Global stable splittings of Stiefel manifolds (Q2672204)

\textit{H. Miller} [Topology 24, 411--419 (1985; Zbl 0581.55006)] produced stable homotopy equivalences between the Stiefel manifold \(V_{n,q}\) of orthonormal \(q\)-frames on \(F^n\), where \(F\) is one of the skew fields \(\mathbb{R}\), \(\mathbb{C}\), or \(\mathbb{H}\), and a wedge of Thom spaces over Grassmann manifolds of \(k\)-planes in \(F^q\), where \(1\leq k\leq q\). Upon taking a limit, this produces stable splittings for the spaces \(O\), \(U\), and \(Sp\), and more generally for \(O/O(m)\), \(U/U(m)\), and \(Sp/Sp(m)\) where \(m\geq 0\). In the article under review, these stable splittings of Miller are refined into the realm of \textit{global equivariant homotopy theory} -- an equivariant homotopy theory with compatible actions of all compact Lie groups; see \textit{S. Schwede}'s book [Global homotopy theory. Cambridge: Cambridge University Press (2018; Zbl 1451.55001)]. The spaces above all have meaningful global homotopy types \[ \mathbf{O}/m\quad \mathbf{U}/m\qquad \mathbf{Sp}/m\quad m\geq 0 \] refining their classical definitions, and the same goes for Grassmann manifolds and Thom space constructions. The main result states that \(\mathbf{O}/m, \mathbf{U}/m\), and \(\mathbf{Sp}/m\) all have globally equivariant stable splittings in terms of wedges of global Thom spaces over global Grassmann manifolds. In particular, this recovers the aforementioned classical result of Miller. Moreover, the author provides a framework of \emph{\(G\)-global homotopy theory} (other treatments of which can be found in work by \textit{M. Barrero} [``Operads in unstable global homotopy theory'', Preprint, \url{arXiv:2110.01674}] and \textit{T. Lenz} [``\(G\)-global homotopy theory and algebraic \(K\)-theory'', Preprint, \url{arXiv:2012.12676}]), where \(G\) is any topological group. With this framework, it is shown that the global stable splittings of \(\mathbf{U}/m\) and \(\mathbf{Sp}/m\) are compatible with the additional action of the Galois group of \(\mathbb{C}\) and \(\mathbb{H}\) over \(\mathbb{R}\), respectively. As the author states in the introduction, the global splitting results are not a formal consequence of Miller's result. In particular, the global spaces above are not in the essential image of the left adjoint to the forgetful functor from the global stable homotopy category to the classical stable homotopy category. Moreover, the ``genuine'' equivariant nature of global homotopy theory is critical for the main result. Indeed, the global splitting results above are false in the naïve stabilisation of global unstable homotopy theory -- it is crucial that suspensions indexed by nontrivial representations are invertible.