Explicit computations in Iwasawa theory (Q6165861)

The goal of this paper is calculating initial layers \(K_n\) of the anticyclotomic \(\mathbb Z_p\)-extension of an imaginary quadratic field \(K\), where \(p>2\) is a fixed prime. The somewhat generic title does not give away this interesting specification. For the purpose of the paper, ``calculating'' means finding a generating polynomial over the rationals with coefficients of manageable size. A well-known representative of this nice line of research is the early article of \textit{J. E. Carroll} and \textit{H. Kisilevsky} [Compos. Math. 32, 157--168 (1976; Zbl 0357.12003)], which pursues a similar goal. The present article is very well written and full of nice examples. The last of these exhibits the second layer \(K_2\) of the 3-anticyclotomic extension of \(K = \mathbb Q(\sqrt{-87})\); this is a cyclic degree 9 extension of \(K\), dihedral over \(\mathbb Q\), and is given by a degree 18 polynomial that needs four lines of print. Another attractive feature of the paper is that it presents two rather different methods and compares their relative advantages and drawbacks. The first is based on the theory of CM (= complex multiplication). The fields \(K_n\) that one wants to determine are subfields of certain ring class fields over \(K\), which in turn are defined via class field theory, using certain nonmaximal orders attached to \(K\). (Warning: in the paper one often needs the class group of \(K\) and other fields, and this always means the class group of the maximal order.) The first approach to produce these ring class fields is via special values of the \(j\)-function, but there are better functions, so-called class invariants, which lead to smaller coefficients in the polynomials. The other approach is perhaps even nicer and certainly very refined technically. In short, it is Kummer theory followed by descent. If one wants to find \(K_n\), one adjoins \(\zeta_{p^n}\) to \(K\), looks at extensions of that larger field obtained by extracting a \(p^n\)th roots of a \(p\)-unit \(\eta\), and descends a suitable extension of this kind to produce \(K_n/K\). The point is that the \(p\)-units of the larger field can often be written down in terms of cyclotomic numbers, and by very clever use of Galois actions the authors are able to identify a suitable \(p\)-unit \(\eta\). Once one has this, one is able to find a defining polynomial by existing routines. The calculations are done in Magma plus Andreas Enge's CM package. As the authors aptly say in their introduction: If the class group of \(K\) is large, the Kummer method seems better than the CM method; if \(n\) grows, the Kummer method is potentially hampered by the involved enlargement of \(K\). -- This is an inspiring paper for all readers who are interested in doing calculations of a similar kind. For the entire collection see [Zbl 1416.11009].