A note on the ideal class group of the cyclotomic Z_p-extension of a totally real number field (Q4548889)

Let \(p\) be a fixed prime number, \(k\) a totally real number field, \(k_\infty=\bigcup k_n\) the cyclotomic \(\mathbb Z_p\)-extension of \(k\), \(A_n\) the Sylow \(p\)-subgroup of the ideal class group of \(k_n\), \(A_\infty=\underrightarrow{\lim} A_n\). In a previous paper [J. Number Theory 88, 175--190 (2001; Zbl 0972.11104)] the author studied the cokernel of the natural map \(j_0 : A_0\longrightarrow A_\infty^\Gamma\), where \(\Gamma = \text{ Gal} (k_\infty/k)\), under the extra hypothesis that \(k\) is abelian over \(\mathbb Q\), semi-simple (viz. \(p\) does not divide the degree of \(k/\mathbb Q\)). In the present paper, the only assumption is Leopoldt's conjecture, but the study is restricted to two cases: (i) There is only one prime ideal of \(k\) and it is totally ramified in \(k\): then \(\text{ coker }\,j_0 = (0)\); (ii) The prime \(p\) splits completely in \(k\): then \(\text{ coker }\, j_0 = (0)\) if \(F_n k_\infty=L_n\) for some \(n\), \(F_n\) being the Hilbert \(p\)-class field of \(k_n\), \(L_n\) the maximal pro-\(p\) abelian extension over \(k_n\) containing \(k_\infty\). The proofs essentially use class field theory and norm computations. Strangely enough, the author even bothers to reprove Nakayama's lemma (= proposition 1)! Reviewer's remark: It seems that for general number fields, the above setting is not appropriate, because \(A_{\infty}^{\Gamma}\) could be infinite. The ``right'' problem would be the study of the maps \(j'_n : A'_n \longrightarrow {A'_{\infty} }^{\Gamma_n}\), where \(A'_n\) is \(A_n\) divided by the subgroup generated by the classes of prime ideal above \(p\), because Gross' conjecture for \(k_n\) (which is implied by Leopoldt's conjecture in the totally real case) predicts the finiteness of \({A'_n}^{\Gamma_n}\). In this new setting, the author has also given a description of \(\text{ coker}\, j'_0\) in the totally real abelian semi-simple case, assuming Leopoldt's conjecture [Pac. J. Math. 206, 129 --137 (2002; Zbl 1055.11063)]. But note that the general problem has been completely solved by \textit{M. Le Floc'h, A.~Movaheddi} and the reviewer [``On capitulation cokernels in Iwasawa theory'', preprint (2003)].