scientific article; zbMATH DE number 1302900

From MaRDI portal
Publication:4249386





This article gives, in succinct yet very readable form, a beautiful survey of recent developments in geometric Galois theory. The reviewer feels he should refrain from long-winded paraphrases and just try to give some idea of the principal topics, strongly recommending that everyone interested in Galois module theory look at Erez's paper directly. The starting point is Fröhlich's theory of the \(\mathbb{Z} [G]\)-module structure of \(O_N\) with \(N/K\) an at most tamely ramified \(G\)-Galois extension. As a first classification, later developments either belong to ``tame geometric theory continuing classical tame theory into the multidimensional realm, or to the ``theory of Omega invariants, extending the study of Chinburg's \(\Omega(3,N/K)\) for non-tame Galois extensions \(N/K\) to objects coming from K-theory or from motives. The main emphasis of this survey is on tame geometric theory. A common trait of most results is that they equate, or relate, an algebraic object with an invariant coming from \(L\)-functions, which is in the classical tame case the so-called root number class. In the tame geometric theory, one considers a projective scheme \(X\) either smooth over a finite field (first subcase) or regular and flat over \(\text{Spec}(\mathbb{Z})\) (second subcase), with an action of a finite group \(G\) fulfilling an appropriate tameness condition: approximately this means that the inertia group of every point has order coprime to the residual characteristic. Let \(f: X\to Y=X/G\) be the quotient scheme. One has a ``refined Euler characteristic \(\chi\) which is a group homomorphism between two Grothendieck groups: at the source one takes coherent \(G\)-sheaves on \(X\), at the target cohomologically trivial \(R[G]\)-modules. Here \(R\) is the finite base field (first subcase) or \(\mathbb{Z}\) (second subcase). The existence of this refined characteristic hinges on the fact that for any coherent \(G\)-sheaf \(F\), the stalks of \(f_*F\) are \(G\)-cohomologically trivial, and this is closely related to Noether's theorem concerning the projectivity of \(O_N\) over \(\mathbb{Z}[G]\) for tame extensions \(N/K\). The definition of the algebraic object is slightly different in the two subcases; we only say that the homomorphism \(\chi\) which we just discussed, and sheaves of differential forms on \(X\) are the main ingredients. (If in the second subcase \(X\) has dimension 1, that is, relative dimension 0 over \(\mathbb{Z} \), then only the zeroth sheaf of differential intervenes, and this is just the ring of functions of \(X\).) We do not go into the description of the analytic class whose equality with the algebraically defined class one wants to prove. It is a sum of two terms, the root number class and the ramification class. The theorems in the first subcase are due to \textit{T. Chinburg} [see for instance Ann. Math. (2) 139, 443-490 (1994; Zbl 0828.14007)], and in the second subcase (``generalized Fröhlich conjecture) to \textit{T. Chinburg, B. Erez, G. Pappas} and \textit{M. J. Taylor} [see for instance Am. J. Math. 119, 503-522 (1997; Zbl 0927.14013)]. The article also discusses generalized \(\Omega\)-invariants \(\Omega_n(N/K)\) for \(n\geq 0\). They are an outgrowth of Chinburg's \(\Omega(3,N/K)\) and involve certain 2-extensions arising in algebraic \(K\)-theory. Their construction is due to \textit{T. Chinburg, M. Kolster, G. Pappas} and \textit{V. P. Snaith} [C. R. Acad. Sci., Paris, Sér. I 320, 1435-1440 (1995; Zbl 0840.11048)]. A very general motivic construction of \textit{D. Burns} and \textit{M. Flach} [Math. Ann. 305, 65-102 (1996; Zbl 0867.11081)] produces \(\Omega\)-invariants for various motives including Tate motives, and Burns and Flach (preprint) have shown how to recover the \(\Omega_n(N/K)\) from their own construction. In a short section, Erez presents the nice and very general result of \textit{G. Pappas} [Duke Math. J. 91, 215-224 (1998)] concerning normal bases in elliptic division orders. Previous results [\textit{A. Srivastav} and \textit{M. J. Taylor}, Invent. Math. 99, 165-184 (1990; Zbl 0705.14031)] had to rely heavily on elliptic functions. This rapid overview does not even mention all the topics covered in the article. A lot of deep mathematics goes into all of them. Another attractive feature of this article is that it includes proof sketches for selected special cases; we should also mention the extensive bibliography. For the entire collection see [Zbl 0905.00052].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4249386)