Hilbert modular forms and the Gross-Stark conjecture (Q640790)

This paper concerns \textit{B. H. Gross}'s \(p\)-adic analogue of the Stark conjecture, introduced in [J. Fac. Sci., Univ. Tokyo, Sect. I A 28, 979--994 (1981; Zbl 0507.12010)]. This conjecture describes the derivatives at \(s = 0\) of \(p\)-adic \(L\)-functions attached to totally odd ideal class characters \(\chi\) of totally real number fields \(F\), in cases where \(\chi(\mathfrak{p}) = 1\) for some prime \(\mathfrak{p} \mid p\) (in which case the value of the \(p\)-adic \(L\)-function at \(s = 0\) is forced to vanish, because of the form of the interpolating factor at \(p\)). Gross's conjecture predicts that the derivative of the \(p\)-adic \(L\)-function is equal to the value of the complex \(L\)-function multiplied by an ``\(\mathcal{L}\)-invariant'' defined in terms of a certain \(p\)-unit of the splitting field of \(\chi\). It was shown in the original paper of Gross that the conjecture holds when \(F = \mathbb{Q}\). This paper proves this conjecture for a wide class of \(F\) and \(\chi\). More specifically, the conjecture is proved under the assumption that (i) Leopoldt's conjecture holds for \(F\), and \textit{either} (ii) there are at least two primes of \(F\) above \(p\) or (ii') a certain condition relating the \(L\)-invariants of \(\chi\) and \(\chi^{-1}\) holds. For a sketch of the argument, the reader is referred to the paper's very clear and informative introduction.