Étale cohomology and the Weil conjecture. With a historical introduction by J. A. Dieudonné (Q4039844)

The classical Riemann hypothesis states that all zeros of the Riemann zeta function in \(\mathbb C\) lie on the line \(\text{Re}\,s=1/2\). Due to the work of E. Artin, F. K. Schmidt and H. Hasse, it was shown that a zeta function can be associated to an arbitrary algebraic variety \(X\) over a finite field \(\mathbb F_ q\). These zeta functions turned out to be much simpler than the classical one, and since they are defined in terms of numbers of points of \(X\) in the finite algebraic extensions of \(\mathbb F_ q\), the analogue of Riemann hypothesis has immediate arithmetic implications allowing to give estimates for the number of solutions of algebraic equations in finite fields. In 1936 Hasse proved an analogue of the Riemann hypothesis for the zeta functions of elliptic curves. In the process of proving this conjecture for arbitrary curves, A. Weil developed the modern theory of Jacobian and abelian varieties which led him to the concept of abstract algebraic varieties over arbitrary domains. This work summarized in \textit{A. Weil}'s book ``Foundations of algebraic geometry.'' Providence, R.I.: American Mathematical Society (1962; Zbl 0168.18701)] marked the beginning of a new era in algebraic geometry. By 1949 A. Weil proved the analogue of the Riemann hypothesis for a number of classes of algebraic varieties over finite fields, but he realized that in order to prove the conjecture in full generality it was necessary to think of a substitute for the classical topology which would be defined for varieties over arbitrary fields and would have the usual properties. - Assuming the existence of such a topology and a suitable cohomology theory and using the Lefschetz trace formula for the computation of the number of fixed points of the Frobenius automorphism, A. Weil was able to formulate (in 1949) his famous conjectures. In particular, the analogue of the Riemann hypothesis in the form suggested by Weil says that the eigenvalues of the representation of the Frobenius endomorphism in the (substitute of the) \(i\)-th cohomology space of \(X\) are integral algebraic numbers whose absolute values are equal to \(q^{i/2}\). Weil suggested the Jacobian as a substitute for the cohomology of curves and the Chow ring of algebraic cycles as a substitute for the cohomology of certain rational varieties (e.g. Grassmannians), but it was quite unclear how to proceed further. This circle of problems, where arithmetic, analysis, algebra and geometry come together in such a beautiful way, attracted much attention, but only much later Grothendieck, following a suggestion of Serre, introduced the notion of étale topology and \(\ell\)-adic cohomology and succeeded in establishing some basic properties of this cohomology similar to properties of classical cohomology. From the existence of such cohomology theory it already follows that the zeta function is rational (this is the first of the Weil conjectures). Grothendieck also made the next step by formulating the so called ``standard conjectures'' about algebraic cycles [cf. \textit{S. Kleiman} in: Dix Exposés Cohomologie Schémas, Adv. Stud. Pure Math. 3, 359--386 (1968; Zbl 0198.25901)]. These standard conjectures imply the Weil conjectures and in fact go far beyond them, but at present there still is no idea how to prove some of the standard conjectures. At last, in 1973 P. Deligne proved the remaining Weil conjectures (the independence of the coefficients of the characteristic polynomial of the Frobenius endomorphism \(F\) in \(H^*(X,\mathbb Q_{\ell})\) of the prime \(\ell\) and the fact that the absolute value of the eigenvalues of \(F\) in \(H^i(X,{\mathbb Q}_{\ell})\) is equal to \(q^{i/2})\). The goal of the book under review is to present Deligne's proof of the Weil conjectures and the prerequisites from algebraic geometry, étale topology and monodromy theory used in this proof. The book begins with a historical introduction (from Gauss to Deligne) by \textit{J. Dieudonné} [reprinted from Math. Intell. 10, 7--21 (1975; Zbl 0329.14012)]. The first chapter is devoted to Grothendieck's étale cohomology theory going as far as finiteness theorems and comparison theorems for complex varieties. The exposition here is based on \textit{A. Grothendieck}'s Sémin. géométrie algébrique 4 = SGA 4 (1963/64; see Lect. Notes Math. 269, 270, 305 (1972/73; Zbl 0234.00007; Zbl 0237.00012; Zbl 0245.00002)], but the authors use the language of derived categories. In chapter 2 the authors prove Poincaré's duality theorem, the Lefschetz fixed points theorem, and the rationality of zeta [a more detailed exposition of this material can be found in SGA 7 (1967-69); see Lect. Notes Math. 288 (1972; Zbl 0237.00013) and 340 (1973; Zbl 0258.00005) and Exposé VIII (1971; Zbl 0221.14026)]. It should be noted that another textbook presentation of the contents of chapters 1 and 2 is given by \textit{J. S. Milne} [``Étale cohomology.'' Princeton, New Jersey: Princeton University Press (1980; Zbl 0433.14012)]. The third chapter is devoted to the monodromy theory of Lefschetz pencils which plays a major role in Deligne's proof. The principal reference here is SGA 7 II [Lect. Notes Math. 340 (1973; loc. cit.)]. Finally, in chapter 4 the authors present Deligne's proof of the Weil conjectures and discuss some generalizations and applications. There are three appendices devoted respectively to the fundamental group, derived categories, and descent. The interest in Deligne's proof of Weil's conjectures obtained as a result of an interplay of ideas from diverse parts of mathematics is widespread among mathematicians, and since the proof requires rather complicated techniques, there certainly is a need for a self-contained exposition. Thus the book under review fills in an important gap in mathematical literature. On the other hand, the authors' presentation is still rather abstract and formal, and the book would have benefited from introducing examples, exercises and discussions.