Arithmetic differential equations on \(\mathrm{GL}_n\). II: Arithmetic Lie-Cartan theory (Q268154)

In [Arithmetic differential equations. Providence, RI: American Mathematical Society (AMS) (2005; Zbl 1088.14001)], the first author developed the theory of arithmetic differential equations, based on a kind of differential algebra where the classical derivation is replaced by the operator \(\delta: a \mapsto \frac{a^p - a}{a}\), where \(p\) is some relevant prime number. Although functional equations involving \(\delta\) look a lot like some kind of difference equations (involving the Frobenius operator or the so-called Mahler operators \(f(z) \mapsto f(z^p)\)), the first author constantly took as a model the theory of differential equations. This may be justified by the enormous successes of the latter theory, by far not yet equalled by ``the sister theories'' (as Birkhoff called them). One of the big successes of the theory of \textit{linear} differential equations has been the ability to put some order and structure in the much wilder theory of \textit{non-linear} differential equations. To a vast extent, this is due to the heavy use of differential geometry (which, like its ancestor, infinitesimal calculus, has exactly this aim: to linearize non-linear problems). One of the first success has been to use symmetries of differential equations to better understand them, thus gaining the benefit of the theory of classical groups, and also some infinite dimensional groups. More recent advances along these lines include the use of (linear) differential Galois theory to tackle non-linear integrability problems (Morales-Ramis criterion); and, above all, the Malgrange theory of Galois groupoid, which relies on all the resources of Cartan-Lie differential geometry. In the present paper, the authors are ``motivated by the search of a concept of linearity in the theory of arithmetic differential equations''. To this aim, they ``introduce [...] an arithmetic analogue of Lie algebras, of Chern connections and of Maurer-Cartan connections''. Then they go at length to show the power of these new tools. This, by the way, includes some new results on the linear group over a local field.