Construction of maximal unramified \(p\)-extensions with prescribed Galois groups (Q2430683): Difference between revisions

In the early days of class field theory, Furtwängler used to comment counterexamples to various conjectures on the capitulation of ideal classes by saying that this would only be a counterexample once it was shown that the group in question actually occurs as the Galois group of a second Hilbert class field of number fields. Various authors, for example Emil Artin and Arnold Scholz, constructed such extensions for certain types of groups, and also succeeded in showing that every finite group occurs as the Galois group of an unramified extension of number fields. Showing that every finite group occurs as the Galois group of the class field tower of a number field is a much more difficult problem. The main result of this article is an important step in this direction: the author shows that every finite \(p\)-group is the Galois group of the maximal unramified \(p\)-extension of a suitably chosen number field. Similarly, the groups occurring as the Galois groups of (not necessarily finite) maximal unramified \(p\)-extensions are exactly the pro-\(p\)-groups with countably many generators. These theorems are derived from two basic propositions. Proposition~I basically claims that there exist totally complex number fields \(k\) with large degree and \(p\)-class number \(1\), and follows from a result of \textit{K. Horie} [Invent. Math. 88, 31--38 (1987; Zbl 0585.12004)] and the theorem of [\textit{B. Ferrero} and \textit{L. C. Washington}, Ann. Math. (2) 109, 377--395 (1979; Zbl 0443.12001)]. Proposition~II is a somewhat technical result that allows proving the main theorem by induction on the order of the \(p\)-group; the proof is rather long and uses Chebotarev's density theorem as well as the theory of embedding problems in Galois theory.