Two lectures on number theory, past and present (Q1844249)

In the second of these two Ritt lectures, the author quotes Bourbaki, who (in a historical note) observes that the history of mathematics should proceed like the musical analysis of a symphony. There are a number of themes. It is fairly clear when a new theme appears for the first time. But then it gets mixed up with the other themes, several may appear simultaneously, they are shifted from one instrument to the other, etc. The author uses this approach in analyzing the evolution of number theory since Fermat to our days. His main Leitmotiv is the continuity of number theory throughout these three centuries: ``... what we are doing now is in direct continuation of what has been done by the greatest number-theorists since Fermat started it all in the seventeenth century''. In the first lecture the question is raised: ``what is number theory?'' No dogmatic answer is given, but the author makes a strong case (with examples ranging from Euler and Gauss to Hardy) for his view that analytic number theory is not number theory, while elliptic function theory is (``Gauss is clearly not making any distinction between the two topics''). After a short mention of Fermat's work and his isolation as a 17th century number theorist (the ``theme'' of Fermat's work will reappear often), one has to go straight to Euler. It is, presumably, well known that Euler proved many of Fermat's assertions (Fermat, apparently, failed to write up many proofs that he actually possessed and Euler had to reconstruct them, or find new ones). However, this reviewer doubts that many are aware to what extent Euler anticipated Riemann in the study of the zeta-function (first appearance of this theme) -- including the proof of the functional equation, at least for integer arguments. Most of the first lecture deals with Euler, but Lagrange, Gauss and Jacobi also are mentioned. As ``themes'' one notes the appearance of Gaussian (more generally, algebraic) numbers, of class numbers and the special relevance of curves of genus one. A beautiful interpretation of Fermat's method of infinite descent in terms of complex multiplication of elliptic functions is given. With this in mind, it seems rather unlikely that Fermat could have handled the equation \(x^5 = y^5 + z^5\) (not to speak of arbitrary exponent \(n\)) by ``his'' method. The lecture ends with the mention of further work of Euler on the (``Riemann'') zeta-function \((\zeta(2n) = \pi^{2n}\cdot R\), \(R\) rational; there are infinitely many primes in each congruence class \(p\equiv \pm 1 \pmod 4\), and his proof of the pentagonal number theorem (a theta function theorem), by which a new contact is established with the theory of elliptic functions. \bigskip The second lecture starts by developing the ''theme'' of the zeta-function, with its ``variations'' \((L\)-series, general Dirichlet series, Epstein zeta function, etc.). Other work that anticipated Riemann's (and which, although due to well-known mathematicians like Schlömilch, Malmquist and Clausen, seems to be hardly known by the general mathematical community) is mentioned. Finally, Riemann's own contributions (``Riemann ...was definitely not a number-theorist'') are duely discussed, particularly from the point of view of (1) the zeta function as (a multiple of the) Mellin transform of a theta function; and (2) the Riemann hypothesis. Here we learn that Riemann had anticipated Hardy in the proof that there are infinitely many zeros of \(\zeta(s)\) with \(\Re s = 4\). The ``theme'' is further developed by the discussion of the work of Dedekind (ideal theory, algebraic number fields and their zeta-functions) and by the mention of Hecke. Here new themes appear: quadratic forms and reciprocity laws. A fleeting return to Fermat shows how the class number problem is present in the background of his discussion of binary quadratic forms that do, or do not represent a given prime p, according to the residue class of \(p\) modulo the discriminant of the corresponding quadratic field. This leads to the consideration of the quadratic reciprocity law (Euler, Legendre, Gauss). This ''theme'' is developed by the mention of Gauss' work on the law of biquadratic reciprocity. The study of Gauss' memoirs on this topic led the author to the formulation of his own famous conjectures (on zeta functions and ''Riemann hypotheses'' of algebraic varieties over finite fields), now proved through the work of Dwork, Grothendieck, M. Artin, Lubkin and Deligne (completed after the present lectures). Interesting, little known details from Gauss' diary are mentioned and several ''themes'' (algebraic number fields, binary quadratic forms, class groups, ideal factors, cyclotomic fields, Kummer fields) are present almost simultaneously. A leap brings us to the 20-th century. The themes of the zeta function and of the reciprocity laws (related to abelian extensions of number fields) are still with us Artin, Hilbert, also Furtwängler and Takagi). Dedekind's definition of a zeta function is applied (by E. Artin) to function fields. Gauss' theory of quadratic forms reappears and is developed further (Hermite, Eisenstein, H. Smith, Minkowski, Siegel). Out of the mixing of these themes emerges the arithmetical theory of algebraic groups (a new ''theme'', perhaps?). The last 2 or 3 ''themes'' lead to contemporary work. What does the future hold? A possibly fruitful direction is to consider now the case of non-abelian Galois ex-tensions. Another one is to look upon class field theory, as the theory of \(\mathrm{G}(1)\) over a number field and to attempt to generalize our present results to \(\mathrm{G}(n)\) (see some results of Jacquet and Langlands for \(n=2)\). It is clear that looking at the program of a concert (or even at the write-up of a music critic) is far from conveying the enjoyment of the music itself. For this reason, the interested reader is advised to take advantage of the fact that these Ritt lectures are available in print.