Field embeddings which are conjugate under a \(p\)-adic classical group (Q2449871)

\textit{P. Broussous} and \textit{M. Grabitz} [Commun. Algebra 28, No. 11, 5405--5442 (2000; Zbl 0968.22015)] investigated irreducible admissible representations of GL\(_m(D)\) via certain embeddings. Their work implies that if two embeddings \((E_i , \mathfrak{a})\) of M\(_m(D)\) with \(k\)-isomorphic fields and the same hereditary order \(\mathfrak{a}\) are equivalent then any \(k\)-isomorphism between the two fields can be realized as conjugation by an element of the normalizer of \(\mathfrak{a}\) in GL\(_m(D)\). If \(G\) is a \(p\)-adic unitary group of a signed Hermitian space \((V, h)\) over a central division algebra \(D\) of finite index \(d\leq 2\) over a non-Archimedean local field \(k\) of odd residue characteristic, the author considers in the paper under review embeddings \((E, \mathfrak{a})\) (so that \(E\) is a subfield of End\(_D(V)\) that extends \(k\) and \(\mathfrak{a}\) is a hereditary order normalized by \(E\)) which are invariant under the adjoint involution \(\sigma\) of \(h\). If \((E_i , \mathfrak{a})\) are two such equivalent embeddings of End\(_D(V)\) and \(g\) is an element of \(G\) such that \(E_1^g\) is equal to \(E_2\), then the main result the paper asserts that there is an element \(g_1\) of \(G\cap \mathfrak{a}\) such that \(E_1^{g_1}\) is equal to \(E_2\) under certain conditions. In one case, if the gcd of \(d\) and the residue class degree \(f (E_1|k)\) of \(E_1|k\) divides \(e(E_1|k)\) or if \(\dim_D(V) \mathrm{gcd}([E_1:k],d)/[E_1:k]\) is odd, then there is even a \(g_1\in G\cap \mathfrak{a}\) such that \(e^{g_1}=e^g\) for all \(e\in E_1\). In all other cases it is assumed that \(\sigma|_{E_1}\) is not the identity and that \(E_1|k\) has the extension property with respect to \(D\), that is, the non-trivial Galois element of the unramified field extension of degree two in \(E_1|k\) can be extended to an automorphism of \(E_1|k\). This extension property is necessary, and is investigated in more detail in the last section of the paper. In large parts the proof is geometric and uses the Bruhat-Tits building \(\mathfrak{B}(G)\) of \(G\). The hereditary order \(\mathfrak{a}\) corresponds to a facet of \(\mathfrak{B}(G)\) stabilized by \(E^\times\). The main result can thus be reformulated. If there is a point \(x\) in \(\mathfrak{B}(G)\) which is fixed by \(E^\times_1\) and \(E^\times_2\) in the reduced building of Aut\(_D(V)\), then \(E_1\) is conjugate to \(E_2\) under an element of the stabilizer of \(x\) in \(G\) if \(E_1\) and \(E_2\) are conjugate under an element of the stabilizer of \(x\) in Aut\(_D(V)\) and a weak extra condition holds.