Pro-\(\ell\) branched coverings of \(P^ 1\) and higher circular \(\ell\)- units (Q583310)

Let \(S_ 0\) be a finite subset of the complex projective line \({\mathbb{P}}^ 1\) containing the points \(0, 1\quad and\quad \infty.\) Let \(\ell\) be a fixed prime number. In this situation, the authors study the category \(X(S_ 0)\) of all those finite branched coverings f: \(Y\to {\mathbb{P}}^ 1\), with Y being a complete, smooth, irreducible complex curve, such that the degree of the Galois closure of f is a \(power\quad of\quad \ell\) and f is unramified outside \(S_ 0\). More precisely, they show that there is a smallest common field \(\Omega =\Omega (S_ 0)\) of definition for all objects and morphisms in \(X(S_ 0)\), which implies that each object and morphism admits a canonical \(\Omega\)-model, and the basic aim of their studies is to obtain an explicit description of \(\Omega\) in terms of circular \(\ell\)-units. The main result in this direction is the proof of the fact that the common field of definition \(\Omega\) coincides with an explicitly constructible infinite, non-abelian \(pro-\ell \quad extension\) of the field \({\mathbb{Q}}(\mu_{\ell^{\infty}})\) unramified outside \(\ell\) and containing the group of circular \(\ell\)-units. This very deep result is basically obtained by the following ingenious trick. It is shown that there are ``sufficiently many'' elementary objects of the form f: \({\mathbb{P}}^ 1\to {\mathbb{P}}^ 1\) in \(X(S_ 0)\), and those, together with their cusps, are studied first. This, in combination with investigating the Puiseux expansion at a cusp of an elementary object, is the basic idea for proving the main theorem. Since such interesting curves like the Fermat curve of level \(\ell^ n\), the Heisenberg curve of level \(\ell^ n\), and the modular curve of level \(2^ n\) are contained in the category \(X(S_ 0)\), the author's results imply that the Jacobian of any one of these curves has only \(\ell\)-power division points which are rational over the common field of definition of \(X(S_ 0)\). It is announced that the consequences and generalizations of the latter results will be discussed in a forthcoming paper under the same title [Part II: Int. J. Math. 1, No.2, 119-148 (1990)].