(l\)-extension of \(k\), then \(M/k\) is a finite extension. This conjecture is wrong for function fields. But the authors show that, if we start with an equivalent conjecture and restrict to special cases, it is possible to get some interesting results. In his famous paper about analytic \(l\)-groups [Publ. Math., Inst. Hautes Étud. Sci. 26, 389-603 (1965; Zbl 0139.02302)] \textit{M. Lazard} introduced the notion of a powerful pro-\(l\)-group \(G\), which means that \(G/\overline{G}^l\) resp. \(G/G^4\) is abelian in the case that \(l\neq 2\) resp. \(l=2\). The conjecture above is equivalent to the conjecture that if \(M/k\) is unramified and powerful, then \(M=k\). The author proves the following: Theorem 1. If a function field \(k\) has a subfield \(k_0\) such that \(k/k_0\) is cyclic of degree prime to \(l\) and such that \(l\) does not divide the class number \(h(k_0)\), then any everywhere unramified powerful pro-\(l\)-extension of \(k\), Galois over \(k_0\), with no constant field extension, is finite. The author proves two similar theorems, and he has some results in characteristic 0 as well. The proof of theorem 1 uses ideas of \textit{N. Boston} [Some cases of the Fontaine-Mazur conjecture II, J. Number Theory 75, 161-169 (1999; Zbl 0928.11050)], who proved it in characteristic~0.